Extracellular vesicles as surrogated biomarkers of Prostate Cancer metabolism. A metabolomics approach to study their role in prostate cancer progression

Extracellular vesicles as surrogated biomarkers of Prostate Cancer metabolism A metabolomics approach to study their role in prostate cancer progression Guillermo Bordanaba Florit 2024 Extracellular vesicles as surrogated biomarkers of Prostate Cancer metabolism A metabolomics approach to study their role in prostate cancer progression Guillermo Bordanaba Florit Supervised by: Juan Manuel Falcón Pérez Félix Royo López 2019 - 2024 (cc) 2024 Guillermo Bordanaba Florit (cc by 4.0) Port D’Addaia, Cala Molí raconet meu estimat; em sent més al teu costat com més lluny estàs de mi I record s’estiu passat que semblava sense fi i que em va fer dir en un esclat: S’any que ve, quedaré aquí. V Acknowledgements Cada etapa y proyecto personal es finito, en cada lugar donde nos encontramos es especial y siempre dejamos a personas que han sido y serán, directa o indirectamente esenciales en tu camino. Estas etapas y personas aparecen en un momento y, así sin más, te cambian la vida. De esta misma manera me sentí yo el día que comencé mi aventura con el proEVLifeCycle en el Laboratorio de Exosomas en Bilbao. Por esto quiero empezar dándole las gracias a JuanMa, por la oportunidad que me brindó sin conocerme de nada y por el gran ejemplo de jefe humano que he tenido y que he disfrutado. Gracias también por la tranquilidad y ayuda que me has dado en todos los proyectos que hemos realizado. Seguimos. A Félix por darme siempre en todo momento buenos consejos sean (vamos a llamarle) “políticamente” correctos o no; siempre con esa creatividad y ganas de ayudar que te caracterizan. Insisto en esto porque siempre me has ayudado a sacar las castañas del fuego aún yendo yo un poco a mi bola. Al resto de grupo de Exosomas, los que están y los que se fueron, por todo el apoyo y especialmente a Clara (hubiese puesto Clarita, pero no te gusta ), Jone, Elisabeth, Patri (el Alzheimer’s lab). Mención especial a la Plataforma con Sebastiaan, Diana y Oihane, un remanso de paz y brotes a partes iguales; era pasar la puertecilla con la txartela y pensar: ya está, casa. MarieAzparrenMarirún que te quiero mucho y que eres la mejor; siempre has estado ahí y ya ni recuerdo mi vida sin conocerte. A todas las personas de CICbioGUNE que de una forma u otra han puesto su granito de arena para que esta tesis tenga algo de sentido. Sobre todo, para tres personas que han puesto algo más que un granito de arena en diferentes momentos de forma desinteresada, gracias a Arkaitz, Janire y Ariane por vuestra ayuda en el desarrollo de esta tesis en cáncer de próstata. Praiseworthy people rule all life-changing events and in this case, the proEVLifeCycle consortia is an example of big and collaborative projects being also fun and inspiring. I would like to specially mention Guido, the coordinator action, and his team who are the standard bearers of good science, creativity and team work. To Elena who actually became a good friend. Big thanks to everyone in this consortia, ESRs, PIs, collaborators, lab members, external trainers; you all made this PhD a memorable journey. To Jason and Aled for all their help to push forward part of this thesis’ work. And finally, to Guillaume and his team. They hosted me in Paris and boosted this PhD project. If I grow old in science I’d like to resemble the vibes of your group . Esas personas que se vuelven importantes son aquellas que te facilitan las cosas y te hacen sentir como en casa. Cuando llegué a Bilbo, la bienvenida no pudo ser mejor en Ramón y Cajal 42. La verdad que me sentí como en una segunda casa. Especialmente a Marta y al incansable Trenco, por todas las bajadas al Izeki a contarnos cualquier tontería; por estar ahí y por tener un corazón tan grande como el que tienes. En las alturas en Blas de Otero también se ha quedado un trocito de mí. La casa es un lugar habitado por sureños (1/8 ya te da el certificado) situado a las afueras de Deusto y que lo siento como un hogar. Con mi segunda casa en BilbaoDeusto estoy muy conforme. La Isa es un todoterreno. Te admiro por la capacidad de tirar siempre para adelante y también por anteponer la gente VI que de verdad te importa. La verdad has sido mi sorpresa y ya no es solo que te quiera muchísimo sino lo que confío en ti. Luego está mi Agu, mi medio olivo aquí. Uno más de la familia y así me he sentido yo también siempre contigo. Admiro lo empático y generoso que eres en todas las facetas de la vida, no te haces idea de lo mucho que me has ayudado. Todo tiene su fin, pero yo no quiero que la casa termine nunca. Todo el mundo que se cruza en tu vida puede formar parte de ella de forma fugaz, temporal o quedarse para siempre. Todos ellos pueden, en un solo momento, cambiarte la vida o tu percepción de ella para siempre. Desconocidos cuando llegas y que luego parece que hayan estado siempre ahí. Muchas gracias a todos los que me habéis hecho sentir así en Bilbo, en cuatro años he compartido grandes momentos con todos y sabéis que siempre os llevaré conmigo. Y a María, sabes todo de mí y has estado siempre ahí, eres mi Augenstern. Dicen los sabios que somos cada momento y cada persona que nos haya dejado un pedacito dentro. Agradezco a todas y cada uno de mis amigos que han formado y que forman parte de mí. Els dijous de Mundo, les birres a la Cívica, Festònomes, Colònies biotec, findes de matar-nos a la biblio, Sant Joans, bàsicament totes les primeres vegades de tantes coses i qualsevol plan que acabi en –tec. Also, to all de Bongerd drinks, liberation days, Droevendaal parties, Dutch weird events, Panda points, ICA club, library days, TT with the Greeks; basically to my Wageningen family. A Natalia, porque hiciste una mejor versión de mí y porque básicamente si no fuese por ti nunca jamás hubiese hecho un doctorado y menos en España. To the craziest lab I have ever been and I will ever be, the AG Bar-Evens. Nico, my supervisor back then, the one who made me love science again. Por último, a mi familia. No és ni la familia més extensa del món ni tampoc la més normal del món. Però ens estimem, els estimo i m’estimen. Saber que sempre hi són, de forma incondicional és el major acte d’amor que mai he pogut sentir. Tinc sort de tenir-vos. Y volveré al castellano para mi familia aragonesa. La de sangre, que me trae recuerdos inolvidables y son mi mesa redonda donde voy a buscar opinión. La familia está creciendo y no puedo sentirme más orgulloso de poder formar parte de ello. Luego está mi otra familia aragonesa, mis primos. No haría falta ni mencionaros, siempre habéis estado en mi vida y siempre lo vais a estar. A s’àvia, s’única que tenc i sa que sempre a mirat de ferme feliç i cuidar-me. Bé, tu i l’avi. Adrián José, qué voy a decir de ti si es que estás siendo ahora mismo para mí un gran ejemplo en todo. Sa mamà y papá, lo que soy es únicamente por vosotros. VII Table of Content ACKNOWLEDGEMENTS ......................................................................................................................... V TABLE OF CONTENT .......................................................................................................................... VII ABSTRACT .......................................................................................................................................... X RESUMEN ........................................................................................................................................ XV INTRODUCTION .................................................................................................................................... 1 I. PROSTATE CANCER DISEASE ............................................................................................................ 2 Prostate gland ................................................................................................................................. 2 Prostate cancer progression ........................................................................................................... 4 PCa metabolic rewiring ................................................................................................................... 6 Steroid hormones metabolism .............................................................................................................. 9 II. EXTRACELLULAR VESICLES ............................................................................................................ 11 EVs as biomarkers ........................................................................................................................ 12 Role of Lipids in EVs ..................................................................................................................... 13 EVs in metabolism ........................................................................................................................ 15 III. METABOLOMICS ......................................................................................................................... 17 Biological problem and experimental design ................................................................................ 19 Sample preparation ....................................................................................................................... 19 Data acquisition ............................................................................................................................. 20 Data processing ............................................................................................................................ 20 Statistical analysis ......................................................................................................................... 22 Metabolic pathways association.................................................................................................... 22 IV. MODELS IN RESEARCH ................................................................................................................. 23 V. HYPOTHESIS AND OBJECTIVES ........................................................................................................ 27 RESULTS ................................................................................................................... 29 CHAPTER 1. CURRENT APPROACHES TO STUDY EXTRACELLULAR VESICLES IN CELL PHYSIOLOGY AND ITS APPLICATIONS ...... 31 1.1. Introduction ........................................................................................................................ 31 1.1.1. EV heterogeneity and biomarker discovery. ........................................................................ 33 1.1.2. Single vesicle analysis techniques for biological characterisation of EVs. ............................ 34 1.2. Single-vesicle techniques .................................................................................................. 34 1.2.1. Label-free methodologies. .................................................................................................. 34 1.2.2. Label-based methodologies. .............................................................................................. 39 1.3. Recent advances in the EV field due to single vesicle analysis ....................................... 42 1.3.1. EV characterisation. ........................................................................................................... 42 1.3.2. EV trafficking and signalling mechanisms. .......................................................................... 45 1.3.3. EV biomarkers. .................................................................................................................. 49 1.4. Future outlook ................................................................................................................... 52 1.5. Concluding remarks .......................................................................................................... 55 CHAPTER 2. EVALUATION OF STEROID HORMONE TRANSCRIPTS ASSOCIATED TO URINARY EXTRACELLULAR VESICLES IN PROSTATE CANCER PROGRESSION .......................................................................................................... 57 2.1. Background ....................................................................................................................... 57 2.2. Materials and methods ...................................................................................................... 57 2.2.1. Database resources. .......................................................................................................... 57 EV biomarker datasets ................................................................................................................... 57 PCa biomarker datasets ................................................................................................................. 58 2.2.2. Real time quantitative PCR analyses. ................................................................................. 59 VIII Sample preparation ........................................................................................................................ 59 Real-time quantitative PCR methodology ........................................................................................ 60 Design of primers for SYBR Green and TaqMan qPCR ................................................................... 60 Housekeeping genes analysis ........................................................................................................ 60 2.3. Results .............................................................................................................................. 61 2.3.1. Selection of targets. ........................................................................................................... 61 Study of steroid hormones mRNA related to EVs ............................................................................ 61 Biomarker potential of steroid hormone enzymes expression ........................................................... 62 2.3.2. Selection of cohorts. .......................................................................................................... 63 2.3.3. Normalisation and Housekeeping genes. ............................................................................ 65 2.3.4. Evaluation of mRNA in EVs as targets. ............................................................................... 67 2.4. Discussion and Conclusions ............................................................................................. 69 CHAPTER 3. LIQUID CHROMATOGRAPHY COUPLED TO MASS SPECTROMETRY ASSAY TO MEASURE STEROID HORMONE ......... 71 BIOSYNTHESIS PATHWAY ...................................................................................................................... 71 3.1. Development of a Targeted Metabolomics Assay in Endocrine Tissues of Male Rats and ................................ Human Samples ..................................................................................................................... 71 3.1.1. Introduction ....................................................................................................................... 72 3.1.2. Materials and Methods ....................................................................................................... 76 3.1.2.1. Tissue and biofluid samples ........................................................................................... 76 3.1.2.2. Western blot Analysis .................................................................................................... 77 3.1.2.3. Chemicals and standards .............................................................................................. 78 3.1.2.4. LC-MS sample preparation ............................................................................................ 78 3.1.2.5. Ultra-high performance liquid chromatography ................................................................ 78 3.1.2.6. Mass spectrometry ........................................................................................................ 79 3.1.2.7. Statistical Analysis ......................................................................................................... 80 3.1.3. Results .............................................................................................................................. 81 3.1.3.1. Liquid chromatography and mass spectrometry method .................................................. 81 3.1.3.2. Analyte recovery optimization ........................................................................................ 82 3.1.3.3. Matrix effect .................................................................................................................. 85 3.1.3.4. Semi-quantitation of steroids in animal tissues ................................................................ 87 3.1.3.5. Quantitation of steroid hormones in human urinary samples ........................................... 89 3.1.4. Discussion ......................................................................................................................... 91 3.2. Clinical Evaluation of Metabolic Signatures as Biomarkers of Prostate Cancer Progression in ......................... Patient Urines ........................................................................................................................ 94 3.2.1. Background ....................................................................................................................... 94 3.2.2. Results and discussion ...................................................................................................... 94 3.2.2.1. Clinical cohort characteristics ......................................................................................... 94 3.2.2.2. Signature result ............................................................................................................. 95 3.2.2.3. Approaches to give an outcome ..................................................................................... 98 3.2.3. Conclusions ...................................................................................................................... 99 CHAPTER 4. TRANSFERENCE OF BIOLOGICAL COMPONENTS AND FUNCTIONALITIES DRIVEN BY EXTRACELLULAR VESICLES ... 101 4.1. Labelled Cholesterol Demonstrates Effective EV-Mediated Metabolite Transfer in Prostate Cancer ........ 101 4.1.1. Introduction ..................................................................................................................... 102 4.1.2. Materials and methods ..................................................................................................... 104 4.1.2.1. Cell culture, EV production and drug treatments. .......................................................... 104 4.1.2.2. Isolation of EVs. .......................................................................................................... 104 4.1.2.3. Characterisation of EVs. .............................................................................................. 105 4.1.2.4. EV uptake experiments. ............................................................................................... 105 4.1.3. Results ............................................................................................................................ 107 4.1.3.1. Cholesterol is quantifiable in cell models and extracellular vesicles. .............................. 107 4.1.3.2. Extracellular vesicles can transfer cholesterol to recipient cells. .................................... 110 XV Resumen El cáncer de próstata es una enfermedad que únicamente padecen los hombres y una de las más amenazantes con la edad. Esta afección causa un grave problema socio-económico debido a la falta de sensibilidad y especificidad en el diagnóstico durante fases iniciales de su desarrollo. El cáncer de próstata es multifocal y extremadamente diverso tanto a nivel molecular como histopatológico. También posee una gran variedad de mecanismos de progresión, pero las consecuencias fisiológicas derivadas y manera de evolucionar son todavía una incógnita. El crecimiento tumoral descontrolado y metástasis a tejidos cercanos genera una gran diversidad de perfiles metabólicos, proteómicos y transcriptómicos. Esta heterogeneidad sus diferentes fases proporciona flexibilidad y adaptabilidad a los tratamientos y, por lo tanto, también complica su diagnóstico. La comunicación celular es imprescindible para el progreso de cualquier enfermedad. Fue en 1967 cuándo Peter Wolf describió una nueva estructura subcelular: vesículas expulsadas por células. Actualmente, se ha demostrado que estas vesículas juegan un papel importante en la interacción celular. Estas vesículas extracelulares son estructuras delimitadas por bicapas lipídicas expulsadas por la gran mayoría de células. Su tamaño oscila entre nanómetros y micrómetros de diámetro y, pueden contener lípidos, proteínas, metabolitos y ácidos nucleicos, incluyendo ARN y ADN. Se trata de estructuras extremadamente heterogéneas en tamaño y composición, que está determinada en gran medida por las rutas de biogénesis que las formen. Tanto las moléculas de la membrana como las encapsuladas en el interior pueden activar cascadas de señalización que pueden producir respuestas fisiológicas como la maduración celular, eventos de coagulación o el establecimiento de nichos premetastáticos, entre otros. La reprogramación metabólica mediada por vesículas extracelulares se ha observado en muchos sistemas biológicos donde diferentes células, que pueden tener distintas funciones y orígenes, intercambian recursos metabólicos. Durante la oncogénesis, las vesículas extracelulares de los tejidos cancerosos o los tejidos adyacentes pueden inducir cambios metabólicos al proporcionar recursos metabólicos o desencadenando señales necesarias para su progresión. De hecho, las vesículas extracelulares se describen como mediadores relevantes en XVI los procesos de progresión del cáncer. Hasta la aparición de las tecnologías “omicas”, se utilizaron enfoques de biología molecular y fisiología para obtener datos que pudieran describir la interacción entre componentes y funciones biológicas específicas. La metabolómica tiene como objetivo medir, identificar y cuantificar un gran número de metabolitos en un sistema biológico determinado. Una estrategia de metabolómica dirigida utilizando cromatografía líquida acoplada a un espectrómetro de masas es el más adecuado para los objetivos de esta tesis. La presente tesis se centra en el estudio del contenido de las vesículas extracelulares y la manera en que interactúa con las células receptoras. Con este propósito, se ha dividido según paquetes de trabajo específicos que incluyeron: Un primer capítulo dedicado a la identificación bibliográfica y revisión de las metodologías actuales para caracterizar las vesículas extracelulares. La heterogeneidad de su estructura y composición dificulta su aplicación en diagnósticos médicos y terapias. Dicha diversidad también dificulta el establecimiento de roles fisiológicos, así como las funciones y composición de diferentes (sub)poblaciones de vesículas extracelulares. Los métodos que promedian el contenido de las muestras, como es el caso de las “ómicas”, tienden a diluir estas heterogeneidades. Los desarrollos recientes en el análisis individual de vesículas han revelado nuevas perspectivas en la comunicación entre células, así como la identificación y validación de biomarcadores. Sin embargo, hay que tener en cuenta que los enfoques 'ómicos' siguen siendo relevantes para estudiar procesos fisiológicos asociados a vesículas extracelulares a escala más global dentro de un modelo biológico. La investigación más detallada de la carga de vesículas extracelulares es de gran utilidad para determinar posibles mecanismos de progresión del cáncer en etapas tempranas de su desarrollo. En este sentido, las hormonas esteroides desempeñan un papel vital en la regulación de procesos celulares, y la desregulación de estos metabolitos puede provocar o agravar problemas patológicos, como enfermedades autoinmunes y cáncer. En el cáncer de próstata, son los principales impulsores de la oncogenicidad en las etapas muy tempranas de la enfermedad. Por lo tanto, la cuantificación de ARNm específicos en muestras biológicas relevantes, como la orina, es un recurso poderoso para diagnosticar la enfermedad. En el capítulo 2, se desarrollaron ensayos cuantitativos de ADN para estudiar la expresión de ARNm específicos. Primero, se identificaron 32 posibles marcadores de progresión del cáncer de XVII próstata utilizando CANCERTOOL. Estos candidatos discriminan entre muestras de pacientes con cáncer de próstata e hiperplasia benigna, ya sea incluyendo muestras con presencia de metástasis o ambas. Además, diferentes estudios científicos los han descrito en vesículas extracelulares. Un panel de dichos candidatos, se cuantificó en una cohorte de orinas de pacientes humanos. Lamentablemente, no fue posible describir a ningún candidato como un biomarcador válido en muestras de orina para ser probado más a fondo. Focalizando en estudios funcionales, el estado oncogénico de las células cancerosas puede evaluarse midiendo la expresión de genes asociados al crecimiento dependiente de andrógenos (Capítulo 4.2.). La tasa proliferativa de las células también es un indicador complementario de la oncogenicidad. La línea celular LNCaP fue el modelo más apropiado para estudiar la dependencia de andrógenos y la expresión de KLK3 el mejor indicador del tratamiento de depleción de andrógenos. Tratando el modelo dependiente de andrógenos LNCaP con vesículas extracelulares, se perseguía describir si las vesículas extracelulares podían mantener el fenotipo oncogénico en las células receptoras. El estado oncogénico se evaluó midiendo la expresión de KLK3, que está asociado a la cascada de señalización de andrógenos. También se consideró el estado proliferativo de las células de cáncer de próstata para evaluar los tratamientos con vesículas extracelulares pero este no proporcionó resultados significativos. Tras la depleción de esteroides, la expresión de KLK3 disminuyó drásticamente con el tiempo. Esta pérdida de fenotipo oncogénico fue neutralizada con la presencia de vesículas extracelulares aisladas de cultivos celulares con crecimiento oncogénico. Esto indica la participación de vesículas en la transferencia de fenotipo oncogénico aunque no explica las implicaciones en la fisiología del cáncer. Estos resultados sugieren una transferencia de funcionalidades entre las vesículas extracelulares y las células receptoras. Esta transferencia puede darse a través de proteínas que activen cascadas de señalización o con diferentes recursos metabólicos. En el capítulo 3, para estudiar la presencia de metabolitos relacionados con hormonas esteroides en estos eventos de transferencia, se desarrolló un ensayo de metabolómica. La extracción y cuantificación de 11 miembros clave en la red metabólica de hormonas esteroideas, incluyendo andrógenos, estrógenos, progestágenos y corticoides, se describe en el capítulo 3 de esta tesis. Una característica importante de este tipo de ensayos es que se pueden implementar metabolitos adicionales a posteriori. El ensayo consiste en una extracción bifásica líquido/líquido y la subsiguiente cuantificación mediante cromatografía líquida de alta resolución acoplada a XVIII espectrometría de masas de tiempo de vuelo. Las eficiencias de recuperación oscilan entre el 74.2% y el 126.9%, y el 54.9% y el 110.7%, para los compuestos apolares y polares, respectivamente. En general, la pérdida de intensidad de señal debido al efecto de matriz no supera el 30%. El método se ha probado en diversas matrices, como tejidos de ratas, orinas humanas, líneas celulares humanas y vesículas extracelulares. En resumen, este ensayo puede medir simultáneamente metabolitos relacionados con hormonas esteroideas en un tiempo de ejecución de 6 minutos el cual puede incrementarse a 9 min para incluir el colesterol en el análisis. Además de su uso en estudios de transferencia, otro propósito del método era el de validar una firma metabólica de hormonas esteroides como biomarcador para el cáncer de próstata. Los lípidos y específicamente el colesterol desempeñan un papel fundamental en la progresión del cáncer de próstata. El colesterol es importante en vías metabólicas protumorales ya que es el precursor principal del metabolismo de hormonas esteroideas. Además, las vesículas extracelulares están enriquecidas con colesterol, constituyendo aproximadamente del 50 a 60% de su contenido lipídico. La transferencia efectiva de metabolitos asociados a vesículas extracelulares todavía no se ha explorado. En el capítulo 4.1. se trataron cultivos de células de cáncer de próstata con colesterol marcado en medios desprovistos de lípidos. Luego, las vesículas extracelulares purificadas se suministraron a células receptoras no marcadas para rastrear la internalización de colesterol mediante microscopía confocal y su posterior metabolización mediante el ensayo dirigido UPLC-MS descrito anteriormente. En este trabajo, se detectó colesterol asociado a vesículas extracelulares en células receptoras no marcadas. Además, al marcar las células de cáncer de próstata con un isótopo estable de colesterol, se detectaron vesículas extracelulares marcadas. En resumen, se ha demostrado que un ensayo de metabolómica dirigida combinado con un enfoque de biología molecular es útil para estudiar la internalización mediada por vesículas extracelulares y la transferencia efectiva de metabolitos relevantes como el colesterol. Finalmente, se utilizó una estrategia metabolómica semidirigida para estudiar la interacción entre diferentes compartimentos de la glándula prostática (Capítulo 5). La glándula prostática es un órgano complejo y heterogéneo compuesto por el epitelio y el estroma. El estroma prostático es complejo y, en conjunto con la progresión del tumor, experimenta alteraciones que incluyen la aparición de un fenotipo de fibroblasto asociado al cáncer. Esta población celular tan heterogénea a menudo incluye células con un fenotipo similar al miofibroblasto XIX que no están presentes en tejido prostático sano. Al utilizar biopsias de aguja de tejidos tumorales y normales asociados al mismo paciente, se pudo estudiar la reconfiguración metabólica de los fibroblastos estromales que experimentan una diferenciación hacia miofibroblastos. Se determinó que los fibroblastos asociados al cáncer son metabólicamente más activos y, por lo tanto, se ven aumentadas las vías metabólicas de producción de energía. Además, los fibroblastos asociados al cáncer muestran un metabolismo lipogénico elevado tanto por la presencia de compuestos de reserva lipídica como de vertiente estructural. Así como en estrategias anteriores, se incubaron vesículas extracelulares derivadas de líneas celulares de cáncer de próstata con fibroblastos primarios sanos. Se detectó un aumento en la respiración basal de los fibroblastos, reflejando el fenotipo similar a la enfermedad. Con estos resultados, se puede proponer que el cambio en el perfil metabolómico de los fibroblastos estromales asociados tejido tumoral está impulsado por vías metabólicas dependientes de oxígeno; sin embargo, los mecanismos específicos aún no están claros. INTRODUCTION 2 I. Prostate Cancer Disease The World Health Organisation describes prostate cancer as one of the most frequently diagnosed and deadly types of carcinomas. It is a global healthcare problem with an incidence of approximately a quarter of all cancer diagnoses in Europe. Prostate cancer (PCa) is an exclusive disease of men that misses sensitive and specific diagnostic tools to surveil early stages of its progression. This causes a huge socio-economic problem for families, partners and caregivers, through loss of life or protracted palliative incapacity. In fact, many men initially attend clinic with a tumour that is already in an advanced and incurable state. Prostate cancer is often a multifocal disease and even within a single tumour, several molecular and histo-pathological arrangements are observed. Many of the underlying mechanisms of progression have been proposed as possible targets for diagnostics and therapeutics; however, the physiological drivers and consequences in its evolution are not fully understood. Nowadays, prostate-specific antigen (PSA) blood screening tests are the cornerstone for PCa surveillance at early stages. Up to 40% of men undergo unnecessary biopsies due to its poor specificity. Upon biopsy of the patient, the serum PSA test is combined with the clinical T-stage and Gleason score as standard tests to discriminate patients with low, intermediate or high risk to suffer PCa1. The progression of PCa within prostatic tissue is very diverse, so it is at metastatic stages. Metastases principally occur in bone and lymph node tissues but they are also reported in liver, lung and brain tissues2. This diversity causes variation in therapy response and resistance mechanisms. A high proportion of men exhibit slow growing tumours that are unlikely to progress to a life-threatening stage. Often, initial treatments of androgen-ablation reliably trigger regression but this is temporary, and within 2-3 years, castrate resistant PCa may recur, becoming a particularly intractable and aggressive disease. Yet, discrimination of indolent from aggressive disease is difficult. For these men, the identification of novel targetable pathways to be considered during decision-making process for therapeutic intervention. Furthermore, a better understanding of PCa disease will also help to design improved treatment strategies. Above all, understanding the molecular and physiological mechanisms of PCa disease hold an enormous potential to transform disease diagnosis, patient stratification and novel treatment modalities; providing superior information of wealth. Researchers need to decipher the secrets of PCa, from the spark that unleashes cancer development to the survival strategies improved by tumour cells. Prostate gland The prostate is the main accessory gland of the male reproductive system. The human prostate is a walnut-sized organ whose base is located at the urinary bladder neck and the apex at the urogenital diaphragm (Figure 1). As far back as the mid-sixteenth century, when Andreas Vesalius published his observations of the male accessory glands, prostate anatomy was of medical interest3. Interestingly, the macroscopic anatomy of this organ differs considerably between species. Rodent model organisms such as rats or mice utilised are to model human diseases; therefore, it is important to acknowledge the differences in the structure of human and rodent prostates. Unlike human, rat or mouse prostate is not merged into a unique anatomical structure with three glandular regions but it is composed of four distinct 3 lobular structures (Figure 1). Nonetheless, there is no existing evidence that supports a direct relationship between the specific mouse prostate lobes and the specific zones in the human prostate4. Figure 1. Representation of a human and a mouse prostate. On the left, a diagram of an adult human prostate where the urethra, the bladder and the three major regions of a prostate are indicated. On the right, a mouse prostate is depicted. It is structured in four lobes (similar to rat organization): lateral prostate, dorsal prostate, ventral prostate and anterior prostate. 5Reprinted from Urologic Clinics of North America, 282, LaTayia Aaron, Omar E. Franco, Simon W. Hayward, Review of Prostate Anatomy and Embryology and the Etiology of Benign Prostatic Hyperplasia, Volume 43, Issue 3, Copyright 2016, with permission from Elsevier. In humans, the prostate gland contains three major glandular regions, which differ both histologically and biologically: the peripheral zone, the central zone, and the transition zone. The latter is the main origin of prostate hyperplasia6 and carcinoma is more common in the peripheral zone7. At a histological level, the prostate is a branched duct organ organized in glandular acinis constructed by the epithelium and sustained by the stroma7. The two compartments influence each other via signalling pathways to promote prostate development and correct functioning8. Both tissues are extremely heterogeneous with many types of cells participating in different physiological functions. The epithelium contains secretory epithelial cells around the glandular lumen, where they secrete the prostatic fluid (Figure 2A). These secretory cells, also called luminal cells, are surrounded by basal cells and sparse neuroendocrine cells, both attached to a basal lamina9,10. Basal cells are associated to ductal integrity and secretory cell supportive functions while neuroendocrine cells role is rather unclear. Yet, it is known they secrete specific hormones and lack the expression of androgen receptor. Fundamentally, the prostate epithelial compartment holds the man glandular function: The secretion of prostatic fluid to the lumen (Figure 2A). The prostatic fluid is the major contributor to the ejaculate volume and contains factors to control maturation and motility of sperm or the ejaculate’s fluidity, among others. 4 Beyond the basal lamina, a prominent fibromuscular stroma offers physical support and contraction of the gland (Figure 2A). A major function of the stromal compartment is to ensure the appropriate microenvironment for the epithelial compartment, which is provided principally by surrounding fibroblasts. In addition, it participates in the vascularisation of the prostate but also evacuates prostatic fluid towards the urethra by contraction. Remarkably, the stromal compartment provides many supportive signals to retain or restore gland homeostasis in healthy conditions or during regeneration processes6. Prostate cancer progression The structure and well-functioning of the prostate is often impaired with aging and, finding the molecular or physiological mechanism that triggers PCa to spark is the one-billion dollar question. In general, it is accepted that tumour initiating cells undergo specific biological modifications such as the inactivation of certain pathways or the expression of specific genes. The prostate gland frequently suffers from inflammation due to an overgrowth of stromal cells (and epithelial cells) at transition zone and periurethral areas, which is usually associated to benign prostatic hyperplasia (BPH). However, when the growth is localised in specific loci of the prostate and associated to the epithelium it is prone to be an initiating tumour. The identification on gross inspection of PCa by palpation of the gland is often challenging. According to experts, the foci should be at least 5 mm in diameter for reliable diagnosis as tumours tend to be multifocal6. Besides, much larger areas can neither be accurately identified. As mentioned, tumours are mainly found in the peripheral zone11,12, followed by the transition zone and then central zone. Figure 2. Histological section of a human prostate tissue slice. Hematoxylin plus eosin provides a comprehensive picture of microanatomy of tissues by precisely staining nuclear (hematoxylin) and cytoplasmic (eosin) components. A. Healthy prostate tissue with glandular acini is depicted; a dotted red line delimitates the structure. The three different compartments are indicated with black arrows. B. Prostate cancer tissue in an advanced stage, with Gleason score 7 (3+4) and minor components of cribriform glands. Images are adapted from http://commons.wikimedia.org, and reproduced with permission under Creative Commons Attribution 4.0 International License http://creativecommons.org/licenses/by/4.0/. 5 Table 1. Summary of the current TNM staging classification based on AJCC’s classification. Adapted from 7,13,14. Stage Definition Primary tumour, clinical (T) TX Primary tumour cannot be assessed. T0 No evidence of primary tumour T1 Clinically inapparent tumour not palpable or visible by imaging. Slow growing and PSA levels are low. T1a: Tumour incidental histologic finding in 5 % or less of tissue resected. T1b: Tumour incidental histologic finding in more than 5 % of tissue resected. T1c: Tumour identified by needle biopsy (i.e. because of elevated PSA levels. T2 Tumour confined within the prostatea. PSA levels are medium or low. T2a: Unilateral. Tumour involves half of one lobe or less. T2b: Unilateral. Tumour involves more than half of one lobe but not both. It can be large enough to be felt by DRE. T2c: Bilateral. Tumour involves both lobes. T3 Tumour extends through the prostate capsuleb.Locally advanced tumour with high levels of PSA. T3a: Extracapsular extension (unilateral or bilateral). T3b: Tumour invades seminal vesicle(s). T3c: Tumour cells are poorly differentiated. Perhaps, invasion of bladder or rectum. T4 Tumour is fixed or invades adjacent structures other than the seminal vesicle(s): bladder neck, external sphincter, rectum, levator muscles, pelvic wall, or all the above. T4a: Tumour has spread to regional lymph nodes. T4b: Tumour has invaded distant lymph nodes or other tissues such as bones (i.e. pelvic wall), levator muscles, external sphincter and others. Regional lymph nodes (N) NX Regional lymph nodes cannot be assessed. N0 No regional lymph node metastasis. N1 Metastasis in regional lymph node or nodes. Distant metastasesd (M) MX Distant metastasis cannot be assessed. M0 No distant metastasis. M1 Distant metastasis. M1a: Non-regional lymph node(s). M1b: Bone(s). M1c: Other site(s). a Tumour that is found in one or both lobes by needle biopsy, but not palpable or reliably visible by imaging, is classified as T1c. b Invasion into the prostatic apex or into (but not beyond) the prostatic capsule is classified as T2, not T3. c There is no pathologic T1 classification. d When more than one site of metastasis is present, the most advanced category is used; pM1c is most advanced. 12 categorisation includes not only vesicles but also lipoproteins, viruses, protein aggregates, ribonucleoprotein complexes and exomeres76–78, which provides an additional layer of diversity. This is because such structures can participate in similar (pato)physiological processes along with EVs. Figure 5. Summary of EV biogenesis, EV-mediated cell-to-cell interaction and intracellular trafficking. (a) Cargo of EVs includes different types of proteins nucleic acids and metabolites. (b) Plasma membrane invagination endocytose extracellular constituents and cell surface proteins. This structure can fuse with constituents of the ER, trans-Golgi network and mitochondria, which leads to the formation of early sorting endosomes (ESEs). Then, ESEs maturate to late sorting endosomes (LSEs) forming intraluminal vesicles (ILVs) by a second invagination of the membrane. Finally, multivesicular bodies (MVBs) are generated and they can be sorted to lysosome, undergoing degradation; or transported to plasma membrane, docking on luminal side of cells. Exocytosis of MVBs releases ILVs as EVs to the extracellular space. (c) EVs can be internalised by cells using different pathways: fusion with cell membranes, a receptor-mediated entry, clathrin-coated invaginations or lipid rafts interactions, among others. Reproduced from 79 with permission under Creative Commons Attribution 4.0 International License http://creativecommons.org/licenses/by/4.0/. EVs as biomarkers In 1989, a task group on Biomarkers and Risk Assessment proposed a definition for biomarkers, which are chemicals, metabolites of chemicals, enzymes and other biochemical substances to document their interaction with biological systems. In this line, biomarkers can be defined a broader sense by including chemical, physical and biological interactions between these systems. The use of biomarkers in research has grown exponentially during the last decades seeking to establish direct measurements of disease causes and affections80,81. One of the most important criteria is their accessibility and biological relevance. A biomarker should be present in a minimally invasive source and it should be as much sensitive as possible. Moreover, it has to be meaningful in a biological framework and often provide a better understanding of the disease’s mechanism. This will definitely lead to describe fast 13 response biomarkers upon disease progression or treatment and hence, provide risk stratification and prognosis. Currently, biomarkers for PCa do not guarantee an effective and reliable diagnosis due to the high heterogeneity and plasticity of the disease. On top of that, its diversity is usually assessed relying on tissue biopsies. This strategy is rather invasive and does not provide a reliable prognostic output. For this reason, and considering that current diagnostic tools present several flaws, finding other biomarker sources is fundamental in the diagnosis field. Nearly all cell types expel EVs to the extracellular milleu with signalling or disposal purposes65,82. These physiologically relevant cell goods are exchanged safely thanks to the protection offered by the lipid membrane. As cellular components are not easily degraded, EVs represent a new source of biomarkers that could describe physiological processes or cellular stages in disease progression. Depending on the information provided, biomarkers can be used as predictors when they provide information on the biological response to therapeutic interventions, prognostic assets when they provide information on disease progression and/or recurrence or diagnostic assets if they can distinguish healthy and disease samples83–85. Remarkably, EVs have been identified in wide variety of human tissues or fluids, including blood, urine, saliva, synovial fluid, cerebrospinal fluid, uterine fluid, bile, breast milk or faeces, among others65. In this line, liquid biopsy may be highly informative as it is low-invasive and could reflect the status of a disease better than a conventional biopsy. Therefore, novel biomarkers can be described from EVs and their concomitant compounds to inform about the progression of a disease or a physiological response. Role of Lipids in EVs The biogenesis of EVs commences with processes of microautophagy in late endosomes or outward budding at the plasma membrane86. Historically, the resulting vesicles have been called exosomes and microvesicles, respectively. The biological events governing their formation are different; however, they are all somehow driven by their lipid composition. For instance, the translocation of acid sphingomyelinase generates ceramides in the outer leaflet of the plasma membrane, inducing a curvature in the plasma membrane that triggers microvesicle budding87. Actually, the disruption of lipid organization appears to be critical; not only ceramides affect their formation but also cholesterol/sphingomyelin ratio, and phosphatidylserine asymmetry to the outer leaflet. After microautophagy in late endosomes, multivesicular bodies (MVBs) containing intraluminal vesicles (ILVs) can reach the plasma membrane in a dynamic process regulated by cholesterol. Upon fusion with plasma membrane, ILVs are released to the extracellular space where they are called exosomes. Exosomes are enriched in disaturated molecular species of phospholipids, which accounts for their increased membrane rigidity87. In summary, there are many lipid-related pathways involved in the biogenesis of EVs. The release of vesicles implies more steps previous to budding events in which the composition of membranes plays a critical role. Studies working with cells in vitro show a remarkable enrichment of cholesterol, sphingomyelin, glycosphingolipids and phosphatidylserine in EV membranes88. Moreover, the presence of lipids with small head groups as ceramide, diacylglycerol and phosphatidic acid has been associated with the formation and release of EVs89. Considering the role of lipids in EVs one 14 should consider both their structural role in membranes and the manner they modify membrane fluidity and curvature, but also their potential role as signalling molecules89. Figure 6. Classes of lipids. Different groups are displayed according to their structure and synthesis pathways. 15 EVs in metabolism The membrane-enclosed entities that we baptised as EVs carry proteins, nuclear acids, lipids and metabolites. Indeed, EVs are often considered metabolic machineries since they can contain active enzymes, metabolic substrates or ligands to carry out enzymatic reactions or regulate enzyme activities. There are various mechanisms by which metabolic cargoes are loaded into EVs: identifying specific miRNA90, proteins91 or performing posttranslational modifications, such as acetylation, sumoylation or phosphorylation92,93 on proteins and RNA to increase the likely hood of actively or passively loading EVs with specific compounds. In addition, EVs are entities with a high content of lipids, a wide family of metabolites (Figure 6) with a very diverse chemical structure and hence, several properties physiological function. Although EV membranes are similar to donor cells they are enriched in specific lipids as cholesterol, sphingomyelins, ceramides and glycosphingolipids94–96. This may explain why proteins with affinity to certain lipids are enriched in EVs, i.e. tetraspanins or flotillin. While some of the mechanisms of cargo recruitment have been described, the involvement of metabolic assets in actual physiological processes is often overlooked97. It is known that metabolic resources are transported by EVs yet, the manner their cargo (reaction substrate -metabolitesor activity -enzymes-) interacts with recipient cell is not understood neither comprehensibly studied. Table 2. Glycolytic enzymes and occurrence in exosomes and PCa-derived EVs. Enzymes in the glycolytic pathway and their frequent appearance in EVs are listed. The numbers refer to their placement in the top 100 most frequently identified proteins. H = Human; D = Dog; S = Stallion; B = Bull. Proteomic data was retrieved on 2018-10-16 and the original table is extracted from98. Enzyme EVpedia ExoCarta Protasomal lipid raft Specie Hexokinase - - No H, D, S, B Glucose-6-phosphate isomerase 63 - Yes H, D, S, B 6-Phosphofructokinase - - No H, S Frusctose-biphosphate aldolase 12 18 Yes H, D, S, B Triosephosphate isomerase 20 27 Yes H, D, S, B Glyceraldehyde-3-phosphate dehydrogenase 1 4 Yes H, D, S, B Phosphoglycerate kinase 16 16 Yes H, D, S, B Phosphoglycerate mutase 87 - Yes H, D, S, B Enolase 2 9 Yes H, D, S, B Pyruvate kinase 3 12 Yes H, D, S, B Lactate dehydrogenase-A 7 13 Yes H, D, S, B Glycolytic enzymes frequently appear in EVs (Table 2) and prostasomes – EVs produced by prostatic cells – from living cells98. Several mechanisms can explain the presence of these enzymes in EVs such as the fact they are anchored to lipid rafts or tetraspanin microdomains99. Indeed, a majority of the glycolytic enzymes are related to lipid rafts (Table 2). Up-to-date, the full set of glycolytic enzymes has been identified in EVs100 through the efforts of multiple initiatives as EVpedia101, Vesiclepedia102 or Exocarta103. These databases have been compiling identified proteins in EVs by means of proteomics analysis. Intriguingly, lactate dehydrogenase-A, the enzyme responsible of converting pyruvate to lactate by oxidation of NADH to NAD+, is among the most frequently identified proteins in EVs (Table 2). 16 Nonetheless, EVs do not only convey components but they can transfer their activity. Extracellular ATP production was observed upon addition of fructose to purified prostasomes104. Other studies described arginase activity of hepatocyte-derived EVs105 or L-asparaginase activity in EVs derived from neural stem/progenitor cell106. This suggests EVs may function as independent metabolic units with a potential impact to the composition of the extracellular microenvironment. Metabolic reprograming mediated via EVs is present in many biological systems where cell types with different or the same functionality exchange metabolic assets107. Besides metabolites and enzymes, EVs can modify metabolic pathways by the transfer of nucleic acids, which can further regulates glycolysis108,109 or OXPHOS110, among others. During oncogenesis, not only cancer cells influence the environment by means of EV release but also stromal cells can drive metabolic changes in cancer cells by providing metabolic resources required for cancer progression111. Cancer-associated fibroblasts (CAFs) are defined by specific morphological features or expression of markers including α-smooth muscle actin (α-SMA), fibroblast-specific protein-1 (FSP1/S100A4), and fibroblast activation protein (FAP)112. Recent studies showed the crosstalk between cancer cells and CAFs plays a vital role in tumour growth by regulating their metabolism. Cancer cells promote an enhanced glycolysis of CAFs which, in turn, provide TCAand OXPHOS-related metabolites113. This phenomenon was recognized as the “Reverse Warburg Effect”. Interestingly, EVs derived from CAFs have been described as partially responsible for tumour cell survival in the hostile, nutrient-deprived, and hypoxic environment of PCa. These EVs contain several metabolites, including lactate, acetate, aminoacids, lipids, and TCA cycle intermediates, that further favours glycolysis, reductive glutamine metabolism, and proliferation of cancer cells67. Reductive carboxylation of glutamine replenishes TCA intermediates through glutamate and citrate, which are also precursors of major macromolecules related to anabolism and catabolism. In addition, lactate facilitates tumour cell survival under hypoxic and nutrient-deficient conditions114. These results further demonstrate that the existing metabolic symbiosis between PCa cells and CAFs. Increasing evidence are piled up reporting EVs as relevant mediators between CAFs and cancer cells. PCa cells releasing EVs can trigger normal fibroblast differentiation towards myofibroblasts or CAFs. In 2010, Webber et al. described that TGF-β1 contained in EVs could activate fibroblasts as soluble TGF-β1 does115. However, in further studies they showed PCa-derived EVs triggered TGF-β1dependent fibroblast differentiation that resembles cancerous stromal cells while soluble TGF-β1 did not deliver pro-angiogenic or tumour-promoting phenotype116. Other studies performed in neonatal fibroblast cells noted that colorectal cancer-derived EVs could drive the expression of the activated fibroblast marker α-SMA. In the same study, the authors described metabolic modifications such as aminoacid biosynthesis, enhanced glycolysis via glucose and lactate transport or upregulation of glycogen metabolism, upon EV treatment117. In breast cancer models, a metabolic remodelling of fibroblasts and increased glycolysis through cancer-released EVs was described117,118. These results contribute to the idea of an EV-mediated signalling that sustains tumour growth and the activation of fibroblast. It further suggests that in specific systems or conditions, EVs could be rather relevant contributors for cancer progression. 17 III. Metabolomics Until the rise of the omics era, molecular biology and physiology approaches were utilised to acquire data that could describe biological components and functions, separately. Nowadays, the interaction between those is an essential focus in the systems biology paradigm, which aims to explore physiological processes of an organism as a whole in a holistic manner. Omics strategies hold the potential to identify the entire set of biomolecules (Figure 7) contained in biological samples. This generates a bast amount of data that can be correlated to describe a biological system in a complete and integrated way. Metabolomics is defined as the comprehensive and quantitative analysis of small molecules within a biological system including cells, biofluids, tissues or entire organisms, commonly known as metabolites. In literature, the term metabolome first appeared in 1998 referring to the entire set of metabolites present in an organism of any kind. Metabolites are biologically active compounds - smaller than 1500 Da - which participate in all biological processes of a living cell as i.e. building blocks for macromolecules, energy carriers, signal effectors or inhibitors85,119. Distinct classes of compounds constitute the metabolome: lipids, aminoacids, inorganic species or nucleotides, among others. Metabolites are difficult to study due to their extreme variety of chemical and physical properties, such as molecular weight, polarity, solubility and volatility. Moreover, the number of metabolites considered in a metabolome is highly variable and tightly dependent on the studied organism. It ranges from 600 metabolites estimated in Saccharomyces cerevisiae to approximately 200,000 metabolites annotated in the plant kingdom. It is accepted human’s metabolome is smaller than plant’s one but more than 4,000 compounds have already been annotated120. There are two groups of metabolites regarding their implication in the biological system. Primary metabolites are those involved in biological processes, thus making them essential to life. This is the case of amino acids, organic acids, lipids, etc. Instead, secondary metabolites are those metabolites not essential for the cell to live. This is because they do not have a role in any essential biological process. Therefore, they are restricted to a selected set of cells, which synthetize them for specific biological functions. Metabolomics aims to identify and quantify a large number of metabolites in a biological system to resolve specific scientific hypotheses121. In general, the methods of choice are Nuclear Magnetic Resonance spectroscopy (NMR) or Mass Spectrometry (MS). The most widely used mass spectrometers are Orbitrap and Time-Of-Flight (TOF) based systems because they provide the advantage of analysing a complete mass range with relatively high acquisition rates122. Triple quadrupole (QQQ) or Quadrupole-TOF (Q-TOF) provide a higher sensitivity and the opportunity to analyse fragmentation patterns, which is very useful to identify known analytes in certain metabolomics approaches. MS is usually coupled to a separation system such as Gas Chromatography (GC-MS), Liquid Chromatography (LC-MS) or Capillary Electrophoresis (CE-MS) to deliver the identification of nay metabolite. To increase the separation power, the combination of several separation approaches into multidimensional systems is becoming common119. All these analytical platforms and methodologies generate large amounts of highdimensional and complex experimental raw data when used in a metabolomics context. The amount of 18 data, the need for reproducible research, and the complexities of the biological problem under investigation necessitate a high degree of automation and standardise analytical workflows. Many tools and methods have been developed to facilitate the processing and analysis of metabolomics data; most seek to perform reproducible data analysis and to work with different types of raw data. Figure 7. Scheme of the main omics strategies utilised in systems biology approach. The scheme represents the correlation of each omic approach to genotype or phenotype in any biological systems. Currently, there is no actual consensus regarding the classification of metabolomics studies. For this reason, we prefer to use a classification based on whether the researcher knows a priori which metabolites will be quantified119. A general metabolomics workflow is represented in Figure 8 considering this classification. A targeted metabolomics approach is the quantitative analysis – concentrations in the samples are determined – or semiquantitative analysis – relative intensities are registered – of few upfront-known metabolites that are associated to common chemical classes or linked to selected metabolic pathways119. This approach provides a metabolic profiling of specific biochemical pathways with biological relevance to the hypothesis. The use of isotope-labelled metabolites (with 13C, 15N or deuterium) in experiments permits to trace metabolic reactions and pathways. It also allows assessing the utilisation of a substrate by a metabolic network. This type of analysis can also be performed in an untargeted manner. The term untargeted metabolomics was coined to define the qualitative or semiquantitative analysis of the largest possible number of metabolites from different chemical and biological classes which are contained in a biological specimen119. This approach enables a rapid classification of samples according to their origin or status but also the analysis of different compartments of a system, for instance, cell content (fingerprint) and secretome (footprint) under controlled conditions. In this way, an alteration of a small subset of metabolites can be pinpointed in a broader context of analytes and describe the potential interactions with other metabolites and/or pleiotropic effects considering several pathways (or modules) of the entire metabolomics network. In contrast, identification reliability is loss. 19 Ideally, one would start with an untargeted analysis to provide a potential answer to the hypothesis and finally, confirm it by using a targeted approach. In this thesis, we will work with liquid chromatography coupled to mass spectrometry with a targeted metabolomics manner. In this approach, sample preparation is adapted to the chemical properties of the selected compounds and the sample matrix hence, reducing signal alteration of targeted analytes by non-desired compounds123,124. As mentioned, some hypotheses may require a broader set of metabolites to properly describe the impact to the biological system. As the inclusion of the whole metabolome has technical limitation and complications, different extraction and sample preparations have been combined in so-called analytical platforms. These platforms enable the extraction of chemically similar metabolites in different sets to analyse them separately. Therefore, we can analyse polar and apolar compounds separately yet maintaining a strong resolving and identification power without compromising the selectivity and sensitivity of the assay. For targeted metabolomics, the typical metabolomics workflow (Figure 8) is organised in six general steps: establish the biological problem and experimental design (1), ideally within the input from a statistician, sample preparation (2), further data acquisition (3), data processing (4), statistical analysis (5) and pathway functional analyses (6). Biological problem and experimental design A clear formulation of the biological problem is crucial because it will govern further decisions during the experiment design. The integration of all subjects related to the workflow is essential to both consider all the potential confounding variables and obtain readily interpretable results. At this point, the type of metabolomics approach is defined considering the sample size and type – cells, fluids, EVs, tissues or intact organisms -, the experimental conditions, frequency of collection, storage conditions, metabolic quenching to interrupt metabolism and, the analytical platforms and preparation strategies. Sample preparation Once the researchers of the roundtable ruled the most appropriate experiment design, the appropriate sample preparation must be defined. Because of the wide variety of physicochemical properties of metabolites, the extraction procedure is usually optimised for a specific set of compounds or chemical classes in the analytical platforms. It usually involves clean-up steps to remove sample matrix interferents (i.e. protein precipitation with methanol), preconcentration strategies to boost detectability and the selection of appropriate extraction solvents125. An important consideration to make is that metabolism is a constant flux; thus, stopping any potential metabolic reaction is imperative in sample collection, preservation and manipulation to avoid any potential loss of metabolite signals. Metabolite extraction methods from the collected samples have been developed to be effective for specific compound classes. This implies each method will lead to the loss of metabolites not specific of the assay. Considering all potential issues in extraction methods, an ideal preparation should: (i) incorporate a preservative so that the metabolite composition reflects the original one at the sampling moment; (ii) be as non-selective as possible in order to incorporate the broader range of metabolites. To 20 this aim, a combination of distinct extraction protocols may be considered; (iii) be simple and fast, to avoid metabolite loss; (iv) be reproducible. Data acquisition The chemical diversity of the metabolome, as well as its wide dynamic range, is a drawback to achieve a comprehensive identification and quantification of an entire metabolite set in a biological system. It is important to emphasize that not only distinct chemical properties, structural arrangements and functionalities differentiate metabolic compounds but also their concentration levels. For instance, sugars and concomitant intermediates range between micromolar or millimolar depending on the matrix126,127 but steroid hormones are at the nanomolar range. Typically, the analytical techniques employed for data acquisition are NMR and MS. Their strengths and limitations must be considered to decide the best approach. While NMR spectroscopy is considered as a universal metabolite detection technique, with a low sample manipulation and a wide variety of chemical classes detected simultaneously, MS is much more sensitive and specific, however, it usually requires a separation technique to reduce sample complexity and ion suppression128–130. Nowadays, MS is usually coupled to high-pressure liquid chromatography (HPLC) and ultra-high pressure liquid chromatography (UPLC). The mass spectrometer is composed of an ion source, which is the entry point for the ionized metabolites into the equipment, a mass analyser to separate the analytes by mass-tocharge ratio (m/z) and the detector. The m/z abbreviation stands from the value resulting from dividing the mass number (m) of an ion by the corresponding charge number (z) and it is characteristic of each chemical formula131. Analytical platforms often exhibit stability issues over time, which can be circumvent with the use of a quality control (QC). It is prepared by mixing small volumes of all running samples. QC is normally run several times upfront samples to check for instrumental stability and intercalated during samples run to ensure data reliability. Data processing A correct analyte quantitation or semiquantitation is imperative in targeted metabolomics. Usually, an assay is developed under specific and optimised conditions for a specific set of metabolites. Therefore, further validation of the parameters specificity/selectivity, accuracy, linearity, resolution, limits of detection and quantification and, reproducibility is required to establish a methodology119. Raw data collected from the mass spectrometer is colossal and unmanageable. Henceforth, the aim of this step is to generate a 2D data table of features or metabolites defined as m/z - RT pairs where data has been already corrected according to QC, normalised and, if required, transformed, centered and/or scaled. For more detailed information relate to Chapter 6 of the book 119. 21 Figure 8. Analytical workflow for studies in metabolomics. This metabolomics workflow comprises the sequential steps that underline both targeted and untargeted analyses. Starting from a problem or hypothesis formulation it provides metabolic pathway relationships. Reprinted from 119. RESULTS 31 Chapter 1. Current Approaches to Study Extracellular Vesicles in Cell Physiology and Its Applications This chapter focusses in assessing current approaches in the EV field including recent advances derived from the access to new technologies. Their application in cancer physiology and biomarker discovery is also discussed. It has been published as a Review Article in Nature Protocols journal and it is appended in the supplementary material. Bordanaba-Florit, G. et al. Using single-vesicle technologies to unravel the heterogeneity of extracellular vesicles. Nat Protoc 16, 3163–3185 (2021). DOI: 10.1038/s41596-021-00551-z 1.1. Introduction Since the description of minute bodies found in a piece of cork by Robert Hooke in 166561, both our scientific knowledge and technical abilities have increased enormously. As techniques have become more accurate and intricate, so has our understanding of biological processes and structures. Technological advances in the field of imaging have resulted in the identification of cell structures, such as the mitochondria62 and nuclei63, and the discovery of different levels of cellular complexity. In 1967, Peter Wolf visualised platelet dust in fresh platelet-free blood plasma using an electron microscope64; thus, a mammalian vesicle-like structure was described for the first time. Gradually, these vesicles were characterised in more detail. It has been established that they are released by all kinds of cells (prokaryotic and eukaryotic) into the extracellular milleu172–175. The process of vesicle secretion is conserved throughout evolution, suggesting that such vesicles are likely to have specific roles in the cellular and organismal development and survival176. Indeed, later discoveries showed that secreted vesicles participate actively in many physiological processes in mammals, for example coagulation, inflammatory response, cell maturation, adaptive immune response, bone calcification and neural cell communication, among others65,68. In addition to their critical functions in normal physiology69,70, secreted vesicles mediate in several pathological processes71,72, such as the establishment of pre-metastatic niche during cancer progression73,74. Nowadays, these secreted vesicles are extensively reported and widely known as extracellular vesicles (EVs). EVs are heterogeneous, nanoto micrometre-sized, bilayer lipid containers secreted by most cell types. They are multi-purpose carriers that can contain a wide variety of cargos such as lipids, proteins, metabolites, sugars, RNA (mRNA, miRNA, siRNA) and even DNA65. When they are taken up by recipient cells, they trigger intracellular signalling through EV surface molecules or by the release of cargo into cell compartments via endocytic pathways. These processes can further activate downstream genetic or metabolic pathways in the recipient cell67,68. In mammals, EVs have been found in body fluids like plasma, urine, saliva, breast milk and seminal fluid, among others. They are classified into three groups according to their biogenesis mechanisms and biophysical properties:  exosomes: typically 30–150 nm in diameter, derived from intracellular endosomal compartments, 32  microvesicles: 100–1000 nm in diameter, produced by outward budding and pinching-off the plasma membrane, and  apoptotic bodies: 50–5000 nm in diameter, released as blebs by cells undergoing apoptosis64,65. In a systematic review of guidelines for this field, the International Society for Extracellular Vesicles (ISEV)75 endorsed a categorisation of EVs isolated using ultracentrifugation into large, medium and small EVs. However, it is important to note that ultracentrifugation precipitates not only vesicles but also lipoproteins, viruses, protein aggregates, ribonucleoprotein complexes and exomeres76–78, which provides an additional layer or diversity but also a bias upon analysis of EV samples. Furthermore, there is evidence that exosomes, microvesicles and apoptotic bodies contain subpopulations with unique roles in biological processes177,178. These subpopulations are tightly integrated with a broad range of biological processes and exhibit a wide range of functionalities, which makes them an outstanding source of potential biomarkers for early diagnosis, drug delivery systems for therapeutics, or vaccine production systems72,179–181. During the last few decades, the interest in EVs and their applications has grown considerably. Many articles and reviews have focused on the functional role of EV heterogeneity. Their role in specific biological processes, such as cargo trafficking or regulation of signalling pathways, and their potential as biomarkers have also been examined68,72,178,182,183. Nonetheless, most of these studies examine the vesicles in bulk and use ensemble-averaging assays. Although such methods have been proven useful in specific cases, it is important to realise that the extensive heterogeneity of structure, composition and function of single vesicles are masked in such assays177,182–184. For example, the inability to detect the heterogeneity of molecular states of reaction pathways, individual proteins or nucleic acids may lead to a misinterpretation of ensemble measurements185,186. Recent developments in single-vesicle analysis (SVA) have opened new opportunities for the examination of heterogeneity within EV (sub)populations at the individual EV level and their characterisation on the nanometre scale176,182. This new information is paramount for understanding the biological functions of EVs and for their potential clinical use. Different EV populations and subpopulations can be isolated according to their physicochemical properties, yet the existing isolation technologies are intricate and still need to be further developed177,182. There are five main groups of techniques for sorting EV populations and subpopulations, based on ultracentrifugation, size, immunoaffinity capture, polymer precipitation and microfluidics65. Since each technique type sorts EVs using a different principle, each method can yield different EV subpopulations from the same sample187,188. Moreover, the highly concentrated EV preparations may contain contaminants, such as large protein aggregates and lipoproteins, left behind by some of the isolation techniques189,190. Interestingly, different approaches can also affect the physicochemical surface characteristics of EVs191. In consequence, some techniques tend to enrich or discriminate against specific EV populations. As sorting EVs into populations is usually based on physical properties only, we have to assume that such classification is largely arbitrary with respect to the composition or function of EVs. Vesicle isolation and 33 enrichment techniques can help to yield more homogeneous EV subpopulations, albeit only for particular technique-specific parameters. In conclusion, although various isolation methods could help driving EV analysis towards a single-vesicle approach176, many different compositions and functionalities are still expected to be found within such EV populations. In this review, we describe the current methods used to study single-vesicles , and their contributions to the understanding of EV biology and biomarker discovery. Single-vesicle experiments can deliver direct information on the heterogeneous composition of EVs. They reveal multiple molecular states that govern EV functionality and transport and provide statistically valid information often lost in large ensemble experiments66,182,192. 1.1.1. EV heterogeneity and biomarker discovery. There are many technological challenges to be met in the development of EV-based diagnostics. The relevant vesicles must be identified and isolated from complex biofluids, and a specific diseaserelated EV population or population mix has to be detected. Biomedical studies of EVs often focus on seeking suitable biomarkers for the diagnosis of various diseases193,194. For example, the functional role of EVs in various types of cancer has been extensively studied71,73,74,180. In this review, we concentrate on prostate cancer (PCa) diagnostics as an example application to which the SVA of EVs has made important contributions. According to the World Health Organization, PCa is among the most frequently diagnosed types of cancer, accounting for approximately a quarter of all cancer diagnoses in Europe195, yet the lack of sensitive diagnostic tools and insufficient knowledge of the mechanisms of cancer emergence and progression are of major concern. PCa is a heterogeneous pathological state, both in the primary tumour in the prostate tissue and at the metastatic stage. It is unfortunately not recommendable to examine PCa diversity relying solely on tissue biopsies, since these are highly invasive procedures and do not guarantee an effective and reliable diagnosis196–199. The serum prostate-specific antigen (PSA) test—still the cornerstone of PCa screening—is particularly questionable. Up to 40% of men undergo unnecessary biopsies as a result of poor specificity of the assay. Remarkably, liquid biopsy has a potential for marker identification and provides better evidence of PCa diversity than the conventional solid tissue biopsy. In particular, prostateand PCa-derived EVs and concomitant markers are highly abundant in urine, blood and ejaculate samples60,200. Hence, these body fluids could be used for detecting and measuring the progression of the disease. For example, the EVs released by PCa cells carry unique prostate-specific membrane proteins (e.g. TMPRSS2, STEAP2, PSMA, PPAP2A, etc.) that enable the detection of pathogenic prostate EVs and their capture for ex vivo characterisation201. Although these data show that liquid biopsies may be highly informative and minimally invasive procedures, the methods for vesicle isolation, characterisation and identification for disease diagnostics remain challenging. Up to date, there are no standardised operating procedures for vesicle isolation and characterization for different types of samples and diseases, which complicates the employment of liquid biopsies as a clinical source for EV biomarkers. 34 1.1.2. Single vesicle analysis techniques for biological characterisation of EVs. Most of the studies reviewed in this article examine the three problems that can be addressed by taking advantage of SVA techniques: (1) characterisation of EV heterogeneity including subpopulations, surface (membrane protein and lipid) composition and vesicle content, (2) structural studies of EV membrane and soluble proteins and the assays to probe the metabolic activity of these proteins in a nativelike environment and (3) characterisation of the EV content and function depending on the cells of origin. The last task presents an interesting dichotomy: do the EVs reflect the properties of their cells of origin, or are they completely independent communication assets? On the one hand, it has been reported that EV surface and content depend on the parental cells202–204. On the other hand, some recent reports describe several EV subpopulations, with a range of different functionalities, originating from the same cell type66,67,176,182,205,206. Intriguingly, another recent article using SVA techniques demonstrates that the T2SS-like family of proteins is, in fact, responsible for selective cargo loading into EVs generated by the microorganism Shewanella vesiculosa207. 1.2. Single-vesicle techniques As increasing numbers of researchers have highlighted the importance of accurate EV (sub)population sorting and phenotyping, so far, more than 20 new SVA techniques have been developed66,176,182. Many of these use microfluidic devices designed to integrate various technologies to improve EV sorting and detection. Moreover, several of these methods have been used for characterising EVs at the singlevesicle level208–219. Some of these techniques can directly provide information on vesicle surface, content, size and shape while other may require an upstream physicochemical characterisation of the selected EV subpopulations to conduct surface profiling, monitor the expression of biomarkers and quantify them in body fluids. These technological advances should help to design new diagnostic devices for small sample sizes, using minimally invasive methods. Twelve different methods are presented in Table 5 and discussed in detail in the following sections. Some of these methods utilise labelling techniques (such as fluorescence or nanoparticle coating) to visualise the EVs, and others work as label-free systems (Figure 9). It is important to note that in some cases label-free approaches may hinder the detection of EVs because they often produce weak signals, which can be enhanced using a supporting labelling technique. 1.2.1. Label-free methodologies. Nanoparticle tracking analysis (NTA) is a technique based on the Brownian motion of microparticles in suspension and it is used to determine the size distribution in particle populations208,220. In this approach, microparticles are detected by scattering the light of a laser beam, which is tracked and recorded at video frame rates. However, this approach has some disadvantages and limitations. For instance, the accurate assessment of particle size distribution requires specific track lengths, a steady temperature and a large number of replicates to provide robust results. Care should be taken when comparing different samples because variations in buffer viscosity and microparticle concentration introduce statistical errors. Moreover, the close proximity of two particles can result in overlap of the 35 scattering signals. Accurate detection of particles with a diameter below 60 nm is challenging, regardless of the NTA machine used221. Furthermore, vesicles cannot be discriminated from other particles, such as protein aggregates or virus particles. The vesicles can be probed specifically, and undesired particles excluded from the analysis only by employing fluorescent markers, yet only a fraction of EVs may carry known markers that can be used for labelling a specific subpopulation. General fluorescent labels (such as lipophilic carbocyanines DiO or DiI) can be used instead. However, it is important any non-attached label is removed since this can mask the fluorescence signal emitted by labelled EVs66. Figure 9. Schematic overview of the main SVA techniques discussed in this review. Data visualisation and singlevesicle interpretation using each SVA methodology are depicted. In the centre of the figure, a (tumorigenic) cell releasing EVs is shown. The techniques can be divided into two groups: label-free (a-e) and label-based (f-i) methodologies. The methods used here are: a) Cryo-electron microscopy222,223, b) AFM, c) NTA176,224, d) RTM223, e) SPIRIS225, f) hrFC226, g) ddPCR227, h) SRM228 and i) fluorescence microscopy (TIRF image of synaptic vesicles is depicted)229. References show the source of the images. Raman tweezers microspectroscopy (RTM), also known as laser tweezers Raman spectroscopy (LTRS), can be employed to examine the chemical content of EVs. This approach can be used to 36 investigate both the surface and the internal volume of single EVs, revealing specific biomolecular signatures of proteins, lipids, nucleic acids and carotenoids as major contributors204–206,216,225–231. RTM is an inelastic scattering-based method. It employs a tightly focused laser beam for both optical trapping of single (or very few) vesicles in aqueous medium and excitation for subsequent Raman scattering, which provides a vibrational fingerprint from the trapped constituent biomolecules. The main inherent advantage of RTM lies in the signal linearity, which allows both qualitative and quantitative biochemical characterisation of single EVs. This method is also label-free and provides data with high information content209,223,230. The main disadvantage is that the scattering efficiency is usually very low and thus provides a rather low level of informative Raman signal. As a result, an extended data collection time is required. Therefore, RTM, with a typical processing capacity of 0.2 particles per min, is not considered a high-throughput methodology230. RTM can, however, be used to obtain interesting, unique information not only for EVs209–211,223,230–236 but also for many other bioparticles like liposomes, lipid layers on synthetic nanoparticles and others237–242. Several methods have been developed to compensate for the low Raman signal strength in RTM. For example, the vesicle concentration can be increased by drop-coating deposition of the sample, followed by drying243–247. Unfortunately, this approach results in loss of information about individual EVs, as does any other analytical study of a bulk sample. Another strategy to increase the Raman signal is to use surface-enhanced Raman spectroscopy (SERS). In this method, EVs can be exposed to various signal-enhancing nanoparticles and/or substrates to obtain a strengthened biomolecular signal212,248– 254. The main problem of label-free SERS is that the enhancement effect depends strongly on the distance between the biomolecule and the nanoparticle/substrate, and vanishes at distances longer than a few nanometers248. Therefore, this method is mainly suitable for characterisation of biomolecules on the outer surface of EVs. In addition, Raman modes corresponding to molecular vibrations perpendicular to the SERS surface are preferably enhanced248. As a result, the overall SERS vibrational spectrum is usually distorted, lacks reproducibility, and is often difficult to interpret. In electron microscopy, a beam of electrons is emitted onto a sample in a vacuum environment. The wavelength of electrons is shorter than the visible light used in optical microscopy; thus, the method gives images of much higher resolution, typically below 1 nm66. Cryogenic transmission electron microscopy (cryo-TEM) is among the electron microscopy methods most commonly utilised for EV characterisation. In contrast to the lengthy sample preparation needed for other TEM methods (usually taking hours), no heavy metals or fixatives are added, and no dehydration steps are required. This also limits sample damage and artefact effects, but yields lower contrast images255. In cryo-TEM, the samples are prepared by rapid freezing, typically with liquid ethane209,256. In this process, the water vitrifies, instead of forming ordered crystals, and the native structure of EVs is preserved257. The first exosome visualisation was achieved using cryo-EM in 2008258. Since then, this technique has successfully revealed EV polymorphism by imaging the membrane bilayers, EV structures and internal features of individual EVs209,213,255,259. Even though cryo-TEM is an extremely useful technique for high-resolution visualisation of EVs, this approach is relatively low-throughput. Cryo-TEM images typically only contain a few EVs (although the throughput could be enhanced by using automated search). In addition, cryo- 37 Table 5. Summary of main SVA techniques for individual EV characterisation. Technique Detection principle Information obtained Throughput Time per analysis (a) Sample preparation (b) Is sample reusable? Loading volume Working concentration (mL-1) High-resolution flow cytometry Elastic light scattering. Auto-fluorescence or fluorescence from external labels. Profiling(c) of EVs in a heterogeneous fluid mixture. Functionalised fluorescent labels characterise specific populations. High ~1 min Immunofluorescence or fluorescent conjugate staining protocol. No (d) 20–100 µL 107–1010 (e) Nanoparticle tracking analysis Imaging of Brownian motion pathways for EVs, using elastic light scattering or fluorescent labels. Particle concentration, size distribution. Moderate ~1 min Dilution or concentration of EVs to the range optimal for the method. Yes 0.3–1mL(f) 108–109 Raman tweezers microspectroscopy Raman scattering from optically trapped single EVs. Biomolecular(g) composition of the surface and the interior of single EVs. Low ~1-5 min(h) per one(i) trapping event Concentration of EVs to the range optimal for the method. Yes ~100 µL 107–1011(j) Surface-enhanced Raman spectroscopy (SERS) Raman scattering enhanced by external active surface/coating. Partial(k) biomolecular(g) composition of the membrane of single EVs. Moderate ~1-10 s per one(i) EV Fixation and coating protocol. No 50–100 µL 109 –1011 SERS with external labels Enhanced Raman signal from SERS nanotags. Number of EVs with specific functionalised SERS nanotags attached. High ~10 s Fixation and SERS-label staining. No 10–100 µL 10 –107(l) Cryo-transmission electron microscopy (Cryo-TEM) Transmission electron microscopy imaging. Morphological EV characterisation(m). Direct visualisation of single EVs and examination of contaminants. Low ~1 hour (h) Vitrification of water in EVs dispersed on carbon grid, using fast-plunge freezing. No 2–10 µL 1010–1012 Atomic force microscopy (AFM) Imaging (raster scanning) exploiting interaction force between the probing tip and immobilised EV. Precise(q) morphological, mechanical and biochemical (o) characterisation of the EV surface. Low ~1 min per image(s) Immobilisation protocol.(p) No 5–25 µL Relative(q) Table continued 44 Table 6. Summary of EV characterisation studies using SVA techniques. Technique Main conclusion Ref AFM The first time that plant exosomes have been visualised in their native state. They have been observed on the internal layers of the cell walls and their cargo assessed. 313 Matrix vesicles initiate changes during the mineralisation of the extracellular matrix. In the course of this process, the matrix vesicles increase their size and crystallinity and change shape. 314 AFM-IR The first attempt to probe the differences between the molecular constituents (proteins, lipids and DNA) and structures of individual vesicles in two subtypes of placenta stem cells. In this work, protein aggregates have been successfully differentiated from vesicle structures. 266 RTM Vesicle shape and size depend on the lipid composition of the membrane. A decrease in cholesterol concentration increases the local membrane curvature and stretches the vesicle. 241 Cancer research hrFC (immunofluorescence) Microenvironment acidity of the prostate tumour increases the release of prostate-specific EVs. 331 EVs from PCa cells are very heterogeneous, but specific populations are not associated with different cancer stages. Mainly microvesicles and several exosome subpopulations are found, according to the surface signature. 332 NTA Increase in the number of released exosomes under acid pH (6.5), independent of tumour histotype. 330 Tumour cells release more EVs than non-tumorigenic cells (most likely due to the acidic environment in tumour cells). 327 Heterogeneity-related research NTA Description of different EV populations in human glioblastoma cells. 322 NTA (with fluorescence) The first attempt to use this technique to determine the concentration and particle distribution in specific EV subpopulations, according to surface markers such as CD9, CD63, vimentin and LAMP-1. 328 Electron microscopy (cryo-TEM) Large diversity of exosome morphology revealed, regardless of the cell type and origin, suggesting that different exosome subpopulations from the same cell line perform different functions. 302 Visualisation of a large spectrum of EVs from cerebrospinal fluid, including multilayer vesicles, single, double and double-membrane vesicles, and internal vesicular structures. These specific subpopulations are suggested to serve as potential biomarkers for Parkinson’s disease. 321 Two different stimuli affecting EV release. Lipopolysaccharides and starvation conditions in human leukaemia monocytic cell line (THP-1) affect the type of EVs shed and, probably, the shedding process. 222 AFM EVs isolated from different species are imaged. Using the tapping mode, the mechanical properties of EVs are assessed, concluding that the rat EVs are more fragile than the mouse vesicles. 326 RTM During cell growth and starvation-induced aggregation, Dictyostelium discoideum produces EVs that differ drastically in their biomolecular composition (nucleic acids, proteins, lipids, carotenoids). 209 Detection of four EV populations with specific protein, phospholipid and cholesterol surface signatures. Their functionality determines their distribution; the populations are shared among several cell types. 230 In EV preparations from rat hepatocytes and human urine, single EVs from the same sample have different biochemical properties, well beyond the “usual” variations for EVs from rat hepatocytes or human urine. A quantitative method developed to measure the concentration of nucleic acids in a single EV. 223 Combined morpho-chemical profiling of individual EVs is proposed based on Raman-enabled nanoparticle trapping analysis. 233 45 SVAs have been extensively utilised in cancer research, for example, to characterise the EVs in PCa. NTA-based research has established cancer cells produce larger amounts of EVs than non-tumorigenic cells194,217,327 and that low extracellular pH increases the release of EVs from cancer cells328,329. Interestingly, NTA studies also support the idea that tumorigenic cells upregulate EV production due to the acidification of the immediate microenvironment327–330. Logozzi et al. have employed a combination of hrFC and immunofluorescence to discover that acidic microenvironment of prostate tumours induces the release of PCa-specific EVs330. These EVs are very heterogeneous, and several populations can be identified, according to their surface composition. However, it is important to note that this heterogeneity is not related directly to different cancer stages333. In summary, the highlighted reports of single-EV characterisation demonstrate that EV morphology and composition are largely independent of cell origin and that certain (sub)populations are involved in various diseases. These discoveries indicate that morphologically different EV populations may be distributed according to specific functions and biogenesis pathway, rather than their original cell-type. 1.3.2. EV trafficking and signalling mechanisms. SVA approaches have advanced EV research by tracking particular molecules and examining the changes in cells under different conditions. Early in 2020, FRET studies in T cells showed that the concentration of free zinc in cells is a major regulator of the maturation process in insulin-storing vesicles334. SVA has also contributed to discoveries reporting an increase in intracellular Ca2+ and/or protein C under certain stimulating and activating conditions in red blood cells. These conditions alter cell morphology and cause an increase in the release of microvesicles335. Table 7 summaries more examples of recent advances in research in (a) vesicle–cell communication and trafficking and (b) vesicle fusion and endocytic pathways. EVs are serve as information carriers for cell pathways and may trigger some diseases due to this activity. Neurodegenerative diseases like Alzheimer’s are caused by the misfolding and aggregation of tau protein. Polanco et al. have described a prion-like spread of tau protein seeds brought about by EVs, employing cryo-TEM336 and SRM (PALM and STORM)228. The contribution of EVs to trans-synaptic tau transmission has been confirmed using cryo-TEM in another study337. This single-vesicle methodology has also been utilised recently to study the near-native 3-D architecture of EVs secreted after infection with poliovirus. Cryo-TEM tomography has generated images of virions and viral structures contained in EVs before cell lysis338. Moreover, FRET microscopy has been used to track the triad protein VOR (VAP-A, ORP3, Rab7), paramount for the transfer of EV-derived components to the nucleus339. This type of research has therapeutic potential for diminishing the progression of neurodegenerative diseases (in the case of the VOR complex, by inhibiting EV-mediated intercellular communication). Cancer mechanisms have also been analysed using single-molecule techniques. Mannavola et al. have performed a ddPCR experiment using osteotropic melanoma cells and observed that EVs could induce the upregulation of genes such as CXCR7340. Thus, EVs may act as chemotaxis agents and, 46 hence, participate in the progression of cancer; however, additional research in this field is still required to achieve better understanding of how EVs contribute to cancer340. Table 7. Mechanistic studies of vesicles using SVA techniques focusing on vesicle–cell communication, trafficking and vesicle fusion and endocytic pathways. Technique Main conclusion Ref Vesicle–cell communication and trafficking research FRET microscopy Tau protein misfolding and monomer aggregates spread in a prion-like manner causing neurodegenerative diseases such as Alzheimer’s. Extracellular vesicles play a role in the transmission of pathological tau seeds. 336 The triad protein VOR is essential in the regulation of the endocytic nuclear transfer of EV-derived components. Targeting VOR might have therapeutic potential by inhibiting EV-mediated intracellular communication. 339 Fluorescence microscopy (SRM) In vivo visualisation of EVs in the zebrafish embryo. Uptake of EVs by endothelial cells and blood patrolling macrophages is shown. The study demonstrates tumour EVs activating macrophages and promoting metastasis. 218 Electron microscopy (cryo-TEM) Multiple virions and unique morphological components forming a mat-like structure are transported via infectious EVs of 100–1000 nm in diameter. These vesicle-containing enterovirus are secreted from host cells prior lysis. 338 ddPCR EVs from osteotropic melanoma cells induce chemotaxis and cancer progression via upregulation of CXCR7 of nonosteotropic melanoma cells. 340 Vesicle fusion and endocytic pathway research AFM Description of mechanical properties of erythrocyte-derived EVs according to their protein–lipid ratio. While a high ratio is associated with soft vesicles, a low ratio corresponds to stiff EVs. These mechanical differences are linked to several vesiculation and budding mechanisms. 341 Electron microscopy (cryo-TEM) Visualisation of the SNARE-mediated membrane fusion-by-hemifusion of small vesicles with cellsshows fusion intermediates where lipid monolayers partially mix en route to complete bilayer merger. 342 FRET microscopy Description of the molecular mechanism of SNAREs during different membrane fusion stages, (docking, hemifusion and full fusion) by tracking the lipid-mixing process at the single-vesicle level. 343 Fluorescence microscopy The fusion pathways are heterogeneous with an arrested hemifusion state predominating. The fusion of two lipid bilayers occurs spontaneously in a single step when they are brought into close proximity. 344 Fluorescence microscopy (TIRF) Heterogeneity of endocytic vesicle behaviours upon internalisation. Prior to scission, vesicles remain proximal to the plasma membrane for variable periods. Clathrin uncoating is also variable. 345 Calcium activates synaptotagmin-1, resulting in SNARE-mediated fusion of synthetic vesicles used as an exocytic model for synaptic events. An optimal distance for the fusion is 5 nm. 346 Alterations in membrane cholesterol content shift hemifusion intermediates to full fusion membrane and affect the stability of fusion pores. A large increase in cholesterol levels boosts individual SNARE-mediated fusion events. 347 Disassembly of the clathrin lattice surrounding coated vesicles is the obligatory last stage in their life cycle. The study visualises the recruitment of auxilin and Hsc70 which essential for this well-studied endocytic process. 348 hrFC (fluorescent conjugates) Characterisation of microvesicles from red blood cells under stimulation conditions. An increase in intracellular Ca2+ or protein kinase C levels leads to alterations in cell morphology and increased release of microvesicles. 335 FRET microscopy Free zinc concentration in insulin-storing vesicles is quantified using a novel FRET-based zinc sensor. The concentration of free zinc is important for insulin-storing vesicle maturity, hence it alters insulin trafficking. 334 47 Vesicle budding and shedding and the mechanical properties of the vesicles are poorly understood. Remarkably, a recent comparative review suggests that biomechanical analysis of single EVs provides key insights into their biological structure, biomarker functions, and potential therapeutic functions325. Sorkin et al. have used AFM to study these properties in erythrocyte and EV membranes under different conditions342. They have established that stiffness is inversely proportional to the protein–lipid ratio and linked it to several different budding mechanisms341. On the one hand, budding of protein-rich soft vesicles is possibly driven by protein aggregation, and on the other, budding of stiff vesicles with low membrane-protein content is likely to be driven by cytoskeleton-induced buckling341. A further investigation comparing EVs originating from healthy erythrocytes and from those with hereditary spherocytosis has supported these observations. It also uncovered mechanical and vesiculation differences between these two EV populations with potential use as diagnostic parameters349. Vesicle endocytic pathways have been investigated primarily using microscopy-based techniques. One major EV endocytosis pathway is mediated by the formation of clathrin-coated vesicles (Figure 10b). In this process, intracellular clathrin interacts with the membrane, producing a membrane invagination that will form an endosome through which the EVs can be internalised. The disassembly of the clathrin lattice surrounding coated endosomes is a mandatory last step in their life cycle. The recruitment of auxilin and Hsc70 (fluorescently labelled) was directly visualized using an inverted fluorescence microscope equipped with TIRF hardware and described as essential for clathrin-based internalisation events348. The clathrin-driven uncoating is a variable process in which the endocytosing clathrin-coated vesicles remain proximal to the membrane for different periods prior to the scission of plasma membrane. The dynamics of clathrin mediated endocytosis were assayed using fluorescently tagged proteins and TIRF microscopy345. Fusion states, dynamics and mechanisms of vesicle internalisation during single-vesicle fusion events have been directly examined using cryo-TEM, FRET and TIRF microscopy. Characteristics and kinetics of individual fusion events can be quantified for the lipids or DNA-lipid complexes involved in the process. Different fusion pathways exist: vesicles and cell membrane merge via a direct fusion of membranes (Figure 10a) or using protein-mediated mechanisms (Figure 10c). These mechanisms are involved in both endocytosis and exocytosis events. Examining individual giant unilamellar vesicles by fluorescence microscopy, it has been shown that during a direct fusion event, the hemifusion state predominates, and the fusion of two bilipid layers occurs in a single step when they are sufficiently close344. During the last decade, the fusion mechanism based upon SNARE-mediated internalisation pathways (Figure 10c) has been investigated employing SVA. At the molecular level, SNARE proteins mediate vesicle fusion with the target membrane and with membrane-bound compartments. Recently, Mattie et al. have visualised SNARE single-fusion events using cryo-TEM342. The sequence of fusion intermediates from lipid monolayers to a complete bilayer merge has been reported in homotypic vacuoles342. Hu et al. have correlated the membrane fusion stages with a molecular mechanism using reconstituted vesicles343. At a single-vesicle level, they have traced the lipid-mixing process using FRET microscopy and correlated it with the docking, hemifusion and full fusion stages343. They also report 48 that an optimal distance for a SNARE-mediated fusion is 5 nm. Interestingly, some regulators of fusion pathways were identified in related studies. Calcium acts as an activator of synaptotagmin-1, leading to the fusion of synaptic vesicles with the presynaptic membrane346. Stratton et al. have identified cholesterol as an important regulator of fusion dynamics, shifting the process from hemifusion intermediates towards full fusion membranes347. Large amounts of cholesterol precluster t-SNAREs, which serve as functional docking and fusion platforms. These clusters substantially affect the stability of pores by increasing the fraction of fully open pores and accelerating fusion events. Consequently, high cholesterol content triggers fast and individual SNARE-mediated fusion events347. Figure 10. Schematic representation of endocytic and fusion pathways recently investigated using SVA techniques. A lipid bilayer from a recipient cell is shown in dotted orange and EVs lipid bilayer in dotted red; the intracellular content is shaded in blue and vesicle content, in red. (a) direct fusion of EVs with the plasma membrane348. (b) Clathrin-based internalisation349,352. TIRF microscopy allows the examination of the clathrin uncoating process. In (c), SNARE-mediated membrane fusion is shown346. FRET microscopy facilitates the analysis of the three main stages of endocytic and fusion pathways: docking, hemifusion and full fusion. Membrane composition enhances this fusion pathway through t-SNARE-machinery recruitment and enrichment351. During the last decade, multimodal imaging platforms have been tested in vitro in a number of different cellular models of disease 350–352. These platforms have considerable potential to be used in vivo, for instance, in mice353–356. However, these systems usually fail to perform single-EV tracking and have been successful only in ex vivo cultures350. Nevertheless, there is an animal model worth mentioning due to its physiological characteristics and transparency. Zebrafish embryo has emerged recently as a prospective model for tracking EVs and assessing their dissemination and uptake in vivo357,358. In 2019, Hyenne et al. reported an approach for tracking individual circulating tumour EVs in the zebrafish embryo218 using confocal microscopy and a combination of chemical and genetically encoded probes to image EVs in vivo. The authors described, for the first time, the hemodynamic behaviour of tumour EVs and their intravascular arrest. The study shows that the endothelial cells and blood macrophages rapidly 49 take up circulating tumour EVs. These EVs activate blood patrolling macrophages and promote metastatic outgrowth218. A back-to-back study performed by Verweij et al. combined the genetic labelling (using a CD63-pHluorin exosomal reporter) of specific tissues with electron microscopy to track endogenous EVs in blood and further unravel their mechanisms of biogenesis, biodistribution, and target cells throughout the zebrafish embrio357. Intriguingly, Sung et al. have recently reported a CD63-pHluorinmScarlett fusion protein that can be used to image several stages of the exosome lifecycle in vitro359. This reporter likely can be used to visualize exosomes in vivo and is a prospective tool for understanding the physiological roles of exosomes. 1.3.3. EV biomarkers. SVA techniques hold the capacity to discover new, specific and effective biomarkers in EVs that have been missed by ensemble methods and can be used in disease diagnostics. While the resolution and sensitivity of SVA techniques still needs improving, they provide an accurate characterisation of EV subpopulations and assessment of biomarkers. Extensive research has been carried out seeking the biological biomarkers for several diseases such as prostate cancer60,224,308, fibromyalgia360, endometrial cancer361, colorectal cancer362,363 and liver-associated diseases364, among others. Figure 3 summarises recent advances in cancer research at the single-vesicle level. Currently, fluorescent assays, which are often used in cancer screening365, can detect miRNAs and correlate their presence with individual EVs and EV populations. Exosome-localised miRNA-21 might be used to differentiate between cancer patients and assess tumour progression and response to treatment365. Moreover, a specific lipid-protein signature may identify tumour-derived EVs as Raman spectra within the range of 2800 to 3100 cm-1 appear to be a distinguishing feature of genuine cancer EVs234. These findings open the way for early cancer detection. Nevertheless, the biomarkers discussed here are either generic (cannot discriminate between different cancers) or come from 2-D cancer model research and may not adequately diagnose clinical cancers. The development of next-generation sequencing SVA methods has optimised the identification of cancer biomarkers. Specifically, targeted sequencing using cancer gene panels has allowed the study of EV-derived and circulating free DNA300, resulting in the discovery of genetic biomarkers for several diseases. For instance, nine miRNAs have been profiled in serum EVs. Using these profiles, chronic hepatitis C patients can be distinguished from healthy individuals with accuracy above 95%366. Diagnostic opportunities presented by EV-specific genetic biomarkers have been widely reviewed367–370. It has been shown that cancer-derived EVs rewire and modify the pre-metastatic microenvironment, supporting tumour growth and metastasis during cancer progression67,71,74,295. This research describes a set of potential marker targets to be used as early diagnosis of PCa. The promotion and proliferation of PCa triggered by EVs produced as a result of DIAPH3 loss or growth factor stimulation have also been reported295. Other studies have shown that miRNA quantification in tissues can identify PCa by detecting the expression of RNU24371 or miR-130b372. Recently, several studies have reported specific and sensitive biomarkers for cancer detection227,251,296,298,373–375. ddPCR has detected and quantified the IDH1 transcript in cerebrospinal fluid- 50 derived EVs of patients with glioma tumours in the brain296. In lung cancer, eleven cancer-specific SERS signals have been obtained, allowing to differentiate between EV populations derived from healthy and lung cancer cells with high sensitivity251. Additionally, the CD147 protein found in EVs has been identified as a biomarker for colorectal cancer (CRC) diagnosis362. Further studies have identified BRAF and KRAS somatic mutations in plasma-derived EV populations from CRC patients227. Intriguingly, Melo et al. have described an explicit biomarker, glypican 1 (GPC1), found in EV populations from pancreatic cells containing different KRAS oncogenic isomers374. Hence, GPC1-EV identification could facilitate early pancreatic cancer diagnosis374. In 2020, the EV-transported HULC lncRNA (long non-coding RNA highly upregulated in liver cancer) was suggested as a chemotaxis agent for cell invasion and migration. This encapsulated HULC is a potential biomarker for human pancreatic adenocarcinoma diagnosis298. Figure 11. Schematic overview of reviewed biomarkers discovered using SVA techniques. The methods shown here are: (A) hrFC226,362,374, (B) RTM211,232,235,243,251, (C) TIRF361 and (D) ddPCR227,296,298,370. More information for specific biomarkers discovered using each method can be found in Table 8. The recent advances in cancer biomarker research are a proof-of-concept for non-invasive diagnostic tools based on EV fingerprinting in combination with multivariate statistical analysis212. The investigations in the field of non-invasive diagnostics for PCa have been stimulated by the discovery of intrinsic biomarkers detected in urinary-derived EVs. As a result of SVA research, several biomarkers have been associated with different stages of PCa. Biggs et al. have measured the levels of circulating prostate microparticles in plasma from PCa patients226 and used these microparticles in a liquid biopsy platform to identify and characterise patients. This study found that subjects with an advanced and aggressive tumor (in Gleason scale, scoring 8 or higher) can be identified independently of their PSA value226. The 51 lipid and surface protein signatures of prostate-derived EVs have been described using RTM210,232,235,243. Several characteristics have been highlighted, indicating potential PCa biomarkers. For example, a shift in the structure of surface proteins from alpha-helix-rich in prostate EVs to betasheet-rich proteins in PCa-specific EVs (isolated from blood samples) has been observed243. Moreover, Otto et al. have detected chemical signatures in the Raman and Rayleigh scattering data232 that efficiently differentiate between the EV populations from normal prostate and PCa cells211,234. The authors have trained a convolutional neural network to predict precisely the cellular origin of EVs for an automated diagnosis of PCa235. In another study, using various SVA techniques such as nanoscale flow cytometry and ddPCR, Joncas et al. have explored castration-resistant PCa370, identifying KLK3 and the androgen receptor variant 7 (AR-V7) as specific biomarkers for this cancer370. Table 8. Summary of recent EV biomarker-related discoveries achieved using SVA techniques, focused on cancer diagnostic biomarkers. Technique Main conclusion Ref Flow cytometry (immunofluorescence) Detection of CD147 as EV biomarker for CRC diagnosis. 362 Detection of GPC1 as EV biomarker for pancreatic cancer diagnosis. Cancer-cell EVs contain oncogenic KrasG12D. 374 Measures circulating prostate microparticles (PMPs) levels in plasma. This PCa "liquid biopsy" can identify patients with Gleason score ≥ 8, irrespective of their PSA. 226 Fluorescence microscopy (TIRF) Employs a fluorescent assay to detect exosome miRNA-21 (miR-21-EX) as a cancer-screening assay. miR-21-EX can be used to distinguish between cancer patients, tumour progression stages and treatment responses. 365 Raman (dry EVs) The surface protein signature shifts from alpha-helix-rich proteins to beta-sheet-rich proteins in prostate cancer-specific blood EVs. 243 RTM Chemical signatures from Raman spectra can be used to differentiate between EV populations derived from healthy and prostate cancer cells. A trained convolutional neural network can identify the cellular origin of EVs. 232,2 35 Raman analysis coupled with Rayleigh scattering distribution is used to detect specific lipid and protein signatures of tumour-derived EVs. 234 Label-free SERS Report of real-time and label-free diagnosis of lung cancer by detecting 11 cancerspecific SERS signals (proteins and lipids) that distinguish EV populations from healthy and lung cancer cells with high sensitivity. 251 ddPCR Detection and quantification of mutant and wild-type IDH1 RNA transcripts in EVs from cerebrospinal fluid of patients with glioma tumours. 296 Identifies somatic BRAF and KRAS mutations in plasma-derived EVs from CRC patients. Probes a wide range of cancer cells and the derived EV populations. 227 Identifies KLK3 and AR-V7 (androgen receptor variant) as biomarkers for castration-resistant prostate cancer (which progresses even under steroid deprivation therapy). 370 Extracellular vesicle-transported HULC promotes cell invasion and migration. This encapsulated HULC is identified as a biomarker for human pancreatic ductal adenocarcinoma (PDAC). 298 52 1.4. Future outlook Individually analysed EVs provide excellent prospects for future basic and practical research with a view to halt disease progression and control cell–to–cell communication processes. To exploit the full potential of SVA techniques, biological validation and reproducibility must meet the demands of clinical applications. Each technique has its specific advantages and disadvantages, and the exact choice of the method of analysis depends on the research question, the nature of the samples and EV characteristics. These techniques still need to improve their quantitative detection power, lower their cost and increase their reliability, resolution and throughput. In addition to technologies in use for SVA detection, we highlight several promising approaches that have yet to prove their potential in SVA. Conventional methodologies have the potential to be applied in SVA, and flow cytometry is a good example. Its implementation using innovative approaches can provide new features and capabilities, as shown by vesicle impact electrochemical cytometry (VIEC). This electrochemistry-based flow cytometry technique uses single ruptured vesicles whose content is detected and quantified based on Faraday’s law exploiting the produced oxidation current376,377. Extensive studies of the regulation of neurotransmitter trafficking by Ewing and colleagues, focusing on catecholamine exocytosis377–384 have demonstrated the prospective application of this approach in the EV field, highlighting electrochemical flow cytometry as prospective asset in the studies of EV functions and biology in the near future. The VIEC-based experiments have examined neurotransmitter content at a single-vesicle level in a pheochromocytoma cell line. Several studies have established that the neurotransmitter catecholamine is only partially released from the vesicles during an exocytosis event377. Moreover, catecholamine concentration is a key factor in regulating vesicle size since vesicular transmitter content is relatively constant and independent of vesicle size381. Nonetheless, many different agents may regulate exocytosis events. Using VIEC, the group has later verified that the zinc and cisplatin content serve as major regulators in these processes382,385. Video microscopy is another common technique that is potentially useful in EV analysis. In 2010, Zupanc et al. developed an efficient algorithm to transform video sequences into quantitative data383. Their work was a crucial step towards the creation of automated computer analysis and led to the development of another, more popular, methodology, NTA. In 2008, the-then emerging fluorescent ratiometric image analysis (FRIA) method was used to determine the post-endocytic fate and transport kinetics of internalised cargo384. FRIA presented a breakthrough in this field at the time and its application led to colocalization of EV cargo with organelle markers. However, the technique has not been very successful in further EV research possibly due to the emergence of other microscopy approaches such as TIRF or SRM that enable the study of EV internalization and fate with a better resolution in fluorescence images and a more straightforward analysis. The last decade has seen the development of several techniques with primary applications in analytical fields other than EV analysis. However, some, most notably radio-frequency analysis and SPIRIS, are applicable to single-vesicle research and could play an important role in scanning and evaluation of specific EV populations or characterising several EV parameters in a single experiment. 53 Radio-frequency analysis, also known as electrically controlled tuneable broadband interferometric dielectric spectroscopy, has only been presented in several conference papers386–389 after its first publication. In this method, specific sensors are used to perform a highly sensitive and tuneable broadband radio-frequency analysis. One study applying this method to EVs showed that the highly concentrated radio-frequency fields stimulates strong interactions between vesicles, which can be detected and quantified390. Specifically, the authors could detect and scan a type of EVs, giant unilamellar vesicles, at multiple frequency points and determine their molecular composition. In 2017, Wu et al. reported a separation method based on acoustofluidics and created a platform employing these acoustic trapping or tweezers phenomenon391. This technique isolates EVs from whole blood in a label-free and contactfree manner391. An acoustic wave falls upon a vesicle, and its scattering acts as a driving force to retain it. A year later, Ku et al. demonstrated isolation and enrichment of EVs using a similar acoustic trap technology392. In follow-up research, the method was used to isolate RNA and sequence miRNAs from EVs. Since then, an acoustic-based microfluidic platform has been released, coupling EV trapping technology with next-generation sequencing techniques. Together, these platforms form a robust and automated strategy for biomarker discovery in small sample volumes393. SP-IRIS, has so far found limited use in EV research, but has the potential to fill a unique experimental niche. SP-IRIS can characterise the size and phenotype (surface biomarkers) of EVs with no need to correlate two separate measurements214,394. This feature provides SP-IRIS with a high throughput and substantially reduces the amount of false negatives and positives compared to techniques that assess two characteristics in individual measurements. Due to the lateral resolution of microscopy (340–435 nm), highly concentrated samples cause signal overlap and a subsequent shift in the apparent vesicle size. Strikingly, despite the microscope resolution drawback individual Flaviviridae particles of approximately 40 nm have been identified and characterised using SP-IRIS269. Besides the examples presented in Table 9, there is a commercialised platform for EV phenotyping developed by Nanoview Biosciences using SP-IRIS, highlighting its potential application in characterising limited input EV samples. Several papers have proposed other highly-promising automated on-chip platforms using SP-IRIS for EV characterization214,269,395–397. These platforms have been combined with immunoblotting to sort and characterise EV populations from a sample and can detect size and phenotype at a single-particle level, visualising and quantifying either viruses or single EVs or both, in an uncharacterised sample. One of the greatest advantages of such microfluidic platforms is that they only need very small sample volume (~20 µL) for an effective analysis396,398. Other techniques different from SP-IRIS have been implemented in such devices to characterise EVs from EV-regulated diseases and examine their future use as diagnostic tools. EVs derived from transfusion-related acute lung injury (TRALI) have been investigated using SP-IRIS coupled with AFM mechanics. Obeid et al. used this approach to determine that certain types of EVs trigger neutrophil extracellular traps (NETs) and that these NETs are likely to mediate in TRALI399. Fluorescence microscopy coupled with on-chip nanoflow cytometry enables automated quantitative SVA of body fluid samples398. According to Yokota et al., the morphology and deformability of EVs from different cell lines can be investigated using nanopatterned tethering of EVs in combination with AFM400. 60 Urinary EVs (uEVs) were obtained by ultracentrifuging urine samples as described elsewhere59. In brief, approximately 50 to 100 mL of urine were collected of which aliquots of 50 mL were centrifuged at 2,000g for 10 min, filter sterilised (0.22 μm pore size) and immediately frozen at -80 ºC for further processing. When collection was higher than 50 mL, aliquots of either 1 mL or 10 mL were generated prior freezing. In these experiments, urine samples of 1 mL or 10 mL were thawed at room temperature and centrifuged for 5 min at 2,000g to remove any precipitate. Then, RNA associated to uEVs was isolated by using a Urine Exosome RNA Isolation Kit (Product #47200, Norgen Biotek Corporation, Canada). Protocol is extensively described elsewhere188. The isolated RNA was quantified in nucleic acid concentration using a NanoDropTM (ND-1000 Spectrophotometer, Thermo Fisher Scientific, MA, USA) device in order to optimize the retrotranscription to cDNA by means of SuperScriptTM VILOTM polymerase (#11754-050, InvitrogenTM, MA, USA). An approximately 100% cDNA conversion efficiency was considered to proceed with further assays. Real-time quantitative PCR methodology Real-time quantitative PCR (qPCR) was utilised to quantify the expression of selected target genes in uEV-derived RNA. RNA derived from PCa cell lines (PC-3, DU145, 22Rv1, LNCaP, BPH-1) and a universal human RNA template (QS0639, ThermoFisher Scientific, MA, USA)) were utilised to optimise qPCR methodology. To perform qPCR assays, SYBR® Green Select Master Mix was purchased from Applied Biosystems (#4472897, MA, USA). Reaction mix was prepared following manufacturer’s instructions and qPCR was run as specified in Table 10. Melting curves of each amplicon were determined to evaluate the efficiency and specificity of the amplification. A QuantStudio™ 5 Real-Time PCR Instrument (Thermo Fisher Scientific, MA, USA) was utilised to run the reaction and QuantStudio™ Real-Time PCR Software v1.3 for the analysis. Unless specified, all experiments were run with 1 ng of DNA per reaction. Design of primers for SYBR Green and TaqMan qPCR Primers for the selected target genes were designed following a protocol shared by ExosomeDx for clinical trials in urine. The specific procedure and parameters cannot be specified as it is their industrial property. Remarkably, the design was performed considering a future application in a clinical trial. This means the screening of potential biomarker targets was designed to work with two flanking primers when using SYBR Green polymerase – that requires less optimisation – with a TaqMan probe in case we could move forward to a clinical trial, which will require a higher sensitivity. The target genes tested in this chapter are: SRD5A2, CYP1A2, CYP3A5, CYP3A7, CYP3A4, SULT1A, STS, SULT2B1, SRD5A1, SRD5A3, AKR1C2 and CYP11B1. As housekeeping genes (HKG) GAPDH, ACTB, 18srRNA, SPEDF, EEA1F1, RPL6 and KLK3 were tested. Housekeeping genes analysis To ensure an adequate normalisation of quantitative RNA content, appropriate HKG are required. In general, cell lines and other well-established matrices already have HKG described in literature. Due to the heterogeneity of human samples, the establishment of a steadily expressed transcript is required. To perform such analysis there are many normalisation algorithms available405. As in every modelling strategy, when many algorithms are available the use of several approaches provides a more robust 61 and trustable outcome. For this reason, several reference genes were tested in these urine samples with NormFinder406 and BestKeeper407 algorithms. Both with their limitations and strengths found out the most robust HKG in PCa cell lines was GADPH combined with ACTB (data not shown). EE1AF1 is also a top candidate as a HKG but in some samples it does not amplify, therefore, should not be considered as a HKG a priori. They were also applied to urine samples (Figure 16). Table 10. Real-time qPCR conditions. These qPCR conditions were optimised for housekeeping genes and target genes in uEVs. # Temperature Time Temperature gradient Hold stage 1 50.0 ºC 120 s 1.6 ºC per s 2 95.0 ºC 120 s 1.6 ºC per s PCR stage (40 cycles) 1 95.0 ºC 15 s 1.6 ºC per s 2 60.0 ºC 60 s 1.6 ºC per s Melt Curve Stage 1 95.0 ºC 15 s 1.6 ºC per s 2 60.0 ºC 60 s 1.6 ºC per s 3 95.0 ºC 15 s 0.5 ºC per s 2.3. Results 2.3.1. Selection of targets. Study of steroid hormones mRNA related to EVs The luminous darkness of PCa works in mysterious ways in advanced stages, however, it is welldefined steroid hormones are crucial in early PCa progression. For this reason, gene transcripts related to this pathway associated to EVs were the objective in this section. Over the decades, mRNA encapsulated inside the EV lipid bilayer have been increasingly utilised as progression markers for several diseases. In fact, mRNA associated to EVs emerged as an outstanding source of information to characterise biological samples and hence, to describe specific signatures of their composition as biomarkers. A web-based database containing over 500 datasets related to EV studies – Vesiclepedia – was used to discover potential markers of early stage progression of PCa associated to EVs. Vesiclepedia compiles more 18,000 different proteins associated to EVs in more than 500 different studies (Retrieval: November, 2021). Approximately 8,000 proteins have been linked to prostate research with studies using either cell lines or urine samples. Interestingly, there are unique sets of proteins related to cell lines, which are also specific to different types of cells. In this unmanageable database, there is a set of 61 enzymes participating in the biosynthesis of steroid hormones (Extracted from KEGG) and it concerns this chapter. Figure 13 shows a total of 43 mRNAs associated to EVs considering only datasets uploaded to Vesiclepedia; this also means 18 enzyme-related protein or mRNA have not been ever described in EVs. Remarkably, 24 of these mRNAs have been linked to EVs but none was identified in prostate related research. There are 10 mRNAs – AKR1C3, AKR1C2, AKR1C1, 62 CYP17A1, SULT2B1, SULT1E1, HSD17B8, HSD17B12, DHRS11 and COMTthat are associated to both cell lineand urine-related research while 9 mRNAs - HSD17B2, HSD17B6, CYP1A1, UGT1A6, UGT1A9, UGT2B7, UGT1A8, HSD11B2 and CYP7B1 - are unique to studies performed with uEVs. Figure 13. Venn Diagram including the enzymes of steroid hormone biosynthesis detected in EVs and in different datasets in silico. Datasets included were all filtered to all enzymes related to steroid hormones biosynthesis. Moreover, it reports the names of the target genes we selected to test in this chapter. Biomarker potential of steroid hormone enzymes expression The next step was to evaluate which of the 61 hormone steroid-related genes could be relevant in categorising patient samples. The statistical appliance CANCERTOOL was utilised to report the expression of these genes of interest along groups of samples in different study cases. In the Y-axis a Log2normalized gene expression is represented and depending on the groups of samples compared it informs about the status of the gene: A. In prostate cancer: comparing non-tumoral vs PCa specimens. B. By progression: comparing non-tumoral, primary tumor and metastatic PCa specimens. C. By Gleason Grade: it is indicated as GS6, GS7, GS8, GS8+9, GS9 and GS10 and; it informs about the presence of a transcript in a specific Gleason architecture. D. Disease Free Survival: Kaplan-Meier curves representing the disease-free survival (DFS) of patient groups according to the expression of target genes. It indicates whether the gene prognoses PCa aggressiveness. The tool provides a report for each single gene in every definition above. These reports were checked manually to assess whether each specific transcript expression could define cancer progression. The discrimination between normal and PCa specimens (Status in PCa) and, including metastatic stages (Status by progression) were considered the most relevant aspects to aim for early stage biomarkers. 63 In Table 11, a summary of this analysis is reported. Only the transcripts able to significantly discriminate patients according to definitions A and B are shown. The plus sign (+) indicates a significancy close to 0.05 and double sign (++) close or over 0.01; unsure means that there missing results or that they are not conclusive and a minus sign (-) indicates no significance in discrimination. The stratification by GS was not robust and in many cases expression data was spread out due to a short number of samples per group. This categorisation is of importance because including very aggressive PCa samples to compute statistical differences may hinder relevant information for early stages of progression. Finally, DFS indicates aggressiveness of the cancer followed over time and considering death by cancer as a measure. This indicator is not much relevant to consider disease recurrences at this stage. Among the 32 candidates we have filtered using CANCERTOOL (Table 11), the first 12 most relevant entries in this classification were selected to be tested as mRNA biomarkers associated to uEVs. Figure 13 summarises the analysis and further indicates whether candidates have been already identified. According to vesiclepedia, 10 of them have not been ever detected in urine samples. There are two interesting candidates, AKR1C2 and SULT2B1, reported in EVs and urine samples that are also relevant as a patient classifier. Anyway, the 12 candidates will be assessed once the cohorts are designed. 2.3.2. Selection of cohorts. One of the most important hallmarks of PCa is its poor prognosis. Although the classification of PCa patients is becoming more specific, the tools or parameters utilised to categorise them are not backing up. The diagnostic tools are highly invasive, time consuming and require high expertise in examination. Nowadays, the gold standard biomarker before undergoing a biopsy is PSA concentration in blood. Its poor prognosis at early stages of development has already been discussed in this thesis (see in INTRODUCTION), demonstrating the existence of a research niche. Therefore, biomarkers with a superior capacity to stratify patient samples claim the spotlight of cancer diagnosis. Indeed, an analysis with a binary classifier model demonstrated PSA is not a good classifier of PCa patient groups in this urinary cohort of 646 patients (Figure 14). A receiver operating characteristic (ROC) curve illustrates the true positive fraction against the false positive fraction of a sample population at different threshold values. Thus, by incrementing the value of the classifier the successfulness in classification is assessed. The area under the curve (AUC) estimates how well the classifier discriminates between two groups of samples. In Figure 14, ROC curves and AUC values for each comparison between groups of patients are presented. To note, an AUC value close to 0.5 or lower indicates random-driven classification by the analysed parameter; the higher the value, the better the classifier is. From this data, one can conclude PSA is not a good classifier of patient groups in this cohort. Only when comparing rather advanced PCa (GG4 and GG5) to very low developed PCa (GG1) the ROC curves discriminate groups reasonably. This means PSA could be appropriate to classify low developed cancer and advanced cancer patients. In this regard, the lack of an appropriate classifier at early stages of development highlights the importance on patient stratification research. 64 Figure 14. ROC curves comparing all groups of patients in the sample cohort (n=646). The AUC are indicated to assess PSA performance as a biomarker. Table 11. Summary of CANCERTOOL results of significant genes of interest in discriminating PCa specimens. The columns indicate discrimination of status: A. in PCa B. by progression C. by Gleason grade D. Disease free survival. Reports of all genes are included in the Supplementary material. A plus sign (+) indicates significancy 65 close to 0.05 and double sign (++) close or over 0.01; unsure means that there are no conclusive and a minus sign (-) indicates no significance in discrimination. Gene Description A B C D SRD5A2 Steroid 5 alpha-reductase 2 ++ ++ + + CYP1A2 Cytochrome P450 family 1 subfamily A member 2 ++ ++ + - CYP3A5 Cytochrome P450 family 3 subfamily A member 5 ++ ++ + + CYP3A7 Cytochrome P450 family 3 subfamily A member 7 ++ + - - CYP3A4 Cytochrome P450 family 3 subfamily A member 4 ++ ++ - Unsure SULT1A1 aryl sulfotransferase ++ ++ - - STS Steroid Sulfatase + + + - SULT2B1 Sulfotransferase family 2B member 1 + + Unsure + SRD5A1 Steroid 5 alpha-reductase 1 + + - Unsure SRD5A3 Steroid 5 alpha-reductase 3 + Unsure - Unsure AKR1C2 Aldo-keto reductase family 1 member C2 + + - Unsure CYP11B1 Cytochrome P450 family 11 subfamily B member 1 + + - - AKR1D1 Aldo-keto reductase family 1 member D1 + - Unsure - CYP7B1 Cytochrome P450 family 7 subfamily B member 1 + + - Unsure HSD17B6 Hydroxysteroid 17-beta dehydrogenase 6 + ++ - - HSD17B7 Hydroxysteroid 17-beta dehydrogenase 7 + + - Unsure HSD17B8 Hydroxysteroid 17-beta dehydrogenase 8 + + - + DHRS11 Dehydrogenase/Reductase 11 + + - - CYP2E1 Cytochrome P450 family 2 subfamily E member 1 + Unsure - Unsure CYP1B1 Cytochrome P450 family 1 subfamily B member 1 + + - Unsure CYP19A1 Cytochrome P450 family 19 subfamily A member 1 + Unsure - + UGT2B11 UDP glucuronosyltransferase family 2 member B11 + Unsure + Unsure UGT2B28 UDP glucuronosyltransferase family 2 member B28 + Unsure Unsure - UGT2B10 UDP glucuronosyltransferase family 2 member B10 + - - Unsure UGT2B7 UDP glucuronosyltransferase family 2 member B7 + + - - UGT2B4 UDP glucuronosyltransferase family 2 member B4 + Unsure - - COMT Catechol-O-methyltransferase + + Unsure - LRTOMT Leucine-rich transmembrane and O-methyltransferase domain containing + + - - HSD11B1L Hydroxysteroid 11-beta dehydrogenase 1 like - + - - HSD11B1 Hydroxysteroid 11-beta dehydrogenase 1 Unsure + - Unsure SULT1E1 Sulfotransferase family 1E member 1 Unsure + + - CYP1A1 Cytochrome P450 family 1 subfamily A member 1 Unsure ++ Unsure Unsure As a consequence, small cohorts mimicking demographic parameters of the whole population were designed to test the proposed biomarkers. Two cohorts of 15 urine samples (of either 1 mL volume or 66 10 mL volume) with BPH, low aggressive and highly aggressive PCa samples were established. The objective was to find good HKG to be used in hypothetical clinical trials as well as assessing which specific biomarkers are associated to EVs and still hold potential to discriminate between patient groups. 2.3.3. Normalisation and Housekeeping genes. The genes 18srRNA, ACTB, EE1AF1, GAPDH, KLK3, RPL6 and SPDEF were examined as HKG to normalise urine cohorts in upcoming assays. Once primers were designed, a qPCR assay was run using RNA extractions from the 1 mL volume cohort. Two control RNA extractions from water and PBS were included. None of the HKG amplified using these samples in exception to 18srRNA that showed quite convincing amplification curves with amplification cycle (Ct) values around 34 cycles (data not shown in Figure 15). Yet, the HKG could still be used as its Ct values using sample RNA extraction range from 17 to 22 cycles, approximately. Figure 15 depicts the mean (± SD) of Ct values of each HKG in all RNA samples of the cohort. The qPCR assay determined SPDEF and RPL6 are not amplifiable in all samples and they exhibit a high variability in Ct values along samples. Hence, they may introduce more uncertainty rather than normalisation capacity. Figure 15. Representation of mean (± SD) amplification cycle (Ct) values of HKG testes in 1 mL volume cohort. It is expressed as an average of technical replicates (n=3) and the error bars correspond to standard deviation of these technical replicates. In this type of assays, Ct values are a quantification measurement. The bigger the Ct is, the lower the concentration of DNA/RNA. In general, Ct values of these HKG candidates are somewhat high (except for 18srRNA). To determine the best candidate, a comprehensive analysis of HKG expression along samples and between groups is required. The establishment of a gene or combination of genes whose expression is treatmentand group-independent is imperative to find a gene that only provides RNA/DNA content information. BestKeeper and NormFinder algorithms were selected to perform such evaluation. 67 Figure 16. Outcome of the HKG analysis by NormFinder algorithm. Stability values determine the likeliness of each gene to be an appropriate HKG in these set of samples. BestKeeper describes the general variability of samples for the analysed gene. This algorithm determined EE1AF1 and GAPDH showed the lowest variability. In addition, it described that the best combination of genes where those including ACTB and GAPDH. NormFinder results are reported in Figure 16. This algorithm considers sample variability but it also includes the standard deviation along samples and variation between groups to report a stability value. The lower the stability value is the better the HKG. NormFinder pointed out ACTB, EE1AF1 and GAPDH as the best HKG but also the combination of them. For this reason, we have selected ACTB, EE1AF1 and GAPDH as HKG to analyse these cohorts and use them for normalisation purposes upon RNA quantification. 2.3.4. Evaluation of mRNA in EVs as targets. The in silico approach issued 12 candidates of which only the expression of six could be reported in cell lines (Table 12). At this stage, a PCa cell line set was utilised to assess primers performance. Remarkably, SULT2B1 has been reported in PCa cell lines while vesiclepedia only describes it in urine studies. The use of cell line based templates has its own limitations since several genes may not be expressed in all of them. The use of a universal RNA derived from tissues could overcome this limitation and primer design can be validated reliably. Table 12. Summary of target hormone steroid genes tested in cell line-derived RNA. The table reports whether the target gene was detected in the two available templates. Universal RNA stands for a commercial template composed of a mixture of 10 different human cell lines (ThermoFisher Scientific, MA, USA). In house PCa cells are RNA extracted from PCa cell lines cultured in our lab. Gene target Universal RNA In house PCa cells Gene target Universal RNA In house PCa cells SRD5A2 Non detected Non detected STS Detected Detected CYP1A2 Non detected Non detected SULT2B1 Detected Detected* CYP3A5 Detected Detected SRD5A1 Non detected Non detected CYP3A7 Detected Detected* SRD5A3 Non detected Non detected CYP3A4 Non detected Non detected AKR1C2 Detected Non detected SULT1A1 Detected Detected CYP11B1 Non detected Non detected * Detection in only one type of PCa cell line Another limitation of the study are the low yields in RNA obtained in urine samples. Upon RNA extraction, 1 mL volume cohort yielded 0.388±0.301 μg of RNA while 10 mL volume cohort yielded 0.541±1.100 μg of RNA in total. These are, in general, low yields except for few 10 mL volume samples of which high concentration of RNA was obtained after isolation. The fact that 10 mL extraction were only 2-fold the yield of the 1 mL cohort already indicates a loss efficiency upon volume increase or a poor extraction specificity of EV-associated RNA from urine. 68 In Figure 17, the selected HKG and target genes that showed amplification in our templates are analysed. The Ct values of each gene of interest per sample are represented. It is observed HKG behave similar to normalisation analysis. Strikingly, RNA extractions from 10 mL volume samples measured lower RNA expression of normalising genes. This already hints that either the RNA extraction or the qPCR assay might have been inefficient. To assess whether a larger volume of initial sample may carry salts or other contaminants hampering the extraction or qPCR assay, a serial dilution (1:5 and 1:10 of the extracted RNA) and a higher amount of template (4 ng instead of 1 ng) were analysed by qPCR. Values of Ct varied according to serial dilution of template in each reaction: a higher amount of template improved qPCR outcome by a decrease of approximately 2 Ct and the dilution increased Ct values as expected. This suggests that there are no compounds impairing qPCR reaction and hence, issues may be due to RNA extraction or a low initial quantity of EV-associated RNA in samples. Perhaps a higher amount of extracted RNA would increase reliability in the assay or could improve the sensitivity of quantification. In fact, quantification sensitivity could be improved in future works by screening target genes directly using TaqMan quenching approach. Figure 17. Average Ct values (± SD) of each RNA sample extracted from urine in analysed genes. The top graph depicts the results from 1 mL volume samples and the bottom the results from 10 mL volume sample. Each dot corresponds to a single urinary RNA preparation analysed for one single gene. They are coloured as follows: in blue, GG5 patient RNA samples; in green, GG2 patient RNA samples; in red, BPH patient RNA samples. Focussing on the six tested genes with confirmed working primers only CYP3A5, SULT2A1 or SULT2B1 amplified in urine samples. They were measured only occasionally and in few samples. Also, Ct values were higher than 38 in general. Non-specific amplifications start occurring from this standard threshold. Indeed, a melting curve analysis confirmed that melting temperature of primers - a characteristic that indicates an adequate and specific annealing to the template – was incorrect, suggesting unspecific amplifications. 69 2.4. Discussion and Conclusions The exploration of new biomarkers in cancer research rose interest due to the limitations that suppose a poor diagnosis and prognosis of cancer disease. The appearance of novel biomarkers carriers such as EVs combined with cutting-edge methodologies as single-vesicle approaches can provide alternatives to discover disease markers. In this line, the bioinformatics analysis performed in this work aimed to describe potential candidates associated to EVs to discriminate between patient groups. Webbased databases compile several proteins or mRNA related to steroid hormones and associated to EVs. These EV datasets derive from hundreds of studies which include cell lines, tissues and body samples. Besides, EV-associated candidates were further filtered using CANCERTOOL to consider their relevance in discriminating cancer groups. To note, CANCERTOOL databases compile transcriptomic analyses of prostate primary samples. One could argue that specific mRNAs found in PCa primary samples would not necessary be present in EV samples of any kind. Up-to-date, concentration of PSA in blood is measured routinely to diagnose a patient with prostate issues. Its poor prognosis at early stages is well documented but also its poor patient stratification; this highlights the need to discover novel biomarkers. Interestingly, Exosome Dx released in 2020 a PSAindependent tool to diagnose PCa. By combining the gene expression of PCA3, ERG and SPDEF in uEVs they could predict the likelihood of high grade PCa occurrence408. The launch of such test demonstrates the potential using uEVs content as biomarkers. However, I also points out the challenge in establishing biomarkers for very early stages. According to SCOPUS, over 18,000 studies have been released describing biomarkers associated to PCa disease; however, only 35 consider EVs and steroid hormone enzymes. Joncas et al.370 is the unique work one could find describing an EV-associated mRNA that could discriminate plasma samples at different stages of disease progression using ddPCR. Moreover, section 1.3.3. EV biomarkers. of Chapter 1 demonstrates the potential of SVA techniques to optimise the identification of cancer biomarkers. Specifically, the use of ddPCR may enable the detection of transcripts that are present in a small subset of EVs in body fluids. This approach is able to distribute single EVs into individual droplets, which allows the identification of their genetic cargo of interest. In this chapter, the collaborative work with Exosome Dx is summarised. A clinical trial after validation of assays was projected, however, it turned out to not happen. Still one was able to: i. Design an in silico study from which markers present in EVs were determined. Moreover, one analysed which biomarkers hold a potential in early classification of patients. ii. Demonstrate PSA was not a good classifier in a large urine cohort. PSA could only be acceptable to discriminate GG1 patients to highly aggressive PCa patients. iii. Design a qPCR assay for the screening of EV-associated mRNA targets. This assay found out appropriate HKG and normalising genes and; could test some of the potential biomarkers. Sadly, no positive results were obtained but we included several suggestions in the chapter to optimise this approach in future work or studies of the same kind. 76 We describe a method for the detection of endogenous steroid hormones and their intermediates using liquid/liquid extraction and ultra-performance liquid chromatography (UPLC) coupled to high resolution time-of-flight mass spectrometry (hrLCMS). UPLC provides fast cycling times and a high chromatographic resolution. The high mass resolution obtained with time-of-flight mass spectrometry results in high specificity while sensitivities are on par with triple quadrupole methods. This method was applied to metabolically profile several animal tissues and urinary EVs (uEVs). Different biological matrices including prostate, adrenal gland, testicle, brain and liver of Wistar male rats but also human urinary samples were tested in this assay. To our knowledge, the present work presents for the first time a reliable and optimized hrLCMS assay to analyse key endogenous steroid hormones in endocrine tissue, bioliquids and EVs. 3.1.2. Materials and Methods 3.1.2.1. Tissue and biofluid samples Tissues and serum were obtained from three wild-type (Wistar, RjHan:WI) rats obtained from Janvier Labs, Le Genest-Saint-Isle, France. All urine samples were obtained from a healthy male on either the morning or the afternoon. uEVs were obtained by ultracentrifuging urine samples as described elsewhere60. Urine samples and uEVs were characterized in several physicochemical parameters and protein markers, respectively. Tissue and serum sample preparations from rats Tissues and serum were obtained from three wild-type (Wistar, RjHan:WI) male rats which are 17 weeks old. Rats were fed with sterile water and a standard diet ad libitum in a temperatureand light cycle-controlled animal facility following the Spanish Guide for the Care and Use of Laboratory Animals (RD 53/2013 - BOE-A-2013-1337). All rats underwent a liver perfusion procedure and were sacrificed by bleeding. Immediately after the perfusion, brain, prostate, testicles and adrenal gland tissues were removed and directly frozen at -80 ºC in dry ice for further use. Liver tissue was collected from animals in which liver perfusion could not be completed. Liver, prostate and testicle tissues were cut on dry ice and aliquoted into portions between 50 to 80 mg prior metabolite extraction. Adrenal glands were stored individually as each individual weighted approximately 80 mg. The exact weight of all tissue samples was calculated and utilised normalise the detected metabolites. Blood was collected during the perfusion procedure into a BD Microtainer® blood collection tube with separator gel (BD, Franklin Lakes, NJ). Then, it was centrifuged at 8,000g for 15 min and the serum fraction transferred into a clean Eppendorf® tube prior storage at -80 ºC for further processing. Human urine samples All urine samples were obtained from a healthy male on either the morning (first urine of the day after fasting) or the afternoon (after lunch, approximately at 4pm) to observe circadian variations intraindividual. Each sample group (morning and afternoon samples) has three replicates collected in independent days by spontaneous micturition. Approximately 80 mL of urine were collected of which 50 mL 77 were centrifuged at 2,000g for 10 min, filter sterilised (0.22 μm pore size) and immediately frozen at - 80 ºC for further processing. An aliquot of the original urine was also stored at -80ºC for further analysis. The six different urine samples were characterized regarding the following parameters: blood and ketone bodies presence in urine, glucose concentration, pH value and density (Table 19). Dip-and-read stripes were introduced into thawed urines samples and physicochemical parameters were measured with an OneStepTM Plus Stripe Urine Analyser (Henry Schein Inc., Melville, NY). EV isolation procedure Urine samples were thawed at room temperature and centrifuged for 5 min at 2,000g to remove any precipitate . Then, they were centrifuged at 10,000 g for 30 min to obtain a pellet (P10K fraction) containing EVs of bigger size than in the remaining supernatant. In a next step, the urine supernatant was ultracentrifuged at 100,000g for 90 min; the resulting pellet (Fraction P100K) contains EVs of smaller size than P10K. Fraction P10K and P100K were both washed in 50 mL of phosphate-saline buffer (PBS) and ultracentrifugated at 100,000g for 90 min. Afterwards, both fractions were resuspended in 50 μL of PBS and stored at -80 ºC together with the urine supernatant (SN100K) of P100K for further analysis. 3.1.2.2. Western blot Analysis An aliquot of 6 µL of each urinary EV (uEV) preparation was loaded and separated under nonreducing conditions in 4–12% Bis-Tris Protein gels (Invitrogen Inc., Waltham, MA). Western blotting was performed to determine the presence and relative amount of uEVs in each sample and fraction, for this reason only the approximately 10% of each uEV isolated fraction was utilized. In brief, the proteins were transferred to nitrocellulose membranes and then, they were blocked for 1 h (in 5% non-fat milk and 0.1% Tween-20 PBS solution). Then, the primary antibody was incubated overnight (approximately 16 h) at 4 ºC, washed and incubated for 1 h with a secondary HRP-conjugated antibody at room temperature. The primary antibodies used in this study were: MoαCD63 (clone H5C6) purchased from Developmental Studies Hybridoma Bank (Iowa, IA), MoαCD9 (clone 209306) from R&D Systems (Minneapolis, MN), RbαAQP2 (clone A7310) obtained from Sigma-Aldrich (St. Louis, MO), RbαCOX-IV (clone 3E11) from Cell Signaling Technology (Danvers, MA), MoαCD10 (clone F-4) from Santa Cruz Biotechnology Inc. (Dallas, TX), and finally, RbαAnnexin V (ab14196) and RbαLAMP2A (clone EPR4207(2)) were both purchased from Abcam (Cambridge, UK). Jackson ImmunoResearch, Inc. provided the mouse and rabbit HRP-conjugated secondary antibodies. A Clarity Western ECL kit from BioRad (Hercules, CA) was utilized for the chemiluminescence detection of bands. Either by scanning AmershamTM HyperfilmTM MP photosensible films (Cytiva, Uppsala, Sweden) or using a luminescent image analyser ImageQuantTM LAS 4000 (GE Healthcare, Chicago, IL), the proteins in the nitrocellulose membranes coming from different urine samples and fractions were identified. 78 3.1.2.3. Chemicals and standards DHEA, DHT, cortisol (in methanol solution) and the sodium salt of androsterone sulfate were obtained from Cerilliant Corporation (Round Rock, TX). Supelco (Bellefonte, PA) procured androstenedione. The sodium salts of DHEAS and pregnenolone were obtained from Avanti Polar Lipids, Inc. (Alabaster, AL). Testosterone, aldosterone, corticosterone, estrone, pregnenolone 3-sulfate (sodium salt form), leucine-enkefaline (Leu-Enk), chloroform (>99.8% pure; of chromatography grade) and ammonia solution were purchased from Sigma-Aldrich (St. Louis, MO). LC-MS grade water, acetonitrile, formic acid and methanol were purchased from Fisher chemical (Fair Lawn, NJ). 3.1.2.4. LC-MS sample preparation Steroid metabolites were extracted by liquid-liquid extraction using a methanol/water mixture and chloroform as extraction liquids. EV fractions were sonicated for 15 min in a total volume of 400 µL 50% v/v methanol/water mixture containing 1 mM ammonia to lysate EVs. The cell culture (DU145 cell line), fixed on culture well plates, was scrapped after 5 min incubation with 500 μL 50% v/v methanol/water mixture containing 1 mM ammonia. Tissue aliquots - approximately 50 mg - were lysed using 1.4 mm zirconium oxide beads into standard 2 mL homogenizer tubes (Precellys, Montigny, France). Each sample was homogenized in 500 μL 50% v/v methanol/water mixture containing 1 mM ammonia by performing two cycles of 40 s at 6,000 rpm in a FastPrep-24TM 5G bead beating grinder (MP Biomedicals, Solon, OH). After lysis, 400 μL of the homogenate – either tissue, EV fraction or DU145 cell culture - was transferred to a clean Eppendorf® tube. Subsequently, 400 μL of LC-MS grade chloroform was added on top of the 400 μL of any lysated sample and shaken for 60 min at 1,400 rpm at 4 ºC. Then, the samples were centrifuged for 30 min at 14,000 rpm at 4 ºC in order to precipitate proteins and to separate the organic from the aqueous phases. The aqueous (top) and organic (bottom) phases were separated. The protein fraction precipitated on the meniscus between these two immiscible phases. Then, 250 μL of each fraction was transferred to clean Eppendorf® tubes and evaporated using a centrifugal vacuum concentrator. The pellets from the organic fraction were dissolved in 100 μL pure methanol and the pellets from the aqueous fractions were dissolved in 50% v/v methanol/water. All resuspended pellets were centrifuged for 30 min at 13,000 rpm and 4 ºC. Finally, 80 μL of the resuspended pellets were transferred to deactivated glass vials or 96-wells plates for injection into the hrLCMS system. 3.1.2.5. Ultra-high performance liquid chromatography The chromatographic separation of the analytes was performed with an ACQUITY UPLC I-Class PLUS System (Waters Inc.). This system was equipped with a cooled (10 ºC) Process Sample Manager with a sample loop of 10 µL and a Sample Organizer, a Binary Solvent Manager and a High Temperature Column Heater. A reversed-phased 1.0 mm x 100 mm BEH C18 column (Waters Inc.), thermostated at 40ºC, was used for separating the analytes. Samples were injected from either 2 mL deactivated glass vials or 700 µL Round 96-well polypropylene plates. 79 Chromatographic behaviour was optimized with respect to peak intensity and an adequate separation of the 11 analytes along the run. Gradient elution was accomplished with an aqueous mobile phase (eluent A) consisting of 99.9% water with 0.1% formic acid and an organic mobile phase (eluent B) consisting of 99.9% acetonitrile with 0.1% formic acid. The flow rate was 140 µL per min. Several gradients were tested during the optimization process (Table 13) in order to avoid break-through (elution of analyte in the injection peak) and obtain good peak separation. The optimal gradient was as follows: start at 30% B, a linear increase to 80% B in 3.8 min., a step increase from 80% to 99%, constant at 99% for 1.0 min and back to 30% B in 0.2 minutes. The total cycle time from injection to injection was 6 minutes. The injection volume for all samples was 2 µL. Table 13. Summary of the six gradient utilised in optimization rounds to establish the best methodology with respect to peak intensity and analyte separation. Time (min) Gradient 1 Gradient 2 Gradient 3 Gradient 4 Gradient 5 Gradient 6 % A % B % A % B % A % B % A % B % A % B % A % B 0 95 5 90 10 85 15 80 20 80 20 70 30 3.8 - - 10 90 10 90 10 90 20 80 20 80 4 1 99 1 99 1 99 1 99 1 99 1 99 4.8 1 99 1 99 1 99 1 99 1 99 1 99 5 95 5 90 10 85 15 80 20 80 20 70 30 3.1.2.6. Mass spectrometry A Time-Of-Flight mass spectrometer SYNAPT G2-S (Waters Inc.) was utilized for the detection of analytes. The instrument was operated in either positive (ESI+) or negative (ESI-) electrospray ionization mode and in full-scan mode with a scan range between 50 Da and 1200 Da and scan time of 0.2 seconds. Table 14. Summary z-spray source parameters optimized for m/z 556.2771 in ESI+ and ESI-. Parameter ESI+ ESICapillary voltage 1.00 kV 2.00 kV Sampling cone voltage 25 40 Source Offset 80 80 Source temperature 120 ºC 120 ºC Desolvation temperature 450 ºC 450 ºC Desolvation gas flow 5.00 L/h 5.00 L/h Cone gas flow 1,000 L/h 600 L/h Nebuliser 6.00 bar 6.00 bar The z-spray source parameters: temperatures, gas flows, capillary position and voltages were tuned as detailed elsewhere. Optimal source parameters for this assay in either ESI+ or ESIare summarised in Table 14. Ion optics were fine-tuned by spraying Leu-Enk (100 ppb), at a rate of 10 µL per min, to a 80 resolution over 20,000 (FWHM) for m/z 556.2771. The same Leu-Enk solution was sprayed as a lock mass to correct for m/z fluctuations along the assay. The lock mass solution was introduced into the source every 90 seconds using a second ESI probe and it was recorded for 0.5 s. Mass spectrometer spectra was corrected according to fluctuations detected in the lock mass. 3.1.2.7. Statistical Analysis Analyte recovery study The extraction step efficiency was assessed by performing a recovery assay with various mixtures of organic solvents and water. Five different extraction buffers were tested in this assay: 25/75% v/v and 50/50% v/v of methanol/water mixture, 25/74.9/0.1% v/v/v and 50/49.9/0.1% v/v/v of methanol/water/formic acid mixture and 50/50% v/v of methanol/water mixture with 1mM ammonia. To compare and calculate the recoveries of 10 different analytes, a culture of a prostate cancer cell line - DU145 - was spiked with analyte standards. Each well containing 5·105 cells was spiked with a mix of standards at 2 µM before lysis (pre-spiked) and at the resuspension stage (post-spiked) with a standard mix at 10 µM. Thus, the pre-spikes contained 1 nmol in 500 µL and post-spikes (aqueous and organic fractions) contained the same total amount in 100 µL, which would be the theoretical maximum absolute if no loss during the extraction. In addition, for each extraction solution, non-spiked samples were prepared in order to correct for endogenous metabolites in the matrix. Samples for pre-spiked, post-spiked and nonspiked conditions and the five different extraction buffers were prepared in biological triplicates. Only the absolute peak areas were taken into consideration to establish the recovery efficiency in the extraction step. Average peak areas were obtained by mean smoothing the raw signals of triplicates. The recovery (R) was determined by dividing the corrected pre-spike average by the corrected postspike average and represented as a percentage (Eq. 1). Both pre-spiked and post-spiked raw signals ought to be corrected by subtracting the endogenous analytes signal in DU145 culture matrix (Snon-spike). However, as Snon-spike of DU145 culture matrix was less than 0.05% of the signal, endogenous correction was neglected during calculation. Importantly, pre-spikes were corrected with respect to analyte loss (α) during extraction procedure. Moreover, raw signals of each sample did not have to be corrected by the amount of initial samples because every well contained the same amount of cells. 𝑅 (%)=𝛼(𝑆𝑝𝑟𝑒−𝑠𝑝𝑖𝑘𝑒 − 𝑆𝑛𝑜𝑛−𝑠𝑝𝑖𝑘𝑒) 𝑆𝑝𝑜𝑠𝑡−𝑠𝑝𝑖𝑘𝑒 − 𝑆𝑛𝑜𝑛−𝑠𝑝𝑖𝑘𝑒 𝑥100 (1) Study of matrix effect in analyte quantification 𝑀𝐸 (%)=𝑆𝑝𝑜𝑠𝑡−𝑠𝑝𝑖𝑘𝑒 − 𝑆𝑛𝑜𝑛−𝑠𝑝𝑖𝑘𝑒 𝑆𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑𝑠 𝑥100 (2) In order to assess the matrix effect (ME) in the quantification of the analytes, the post-spiked raw signal was compared to an equivalent raw signal of a mixture of analytes (10 µM) in solution. Postspiked raw signals were corrected by subtracting the endogenous analytes detected in the non-spiked 81 DU145 culture samples. Then, the numerator was divided by the average peak areas of the standards and expressed as a percentage (Eq. 2). Analyte semi-quantification In this work, a calibration curve was prepared in solution with 50% v/v methanol/water for the semiquantification of analytes. This calibration curve consisted of a serially diluted mixture containing all the analytes starting at a concentration of 10 µM. The initial concentration was diluted to half concentration twice, resulting in 5 µM and 2.5 µM concentration is the curve. Then, this set of triplets was diluted in five decades; it resulted in the following 15 different concentrations per analyte: 10, 5, 2.5, 1, 0.5, 0.25, 0.1, 0.05, 0.025, 0.01, 0.005, 0.0025, 0.001, 0.0005 and 0.00025 µM. The calibration samples were injected at the beginning and at the end of each experiment; the average of these two points was used to semi-quantify of metabolites in tissues. The limit of detection (LOD) for each analyte was set to be the lowest concentration at which the signal-to-noise (S/N) ratio was above 3. The LOQ was defined as the lowest concentration at which S/N ratio was above 10. The highest quantifiable concentration is the highest concentration per analyte that fits the calibration curve with an acceptable accuracy and precision (CV ≤ 15%)413. In general, the data of a calibration curve ranges over several orders of magnitude, it is not linear and tends to be heteroscedastic426. For this reason, the relation between the peak area and sample concentration was determined by power-fitting427. Power fitting resulted in a calibration curve (Eq. 3) with α and b as the fitted parameters. Once the sample concentrations were calculated using a calibration method in solution, the amount (in nanomole) per gram of tissue weight was estimated. 𝑃𝑒𝑎𝑘 𝑎𝑟𝑒𝑎 = α[𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛]𝑏 (3) 3.1.3. Results 3.1.3.1. Liquid chromatography and mass spectrometry method We have compared six different chromatographic methodologies (Table 13) to separate the analytes satisfactorily. The gradient 6 (30% B to 80% B in 3.8 min; detailed steps in Table 13) showed the best peak separation along this run time compared to other tested gradients (data available in 430). Due to the nature of the stationary phase, analytes elute in order of increasing hydrophobicity. The resulting extracted ion current (XIC) chromatograms of a standard mixture at 10 µM are depicted in Figure 21. In brief, aldosterone (m/z 361.2015; ESI+) elutes at 0.99 min, cortisol (m/z 363.2171; ESI+) at 1.20 min, DHEAS (m/z 367.1579; ESI-) at 1.60 min, corticosterone (m/z 347.2222; ESI+) at 1.68 min, androsterone sulfate (m/z 369.1736; ESI-) at 1.85 min, pregnenolone sulfate (m/z 395.1892; ESI-) at 2.23 min, estrone (m/z 271.1698; ESI+) at 2.39 min, androstenedione (m/z 287.2011; ESI+) and DHEA (m/z 289.2168; ESI+) coelute at 2.40 min, DHT (m/z 291.2324; ESI+) at 2.65 min, pregnenolone (m/z 317.2481; ESI+) at 3.25 min. 82 Table 15. Summary of the five different extraction buffers tested in the recovery efficiency experiments. Extraction buffer % methanol % H2O pH modifier Buffer 1 50 50 None Buffer 2 50 49 1% formic acid Buffer 3 25 75 None Buffer 4 25 74 1% formic acid Buffer 5 50 49.975 1 mM ammonia The buffers differ in their methanol content and the compound utilised to tune the pH of the buffer (pH modifier). Regarding the mass spectrometry method, Leu-Enk signal (m/z 556.2771) was aimed at a resolution of over 20,000 (FWHM) and provided the necessary mass accuracy to evaluate assay analytes. Isotope pattern matching and the use of chemical standards confirming elution times further ensured specificity. In general, mass accuracies for the analytes in solution were between -1 to 1 mDa. Note-worthy, several analytes were not adequately separated during chromatographic elution. Corticosterone and DHEAS elute at similar retention times - 1.60 min and 1.68 min -, however, the MS could properly distinguish them by their m/z difference and their fragmentation pattern. Moreover, DHEAS was not detected with a high intensity signal in ESI+ mode. For this reason, corticosterone was measured in ESI+ and DHEAS in ESImode. Likewise, estrone, DHEA and androstenedione eluted in approximately 2.40 min. In this case, one could only rely in MS sensitivity (estrone m/z 271.1698, DHEA m/z 289.2168, androstenedione m/z 287.2011) and fragmentation pattern that was sensitive enough to distinguish and quantify them separately. 3.1.3.2. Analyte recovery optimization Afterwards, we evaluated the recovery of 11 analytes using a biphasic liquid-liquid method and analysing them with the optimized hrLCMS method. Extraction was performed using DU145 cell line as matrix. Five different mixtures of organic solvents and water, containing either formic acid or ammonia to modify the pH of the extraction buffer or no pH modifier were assessed (Table 15). Addition of formic acid strived for lowering the pH approximately to 3 while 1mM ammonia modified the extraction buffer to pH 8-9 in order to chemically neutralize functional groups of steroid compounds. From previous experiments in our metabolomics platform, we observed that in liquid-liquid extraction requires at least 25% organic solvent during the extraction step to precipitate the proteins. This is important to avoid clogging the chromatographic system427. Moreover, the effectivity of tissue homogenization using beads has been reported as high and does not differ much to the homogenization of other matrices such as urine or cell cultures427,428. Therefore, the calculated recoveries are ultimately dependent on the extraction buffer utilized regardless the homogenization methodology. During the optimization process, it was determined steroid sulfate compounds are recovered completely in the aqueous fraction whilst steroids without sulfate group are found in the organic fraction. Notably, only cortisol was detected systematically in both fractions (Figure 22); however, it was majorly recovered in the organic (80% or higher) rather than in the aqueous (approximately 20%) fraction. 83 Figure 21. Extracted ion current chromatograms of the analytes from a mixture at 10 uM concentration of each standard. In each chromatogram, the optimal ES (+ or -), m/z value and signal intensity of detection are indicated. They are ordered by decreasing retention time. From top to bottom pregnenolone, dihydrotestosterone, DHEA, androstenedione, estrone, pregnenolone sulphate, androsterone sulphate corticosterone, DHEAS, cortisol 84 and aldosterone. In green, the parameters of the analytes detected in ESmode; in dark orange, the parameters of the analytes detected in ES+ mode. Moreover, the addition of formic acid to the extraction buffer led to a dramatic decay of recoveries in sulfate compounds and a slight decrease in the rest of steroid analytes (Figure 22). One can infer that the presence of protons in the buffer do not stabilize steroid charges and severely hampers the extraction of sulfate steroids in a polar environment. Supplementation of 1mM ammonia outperformed the extraction in terms of recovery and robustness com-pared to the other extraction liquids. Notably, recovery values using different percentages of methanol in the ex-traction buffer do not differ much. However, the extraction efficiency of sulfate compounds using 25% v/v methanol underperforms 50% v/v methanol with a recovery loss of 40 to 50%. Figure 22. Recoveries (± standard deviation) of the selected panel of standard analytes are shown (n=6). For each analyte, the recoveries using different extraction buffers are depicted. In green, the results of an extraction using only a mixture with the solvent (50% methanol dark green and 25% methanol light green). In blue, the results with extraction buffers containing 1% of formic acid (50% methanol dark blue and 25% methanol light blue). Recoveries obtained with a 50% methanol (1mM NH3) extraction are depicted in red. a, analyte recoveries of the organic fraction. b, analyte recoveries of the aqueous fraction. In Table 16, the recoveries of 11 selected analytes using a mixture of 50/50% v/v methanol/water with 1mM ammonia as extraction buffer are reported. In general, the present methodology is able to recover and detect over 90% of the initially spiked analyte. Only DHT was detected in a lower percentage, approximately 80% of the initially spiked DHT was recovered. As expected in a biphasic extraction, hormone steroids were retrieved in an apolar environment and sulfated steroids in a polar solvent. Besides cortisol, pregnenolone sulfate was also reported in both fractions; it was mainly recovered in the more polar solvent and a derisory amount in the organic fraction. Using this methodology, the recoveries for 10 µM of analyte ranged from 74.2% to 126.9%. These values are acceptable for routine muti-analyte hrLCMS analysis since all the results are reproducible429. Thus, extraction using 50/50% v/v of methanol/water mixture with 1 mM ammonia was selected for further experiments in different biological matrices. 85 Furthermore, the performance of the optimized methodology was tested using urine as matrix since it has a high interest for clinical applications. Six urines from a male individual were pooled and aliquoted in different two volumes to assess matrix effect to recovery efficiency. Table 17, the recoveries of 10 analytes are reported; DHEA recovery has not been retrieved because its peak was masked by testosterone’s signal. In general, over 85% of the initially spiked analyte is recovered and detected in 50 µL urine matrix. Importantly, sulfated steroids are not recovered with the same efficiency; DHEAS and pregnenolone sulfate report a recovery efficiency of 75.7% and 54.9%, respectively. Recoveries of the analytes using 250 µL urine as matrix describes a slight decrease in non-sulfated steroids while the efficiency decay is dramatic in sulfated species. Table 16. Summary of the optimized method characteristics. The recoveries (± standard deviation) and matrix effect as signal loss (± standard deviation) of the extraction procedure in two different biological matrices (n=6; biological matrix: DU145 cell) are reported. In addition, LOD and LOQ values of the analytes in the adequate fraction are compiled. LOD: Limit of detection; LOQ: Limit of quantification. Analyte Fraction Recovery [%] Matrix effect [%] LOD [nM] LOQ [nM] Pregnenolone Organic 97.2 (± 1.9) 25.2 (± 3.1) 2.5 nM 10 nM Aqueous - 24.0 (± 2.8) DHEA Organic 122.7 (± 2.9) 37.7 (± 5.7) 5.0 nM 50 nM Aqueous - 28.0 (± 6.2) - - Androstenedione Organic 102.2 (± 3.2) 30.8 (± 4.6) 0.25 nM 0.5 nM Aqueous - 23.2 (± 4.5) Estrone Organic 103.7 (± 3.8) 25.5 (± 4.8) 5.0 nM 10 nM Aqueous - 25.7 (± 4.0) DHT Organic 74.2 (± 3.4) 23.1 (± 3.9) 0.25 nM 1.0 nM Aqueous - 23.4 (± 2.9) Cortisol Organic 114.3 (± 3.8) 25.9 (± 4.2) 0.5 nM 1.0 nM Aqueous 22.28 (± 4.5) 17.6 (± 4.7) Aldosterone Organic 99.8 (± 1.77) 18.7 (± 4.3) 0.5 nM 2.5 nM Aqueous - 17.7 (± 5.1) Corticosterone Organic 109.4 (± 3.1) 25.1 (± 3.6) 0.25 nM 1.0 nM Aqueous - 20.2 (± 3.2) Testosterone Organic 126.9 (± 1.7) 14.3 (± 1.9) 0.25 nM 0.25 nM Aqueous - 8.0 (± 2.1) Pregnenolone sulfate Organic 6.9 (± 2.7) 25.2 (± 3.1) 0.25 nM 1.0 nM Aqueous 94.8 (± 1.9) 24.0 (± 2.8) DHEAS Organic - 42.6 (± 1.1) 0.25 nM 0.5 nM Aqueous 108.0 (± 1.4) 42.5 (± 0.1) 3.1.3.3. Matrix effect It is well known that phospholipids and other lipids typically enriched in biological matrices such as tissues, body fluids or cell cultures can cause ion suppression in mass spectrometry, thereby hampering the analyte signal129,130. This phenomenon negatively influences the detection of analytes and may underestimate their quantification. For a specific matrix, the higher the ion suppression effect is the higher 92 assay is very convenient as all metabolites (except cortisol) are recovered in one fraction. This permits a faster measurement of steroid hormones in diverse biological matrices. Existing quantitation methods for steroid hormone compound have a wide span of LOQ, ranging from 0.002 to 10 ng per mL. However, it is highly dependent on the analysed matrix, i.e. a urine matrix shows a range from 0.002 to 0.2 ng per mL415,416, whilst cell matrices display a higher LOQ up to 10 ng per mL413. This suggests the matrix effect also depends on the specific matrix where the metabolites are contained. Comparing these studies, cell matrices report a lower sensitivity compared to urine; this is important when applying this method in future experiments or assays. In fact, this observation spotlights the major limitation of this study: The quantitation has been performed semiquantitavely. Ion suppression in mass spectrometry affects negatively the analyte signal, and subsequently underestimates its quantitation or it simply hampers its detection. Moreover, ion suppression may be limiting the detection of certain steroid compounds in several matrices, i.e. EV preparations. In consequence, this method should be utilised in matrices that facilitate steroids detection. A matrix-spiked calibration is usually the appropriate method to quantify absolute amounts of analytes in samples427. In this work, a calibration curve of the analyte standards was prepared in solution with 50% v/v methanol/water as solvent. Such approach cannot compute absolute amounts of the analytes in tissue since the matrix effect is not considered, however, a semi-quantitative approximation of the metabolites in tissues can be calculated. In this assay, the reported LOQ range lays between 0.50 and 50 nM (equivalent to 0.14 and 14.42 ng per mL) in solution, similarly to previous studies. However, it is advised to use matrix-spiked curves in further experiments using this assay. The time required to perform the chromatographic separation is typically long in literature; they report runtimes from over 10 min up to 45 min59,409,416–420,425. Only the work of Quanson et al.413 and Indapurkar et al.414 described a methodology with a short runtime (4 to 5 min); however, they tested and applied the method solely in cell matrices: PCa and induced pluripotent stem cell lines, respectively. Indapurkar et al.414 developed a methodology specific for estradiol-related metabolites and Quanson et al.413 measured androgens with ultra-performance convergence chromatography. In 2012, Maeda et al.418 accomplished the separation, detection and quantification of a panel of steroids in rat organs except in the liver but using an HPLC system. For this reason, sample preparation strategy demanded high volumes of extraction buffer – 15 mL of acetonitrile per sample – and required a total run time of 11 min. In this work, the volumes and run time are lower than 1 mL and the 10 min. In order to test the performance of our methodology, we have measured steroid hormone analytes from several rat tissues: adrenal glands, testis, prostate, liver and brain. The data shown in Table 18 is in accordance with the fact that the pathway is tissue-dependent in regular physiological conditions. Two metabolites upstream the pathway, pregnenolone and androstenedione, were quantified in adrenal glands but could not be quantified in prostate or brain. This hints that adrenal glands are in charge of metabolising cholesterol into steroid compounds in complex organisms such as rats; it is in line with previous findings in literature435–437. Likewise, adrenal glands are known to produce corticoid hormones. Our data confirms this since corticosterone is quantified in a higher amount – three to four orders of 93 magnitude – when compared to prostate, brain and testicles. Adrenal glands also seem to accumulate androgens (Table 18); however, the presence of active androgens (DHT) is 2-fold higher in prostate compared to other tissues. Importantly, the ratio DHT/testosterone, which are the active and non-active paired androgens, was approximately 11 in prostate while adrenal gland and testis were below 1. Because the presence of active androgen plays a physiological role in prostate, the ratio DHT/testosterone was also higher in this tissue. Since the first urinary metabolomics attempts to analyse urinary samples and other biofluids, several methodologies have been developed during the last years416–418. Nevertheless, none of the reported methodologies was optimal to assess steroids in EV sample preparations, tissues or body fluids in a fast and simple manner. Up to date, many studies have shown metabolomics in EVs60,417,438, but none of them has reported the detection of steroid hormones in a targeted approach. A plausible explanation is that the identification and detection of compounds similar in molecular mass – even the same one in some cases – hampers the allocation of mass signals with the corresponding chromatographic peak. For those steroids, i.e. DHEA and testosterone, which share empirical formula, the identification of each specific compound remains challenging using MS and it relies on chromatographic separation. Importantly, we have been able to quantify steroid hormones in urine samples and derived uEV in a fast and simple manner. However, only one DHEAS was detected in uEVS and cortisol, androstenedione and DHEAS were detected in urine samples. These EVs were isolated by ultracentrifugation including a washing step to avoid any contamination from the soluble fraction. Urine samples from a healthy man were collected in different days and in time collection (morning and afternoon). Time collection was a parameter to be assessed from a metabolomics perspective but we found out that interday variability had also a high impact in the analysis. Morning samples are considered to contain a higher concentration of steroids coming from the prostate, possibly due to accumulation and leakage towards the urinary tract at night. However, this trend was not described in our morning samples. The reason may be that urine sample U003 (Table 19) was not available for metabolomics analysis since the analysis of the soluble fractions of urine (after uEV isolation), which includes U003, morning samples had a higher concentration of DHEAS. This highlights the importance of analysing a larger cohort to obtain significant results non-dependent of a unique high concentrated sample. In the end, this is a fast and sensitive method that was successfully applied for the detection and quantification of a panel of steroid hormone compounds in biological samples in 6 min runtime per sample. The sensitivity of this method makes it ideally suited for multiple in vivo applications. In this manuscript, we explore the analysis of steroids in several rat tissues and also human urine and uEV samples. This has evident applications in profiling the metabolic status of patients suffering any hormone-dependent disease. To note, the assay requires a longer cleanse step to wash the column out of lipids and peptides when running a long experiment with many tissue samples. To our knowledge, this is the first hrLCMS-based method able to detect and quantify steroid hormones associated to EVs isolated from body fluids in a targeted approach. 94 3.2. Clinical Evaluation of Metabolic Signatures as Biomarkers of Prostate Cancer Progression in Patient Urines 3.2.1. Background The use of body fluids as a source of biomarkers have been intensified over last years. Currently, urine samples are one of the most important sources to identify PCa biomarkers since it is near the prostate and they can provide biological information. Documented studies have shown metabolites and metabolic signatures contained in urine samples can discriminate healthy and disease patients439–441. Moreover, a metabolomics study in uEVs described several altered metabolites of which few steroid hormones were heightened60. In general, these type of approaches aim to describe metabolic biomarkers in already distinct samples. This provides a valuable diagnosis information; however, they often present a poor prognosis, as they cannot predict the likeliness of a patient to develop the disease or how it will progress. In this section, a method for the detection and quantification of steroid hormones in biological samples has been developed and optimised as these hormones are the major drivers of progression at early stages. This approach will be used in upcoming chapter to describe the transference of metabolites via EVs. Nonetheless, another purpose of the method was to validate a steroid hormone metabolic signature as a biomarker for PCa. 3.2.2. Results and discussion 3.2.2.1. Clinical cohort characteristics In Chapter 2. Evaluation of Steroid Hormone Transcripts Associated to Urinary Extracellular Vesicles in Prostate Cancer Progression a cohort of samples obtained from Basurto’s Hospital has been presented. From that cohort, a clinically relevant cohort was designed to seek for specific signatures to define the status of disease in patients. In this section, it has been utilised to assess the hormone steroid metabolic signature of patients in urine. This cohort consists of 85 urine samples of 1 mL volume selected to cover the 5 groups of PCa patients (GG1, GG2, GG3, GG4 and GG5) and control BPH. Relevant information about the patients and samples is compiled in Table 21. To note, all urine samples were collected prior any biopsy hence, metabolic signatures of patients inform about their status at first visit to clinician. Diagnosis in terms of GS, pT, Pn and pN were reported after undergoing a second biopsy, a so-called surgical specimen biopsy, which was performed after 1 to 3 years. The staging mentioned in Table 21 is well explained in Prostate Cancer Disease section in the INTRODUCTION. Importantly, whether perineural invasion was observed clinically (Pn) or validated anatomopathologically (pN) is specified. 95 Table 21. Summary of clinical characteristics of urinary samples included in the study. Patient groups, pT and Pn stages were determined upon surgical specimen biopsy. 3.2.2.2. Signature result Metabolomics assays require an adequate performance in the specific biological matrices to be analysed. In terms of accuracy and sensitivity, the utilised matrix may influence on the outcome; however, the most important parameter to test are the efficiencies of analyte recovery. This parameter will inform whether the assay is working successfully. Similar to section 3.1. Development of a Targeted Metabolomics Assay in Endocrine Tissues of Male Rats and Human Samples, recoveries of the analytes in different volumes of urine were calculated (Figure 25). Non-sulphated metabolites were recovered in Parameter Patients (%, n=85) Median Age (IQR) 67(63,73) BMI (IQR) 26.6 (25.3,28.1) PSA value, ng/mL < 4 41 (48) 4-10 35 (41) > 10 9 (11) ADT treatment not available Prior prostatectomy Yes 84 (99) Unknown 1 (1) Family history PCa 6 (7) Other cancer 39 (46) None 33 (39) Unknown 7 (8) Patient group (GS) BPH 35 (41) GG1 (3+3) 10 (12) GG2 (3+4) 10 (12) GG3 (4+3) 10 (12) GG4 (8) 6 (7) GG5 (9-10) 14 (16) pT stage None 37 (43) 2 26 (31) 3 a 10 (12) b 12 (14) Pn stage Not reported 38 (44) Pn0 3 (4) Pn1 44 (52) pN stage Not reported 44 (52) pN0 10 (12) pNx 30 (35) pN1 1 (1) ADT, androgen deprivation therapy; GS, Gleason score; IQR, interquartile range; pT, PCa progression stage; Pn, Perineural invasion (Clinical); pN, positive lymph nodes (Anatomopathological) 96 a range between 70 to 110%, approximately. In contrast, sulphated metabolites - DHEAS - range between 45 to 75% approximately, exhibiting a decreased recovery values compared to previous section. In general, an increase in urine volume leads to a lower recovery of analyte; in sulphated metabolites they suffer a dramatic decay. This result indicates that urine matrix negatively affects the performance of the assay. In the case of polar analytes, the loss is rather relevant and it is something to consider in the decision making process upon designing an experiment. Anyway, recoveries are still adequate so we decided to proceed quantifying the cohort. Figure 25. Recoveries (± standard deviation) of a selected panel of standard analytes with urines as matrix are shown (n=6). For each analyte, the recoveries using a different volume of urine to undergo analyte extraction are depicted. Either starting with 250 µL or 50 μL of urines was tested. In blue, the analytes recovered in the aqueous organic fraction. In red, the analytes recovered in the aqueous organic fraction. Extraction with 50% methanol (1mM NH3) was utilised for urines. Thence, a total of 85 urine samples were analysed and quantified in steroid hormones content. The urine volume utilised in this approach was 250 μL since apolar hormones are present in low levels and one would like to maximise detection even compromising analyte efficiency of recovery. The analytes aldosterone, androstenedione, testosterone, cortisol, cholesterol, androsterone sulphate and DHEAS were successfully detected in this cohort. Prior analysis, analyte signal was normalised using median fold change442,443 to correct for metabolite abundance intrinsic of the urine sample. Hormone steroid levels in urines failed to discriminate between PCa and BPH patients (Figure 26). The score plot indicates a tendency to spread out but the two groups of samples are not separated. Further univariate analysis confirmed this result. Interestingly, few significant differences were measured when PCa samples were categorised according to their disease stage (Figure 27). Figure 27 depicts the four most relevant metabolites found back in urine samples. It shows that androsterone, DHEAS and cortisol levels are differentially quantified between BPH and some PCa groups but also, within PCa groups. 97 The major issue of these results is that no signature can determine whether a sample is diagnosed with PCa or BPH. Documented reports using EV preparations described DHEAS as a potential marker of PCa60; however, analysed urine samples do not support this hypothesis. The fact the metabolite DHEAS is readily soluble in biofluids and it is residual in EV samples derived from urines444 suggested that its evaluation in urine could have been more informative of the status of the disease. Nonetheless, the new data offers a better classification of PCa sample groups and shows few significant differences between them. A limitation of this subgrouping is the sample size because it is rather small at group level. Moreover, steroid hormone levels do not describe a classification pattern nor describe all the different subgroups. Notably, the classification of samples is very relevant in proceeding with these approaches, therefore differences in classification or building up cohort may hinder significant results. This makes the data quite difficult to interpret. Hence, predictive tools or validation models might be of interest to treat this data. Figure 26. Score plot corresponding to PCA analysis of PCa and BPH urine samples. This PCA model was generated with all the samples in order to differentiate both groups. In the axes, top two most representative features to distinguish samples groups are plotted. In purple triangles PCa samples and in green circle, BPH samples. Red arrows indicate the variables that explain variability. 98 Figure 27. Normalised quantification of the relevant analytes detected majorly in the cohort of urine samples. The BPH and PCa levels of androstenedione, androsterone sulphate, cortisol and DHEAS are reported as well as the significant differences between the groups. 3.2.2.3. Approaches to give an outcome As urine samples of this cohort were collected at first visit, metabolomics data represents the status of the patient at that time. This metabolomics data is of use if it can discriminate between patient groups. A logistic regression is a widely used binary classification model, which could provide a decision system by giving a matrix with data as an input. This means that a matrix with steroid hormones data, a multivariate logistic regression can be written to decide the output in a binary manner i.e. PCa or BPH (Eq. 4). However, the decision can also be triggered between PCa groups. In this approach, the idea was to build up a model, a logistic regression with diagnosis data. Then, challenge the model with a number of random samples to test whether it classifies the samples successfully. Because no metabolite was differentially present in PCa vs BPH, no valid model could be built. One could neither apply it successfully to classify PCa groups. F(x)= 𝛽0+ 𝛽1𝑥1+ 𝛽2𝑥2+ ⋯ + 𝛽𝑝𝑥𝑝 𝑤ℎ𝑒𝑟𝑒 𝑥 𝑎𝑟𝑒 𝑚𝑒𝑡𝑎𝑏𝑜𝑙𝑖𝑡𝑒𝑠 𝑝𝑎𝑟𝑡𝑖𝑐𝑖𝑝𝑎𝑡𝑖𝑛𝑔 𝑖𝑛 𝑡ℎ𝑒 𝑠𝑖𝑔𝑛𝑎𝑡𝑢𝑟𝑒 (4) Such binary approaches are not useful in assessing the classification of several groups using a metabolic signature. For this reason, other predicting model found in literature are suggested: Gaussian 99 Naïve Bayes, Random Forest, XGBoost, deep neural network or support vector machines. However, the development of a proper pipeline to evaluate a metabolic signature for clinical purposes lays beyond the scope of this thesis. Until diagnosis (via cylinder and/or surgical specimen) is completed, a couple of years could have elapsed hence, the data can also be considered longitudinal. The approaches already described do not consider longitudinal data and it could also be an interesting approach. One possibility would be the usage of a machine learning framework, which requires a longitudinal collection of data. This means the analysis of a urine sample is required in each time point to determine the variation of signature. The R package ‘MetabolicSurv’ identifies biomarker signatures by discovering predictive metabolites to classify the expected progress into groups. These approaches are interesting but do not quite fit the data gathered in this clinical cohort. One has to acknowledge his limitations and perhaps propose to build up a model with patient diagnosis outputs. In this manner, known metabolic signatures lead to specific trajectories of diagnosis over time. Hence, a predictive model can calculate probabilities of specific outcomes. Nevertheless, one is hypothesising and proposing here a possible follow up or an interesting research field for computing sciences. 3.2.3. Conclusions In this section, the metabolomics assay has been further described and evaluated for urine samples. It has been further utilised with clinical purposes to test whether previous results can be reproduced. One can conclude: i. Steroid hormones can be measured and quantified in urines from patients. The efficiencies of their recoveries are acceptable for this type of analysis. ii. Significant differences have been reported between PCa groups and, between BPH and PCa groups in cortisol, androstenedione and DHEAS levels. Unfortunately, no steroid hormone signature could distinguish between BPH and PCa urine samples. This suggests that a correct evaluation and classification of the sample is also required to find out biomarkers and compare studies. iii. No model could have been successfully built up with the provided data. One has suggested follow up opportunities; however, to continue with this type of research was beyond the scope of this thesis. 101 Chapter 4. Transference of Biological Components and Functionalities Driven by Extracellular Vesicles This chapter focusses in describing two approaches to assess the transfer of metabolites or functions to recipient cells by means of EVs. The assay described in Chapter 3.1 has been further developed to trace labelled metabolites in recipient cells and hence, determine whether EVs may participate in feeding metabolic pathways important in PCa early progression. This work has been compiled in a manuscript, which is under a submission procedure, and has not been accepted yet. Chapter 4.1 includes the last version of the manuscript. Furthermore, the transfer of oncogenic signalling driven by EVs has been explored. This approach consists in establishing a PCa model with a readout responsive to androgen signalling. Thence, oncogenic markers were measured upon treatment with EVs. This part has been included in Chapter 4.2. Bordanaba-Florit, G., van Liempd, S., Cabrera, D. et al. Labelled-cholesterol demonstrates effective EV-mediated metabolite transfer in a prostate cancer model. Under submission. 4.1. Labelled Cholesterol Demonstrates Effective EVMediated Metabolite Transfer in Prostate Cancer Lipids and specifically cholesterol play a pivotal role in prostate cancer progression. Cholesterol serves as signal for protumorigenic metabolic pathways because it is the main precursor of steroid hormone metabolism. Typically, the source of cholesterol is extracellular (apo)lipoparticles; however, extracellular vesicles could also provide cholesterol. The effective transfer of extracellular vesicle-associated metabolites to recipient cells has not been explored yet. In this work, prostate cancer cell cultures were treated with labelled cholesterol in lipid-depleted media. Then, purified extracellular vesicles were fed to non-labelled recipient cells to trace the internalisation of cholesterol by confocal microscopy and its further metabolization by targeted UPLC-MS assay. Cholesterol associated to extracellular vesicles was detected in unlabelled recipient cells. Upon treatment with an inhibitor of cholesterol trafficking, an accumulation of cholesterol was observed but the regular transport to mitochondrial compartment was not fully impaired. Moreover, we show stable isotope labelling in prostate cancer cells, which produce labelled extracellular vesicles. Upon treatment with these vesicles, stable isotope was detected in prostate recipient cells. In summary, we report an optimized and rapid UPLC-MS assay for detection of steroid-related metabolites from cells and extracellular vesicles. Furthermore, we demonstrated that a targeted metabolomics assay combined to a molecular biology approach is useful to study extracellular vesicle-mediated internalisation and the effective transfer of relevant metabolites as cholesterol. 108 TopF uptake. Notably, labelled steroid-related hormones were not detected after 24 hour incubation with D6-cholesterol in these prostate cancer cells. Figure 29. Cholesterol uptake by recipient LNCaP and hNAF cells when labelled cholesterol is supplemented in LPDS media. (A) Confocal images of both recipient cells upon treatment with 1 µM TopF (in the left) and the normalised fluorescence quantification of TopF signal to the basal signal detected in recipient cells not treated with TopF (in the right). In confocal panel, from left to right the columns show: DAPI staining of nuclei (in blue channel), TopF signal (in green channel) and the result of merging the two channels. (B) Relative quantification of cholesterol isotopes in LNCaP and hNAF recipient cells (n=6) upon treatment with cholesterol supplemented media, D6-cholesterol supplemented media and, D6-cholesterol and U18666A inhibitor supplemented media. Next, EVs were isolated and purified after incubating the producing LNCaP and hCAF cells with 1 µM BODIPY-cholesterol (TopF) or 70 µM D6-cholesterol (Scheme of production in Figure 31A). EV preparations were characterised and quantified to use them in following treatments. As an example, in Figure 30 we compiled the typical characterisation experiments done in any EV production. Figure 30B shows that EV fractions in both producing cells are the first four fractions of the SEC protocol used in this work according to typical EV markers. Also, in the first fractions it shows no presence of ApoB100, a marker of VLDL apolipoparticles. The bulk of soluble proteins are eluted in the latter fractions of the SEC purification (Figure 30A and B). Importantly, cholesterol content is more prominent in the first 109 fractions with an approximately 2to 3-fold higher concentration of cholesterol in LNCaP-derived EVs compared to hCAF EVs (Figure 30A). EV fractions were further pooled together and quantified by NTA prior treating recipient cells. NTA shows that particle populations contained in EV fractions are homogeneous since they exhibit a similar distribution of particle diameter (Mode of 157 nm in LNCaP and 177 nm in hCAF) with a unique peak (Figure 30C). The quantification was used in the following section to treat recipient cells with a similar number of EVs. Figure 30. Characterisation of cholesterol-containing EVs (EV preparations used in Figure 32) in our production approach and experiments. (A) Cholesterol content and protein quantification of the 13 fractions obtained by SEC separation of conditioned media in producing LNCaP (in blue) and hCAF (in purple) cells. (B) Ponceau and Western blotting results for CD9, CD63 and ApoB antibodies in the first 12 fractions obtained by SEC separation of conditioned media. (C) Nanoparticle tracking profile of the pooled first four fractions obtained in SEC (EV-enriched fractions). In blue, the modes in nm of particle diameter populations are indicated. 110 4.1.3.2. Extracellular vesicles can transfer cholesterol to recipient cells. Once producing cells were labelled and their released EVs purified, we obtained cholesterol-labelled (fluorescently or isotopically) EVs ready to treat recipient cells (Figure 31A). By UPLC-MS we were able to identify and quantify cholesterol and D6-cholesterol upon treatment with EV preparations. Figure 31. Transfer of isotope labelled cholesterol mediated by EVs using a UPLC-MS assay for detection. (A) Schematic overview of the production and purification of EVs from cholesterol-labelled LNCaP and hCAF cells that are used to treat recipient LNCaP and hNAF cells. (B) D6-cholesterol detected in recipient LNCaP cell as a percentage of total detected cholesterol upon treatment with: EVs produced in cholesterol supplemented media (grey), EVs produced in D6-cholesterol supplemented media (light green) or EVs produced in D6-cholesterol supplemented media and U18666A supplementation (dark green). EVs from producing LNCaP and hCAF cells were used in this experiment. (C) D6-cholesterol detected in recipient LNCaP and hNAF cells as a percentage of total detected cholesterol upon treatment with: EVs produced in cholesterol supplemented media (grey), EVs produced in D6-cholesterol supplemented media (light green) or EVs produced in D6-cholesterol supplemented media and U18666A supplementation(dark green). Only EVs produced by LNCaP cells were used in this experiment. In Figure 31BC, we show the results of two independent experiments in which recipient cells were treated with EVs carrying either cholesterol or D6-cholesterol. Recipient LNCaP cells incubated with similar number of EVs produced by cholesterol and D6-cholesterol treated hCAF or LNCaP retained D6-cholesterol associated to EVs (Figure 31B). It is observed that D6-cholesterol associated to EVs was 111 uptake by recipient hNAF and LNCaP cells in minute relative amounts, representing less than 0.1% of total cholesterol in recipient cells. Importantly, D6-cholesterol uptake was consistent in both experiments and specific to labelled EVs because there was no trace of the isotope in recipient cells treated with cholesterol-EVs. Figure 31B describes a higher uptake of D6-cholesterol by recipient LNCaP cells when treated with LNCaP-derived EVs rather than hCAF-derived EVs. Similarly, when treating recipient hNAF and LNCaP cells with approximately half the number of LNCaP-derived EVs (Figure 31C) compared to Figure 31B, hNAF showed a higher uptake of cholesterol. To note, the observed relative quantifications are very small so results may be highly variable from every other experiment. Ultimately, one can conclude D6-cholesterol associated to EVs was transferred to (non-labelled) recipient cells. 4.1.3.3. Cholesterol associated to extracellular vesicles is transported to mitochondrial compartment. Tracing labelled metabolites demonstrated the ability of EVs to transfer their associated cholesterol to the recipient cells LNCaP and hNAF. However, such a bulk ensemble approach cannot image the uptake of EVs to different cellular locations. To image the uptake of TopF associated to EVs by recipient LNCaP and hNAF cells, we used a dual tagging workflow to load EVs from producing cells with TopF and label them with the membrane dye MemBright 472. Then, unlabelled recipient cells underwent different treatments while fed with these labelled EV preparations. We first confirmed the uptake of cholesterol was associated to EVs by monitoring the cholesterol as TopF signal and EVs associated to MemBright signal (Figure 35). Recipient LNCaP treated with dual tagged EVs showed partial colocalisation of both signals (Figure 35A) while the treatment with free cholesterol followed by a MemBright staining of cells did not show a similar colocalisation phenotype (Figure 32B). Co-tracking of cholesterol and EV-associated MemBright in similar compartments suggests cholesterol signal is associated to EVs. Then, we treated recipient cells with the NPC1 inhibitor U18666A, which hampers intracellular trafficking of cholesterol in endosomal compartment473. The treatment did not affect recipient LNCaP uptake of cholesterol (Figure 32B) but it induced an accumulation phenotype of cholesterol not observed in absence of the inhibitor (Figure 32A). MitoTracker staining and colocalisation analysis confirmed that the intracellular transport of cholesterol was slightly impaired (Figure 32C) suggesting cholesterol from EVs is transported by the endocytic pathway to other cellular compartments. Importantly, cholesterol shows a high degree of colocalisation in mitochondrial compartment (Figure 32C), ranging from 50% to 70% approximately and depending on the image analysis approach, denoting a high presence of cholesterol in mitochondrial membranes. The fact that a well characterised cholesterol trafficking inhibitor as U18666A did not dramatically hampered cholesterol colocalisation to mitochondria (still up to 50%) indicates that there might be other pathways in recipient LNCaP overtaking the transport of cholesterol to make it available in mitochondria. 112 Identical experiments in recipient hNAF - prostatic stroma fibroblast – did not show a similar accumulation phenotype by treating with U18666A inhibitor (Figure 33A). Co-tracking of TopF and MitoTracker confirms that the transport towards mitochondrial compartment was not significantly impaired (Figure 33C). Nonetheless, the treatment affected cholesterol uptake from EVs (Figure 33B). Figure 32. EV-associated cholesterol is transported to mitochondrial compartment in LNCaP recipient cells. (A) Confocal images of LNCaP recipient cells upon treatment with LNCaP EVs (Mock), LNCaP-TopF EVs (BODIPY-cholesterol loaded EVs) and LNCaP-TopF EVs and U18866A inhibitor. All treatments were incubated with MitoTracker. From left to right the columns show: DAPI staining of nuclei (in blue channel), TopF signal (in green channel), Mitotracker signal (in red channel) and the result of merging the three channels. (B) Normalised fluorescence quantification of TopF signal to the basal signal detected in Mock. (C) Co-tracking quantification of TopF signal to MitoTracker signal using a corrected Mander’s coefficient and object-based colocalisation. 113 Figure 33. EV-associated cholesterol is transported to mitochondrial compartment in hNAF recipient cells. (A) Confocal images of hNAF recipient cells upon treatment with LNCaP EVs (Mock), LNCaP-TopF EVs (BODIPYcholesterol loaded EVs) and LNCaP-TopF EVs and U18866A inhibitor. All treatments were incubated with MitoTracker. From left to right the columns show: DAPI staining of nuclei (in blue channel), TopF signal (in green channel), Mitotracker signal (in red channel) and the result of merging the three channels. (B) Normalised fluorescence quantification of TopF signal to the basal signal detected in Mock. (C) Co-tracking quantification of TopF signal to MitoTracker signal using a corrected Mander’s coefficient and object-based colocalisation. In line with describing the transfer of metabolites associated to EVs from PCa models to other PCa models, we performed experiments with a similar layout but using producing LNCaP and hCAF cells. Both EV preparations were given as a treatment – with similar number of EVs – to recipient LNCaP cells. In Figure 34A one can observe the transfer of cholesterol associated to EVs, showing a similar phenotype even with different EV preparations. Moreover, co-tracking of TopF and MitoTracker shows no significant quantitative differences in the localisation of cholesterol in the mitochondrial compartment 114 (Figure 34C). Interestingly, the treatment with different EV preparations affected significantly the uptake of cholesterol by recipient LNCaP cells. Uptake variations are similar to the differences in cholesterol content described in Figure 29B, therefore, the fact that EV-associated fractions from LNCaP carry three times more cholesterol than hCAF EVs (Figure 30A.) may explain these uptake differences. This supports the idea that the transferred cholesterol is associated to EVs since uptake quantifications are dose dependent on EV preparations. Figure 34. EV-associated cholesterol from LNCaP and hCAF producing cells is transported to mitochondrial compartment in LNCaP recipient cells. (A) Confocal images of LNCaP recipient cells upon treatment with LPDS media (Mock), LNCaP-TopF EVs (BODIPY-cholesterol EVs) and hCAF-TopF EVs. All treatments were incubated with MitoTracker. From left to right the columns show: DAPI staining of nuclei (in blue channel), TopF signal (in green channel), Mitotracker signal (in red channel) and the result of merging the three channels. (B) Normalised fluorescence quantification of TopF signal to the basal signal detected in Mock. (C) Co-tracking quantification of TopF signal to MitoTracker signal using a corrected Mander’s coefficient and object-based colocalisation. 115 Figure 35. LNCaP recipient cells uptake TopF and MemBright signals associated to EVs. (A) Confocal images of LNCaP recipient cells upon treatment with TopF supplemented media and no MemBright, LNCaP-TopF EVs labelled with MemBright or LNCaP-TopF EVs labelled with MemBright and U18666A inhibitor. (B) Confocal images of LNCaP recipient cells treated with TopF and MemBright supplemented media and TopF, MemBright and U18666A inhibitor supplemented media. In confocal panels, from left to right the columns show: Mitotracker signal (in red channel), TopF signal (in green channel) and the result of merging the two channels. 116 4.1.4. Discussion The relevance of metabolomics approaches beyond the discovery of biomarkers and the metabolic profiling of tissues in different disease stages is in the spotlight. Many efforts have been done in describing metabolic differences between the epithelium and stroma in PCa but also between those tissues in different progression stages of the disease. Further to that, this work describes a metabolomics approach to study the transfer of metabolites in cell-to-cell communication events mediated by EVs. The detection of D6-cholesterol recipient cells treated with EV preparations containing labelled cholesterol demonstrated the transfer of cholesterol between physiologically relevant cells in prostate cancer disease. On one hand, fluorescently-labelled cholesterol showed its transport to mitochondria where it can be further metabolised into pregnenolone, the first precursor to produce other steroid hormones. This suggests that cholesterol carried by EVs can potentially be metabolised into other steroid hormones relevant to PCa progression. On the other hand, isotope labelled cholesterol associated to EVs was uptake by recipient hNAF and LNCaP cells. The fact only a small percentage of cholesterol was labelled – less than 0.1% of total cholesterol in recipient cells – indicates the treatment with EVs was not sufficient. For metabolomics, we require ten times more cells per treatment compared to fluorescence assays due to reproducibility and, the production of labelled EVs is limiting. Even so, D6-cholesterol was quantified consistently through experiments and its detection is specific to labelled EV preparations since it could not be detected upon treatment with unlabelled EV preparations. Besides the detection of labelled cholesterol in recipient cells, its usage to metabolise other steroid could not be demonstrated yet. In this study we showed the transferred cholesterol is related to extracellular material derived from labelled (producing) cell cultures. The results of co-tagging EVs with TopF and MemBright and, experiments using EV preparations with different cholesterol content confirmed this hypothesis. Furthermore, experiments inhibiting the transport of cholesterol via endocytic pathway strengthened the hypothesis labelled cholesterol came from an external source and it undergoes internalization towards lysosomes prior distribution to other compartments. By using the well-known endosomal trafficking inhibitor U18666A, we would have expected an accumulation of cholesterol in the endocytic compartment and therefore, a dramatic decay of cholesterol in the mitochondrial compartment. Even though colocalisation experiments showed a decrease of cholesterol uptake or a decrease of cholesterol in mitochondrial compartment and, the characterising intracellular accumulation in confocal images, cholesterol was still transported to mitochondria. U18666A impairs NPC1 activity, in charge of translocating cholesterol captured by NPC2 to vesicular membranes. Thus, this finding indicates that recipient cells may be using other pathways different to NPC complex to overtake cholesterol trafficking towards mitochondrial compartment. Mitochondria are known to not participate in vesicular trafficking as other organelles and, it is STARD3 the protein in charge of translocating cholesterol from vesicular to mitochondrial compartment upon generation of contact sites474,475. Therefore, other proteins than NPC1 and NPC2 might be trans- 117 locating cholesterol towards endocytic compartment membranes prior transport to mitochondria. Perhaps other protein transporters such as LIMP2 and LAMP2 can make cholesterol available to be transported to mitochondria46. Cholesterol is essentially transported and delivered by (apo)lipoparticles via bloodstream. Lipids (and cholesterol) are processed in the liver, the organ in charge of manufacturing these lipid-containing particles for further delivering to other tissues. Therefore, one would expect that only liver-derived models exhibit lipid reservoirs and majorly expel lipid-containing particles instead of EVs. Indeed, the main limitation of this study is the possible presence of lipid-containing particles in the EV preparations because EV isolation techniques are mainly based on size or precipitation parameters and the ones based on protein markers or other parameters are not well-standardised yet. For this reason, we purified EVs with SEC because it is cleaner than other techniques such as ultracentrifugation and the main contamination we could find is of VLDL because they have a similar size compared to EVs. Nonetheless, the source of LNCaP, hCAF and hNAF are human excised prostate tissues and they have not been reported to produce any lipoparticles. Also, apolipoproteins characteristic of VLDL particles – ApoB100 and ApoE (data not shown) - were not found in EV preparations. The fact we observe a degradation pattern of ApoB100 in soluble protein fractions by Western blotting indicates the cells are not using apolipoproteins to manufacture particles and are eliminating them instead. 4.1.5. Conclusions In conclusion, the UHPLC-MS assay developed in this work demonstrated to be suitable for tracing isotope labelled metabolites of the hormone steroids pathway. We have been able to detect D6-cholesterol in cell cultures and thence, to produce EVs carrying D6-cholesterol. Furthermore, EVs were able to transfer cholesterol to recipient cells. Although this transfer was confirmed, no production of other steroid hormones could be reported using LNCaP and hNAF as reference cell lines. In addition, treatment of recipient cells with EVs labelled with fluorescent cholesterol demonstrated the internalisation of cholesterol to cellular compartments. Also, the quantification of TopF was dose dependent upon cholesterol concentration in EV preparations and; the quantification of D6-cholesterol in recipient cells decreased with a treatment of a minor number of EVs per cell. This suggests that EV preparations are the source of D6-cholesterol detected in recipient cells. Finally, co-tracing cholesterol and MitoTracker showed EV-associated cholesterol is transported partially to mitochondrial compartment. 124 mRNA and it usually leads to a very high variability because of the loss in sensitivity. Actually, the error bars are not included in the graph because they are large and lay beyond the graph range, which makes meaningless any analysis Therefore, the assay would not confidently provide an output for the androgen depletion treatment. Oncogenicity markers Although KLK3 is described in literature as the major marker of oncogenicity in androgen-dependent progression476, other markers were also tested in LNCaP model to strengthen the evaluation of the oncogenic phenotype. The selected genes have been presented in documented reports as biomarkers in PCa diagnostic tests: SPDEF478, HOXC6477, DLX1477 and PCA3439. SPDEF is a transactivator of PSA promoter and is expressed in men with advanced PCa. HOXC6 is a homebox gene involved in organ development and cell proliferation. PCA3 is long non-coding RNA highly overexpressed in PCa cells. DLX1 may function as a transcriptional regulator of multiple TGF superfamily members. Figure 39. Fold change expression of target genes (mean ± SD; n=3) over time (in days) in recipient LNCaP cell line. Target gene expression is calculated as 2-ΔΔCt compared to expression at day 0. ACTB and GAPDH expression levels were considered for normalisation. In dark red, treatment with androgen-containing media and, in light red, androgen depleted media. Solid lines do not represent experimental data, they are added just as a guide for the eye. Using a similar approach, Figure 38 represents the fold change over time of 5 different potential oncogenic markers normalised with 3 HKG upon androgen signalling depletion. Target SPDEF, DLX1 and PCA3 gene expressions were not responsive to androgen depletion. One could expect this result because those transcripts were described in highly grade PCa, usually androgen independent. However, HOXC6 appeared as an adequate marker of oncogenicity when normalised with ACTB and 18srRNA expression. No error bars are shown until now because experiments were not yet performed in biological triplicates, only technical triplicates were included. 125 According to the results, LNCaP was a suitable model to follow androgen-dependent oncogenicity and, the transcripts KLK3 and HOXC6 the oncogenic markers responding to this hormone signalling. Thence, a fold change expression was analysed over time including appropriate biological replicates and the two best normalisers of previous experiments in terms of reproducibility and stability (Figure 39). This experiment confirmed KLK3 expression was dependent on androgen signalling, however; HOXC6 expression was rather variable and, although its expression over time displays a similar tendency as in Figure 38, the potential response to androgen signalling was not significant. Therefore, LNCaP was selected as the cell model to be treated with EVs and KLK3 expression the marker to follow oncogenicity over time. EV-mediated transfer of oncogenicity Figure 40. Summary of oncogenic markers over time (in days) under presence or absence of steroid hormones and upon treatment with EV preparations. A. Fold change expression of target KLK3 gene (mean ± SD; n=3) over time (in days) in recipient LNCaP cell line. Target gene expression is calculated as 2-ΔΔCt compared to expression at day 0. ACTB and GAPDH expression levels were considered for normalisation. B. Positive events of nuclear KI-67 (%) over rime (in days) in recipient LNCaP cell line. Positive events were counted using a citometer. The colours correspond to different treatments as indicated in the legend. Addition or depletion of androgens are depicted in dark or light red (+/- DHT); BPH-1 derived EVs (Ctrl EVs) in green, LNCaP derived EVs (AD EVs) in dark or light blue and, in yellow, PC-3 derived EVs (AI EVs). Once the model was evaluated, it was challenged with EV treatments in absence of androgens. EVs were isolated from three recognised PCa models that were selected according to their status and re- 126 sponse to androgen signalling. Control (Ctrl) stands for the non-tumoral cell line BPH-1, androgen dependent (AD) LNCaP line requires androgen signalling to sustain tumoral growth and androgen independent (AI) PC-3 uses non-androgen signalling to promote outgrowth. In order to generate EV samples, cells were cultured in complete media and LNCaP was also stimulated with 10 mM DHT. Protein titter of EV preparations was similar in all treatments and those were calculated as 3 producing cells providing EVs to 1 recipient cell. Expression of KLK3 (as fold change) was calculated using different normalisers over time (Figure 40). As in previous experiments, a higher fold change of KLK3 describes oncogenic growth of the cell culture. In Figure 40, the fold change of KLK3 using two different pairs of primers and two different normalisers at days 3, 6 and 9 is represented. Similar to previous sections, LNCaP responded with a decrease in oncogenic growth upon depletion of androgens and, when androgens were supplied, oncogenic markers increased their transcription levels. This confirms an appropriate performance of the model to further analyse EV treatments depicted in the same figure. In general, the impact of EV treatments in KLK3 expression is rather low; and in some cases, it is negligible. Even though, the treatment with AI-derived EVs tends to show a lower KLK3 expression compared to AD-derived EVs the differences are not quite relevant. Moreover, the expression of KLK3 increased in LNCaP cells growing in androgen-depleted media at treatment days 6 and 9. This finding suggests that treatments with EVs have no long-term effect in sustaining oncogenic growth. Undoubtedly, the layout of this experiment does not represent a reliable picture of a physiological context. EVs are continuously released by producing cells and so they are provided to recipient cells. Perhaps, a continuous supply of EVs would result in an oncogenic phenotype sustained over time. The limitation turns into technical because EV production is highly variable, and it requires time and a high volume of conditioned media to outline such approach. Nonetheless, Figure 40 depicted a relevant effect of EV treatments at day 3 (short-term). Although EV treatments cannot mimic results observed with androgen-supplemented media, the dramatic decrease in KLK3 expression is slightly neutralised. While androgen depletion decreases its expression by approximately 75%, EV treatments neutralise this decay to 50%. In other words, the fold change expression of KLK3 at day 3 is approximately 2to 2.5-fold the expression in androgen-depleted media. As already suggested, the supply of EVs over time is relevant to assess whether they can support the oncogenic growth observed in previous experiments with androgen ligands in the media. However, EV samples and their production is highly variable and so are EV treatments; hence, one would expect a high variability in the oncogenic marker expression caused by the treatments in recipient LNCaP cells. An EV migration assay, using Transwell® migration chambers of 400 nm pore size, is of interest to solve issues in EV production and to assess this functional transfer in a more physiological approach. 127 4.2.4. Conclusions In the previous section of the chapter, the transfer of metabolites contained in EVs has been evaluated. In the present section, the transfer of biologically relevant functionalities mediated by EVs is shown. The main purpose was to establish a functional interaction of EVs with recipient cells and, study the capability of EVs to sustain androgen-dependent oncogenicity. Since the actual composition of EVs driving oncogenic phenotype have not been evaluated, a mechanistic explanation is missing. However, one can conclude that: i. LNCaP is an appropriate model to study androgen-dependent signalling in PCa. Moreover, it was the only cell line with oncogenic markers responding to androgen treatments; according to the oncogenic markers tested, 22Rv1 and BPH-1 did not show androgen-dependency to sustain growth. ii. Not all the oncogenic reported genes were appropriate markers to inform about androgen dependent progression. Measuring the expression of the genes SPDEF, DXL1 and PCA3 could not inform about oncogenic growth of PCa cells. The expression of transcripts HOXC6 and KLK3 was altered upon androgen depletion, suggesting them as markers of oncogenicity. However, HOXC6 was rather variable and hence, only KLK3 was utilised in this work. iii. EV treatment provoked an effect regardless the type of producing cells from where they were isolated. At day 3, the fold change expression of KLK3 was approximately 2to 2.5-fold the expression in androgen-depleted media. iv. EVs showed a neutralisation in the loss of oncogenic phenotype. However, a one shot treatment with EVs could not sustain this oncogenic phenotype over time. 129 Chapter 5. Metabolic Alterations of Normal and CancerAssociated Fibroblasts from Human Stroma Samples This chapter focusses in studying the interaction between the two main compartments of prostate. The metabolic rewiring of stromal-derived fibroblasts from a normal region of the prostate compared to a region with the presence of a tumour was evaluated. It is well-known that epithelium interacts with stroma and viceversa. Documented reports show EVs are one of the players of this communication driving differentiation of fibroblasts to cancer-associted fibroblasts. In this chapter, we also discuss metabolic alterations in normal fibroblasts caused by PCa-derived EVs. The work has been published as an Original Article in Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease Journal and it is appended in the supplementary material. The original manuscript utilised in the publication is included and formatted in this section. Bordanaba-Florit, G., Royo, F., Albóniga, O.E. et al. Integration of proteomics and metabolomics reveals metabolic alterations of prostate cancer fibroblasts from patient’s stroma samples. Biochim. Biophys. Acta. Mol. Basis. Dis. Under revision Figure 41. Graphical abstract of Chapter 5. Metabolic Alterations of Normal and CancerAssociated Fibroblasts from Human Stroma Samples. Matched-needle biopsies from PCa patients were obtained and associated fibroblasts were obtained. Metabolomics analysis was performed to determine the distinct metabolic profile of normal and tumour fibroblasts. Seahorse analysis was utilised to determine any central metabolism changes related to TGF-β and EVs. 130 The prostate gland is a complex and heterogeneous organ composed of epithelium and stroma. Prostate cancer is most commonly seen emerging from luminal epithelial cells and requires the assistance of adjacent stroma. Prostatic stroma is compositionally complex, and in concert with disease progression undergoes many alterations which include the emergence of cancer associated fibroblasts (CAFs). This heterogeneous cell population often contains cells with a myofibroblast-like phenotype that are not normally present in healthy prostate tissue. In this work, we studied the metabolic rewiring of stromal fibroblasts following myofibroblast differentiation. First, the metabolic abundances of normaland cancer-associated fibroblasts derived from needle biopsies of the same patient was analysed using UPLC-MS. It was determined CAFs were metabolically more active and, therefore, energy producing metabolic pathways were enhanced. Also, CAFs showed a heightened lipogenic metabolism as both reservoir species and building block compounds. Interestingly, lipid metabolism affects mitochondria functioning yet the mechanisms of lipid-mediated functions are unclear. The fact oxidised fatty acids and glutathione system are elevated in CAFs strengthens the hypothesis that increased metabolic activity is related to mitochondria. In further experiments measuring the metabolic flux with a Seahorse bioanalyser, we studied whether TGF-β1 and extracellular vesicles (EVs) could stimulate myofibroblast differentiation in normal fibroblasts. An increase of basal respiration in normal fibroblasts was reported, mirroring the disease-like phenotype. This indicates an altered metabolism associated to mitochondria in CAFs and treated fibroblasts. Hence, one proposes that the change in the metabolomics profile of tumour-associated stromal fibroblasts is driven by oxygen-dependent metabolism, possibly associated to mitochondria; however, the specific mechanisms are still unclear. 5.1. Introduction The prostate gland is globally composed of epithelium and stroma, which are extremely heterogeneous tissues. The epithelium is organized as glandular acini and it contains cuboidal to columnar secretory epithelial cells with apical junction complexes, a continuous layer of basal cells10, and sparse neuroendocrine cells, each attached to a basal lamina9. Beyond the basal lamina, a prominent fibromuscular stroma composed of smooth muscle, fibroblasts, blood vessels, autonomic nerve fibres, inflammatory cells, and extracellular matrix components offers physical support and contraction of the gland479. In prostate carcinoma (PCa), the interstitial stroma is often abnormally rich in myofibroblastic cells467,468, capable of supporting tumour growth, vascularization, angiogenesis, and metastasis in vivo115. Transforming growth factor β1, TGF-β1, remains among the most critical factors for myofibroblastic differentiation and the generation of a tumour reactive stroma. Intriguingly, we have previously shown that cancer extracellular vesicles (EVs) can trigger fibroblast to myofibroblast differentiation in an EV-TGF-β1 dependent manner115. In further studies, the essential role of EVs in directing this stromal cell differentiation towards cancer-associated myofibroblast-like phenotype was described116. However, the mechanisms by which EV-activated stromal cells support tumour growth remain unclear. Our previous studies with biopsy material from patients where tumoral growth was located in one half of the prostate and not the other, revealed clear differences between normal and disease regions. Histological examination showed a smooth muscle stromal architecture around glandular structures in 131 normal tissue while the glands were disorganised and showed an altered, fibrosis-like interstitial stroma observed in disease-associated tissue480. A panel of antibody markers confirmed the higher abundance of α-Smooth Muscle Actin (α-SMA)-positive myofibroblast cells in tumour-associated tissue. Strikingly, EVs isolated from prostate cancer epithelial cell lines generated a myofibroblast-like phenotype in normal fibroblasts116. In that work, we demonstrated the essential role of EVs in directing stromal differentiation to a pro-tumorigenic phenotype, exhibiting pro-angiogenic properties and enhancement of tumour growth in xenograft models. Further functional assays and proteomics profiling work highlighted that stroma activation mediated by EV stimulation mirrors the naturally occurring fibroblast differentiation during disease480. Altered metabolism is a hallmark of PCa and several metabolites and metabolic pathways are already distinctive in different prostate types of tissue481,482. In this line, Andersen et al. reported higher levels of energy-related pathway metabolites, such as ADP, ATP, and glucose as well as higher levels of the antioxidant taurine in stromal tissue compared to cancer and non-cancer epithelium483. Besides, increased levels of crucial metabolites for fatty acid oxidation and building blocks in lipid synthesis were described in cancer tissue. Other studies also reported metabolic rewiring of reactive stroma, showing different levels of certain metabolites between highly differentiated stroma compared to poorly differentiated stroma484. The stromal-epithelial interactions have a dominant role in tumour growth, invasion and metastasis. Actually, many reports over the last decades showed the interaction of reactive stroma with PCa9,469,484–488 and EVs influencing aspects of cancer biology such as angiogenesis116,480,489,490 and tumour progression488,491–493. Yet, few studies have investigated the role of tumour EVs in altering metabolic processes in stromal cell compartments. With the rise of omics era, entire sets of biomolecules – genes, proteins or metabolites - contained in a biological tissue, cell, fluid, or organism can be identified. A proteomics analysis of the same matchednormal and disease stroma tissues was able to distinguish both phenotypes and describe a disease-like phenotype480. Similarly, a transcriptomic analysis of derived EVs from the same type of samples discriminated normal from disease stroma samples494. This manuscript presents a broad semitargeted metabolomics approach, which analyses the metabolome of normal and disease stroma tissues, highlighting the metabolic differences of matched normal fibroblasts and cancer-associated fibroblasts within individual PCa patients. Furthermore, we show an altered metabolism of normal fibroblasts treated with EVs from a PCa cell line. 5.2. Material and methods 5.2.1. Stromal primary cell cultures. Six patient-matched normal and tumour-associated needle biopsies were isolated from radical prostatectomy. These were taken from sites of palpable disease and also from apparently normal tissue from the opposite side of the same prostate. Tissue collection and consenting was managed through the Wales Cancer Bank. Cores were manually dissected into 1mm3 pieces and subjected to mechanical homogenization followed by 200 U per mL collagenase-I digestion for 15 to 20 hours at 37 °C. Cells 132 were cultured in Stromal Cell Basal Medium (SCBM) supplemented with human fibroblast growth factorB, insulin, fetal bovine serum (FBS) and GA-1000 (Lonza, Wokingham, UK). for around two weeks until only stromal cells were retained. 8,23Subsequent cultures were maintained in DMEM/F12 media (Lonza) with 10% FBS depleted of bovine EVs. Cultures were confirmed free of epithelial cells by immunofluorescence staining for cytokeratin prior seeding stromal cells in 96 well plates 480,494. A sample of each patient-matched prostatic stromal cell culture was collected and frozen for metabolomics analysis using liquid chromatography coupled to mass spectrometry. 5.2.2. Metabolite extraction. In metabolomics, there is no single platform or method able to analyse the entire metabolome of a biological sample. Therefore, metabolites were extracted by fractionating the cell samples into pools of species with similar physicochemical properties. In brief, proteins were precipitated by adding methanol to the cell lysate. Chloroform solvent was added to the methanol extraction mixture and this biphasic mixture was incubated at -20 ºC for 30 min. Then, three different fractions were collected: (1) fatty acyls, bile acids, steroids and lysoglycerophospholipids, were obtained after centrifuging the supernatant at 16,000g for 15 min, drying and reconstituting in methanol, (2) for aminoacids, aliquots of 5 μL from the first fraction were derivatised and dried, and (3) glycerolipids, cholesteryl esters, sphingolipids and glycerophospholipids were obtained by mixing the chloroform extraction mixture with H2O (pH 9) and incubating at -20 ºC for 60 min. After centrifuging at 16,000g for 15 min, the organic phase from this third fraction was collected then, dried and reconstituted in 50/50% v/v acetronitrile/isopropanol. The aqueous phase that contains polar metabolites, including central carbon metabolism, was collected, dried and reconstituted in H2O. Quality control (QC) sample for calibration and validation were included in this workflow to correct for response factors between and within batches; and to assess the quality of data. 5.2.3. LC-MS analysis. An appropriate UPLC-MS method was used for each platform. The instruments and the conditions for the chromatographic separation and mass spectrometric detection are summarized in Table 23. A test mixture of standards was analyzed before and after the entire set of randomized, duplicated sample injections to check for retention time stability, mass accuracy and sensitivity. All data were processed using the TargetLynx application manager for MassLynx 4.1 software (Waters Corp., Milford, USA). A set of predefined features, defined as retention time - mass-to-charge ratio pairs, Rt-m/z, corresponding to metabolites included in the analysis are fed into the program. Associated extracted ion chromatograms (mass tolerance window = 0.05 Da) are then peak-detected and noise-reduced in both the LC and MS domains such that only true metabolite related features are processed by the software. Then, a list of chromatographic peak areas is generated for each sample injection. 133 Table 23. UPLC-MS Analysis platforms for metabolomics analysis. Platform 1 Platform 2 Platform 3 Platform 4 Column type UPLC BEH C18, 1.0 x 100 mm, 1.7 μm UPLC BEH C18m 2.1 x 100 mm, 1.7 μm UPLC BEH C18, 1.0 x 100 mm, 1.7 μm UPLC HSS T3, 1.0 x 150 mm, 1.8 μm Flow rate 0.14 mL per min 0.40 mL per min 0.14 mL per min 0.10 mL per min Solvent A H2O + 0.05% Formic acid H2O + ACN + 10 mM Ammonium Formate 10 mM Ammonium Bicarbonate (pH = 8.8) 10 mM Tributylamine + 15 mM Acetic Acid + 2% Methanol (pH = 5.0) Solvent B ACN + 0.05% Formic acid ACN + Isopropanol + 10 mM Ammonium Formate ACN Methanol (%B), time 0%, 0 min 40%, 0 min 2%, 0 min 0%, 0 min (%B), time 50%, 2 min 100%, 10 min 8%, 6.5 min 4%, 1.5 min (%B), time 100%, 13 min 40%, 15 min 20%, 10 min 20%, 3 min (%B), time 0%, 18 min 40%, 17 min 30%, 11 min 25%, 8 min (%B), time - - 100%, 12 min 50%, 10 min (%B), time - - 2%, 14 min 45%, 15 min (%B), time - - - 100%, 16-20 min (%B), time - - - 0%, 21-25 min Column temperature 40 ºC 60 ºC 40 ºC 40 ºC Injection volume 2 μL 3 μL 2 μL 2 μL Source temperature 120 ºC 120 ºC 120 ºC 120 ºC Nebulisation N2 flow 600 L per hour 1000 L per hour 600 L per hour 600 L per hour Nebulisation N2 temperature 350 ºC 500 ºC 350 ºC 300 ºC Cone N2 flow 30 L per hour 30 L per hour 10 L per hour 50 L per hour Capillary voltage 2.8 kV 3.2 kV 3.2 kV 2.8 kV Cone voltage 50 V 30 V 30 V 100 5.2.4. Data analysis. Data normalisation and quality control After data inspection in terms of reproducibility and peak integration, each metabolite was corrected and normalized using the intensity of an appropriate internal standard included in the analysis and following the procedure fully described by van der Kloet et al.495. Finally, any remaining zero values in the corrected dataset were replaced with missing values prior averaging to obtain a dataset further used for statistical analyses. A final normalization procedure was applied by dividing every sample by its protein content. Multivariate and univariate analysis Once data was normalized and prepare for statistical analysis, a first approach based on multivariate analysis was performed with SIMCA-P (version 13.0). Firstly, a non-supervised principal component analysis (PCA) was utilised to reduce dimensionality and to study data quality, assess reproducibility of Metabolites 2022,12, 714 2 of 17 5α dihydrotestosterone (DHT), are the major ligands in this molecular pathway and cause the progression of PCa at early stages [4,6]. Figure 1. Schematic representation of the steroid hormone biosynthesis pathway in relevant organs and its regulation. CRH stimulates the release of ACTH from the pituitary gland. ACTH stimulates the production of cortisol (exerts negative feedback on CRH and ACTH) and DHEAS in adrenal glands. Pulses of GnRH from hypothalamic neurons stimulate pulses of LH as well as FSH. LH stimulates testosterone production in testis. Liver maintains pathway’s homeostasis and several processes may happen: sulf desulfation makes metabolites available to feed the pathway while processes indicated with a flat end arrow inactivate metabolites that are in circulation. Bold arrows indicate a higher activity of the specific reaction. In bold, the metabolites that are majorly produced in each specific organ are represented. ACTH: adrenocorticotropin; CRH: corticotropin-releasing hormone; FSH: follicle stimulating hormone; GnRH: gonadotropin-releasing hormone; LH: luteinizing hormone; CYP17A1: Steroid 17-alpha-monooxygenase; CYP19A1: aromatase; SULT: hydroxysteroid sulfotransferase; STS: steroid sulfatase; 3 β -HSD: 3 β -Hydroxysteroid dehydrogenase; 17 β -HSD: 17 β -Hydroxysteroid dehydrogenase; DHEA: dehydroepiandrosterone; DHEAS: DHEA sulfate. Metabolites 2022,12, 714 3 of 17 In mammals, the precursor of sterol biosynthesis is cholesterol, which is further utilized in the adrenal glands, gonads and sexual-derived tissues to produce steroid hormones. There are 99 metabolites involved in the steroid hormone biosynthesis pathway and over 100 reactions are catalyzed by 61 different enzymes [ 7 , 8 ]. All of the steroid compounds share a sterane backbone structure. The physiological role of each individual steroid hormone is primarily defined by the layout of double bonds, hydroxyl and keto groups around this basic sterane backbone structure [ 1 ]. The main structural difference between the classes is the carbon atom arrangement i.e., the androgens are C-19, the estrogens are C-18, the progestogens are C-20 and the corticoids are C-21. In the first step of the steroid hormone biosynthesis, cholesterol is internalized into the mitochondria where it is fed as a substrate to produce pregnenolone (Figure S1, Supplementary Materials). This is the main precursor for steroid hormones produced de novo [ 4 ] inside the mitochondria. Pregnenolone can be converted to progesterone or dehydroepiandrosterone (DHEA), which can be further metabolized to glucocorticoids and mineralocorticoids (C-21) or to androgens (C-19), such as testosterone, DHT or androsterone and estrogens (C-18), respectively (Figure S1, Supplementary Materials). Interestingly, this metabolic network is tissue-dependent. Different organs are specialized on particular modules of the pathway that are physiologically relevant to perform their function. For instance, the adrenal glands are the producers of C-21 hormones, while prostate shows a high SRD5A activity, which catalyzes the conversion of testosterone to DHT (Figure 1). Indeed, this is an intricate network of metabolites. Many of these metabolites participate as ligands in a wide span of signaling cascades and biological processes, and their levels vary strongly between different biological compartments. While cholesterol is the unique de novo precursor in steroid hormone biosynthesis, there exists an interchange between cells and tissues that anaplerotically feeds the pathway at the intermediate steps [ 9 ]. This means that the compounds upstream of the pathway can be provided by the cell environment. In this line, sulfated steroids are of interest since they are, unlike their unsulfated counterparts, readily soluble in the cytoplasm and in biofluids, such as blood or urine. Notably, the sulfates of steroids are considered endogenous and active neurosteroids [ 9 , 10 ]. Over the past few decades, it has been established that sulfonation is not only a process to inactivate and excrete steroid hormones; it also acts as a systemic reservoir for peripheral or local steroidogenesis in non-steroidogenic tissues, i.e., the brain or prostate [ 9 , 11 ]. In addition, it has been reported that the secreted vesicles, also known as extracellular vesicles (EVs), participate in many of the physiological processes [ 12 , 13 ] and they can contain a wide variety of cargos, such as lipids, proteins, metabolites, sugars and even DNA [ 12 – 15 ]. The hormone steroids and related cargos are transported by the blood and other body fluids as sulfated species, but they could also be transported by EVs to reach the target tissues. The steroid hormone metabolism and the consequences of dysregulation have gained interest within the biomedical community to understand and diagnose hormone-dependent diseases, rather than the historic usage of steroid hormones in therapeutics. Indeed, a number of methods to detect and quantify steroid hormones have been reported during the last two decades. Many of the studies describe methodologies to detect steroids from several biological sources: cell cultures [ 3 , 16 , 17 ]; urine samples [ 18 – 20 ]; animal tissues [ 21 – 23 ]; human serum [ 24 – 26 ]; human hair [ 27 ] and waste water [ 28 , 29 ]. In general, steroid metabolomics methodologies focus on profiling a specific set of metabolites of interest in targeted tissues (or in circulation) rather than analyzing steroidogenesis status in a system of organs and related fluids. The methods are usually developed for similar non-sulfated steroids that efficiently ionize in the same mode, avoiding the exploration of the detection and quantification of many different steroids simultaneously [ 16 , 23 , 25 , 26 ]. Methodologically, these studies describe a variety of extraction, separation and detection methods. In particular, the solid phase extraction (SPE) and reversed phase liquid chromatographic-based methods are deployed in the isolation and separation of these compounds. The detection is mostly performed with triple quadrupole instruments. In addition, gas chromatography-coupled Metabolites 2022,12, 714 4 of 17 MS methods was also utilized in a few of the studies. All of these methods have their advantages and disadvantages. We describe a method for the detection of endogenous steroid hormones and their intermediates, using liquid/liquid extraction and ultra-performance liquid chromatography (UPLC), coupled with high resolution time-of-flight mass spectrometry (hrLCMS). UPLC provides fast cycling times and a high chromatographic resolution. The high mass resolution obtained with time-of-flight mass spectrometry results in high specificity, while the sensitivities are on par with triple quadrupole methods. This method was applied to metabolically profile several animal tissues and urinary EVs (uEVs). Different biological matrices, including prostate, adrenal gland, testicles, brain and liver of Wistar male rats but also human urinary samples, were tested in this assay. To our knowledge, the present work presents for the first time a reliable and optimized hrLCMS assay to analyze the key endogenous steroid hormones in endocrine tissue, bioliquids and EVs. 2. Materials and Methods 2.1. Tissue and Biofluid Samples The tissues and serum were obtained from three wild-type (Wistar, RjHan:WI) rats obtained from Janvier Labs, Le Genest-Saint-Isle, France. All of the urine samples were obtained from a healthy male on either the morning or the afternoon. uEVs were obtained by ultracentrifuging urine samples as described elsewhere [ 5 ]. Urine samples and uEVs were characterized in several physicochemical parameters and protein markers, respectively. For a more detailed information on sample collection, preparation and characterization refer to Figure S1 (Supplementary Materials). 2.2. Chemicals and Standards The DHEA, DHT, cortisol (in methanol solution) and the sodium salt of androsterone sulfate were obtained from Cerilliant Corporation (Round Rock, TX, USA). Supelco (Bellefonte, PA, USA) procured androstenedione. The sodium salts of DHEAS and pregnenolone were obtained from Avanti Polar Lipids, Inc. (Alabaster, AL, USA). The testosterone, aldosterone, corticosterone, estrone, pregnenolone 3-sulfate (sodium salt form), leucineenkephalin (Leu-Enk), chloroform (>99.8% pure; of chromatography grade) and ammonia solution were purchased from Sigma-Aldrich (St. Louis, MO, USA). The LC-MS grade water, acetonitrile, formic acid and methanol were purchased from Fisher Chemical (Fair Lawn, NJ, USA). 2.3. LCMS Sample Preparation The steroid metabolites were extracted by liquid–liquid extraction using a methanol/ water mixture and chloroform as extraction liquids. The EV fractions were sonicated for 15 min in a total volume of 400 µ L 50% v/vmethanol/water mixture containing 1 mM ammonia to lysate EVs. The cell culture (DU145 cell line), fixed on culture well plates, was scrapped after 5 min incubation with 500 µ L 50% v/vmethanol/water mixture containing 1 mM ammonia. Tissue aliquots—approximately 50 mg—were lysed, using 1.4 mm zirconium oxide beads into standard 2 mL homogenizer tubes (Precellys, Montigny, France). Each sample was homogenized in 500 µ L 50% v/vmethanol/water mixture containing 1 mM ammonia by performing two cycles of 40 s at 6000 rpm in a FastPrep24TM 5G bead beating grinder (MP Biomedicals, Solon, OH, USA). After lysis, 400 µ L of the homogenate—either tissue, EV fraction or DU145 cell culture—was transferred to a clean Eppendorf ® tube. Subsequently, 400 µ L of LCMS grade chloroform was added on top of the 400 µ L of any lysated sample and shaken for 60 min at 1400 rpm at 4 ◦ C. Then, the samples were centrifuged for 30 min at 14,000 rpm at 4 ◦ C in order to precipitate the proteins and to separate the organic from the aqueous phases. The aqueous (top) and organic (bottom) phases were separated. The protein fraction was precipitated on the meniscus between these two immiscible phases. Then, 250 µ L of each fraction was transferred to the clean Eppendorf ® tubes and evaporated using a Metabolites 2022,12, 714 5 of 17 centrifugal vacuum concentrator. The pellets from the organic fraction were dissolved in 100 µ L pure methanol and the pellets from the aqueous fractions were dissolved in 50% v/v methanol/water. All of the resuspended pellets were centrifuged for 30 min at 13,000 rpm and 4 ◦ C. Finally, 80 µ L of the resuspended pellets were transferred to deactivated glass vials or 96-well plates for injection into the hrLCMS system. 2.4. Ultra-High Performance Liquid Chromatography (UPLC) The chromatographic separation of the analytes was performed with an ACQUITY UPLC I-Class PLUS System (Waters Inc., Milford, MA, USA). This system was equipped with a cooled (10 ◦ C) Process Sample Manager with a sample loop of 10 µ L and a Sample Organizer, a Binary Solvent Manager and a High Temperature Column Heater. A reversedphased 1.0 mm × 100 mm BEH C18 column (Waters Inc., Milford, MA, USA), thermostated at 40 ◦ C, was used for separating the analytes. The samples were injected from either 2 mL deactivated glass vials or 700 µL round 96-well polypropylene plates. The chromatographic behavior was optimized with respect to the peak intensity and an adequate separation of the 11 analytes along the run. The gradient elution was accomplished with an aqueous mobile phase (eluent A) consisting of 99.9% water with 0.1% formic acid and an organic mobile phase (eluent B) consisting of 99.9% acetonitrile with 0.1% formic acid. The flow rate was 140 µ L per min. Several gradients were tested during the optimization process (Table S1, Supplementary Materials) in order to avoid break-through (elution of analyte in the injection peak) and to obtain a good peak separation. The optimal gradient was as follows: start at 30% B; a linear increase to 80% B in 3.8 min.; a step increase from 80% to 99%; constant at 99% for 1.0 min and back to 30% B in 0.2 min. The total cycle time from injection to injection was 6 min. The injection volume for all of the samples was 2 µL. 2.5. Mass Spectrometry A time-of-flight mass spectrometer SYNAPT G2-S (Waters Inc.) was utilized for the detection of the analytes. The instrument was operated in either positive (ESI+) or negative (ESI-) electrospray ionization mode and in full-scan mode with a scan range between 50 Da and 1200 Da and scan time of 0.2 s. The z-spray source parameters: temperatures; gas flows; capillary position and voltages were tuned, as detailed elsewhere [ 30 ]. The optimal source parameters for this assay in either ESI+ or ESI − are summarized in Table S2 (Supplementary Materials). The ion optics were fine-tuned by spraying Leu-Enk (100 ppb), at a rate of 10 µ L per min, to a resolution over 20,000 (FWHM) for m/z556.2771. The same Leu-Enk solution was sprayed as a lock mass to correct for m/zfluctuations along the assay. The lock mass solution was introduced into the source every 90 s using a second ESI probe and it was recorded for 0.5 s. Mass spectrometer spectra was corrected according to fluctuations detected in the lock mass. 2.6. Statistical Analysis 2.6.1. Analyte Recovery Study The extraction step efficiency was assessed by performing a recovery assay with various mixtures of organic solvents and water. Five different extraction buffers were tested in this assay: 25/75% v/vand 50/50% v/v of methanol/water mixture; 25/74.9/0.1% v/v/v and 50/49.9/0.1% v/v/vof methanol/water/formic acid mixture and 50/50% v/v of methanol/water mixture with 1mM ammonia. To compare and calculate the recoveries of 10 different analytes, a culture of a prostate cancer cell line-DU145-was spiked with the analyte standards. Each well containing 5 ·× 10 5 cells was spiked with a mix of standards at 2 µ M before lysis (pre-spiked) and at the resuspension stage (post-spiked) with a standard mix at 10 µ M. Thus, the pre-spikes contained 1 nmol in 500 µ L and post-spikes (aqueous and organic fractions) contained the same total amount in 100 µ L, which would be the theoretical maximum absolute if there was no loss during the extraction. In ad- Metabolites 2022,12, 714 6 of 17 dition, for each extraction solution, the non-spiked samples were prepared in order to correct for endogenous metabolites in the matrix. The samples for the pre-spiked, postspiked and non-spiked conditions and the five different extraction buffers were prepared in biological triplicates. Only the absolute peak areas were taken into consideration to establish the recovery efficiency in the extraction step. The average peak areas were obtained by mean smoothing the raw signals of triplicates. The recovery (R) was determined by dividing the corrected pre-spike average by the corrected post-spike average and represented as a percentage (Equation (1)). Both the pre-spiked and post-spiked raw signals ought to be corrected by subtracting the endogenous analytes signal in the DU145 culture matrix (S non-spike ). However, as the S non-spike of DU145 culture matrix was less than 0.05% of the signal, endogenous correction was neglected during the calculation. Importantly, the pre-spikes were corrected with respect to analyte loss ( α ) during the extraction procedure. Moreover, the raw signals of each sample did not have to be corrected by the amount of initial samples, because every well contained the same amount of cells. R(%) = αSpre−spike −Snon−spike Spost−spike −Snon−spike ×100 (1) 2.6.2. Study of Matrix Effect in Analyte Quantification In order to assess the matrix effect (ME) in the quantification of the analytes, the post-spiked raw signal was compared to an equivalent raw signal of a mixture of analytes (10 µ M) in solution. The post-spiked raw signals were corrected by subtracting the endogenous analytes detected in the non-spiked DU145 culture samples. Then, the numerator was divided by the average peak areas of the standards and expressed as a percentage (Equation (2)): ME (%) = Spost−spike −Snon−spike Sstandards ×100 (2) 2.6.3. Analyte Semi-Quantification In this work, a calibration curve was prepared in solution with 50% v/v methanol/water for the semi-quantification of the analytes. This calibration curve consisted of a serially diluted mixture containing all of the analytes, starting at a concentration of 10 µ M. The initial concentration was diluted to half concentration twice, resulting in 5 µ M and 2.5 µ M concentration in the curve. Then, this set of triplets was diluted in five decades; it resulted in the following 15 different concentrations per analyte: 10; 5; 2.5; 1; 0.5; 0.25; 0.1; 0.05; 0.025; 0.01; 0.005; 0.0025; 0.001; 0.0005 and 0.00025 µ M. The calibration samples were injected at the beginning and at the end of each experiment; the average of these two points was used to semi-quantify the metabolites in the tissues. The limit of detection (LOD) for each analyte was set to be the lowest concentration at which the signal-to-noise (S/N) ratio was above three. The LOQ was defined as the lowest concentration at which the S/N ratio was above 10. The highest quantifiable concentration was the highest concentration per analyte that fits the calibration curve with an acceptable accuracy and precision (CV ≤15%) [16]. In general, the data of a calibration curve range over several orders of magnitude, the data are not linear and tend to be heteroscedastic [ 31 ]. For this reason, the relation between the peak area and the sample concentration was determined by power-fitting [ 30 ]. The power fitting resulted in a calibration curve (Equation (3)) with α and bas the fitted parameters. Once the sample concentrations were calculated using a calibration method in solution, the amount (in nanomole) per gram of tissue weight was estimated: Peak area =α[concentration]b(3) Metabolites 2022,12, 714 7 of 17 3. Results 3.1. Liquid Chromatography and Mass Spectrometry Method We compared six different chromatographic methodologies (Table S1, Supplementary Materials) to satisfactorily separate the analytes. The gradient 6 (30% B to 80% B in 3.8 min; detailed steps in Table S2, Supplementary Materials) showed the best peak separation along this run time compared to other tested gradients (data available in [ 32 ]). Due to the nature of the stationary phase, analytes elute in order of increasing hydrophobicity. The resulting extracted ion current (XIC) chromatograms of a standard mixture at 10 µ M are depicted in Figure S2 (Supplementary Materials). In brief, aldosterone (m/z361.2015; ESI+) elutes at 0.99 min, cortisol (m/z363.2171; ESI+) at 1.20 min, DHEAS (m/z367.1579; ESI − ) at 1.60 min, corticosterone (m/z347.2222; ESI+) at 1.68 min, androsterone sulfate (m/z369.1736; ESI − ) at 1.85 min, pregnenolone sulfate (m/z395.1892; ESI − ) at 2.23 min, estrone (m/z271.1698; ESI+) at 2.39 min, androstenedione (m/z287.2011; ESI+) and DHEA (m/z289.2168 ; ESI+) co-elute at 2.40 min, DHT (m/z291.2324; ESI+) at 2.65 min, pregnenolone (m/z317.2481; ESI+) at 3.25 min. Regarding the mass spectrometry method, the Leu-Enk signal (m/z556.2771) was aimed at a resolution of over 20,000 (FWHM) and provided the necessary mass accuracy to evaluate assay analytes. Isotope pattern matching and the use of chemical standards confirming elution times further ensured the specificity. In general, the mass accuracies for the analytes in solution were between −1 to 1 mDa. It is noteworthy that several analytes were not adequately separated during the chromatographic elution. The corticosterone and DHEAS elute at similar retention times—1.60 min and 1.68 min-, however, the MS could properly distinguish them by their m/zdifference and their fragmentation pattern. Moreover, the DHEAS was not detected with a high intensity signal in ESI+ mode. For this reason, the corticosterone was measured in ESI+ and the DHEAS in ESI − mode. Likewise, estrone, DHEA and androstenedione eluted in approximately 2.40 min. In this case, one could only rely on the MS sensitivity (estrone m/z271.1698, DHEA m/z289.2168, androstenedione m/z287.2011) and on a fragmentation pattern that was sensitive enough to distinguish and quantify them separately. 3.2. Analyte Recovery Optimization Afterwards, we evaluated the recovery of 11 analytes using a biphasic liquid–liquid method and analyzed them with the optimized hrLCMS method. The extraction was performed, using the DU145 cell line as a matrix. Five different mixtures of organic solvents and water, containing either formic acid or ammonia to modify the pH of the extraction buffer or no pH modifier, were assessed (Table S3, Supplementary Materials). The addition of formic acid strived for lowering the pH approximately to three, while 1mM ammonia modified the extraction buffer to pH 8–9 in order to chemically neutralize the functional groups of the steroid compounds. From the previous experiments in our metabolomics platform, we observed that in liquid–liquid extraction requires at least 25% organic solvent during the extraction step to precipitate the proteins. This is important to avoid clogging the chromatographic system [ 30 ]. Moreover, the effectivity of tissue homogenization using beads has been reported as high and does not differ much from the homogenization of other matrices, such as urine or cell cultures [ 30 , 33 ]. Therefore, the calculated recoveries are ultimately dependent on the extraction buffer utilized, regardless of the homogenization methodology. During the optimization process, it was determined that the steroid sulfate compounds were recovered completely in the aqueous fraction, whilst steroids without sulfate group were found in the organic fraction. Notably, only cortisol was detected systematically in both of the fractions (Figure S3, Supplementary Materials); however, it was majorly recovered in the organic (80% or higher) rather than in the aqueous (approximately 20%) fraction. Moreover, the addition of formic acid to the extraction buffer led to a dramatic decrease in the recoveries of the sulfate compounds and a slight decrease in the rest of the steroid analytes (Figure S3, Supplementary Materials). One can infer that the Metabolites 2022,12, 714 8 of 17 presence of protons in the buffer do not stabilize steroid charges and severely hampers the extraction of sulfate steroids in a polar environment. The supplementation of 1mM ammonia outperformed the extraction in terms of recovery and robustness, compared to the other extraction liquids. Notably, the recovery values using different percentages of methanol in the extraction buffer do not differ much. However, the extraction efficiency of the sulfate compounds using 25% v/v methanol underperforms 50% v/v methanol, with a recovery loss of 40 to 50%. In Table 1, the recoveries of the 11 selected analytes, using a mixture of 50/50% v/v methanol/water with 1mM ammonia as the extraction buffer, are reported. In general, the present methodology is able to recover and detect over 90% of the initially spiked analyte. Only DHT was detected in a lower percentage; approximately 80% of the initially spiked DHT was recovered. As expected in a biphasic extraction, the hormone steroids were retrieved in an apolar environment and the sulfated steroids in a polar solvent. Besides cortisol, pregnenolone sulfate was also reported in both of the fractions; it was mainly recovered in the more polar solvent and a derisory amount in the organic fraction. Using this methodology, the recoveries for 10 µ M of analyte ranged from 74.2% to 126.9%. These values are acceptable for routine muti-analyte hrLCMS analysis since all of the results are reproducible [ 34 ]. Thus, extraction using 50/50% v/v of methanol/water mixture with 1 mM ammonia was selected for further experiments in different biological matrices. Table 1. Summary of the optimized method characteristics. The recoveries ( ± standard deviation) and matrix effect as signal loss ( ± standard deviation) of the extraction procedure in two different biological matrices (n= 6; biological matrix: DU145 cell) are reported. In addition, LOD and LOQ values of the analytes in the adequate fraction are compiled. LOD: Limit of detection; LOQ: Limit of quantification. Analyte Fraction Recovery (%) Matrix Effect (%) LOD (nM) LOQ (nM) Pregnenolone Organic 97.2 (±1.9) 25.2 (±3.1) 2.5 nM 10 nM Aqueous - 24.0 (±2.8) DHEA Organic 122.7 (±2.9) 37.7 (±5.7) 5.0 nM 50 nM Aqueous - 28.0 (±6.2) - - Androstenedione Organic 102.2 (±3.2) 30.8 (±4.6) 0.25 nM 0.5 nM Aqueous - 23.2 (±4.5) Estrone Organic 103.7 (±3.8) 25.5 (±4.8) 5.0 nM 10 nM Aqueous - 25.7 (±4.0) DHT Organic 74.2 (±3.4) 23.1 (±3.9) 0.25 nM 1.0 nM Aqueous - 23.4 (±2.9) Cortisol Organic 114.3 (±3.8) 25.9 (±4.2) 0.5 nM 1.0 nM Aqueous 22.28 (±4.5) 17.6 (±4.7) Aldosterone Organic 99.8 (±1.77) 18.7 (±4.3) 0.5 nM 2.5 nM Aqueous - 17.7 (±5.1) Corticosterone Organic 109.4 (±3.1) 25.1 (±3.6) 0.25 nM 1.0 nM Aqueous - 20.2 (±3.2) Testosterone Organic 126.9 (±1.7) 14.3 (±1.9) 0.25 nM 0.25 nM Aqueous - 8.0 (±2.1) Pregnenolone sulfate Organic 6.9 (±2.7) 25.2 (±3.1) 0.25 nM 1.0 nM Aqueous 94.8 (±1.9) 24.0 (±2.8) DHEAS Organic - 42.6 (±1.1) 0.25 nM 0.5 nM Aqueous 108.0 (±1.4) 42.5 (±0.1) Furthermore, the performance of the optimized methodology was tested, using urine as the matrix since it has a high interest for clinical applications. Six samples of urine from a male individual were pooled and aliquoted in different two volumes to assess the matrix Metabolites 2022,12, 714 9 of 17 effect on the recovery efficiency. In Table 2, the recoveries of the 10 analytes are reported; DHEA recovery has not been retrieved, because its peak was masked by testosterone’s signal. In general, over 85% of the initially spiked analyte is recovered and detected in 50 µ L urine matrix. Importantly, the sulfated steroids are not recovered with the same efficiency; DHEAS and pregnenolone sulfate report a recovery efficiency of 75.7% and 54.9%, respectively. The recoveries of the analytes using 250 µ L urine as matrix describes a slight decrease in the non-sulfated steroids while the efficiency decay is dramatic in the sulfated species. Table 2. Summary of the recoveries using the optimized methodology in urine matrix. The recoveries ( ± standard deviation) of two different volumes (50 µ L and 250 µ L) of pre-pooled urine are reported (n= 3). Analyte Urine Volume Recovery (%) Pregnenolone 50 µL 92.4 (±3.6) 250 µL 99.3 (±4.8) Androstenedione 50 µL 93.0 (±3.9) 250 µL 79.3 (±3.8) Estrone 50 µL 94.2 (±3.3) 250 µL 84.8 (±4.8) DHT 50 µL 76.3 (±4.1) 250 µL 71.2 (±3.76) Cortisol 50 µL 87.0 (±3.0) 250 µL 72.4 (±3.6) Aldosterone 50 µL 110.7 (±2.9) 250 µL 103.1 (±3.2) Corticosterone 50 µL 96.2 (±2.8) 250 µL 84.3 (±3.6) Testosterone 50 µL 104.1 (±2.1) 250 µL 96.3 (±5.1) Pregnenolone sulfate 50 µL 54.9 (±1.5) 250 µL 25.5 (±1.2) DHEAS 50 µL 75.7 (±2.5) 250 µL 44.0 (±4.2) 3.3. Matrix Effect It is well known that the phospholipids and other lipids, typically enriched in biological matrices, such as tissues, body fluids or cell cultures, can cause ion suppression in mass spectrometry, thereby hampering the analyte signal [ 35 , 36 ]. This phenomenon negatively influences the detection of the analytes and may underestimate their quantification. For a specific matrix, the higher the ion suppression effect is, the higher the signal loss. Therefore, the conclusions drawn by detecting and quantifying the analytes under these conditions could be misleading. The matrix effect of each analyte was defined as the signal loss measured at the resuspension step (sample spiked with 10 µ M analyte mix) compared to 10 µ M of each analyte in solution. The signal loss was calculated in five different extraction procedures, because they can influence ion suppression. The matrix effect reported in this work was estimated for a prostate cancer cell line (DU145) culture and urine samples. To note, signal loss is specific for each matrix and each independent experiment. In further experiments, in which quantification is required, the matrix effect should be calculated in every particular assay. From our optimization experiments, one can infer that the matrix effect is fraction-dependent, because there is a significant difference between signal loss comparing organic and aqueous fractions (Figure S4, Supplementary Materials). This phenomenon is likely observed due Metabolites 2022,12, 714 10 of 17 to a differential extraction of the phosphatidylcholine (or other lipid) compounds [ 30 , 35 ]. Strikingly, this fraction dependency was not observed upon the addition of ammonia to the extraction liquid. Moreover, the presence of ammonia resulted in a signal loss of up to half-fold compared to extraction liquids with acidic modifier or no pH modifier addition. This suggests that the ammonia impairs the extraction of the lipidic compounds from the biological matrix, hence, decreasing the ion suppression phenomenon in mass spectrometry. In Table 1, the matrix effect (expressed as signal loss (%)) of a DU145 culture of 11 selected analytes, using a 50/50% v/vof methanol/water mixture with 1mM ammonia for extraction, is reported. In general, the present methodology loses approximately 15 to 40% of the signal of non-sulfated analytes but it mainly lays between 20 to 30% loss. On the other hand, the sulfated steroids display a 40 to 50% loss of signal, regardless of the extraction fraction. The signal loss of the 10 µ M analytes spiked in DU145 cell line were: 25.2% for pregnenolone, 37.7% for DHEA, 30.8% for androstenedione, 25.5% for estrone, 23.1% for DHT, 25.9% and 20.2 % for cortisol in the organic and aqueous fraction, respectively, 18.6% for aldosterone, 25.0% for corticosterone, 46.1% for pregnenolone sulfate and 42.5% for DHEAS. All of the analytes are majorly recovered back in a particular fraction of the extraction procedure, which is the one selected to report the matrix effect. Signal loss of sulfate compounds refer to aqueous fraction measurement and the other steroids refer to signal loss in organic fraction. 3.4. Semi-Quantitation of Steroids in Animal Tissues The hrLCMS method was most sensitive in detecting androstenedione, DHT, corticosterone, pregnenolone sulfate and DHEAS with a LOD (S/N > 3) of 250 pM in a 50/50% v/v methanol/water solution. The detection limit for cortisol and aldosterone was 0.5 nM, and a LOD of 2.5 nM was determined for pregnenolone. The least responsive ions were those for DHEA and estrone with a LOD of 5.0 nM. With regards to the quantification limits, androstenedione and DHEAS were the most sensitive compounds, with a LOQ (S/N > 10) of 0.5 nM in solution. The cortisol, corticosterone, pregnenolone sulfate and DHT were in the second group of the most quantifiable ions showing a LOQ of 1.0 nM. The quantitation limit for aldosterone was 2.5 nM, while a LOQ of 0.01 µ M was estimated for pregnenolone and estrone. The DHEA was the compound with the highest quantitation threshold (0.05 µM). We found that the concentration range of the steroid hormones is typically low in tissues, ranging from picoto nanomole per gram of tissue, and cannot be detected in some tissues (Table 3). Only pregnenolone, androstenedione, DHT, corticosterone, cortisol and testosterone were detected in the tissues or serum of Wistar rats. Pregnenolone and cortisol are only quantified in the adrenal gland tissue, however, pregnenolone is also detected in the brain and testicles. Adrenal gland and testicles reported picomole amounts of androstenedione per gram of tissue. Moreover, DHT was quantified in the prostate, adrenal gland and testicles. In prostate, the amount of DHT was two-fold the quantitation in the other tissues. The testosterone and corticosterone were quantified in all of the measured rat samples. In general, they were reported in the picomole per gram range in tissues. In serum, they were quantified in the nM range. Interestingly, the adrenal gland described nanomole per gram concentrations of corticosterone. Furthermore, testosterone was found in a one order of magnitude higher amount in the adrenal gland and testicles compared to prostate and brain. Metabolites 2022,12, 714 11 of 17 Table 3. Quantitation of three independent Wistar rat tissues: adrenal gland, prostate and brain. Adrenal glands of the same animal were titered independently, also, the prostate lobes of each rat. The averages in nmol per gram of tissue, standard deviations and coefficients of variation (%) of the three groups of samples are reported. Analyte Quantification (nmol/g Tissue) Adrenal Gland Prostate Brain Testicle Serum (nM) Pregnenolone Amount 7.04 - Detected Detected - St. dev. 3.74 %cv 53 Androstenedione Amount 5.97 ×10−3Detected 1.45 ×10−3Detected St. dev. 3.35 ×10−31.38 ×10−3 %cv 56 95 DHT Amount 3.47 ×10−37.57 ×10−3Detected 2.70 ×10−3Detected St. dev. 1.02 ×10−32.40 ×10−37.92 ×10−4 %cv 29 31 29 Corticosterone Amount 18.89 4.01 ×10−32.42 ×10−21.25 ×10−328.01 St. dev. 10.05 5.15 ×10−37.04 ×10−37.98 ×10−43.31 %cv 53 128 29 63 12 Cortisol Amount 0.45 - - - - St. dev. 0.19 %cv 43 Testosterone Amount 4.53 ×10−36.92 ×10−47.02 ×10−49.18 ×10−30.20 St. dev. 1.47 ×10−32.36 ×10−44.29 ×10−44.53 ×10−30.02 %cv 32 34 60 49 The standard deviations and coefficients of the variation are rather large, indicating an important variability among the samples obtained from the same strain but independent animals. One could expect this biological variation and it suggests that treatments, stress or any procedure applied to animals can potentially influence the outcome in further experiments. 3.5. Quantitation of Steroid Hormones in Human Urinary Samples Six different urine samples were characterized in several physicochemical parameters (Table S4, Supplementary Materials) to examine whether the sample collection resulted in homogenous sample groups, regardless of the metabolomics’ analysis. No blood, ketone bodies or glucose were detected in the urine sample, and the pH value and density of the urine were similar in all of the samples. The urine samples were centrifuged in two serial steps at 10,000 × gfor 30 min to isolate the so-called P10K fraction—typically containing vesicles of 150 to 200 nm diameter and above—followed by a 100,000 × gcentrifugation for 90 min to isolate the so-called P100K—typically containing vesicles of 100 to 150 nm diameter and below (up to 50 nm) [ 37 ]. The supernatant of the second centrifugation was also analyzed and referred to as SN100K. In this set of urine samples, the current methodology is able to detect and quantify androstenedione, cortisol and DHEAS (Table 4). The other steroids of the panel were below the LOQ and, in general, also below the LOD. The androstenedione and cortisol were detected only in the urine and SN100K. It was not possible to detect them associated with the EVs, and they are majorly solubilized in the urine. The androstenedione was found in lower concentrations compared to cortisol and the variability between the collection days was high (40 to 60%) regardless of the collection time. Concerning cortisol, the variability was extremely high between the morning collection days (approximately 50 to 85%) whilst the concentration of the afternoon collected samples was stable (approximately 2% variation). DHEAS was the compound detected in the highest concentration ( µ M range) soluble in urine, compared to androstenedione and cortisol (nM range). Similar to androstenedione, the DHEAS showed a high variability over independent collection Title: Integration of proteomic and metabolomic analysis reveal distinct metabolic alterations of 1 prostate cancer-associated fibroblasts compared to normal fibroblasts from patient’s stroma samples 2 Author names: Guillermo Bordanaba-Florit1,†, Félix Royo1,2, Oihane E. Albóniga3, Aled Clayton4, 3 Juan Manuel Falcón-Pérez1,2,3 and Jason Webber5,† 4 Affiliations: 5 1Exosomes Laboratory, Center for Cooperative Research in Biosciences (CIC bioGUNE), Derio, Spain; 6 2Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (Ciberehd), 7 28029 Madrid, Spain; 8 3Metabolomics Platform, Center for Cooperative Research in Biosciences (CIC bioGUNE), Derio, 9 Spain; 10 4Division of Cancer and Genetics, School of Medicine, Cardiff University, Cardiff, UK; 11 5Institute of Life Science, Swansea University Medical School, Swansea University, Swansea, UK. 12 †Corresponding authors: [email protected] (G. B.-F.); [email protected] (J. W.) 13 Highlights: 14  Lipogenic metabolic pathways are more prevalent in CAFs. 15  Energy producing pathways alternative to glycolysis support CAFs metabolic activity. 16  Mitochondrial metabolism is altered in the first steps of fibroblast differentiation. 17  Extracellular vesicles provoke a higher basal respiration associated to mitochondria in normal 18 fibroblasts. 19 Abstract: 20 The prostate gland is a complex and heterogeneous organ composed of epithelium and stroma. 21 Prostate cancer is most commonly seen emerging from luminal epithelial cells and requires the 22 assistance of adjacent stroma. Prostatic stroma is compositionally complex, and in concert with 23 disease progression undergoes many alterations which include the emergence of cancer associated 24 fibroblasts (CAFs). This heterogeneous cell population often contains cells with a myofibroblast-like 25 phenotype that are not normally present in healthy prostate tissue. In this work, we studied the 26 metabolic rewiring of stromal fibroblasts following myofibroblast differentiation. First, the metabolic 27 abundances of normaland cancer-associated fibroblasts derived from needle biopsies of the same 28 patient was analysed using UPLC-MS. It was determined CAFs were metabolically more active and, 29 therefore, energy producing metabolic pathways were enhanced. Also, CAFs showed a heightened 30 lipogenic metabolism as both reservoir species and building block compounds. Interestingly, lipid 31 metabolism affects mitochondria functioning yet the mechanisms of lipid-mediated functions are 32 unclear. The fact oxidised fatty acids and glutathione system are elevated in CAFs strengthens the 33 hypothesis that increased metabolic activity is related to mitochondria. In further experiments 34 measuring the metabolic flux with a Seahorse bioanalyser, we studied whether TGF-β1 and 35 extracellular vesicles (EVs) could stimulate myofibroblast differentiation in normal fibroblasts. An 36 increase of basal respiration in normal fibroblasts was reported, mirroring the disease-like phenotype. 37 This indicates an altered metabolism associated to mitochondria in CAFs and treated fibroblasts. 38 Hence, one proposes that the change in the metabolomics profile of tumour-associated stromal 39 fibroblasts is driven by oxygen-dependent metabolism, possibly associated to mitochondria; however, 40 the specific mechanisms are still unclear. 41 Keywords: 42 Mass spectrometry; Prostate cancer; Metabolism; Extracellular vesicles; Human primary fibroblasts 43 Abbreviations: 44 2-deoxy-d-glucose, 2-DG; α smooth muscle actin, α-SMA; Adenosine triphosphate, ATP; Cancer45 associated fibroblast, CAF; Cholesteryl ester, CE; Ceramides, Cer; Extracellular acidification rate, 46 ECAR; Extracellular vesicle, EV; Fatty acid, FA; Carbonyl cyanide p-trifluoro methoxyphenylhydrazone, 47 FCCP; Oxydised fatty acis, FFAox; Glutathione, GSH; Hydroxyeicosapentanoic acid, HEPE; Liquid 48 chromatography coupled to mass spectrometry, LC-MS; Lysophosphatidylcholines, LPC; 49 Glycerophosphocholines, LPE; Mass spectrometry imaging, MSI; Nicotinamide adenine 50 dinucleotide, NAD; N-acyl ethanolamines, NAE; Oxygen consumption rate, OCR; Prostate cancer, 51 PCa; Principal component analysis, PCA; Phosphate buffered saline, PBS; Reactive oxygen species, 52 ROS; Standard error of the mean, SEM; Tricarboxylic acid, TCA; Transforming growth factor beta 1, 53 TGF-β1; Ultra-high performance liquid chromatography coupled to mass spectrometry, UPLC-MS. 54 Graphical abstract: 55 56 57 Introduction: 58 The prostate gland is globally composed of epithelium and stroma, which are extremely heterogeneous 59 tissues. The epithelium is organized as glandular acini and it contains cuboidal to columnar secretory 60 epithelial cells with apical junction complexes, a continuous layer of basal cells1, and sparse 61 neuroendocrine cells, each attached to a basal lamina2. Beyond the basal lamina, a prominent 62 fibromuscular stroma composed of smooth muscle, fibroblasts, blood vessels, autonomic nerve fibres, 63 inflammatory cells, and extracellular matrix components offers physical support and contraction of the 64 gland3. In prostate carcinoma (PCa), the interstitial stroma is often abnormally rich in myofibroblastic 65 cells4,5, capable of supporting tumour growth, vascularization, angiogenesis, and metastasis in vivo6. 66 Transforming growth factor β1, TGF-β1, remains among the most critical factors for myofibroblastic 67 differentiation and the generation of a tumour reactive stroma. Intriguingly, we have previously shown 68 that cancer extracellular vesicles (EVs) can trigger fibroblast to myofibroblast differentiation in an EV69 TGF-β1 dependent manner6. In further studies, the essential role of EVs in directing this stromal cell 70 differentiation towards cancer-associated myofibroblast-like phenotype was described7. However, the 71 mechanisms by which EV-activated stromal cells support tumour growth remain unclear. 72 Our previous studies with biopsy material from patients where tumoral growth was located in one half 73 of the prostate and not the other, revealed clear differences between normal and disease regions. 74 Histological examination showed a smooth muscle stromal architecture around glandular structures in 75 normal tissue while the glands were disorganised and showed an altered, fibrosis-like interstitial stroma 76 observed in disease-associated tissue8. A panel of antibody markers confirmed the higher abundance 77 of α-Smooth Muscle Actin (α-SMA)-positive myofibroblast cells in tumour-associated tissue. Strikingly, 78 EVs isolated from prostate cancer epithelial cell lines generated a myofibroblast-like phenotype in 79 normal fibroblasts7. In that work, we demonstrated the essential role of EVs in directing stromal 80 differentiation to a pro-tumorigenic phenotype, exhibiting pro-angiogenic properties and enhancement 81 of tumour growth in xenograft models. Further functional assays and proteomics profiling work 82 highlighted that stroma activation mediated by EV stimulation mirrors the naturally occurring fibroblast 83 differentiation during disease8. 84 Altered metabolism is a hallmark of PCa and several metabolites and metabolic pathways are already 85 distinctive in different prostate types of tissue9,10. In this line, Andersen et al. reported higher levels of 86 energy-related pathway metabolites, such as ADP, ATP, and glucose as well as higher levels of the 87 antioxidant taurine in stromal tissue compared to cancer and non-cancer epithelium11. Besides, 88 increased levels of crucial metabolites for fatty acid oxidation and building blocks in lipid synthesis 89 were described in cancer tissue. Other studies also reported metabolic rewiring of reactive stroma, 90 showing different levels of certain metabolites between highly differentiated stroma compared to poorly 91 differentiated stroma12. The stromal-epithelial interactions have a dominant role in tumour growth, 92 invasion and metastasis. Actually, many reports over the last decades showed the interaction of 93 reactive stroma with PCa2,12–17 and EVs influencing aspects of cancer biology such as 94 angiogenesis7,8,18,19 and tumour progression17,20–22. Yet, few studies have investigated the role of 95 tumour EVs in altering metabolic processes in stromal cell compartments. 96 With the rise of omics era, entire sets of biomolecules – genes, proteins or metabolites - contained in a 97 biological tissue, cell, fluid, or organism can be identified. A proteomics analysis of the same matched98 normal and disease stroma tissues was able to distinguish both phenotypes and describe a disease99 like phenotype8. Similarly, a transcriptomic analysis of derived EVs from the same type of samples 100 discriminated normal from disease stroma samples23. This manuscript presents a broad semitargeted 101 metabolomics approach, which analyses the metabolome of normal and disease stroma tissues, 102 highlighting the metabolic differences of matched normal fibroblasts and cancer-associated fibroblasts 103 within individual PCa patients. Furthermore, we show an altered metabolism of normal fibroblasts 104 treated with EVs from a PCa cell line. 105 Materials and methods: 106 Stromal primary cell cultures 107 Six patient-matched normal and tumour-associated needle biopsies were isolated from radical 108 prostatectomy. These were taken from sites of palpable disease and also from apparently normal 109 tissue from the opposite side of the same prostate. Tissue collection and consenting was managed 110 through the Wales Cancer Bank. Cores were manually dissected into 1mm3 pieces and subjected to 111 mechanical homogenization followed by 200 U per mL collagenase-I digestion for 15 to 20 hours at 37 112 °C. Cells were cultured in Stromal Cell Basal Medium (SCBM) supplemented with human fibroblast 113 growth factor-B, insulin, fetal bovine serum (FBS) and GA-1000 (Lonza, Wokingham, UK). for around 114 two weeks until only stromal cells were retained. 8,23Subsequent cultures were maintained in 115 DMEM/F12 media (Lonza) with 10% FBS depleted of bovine EVs. Cultures were confirmed free of 116 epithelial cells by immuno-fluorescence staining for cytokeratin prior seeding stromal cells in 96 well 117 plates 8,23. A sample of each patient-matched prostatic stromal cell culture was collected and frozen for 118 metabolomics analysis using liquid chromatography coupled to mass spectrometry. 119 Metabolite extraction 120 In metabolomics, there is no single platform or method able to analyse the entire metabolome of a 121 biological sample. Therefore, metabolites were extracted by fractionating the cell samples into pools of 122 species with similar physicochemical properties. In brief, proteins were precipitated by adding methanol 123 to the cell lysate. Chloroform solvent was added to the methanol extraction mixture and this biphasic 124 mixture was incubated at -20 ºC for 30 min. Then, three different fractions were collected: (1) fatty 125 acyls, bile acids, steroids and lysoglycerophospholipids, were obtained after centrifuging the 126 supernatant at 16,000g for 15 min, drying and reconstituting in methanol, (2) for aminoacids, aliquots of 127 5 μL from the first fraction were derivatised and dried, and (3) glycerolipids, cholesteryl esters, 128 sphingolipids and glycerophospholipids were obtained by mixing the chloroform extraction mixture with 129 H2O (pH 9) and incubating at -20 ºC for 60 min. After centrifuging at 16,000g for 15 min, the organic 130 phase from this third fraction was collected then, dried and reconstituted in 50/50% v/v 131 acetronitrile/isopropanol. The aqueous phase that contains polar metabolites, including central carbon 132 metabolism, was collected, dried and reconstituted in H2O. 133 Quality control (QC) sample for calibration and validation were included in this workflow to correct for 134 response factors between and within batches; and to assess the quality of data. 135 LC-MS analysis 136 An appropriate UPLC-MS method was used for each platform. The instruments and the conditions for 137 the chromatographic separation and mass spectrometric detection are summarized in Table S1. A test 138 mixture of standards was analyzed before and after the entire set of randomized, duplicated sample 139 injections to check for retention time stability, mass accuracy and sensitivity. All data were processed 140 using the TargetLynx application manager for MassLynx 4.1 software (Waters Corp., Milford, USA). A 141 set of predefined features, defined as retention time - mass-to-charge ratio pairs, Rt-m/z, 142 corresponding to metabolites included in the analysis are fed into the program. Associated extracted 143 ion chromatograms (mass tolerance window = 0.05 Da) are then peak-detected and noise-reduced in 144 both the LC and MS domains such that only true metabolite related features are processed by the 145 software. Then, a list of chromatographic peak areas is generated for each sample injection. 146 Data analysis 147 Data Normalization and quality control 148 After data inspection in terms of reproducibility and peak integration, each metabolite was corrected 149 and normalized using the intensity of an appropriate internal standard included in the analysis and 150 following the procedure fully described by van der Kloet et al.25. Finally, any remaining zero values in 151 the corrected dataset were replaced with missing values prior averaging to obtain a dataset further 152 used for statistical analyses. A final normalization procedure was applied by dividing every sample by 153 its protein content. 154 Multivariate and univariate analysis 155 Once data was normalized and prepare for statistical analysis, a first approach based on multivariate 156 analysis was performed with SIMCA-P (version 13.0). Firstly, a non-supervised principal component 157 analysis (PCA) was utilised to reduce dimensionality and to study data quality, assess reproducibility of 158 the analytical procedure, visualize tendencies between groups and determine the presence of outliers. 159 Afterwards, supervised partial least squares discriminant analysis (PLS-DA) and orthogonal PLS-DA 160 (OPLS-DA) were performed followed by a suitable validation method, a cross-validation analysis of 161 variance (CV-ANOVA), integrated in SIMCA-P software. Supervised models that were validated were 162 further used for variable selection. To this end, a variable importance on projection (VIP) score and 163 absolute value of p(corr) greater than 1 and 0.8, respectively, were used as cut-off points for variable 164 selection. 165 Finally, and as a complementary statistical analysis, univariate analysis was perfomed. In order to test 166 normality, Shaphiro test was used; thence either paired student’s test26 or Wilcoxon signed-rank test 167 was applied to assess comparisons significance. After that, the dataset was expressed as a metabolite 168 fold-change and significance (p-value) and further depicted in a volcano plot. 169 Pathway analysis and pathway enrichment 170 A proper biological interpretation is crucial in any metabolomics study to deliver a comprehensive 171 assessment of experimental conditions. To evaluate the prominence of certain metabolic pathways, a 172 Pathway analysis was computed using MetaboAnalyst 5.0 and inputting a dataset of samples and 173 quantified metabolites. It performs an o-representation analysis that integrates enrichment and 174 pathway topology analysis to visualise specific altered pathways in the human metabolic network27. 175 Then, proteomics data was included in a Joint-pathway analysis to evaluate metabolic alterations 176 considering two sets of physiologically relevant molecules in fibroblast samples. This over177 representation analysis module performs an integrated metabolic pathway analysis by combining 178 metabolomics and proteomics data collected from the exact same samples and methodology27. The 179 MetaboAnalyst 5.0 web-based tool includes the normalisation, transformation and scaling of data to 180 complete data integration. 181 Furthermore, a lipid metabolic network analysis28 performed with LINEX 2.4.1 webapp to observe 182 functional associations of lipid classes. It computes specific lipid networks based on compounds and 183 lipid classes connections using a lipidomics dataset. 184 Extracellular vesicle isolation 185 EVs were purified from conditioned media of DU145 prostate cancer cell (ATCC, Teddington, UK) 186 grown in Integra bioreactor flasks (Integra Biosciences Corp, Hudson, NH, USA) 24. EV samples were 187 collected using the sucrose cushion method and resuspended in PBS. Thence, samples were 188 quantified using the BCA-protein assay (Pierce/Thermo, Northumberland, UK), and stored at -80°C. 189 For treatments of stromal cell cultures, EV were used at 200 µg per mL (approximately equivalent to 190 1.5 ng per mL of EV-associated TGF-β1) for 72 hour. 191 Seahorse assay 192 Oxygen consumption rate and glycolytic activity were assessed using a XF24 Extracellular Flux 193 Analyser (Seahorse Biosciences) to probe O2 and pH, respectively. Fibroblasts were equilibrated in 194 unbuffered media (60 min at 37 ºC in a CO2-free incubator) prior transfer to the XF24 analyser. For 195 mitochondrial respiration, the oxygen consumption was measured over the assay. First, basal oxygen 196 consumption (OCR) was determined, and then oligomycin (1 µg per ml), FCCP (0.3 µM), FCCP (0.6 197 µM), and 2 µM rotenone were sequentially injected to assess maximal oxidative capacity, ATP 198 production, coupling efficiency (OCR percentage dedicated to produce ATP) and basal respiration. To 199 analyse glycolytic activity, the extracellular pH was measured over the assay. First, base-line (non200 glycolytic) extracellular acidification (ECAR) was determined, and then media alone followed by 201 glucose (10 mM), oligomycin (1 µg per ml), and 2-Deoxyglucose (0.1 M) were sequentially injected to 202 assess maximal glycolytic capacity and glycolysis rate. 203 Statistical analyses were performed by paired Student’s t-test using GraphPad PRISM 9.5 software 204 (Graph Pad, San Diego, CA, USA). Each experiment was analysed individually then, the mean ± SEM 205 was represented. All p-values lower than 0.05 are considered significant as: * P>0.05, ** P>0.01, *** 206 P>0.001. 207 Results: 208 CAFs exhibit a differential proteomic, transcriptomic and metabolomics landscapes 209 In PCa, the emergence of myofibroblasts within the interstitial stroma is described as the major 210 difference between normal and tumour reactive stroma12,29. This reactive stroma coevolves with 211 prostate cancer in it is capable of supporting its growth. In previous studies, we have characterised 212 normal fibroblasts and CAFs derived from the same patient’s needle biopsies for the typical markers of 213 reactive stroma, α-Smooth Muscle Actin (α-SMA), Cytokeratine, Desmin and Vimentin7,8,23. CAFs were 214 not a homogeneous population of myofibroblasts but a heterogenous mixture of fibroblasts at distinct 215 differentiation stages, including a proportion of cells which are α-SMA positive7. The normal fibroblasts 216 lacked α-SMA. However, they could be induced to express α-SMA when treated with either soluble or 217 vesicle-associated TGF-β18. 218 In 2016, Webber et al.8 described a set of proteins that could discriminate CAFs and normal fibroblasts 219 isolated from the patient—matched needle biopsies described in this study. A more recent study 220 explored the opportunity to use EV-derived RNA from these fibroblasts as indicators of altered tumour 221 environment23. In this work, Shephard et al. identified 19 differentially expressed transcripts that 222 discriminate disease from normal stromal EVs, indicating transcriptional differences between patient 223 samples. The metabolomics study included in this work adds several findings to our published data 224 obtained by proteomics and transcriptomics. A comprehensive evaluation of data showed a clear 225 tendency of enhanced abundances in most of the metabolites in sample 1161-normal during data 226 analysis and normalization compared to other patient-matched tissues. The behaviour of this sample 227 was also different from the remaining samples as demonstrated by PCA (Fig. S1). Moreover, it is on 228 the line that describe the confidence ellipse based on Hotelling’s T2 (significance level = 0.05). For 229 these reasons, sample 1161-normal was excluded from further statistical analysis. Afterwards, PLS-DA 230 and OPLS-DA models were built and the scores plots are included in Fig. 1A and 1B. A clear 231 separation tendency was observed in both models mainly through PC1. Even with this separation 232 tendency, none of the models was validated (CV-ANOVA p-value >0.05). However, those variables 233 that influence on the most in group separation were selected based on their VIP and p(corr) values. In 234 total 64 and 12 metabolites fulfil VIP greater than 1 and ǀp(corr)ǀ greater than 0.8 in PLS-DA and 235 OPLS-DA, respectively. In order to select those metabolites of relevant importance, a Venn Diagram 236 including these metabolites was built to select only common metabolites in PLS-DA and OPLS-DA 237 models (Fig. 1C). Relevant metabolites include ceramides (Cer), phosphatidylcholines (PC) and 238 cholesteryl esters (CE) (Table S3). 239 240 Figure 1. Summary of multivariate metabolomics analysis of normal fibroblasts and CAFs. A. Score scatter 241 plot of the PLS-DA model of fibroblasts obtained from normal and cancer needle biopsies. Model diagnostics (A = 242 2; R2X = 0.682; R2Y = 0.729; Q2 = 0.337; CV-ANOVA = 0.581). In green, CAF samples and, in purple, normal 243 fibroblasts. B. Score scatter plot of the OPLS-DA model of fibroblasts obtained from normal and cancer needle 244 biopsies. Model diagnostics (A = 2; R2X = 0.682; R2Y = 0.721; Q2 = 0.241; CV-ANOVA = 0.799). In green, CAF 245 samples and, in purple, normal fibroblasts. In green, CAF samples and, in purple, normal fibroblasts. C. Venn 246 diagram of features (metabolites) that influence to the separation of CAF and normal fibroblasts groups in PLS-DA 247 and OPLS-DA models. Results are compiled in Table S3. 248 Disease-associated fibroblasts show lipogenic and energy-producing metabolic alterations 249 Reactive stroma relevance grows upon progression and invasion of tumour cells to neighbour locations 250 and tissues since it supports structural growth and nutrient availability. In this line, cell populations in 251 normal stroma undergo a myofibroblast turnover constituting the so-called CAFs. In order to study the 252 metabolic profile of this two tissue subtypes, four platforms based on UPLC-MS were utilised to 253 analyse metabolites. They are fractionated in pools of species with similar physicochemical properties. 254 The platforms include: (1) Fatty acyls, bile acids, steroids and lysoglycerophospholipids; (2) 255 Glycerolipids, glycerophospholipids, sterol lipids and sphingolipids; (3) Amino acids; (4) Polar 256 metabolites profiling, including central carbon metabolism. 257 258 Figure 2. Summary of metabolomics study (considering individual metabolites) comparing normal and 259 cancer-associated fibroblasts isolated from needle biopsies of radical prostatectomy specimens. A. 260 Volcano plot comparing CAFs and normal fibroblasts as -log10(p-value) against log2(fold change). The different 261 metabolite classes are depicted in different shapes and colours; relevant metabolites are labelled. B. Pathway 262 overrepresentation analysis depicted as –log10 (p-value) against pathway impact. Pathway impact stands for the 263 relative importance of the specific module in the analysed metabolite set. It combines pathway overrepresentation 264 results and centrality measures. Representative pathways were labelled, size of nodes represents pathway impact 265 and their significance ranges from high (in red) towards orange, yellow and white indicating a lower significance. 266 Here, all individual metabolites and their fold changes were considered to rank the enrichment and significance of 267 each pathway. AA: aminoacids; TCA: tricarboxylic acid cycle related metabolites; CHD: Carbohydrates derivatives; 268 Nt: Nucleotides; Redox: electron donor and acceptors; Ns: Nucleosides; Vit: Vitamins; HexCer: Hexosylceramides; 269 LPE: Lysophosphatidylethanolamines; PC: Phosphatidylcholines; PE: Phosphatidylethanolamines; PI: 270 Phosphatidylinositols; SM: Sphingomyelins; LPC: Lysophosphatidylcholines; Cer: Ceramides; NAE: N271 of processes and alterations occurring in each patient. Even so, the outcome of this study described 447 metabolic differences with sparse impact in the entire metabolome. Once combined with proteomics 448 data from the exact same set of samples, the outcome acquired a higher relevance in determining 449 important metabolic processes driving myofibroblast differentiation. The integration of significant 450 proteins and metabolites differentially measured in normal and cancer fibroblasts confirmed a major 451 alteration in lipid metabolism. A variation of glycerophospolipids - i.e. PC(O-22:1/20:4) and PC(P452 18:0/20:4) - indicates an enrichment of the major membrane type of lipids. Interestingly, this class of 453 lipids have been associated to mitochondria dynamics and well-functioning32,33. Although lipids can 454 participate in signalling processes or be used as building blocks, in this study most metabolic 455 alterations point out fuelling or energy storage processes. Major free fatty acid pathways – linoleic and 456 arachidonic - are altered together with an accumulation of cholesteryl esters and few oxidised fatty 457 acids. This indicates a mobilisation of lipids to produce energy but also the production of lipid 458 reservoirs readily available for transport. Moreover, central metabolism pathways – 459 glycolysis/gluconeogenesis and pentose phosphate - are positively altered in CAFs only including the 460 proteome to the analysis. This suggests that as protein queries were the most prominent and 461 overlapped a higher number of pathways compared to metabolites, the output of this integrative 462 analysis could be determined majorly by differences in the proteome. 463 Back in 2016, this proteomics study described many proteins altered in normal and cancer fibroblasts. 464 However, several of those were unique to specific treatments with either sTGF-β1 or EV-associated 465 TGF-β1. For instance, they described a higher presence of Annexin-I, which has a role in regulating 466 VEGF function, in CAFs as well as altered mitochondrial proteins, linking CAFs to mitochondrial 467 rearrangement. In this manuscript, we approached the assessment of metabolic alterations upon 468 sTGF-β1 or EV-associated TGF-β1 treatments by Seahorse stress assays. Even considering the 469 limitation of such assays, where O2 consumption and pH acidification are recorded over time, the 470 glycolytic rate and oxidative mitochondrial activity of cells was assessed. ECAR informs that glycolytic 471 activity associated to CAFs was not increased. This, together with the fact that the integration of 472 proteomics and metabolomics indicates an alteration in central metabolism, suggests that these 473 pathways were not used in the catabolic but in the anabolic direction. Only basal respiration, in the 474 mitochondrial stress assay, was distinctive when comparing CAFs and normal fibroblasts. Moreover, 475 the treatment with sTGF-β1 and EVs containing TGF-β1 showed a similar outcome. This result links 476 the distinctive metabolic abundances to mitochondrial metabolism. Our metabolomics data offers a 477 plausible explanation because elevated levels of GSH are usually correlated to the presence of ROS 478 due to mitochondrial activity while altered levels of FFAox may indicate the burning of lipids in 479 mitochondria seeking for energy and counteracting ROS. Perhaps a longer stimulation or a higher 480 dose of sTGF-β1 or EVs is required to observe further metabolic alterations associated to cancer in 481 normal fibroblasts. Nonetheless, this data demonstrates that metabolic alterations are related to a 482 higher activity of mitochondria, yet the specific role of cancer-associated EVs and the metabolic 483 mechanisms in stromal differentiation and shift towards its reactive state remain unclear. 484 [Document text truncated for crawler view.]