Functional Characterization of Circulating Tumour Cell (CTC) Clusters in Breast Cancer

1 TESE DE DOUTORAMENTO FUNCTIONAL CHARACTERIZATION OF CIRCULATING TUMOUR CELL (CTC) CLUSTERS IN BREAST CANCER Inés Martínez Pena ESCOLA DE DOUTORAMENTO INTERNACIONAL DA UNIVERSIDADE DE SANTIAGO DE COMPOSTELA PROGRAMA DE DOUTORAMENTO EN MEDICINA MOLECULAR SANTIAGO DE COMPOSTELA 2021 2 DECLARACIÓN DO AUTOR/A DA TESE 3 D./Dna. Inés Martínez Pena Título da tese: Functional Characterization of Circulating Tumour Cell (CTC) Clusters in Breast Cancer Presento a miña tese, seguindo o procedemento axeitado ao regulamento, e declaro que: 1) A tese abarca os resultados da elaboración do meu traballo. 2) De ser o caso, na tese faise referencia ás colaboracións que tivo este traballo. 3) Confirmo que a tese non incorre en ningún tipo de plaxio doutros autores nin de traballos presentados por min para a obtención doutros títulos. 4) A tese é a versión definitiva presentada para a súa defensa e coincide a versión impresa coa presentada en formato electrónico. E comprométome a presentar o Compromiso Documental de Supervisión no caso de que o orixinal non estea na Escola. En Santiago de Compostela, 05 de Agosto de 2021. Sinatura electrónica 4 5 AUTORIZACIÓN DO DIRECTOR / TITOR DA TESE Functional Characterization of Circulating Tumour Cell (CTC) Clusters in Breast Cancer D./Dna. Roberto Piñeiro Cid D./Dna. Rafael López López INFORMA/N: Que a presente tese, correspóndese co traballo realizado por D/Dna. Inés Martínez Pena, baixo a miña dirección/titorización, e a utorizo a súa presentación , considerando que reúne os r equisitos esixidos no R egulamento de Estudos de Doutoramento da USC, e que como director desta non incorre nas causas de abstención establecidas na Lei 40/2015. De acordo co indicado no Regulamento de Estudos de Doutoramento, declara tamén que a presente tese de doutoramento é idónea para ser defendida en base á modalidade de Monográfica con reproducción de publicaciones, nos que a participación da doutoranda foi decisiva para a súa elaboración e as publicacións se axustan ao Plan de Investigación. En Santiago de Compostela, 5 de Agosto de 2021 6 7 I, Inés Martínez Pena, author of this thesis, have no conflict of interest to declare. 8 9 The studies described in this thesis were performed within the framework of the Roche-CHUS Joint Unit (IN853B 2018/03), at the University Hospital of Santiago de Compostela (CHUS), Spain. The work of this thesis was financially supported by the Galician Agency of Innovation (GAIN), Consellería de Economía, Emprego e Industria, Roche Pharma. I, Inés Martínez Pena, was financially supported by the Formación del Profesorado Universitario (FPU) fellowship (FPU16/01018), Ministerio de Ciencia, Innovación y Universidades, Spanish Government. 16 17 INDEX ABBREVIATIONS ................................................................................. 23 RESUMO ............................................................................................. 29 RESUMEN ........................................................................................... 42 SUMMARY .......................................................................................... 55 INTRODUCTION .................................................................................. 67 1. Breast cancer ...................................................................... 69 1.1. Epidemiology and aetiology ....................................... 69 1.2. Subclassification of BC tumours ................................. 70 1.3. BC stages .................................................................... 71 2. Metastasis .......................................................................... 73 2.1. Migration and Invasion .............................................. 74 2.2. Intravasation, extravasation, and survival into the bloodstream ........................................................................... 76 2.3. Colonization, and growth in a distant organ .............. 77 3. Circulating tumour cells (CTCs) .......................................... 78 3.1. Relevance of CTCs in tumour progression ................. 78 3.2. CTC enrichment technologies .................................... 79 3.2.1. CTC enrichment based on biological properties 79 3.2.1.1. Positive enrichment ....................................... 79 3.2.1.2. Negative enrichment ...................................... 81 3.2.2. CTC enrichment based on biophysical properties 82 3.2.2.1. Density gradient ............................................. 82 3.2.2.2. Microfiltration ................................................ 83 3.2.2.3. Microfluidics ................................................... 83 3.2.2.4. Dielectrophoresis ........................................... 84 4. Circulating tumour cell clusters (CTC clusters) ................... 85 18 4.1. Origin of CTC clusters ................................................. 85 4.2. Metastatic traits of CTC clusters ................................ 88 4.3. Clinical value of CTC clusters ...................................... 90 OBJECTIVES ......................................................................................... 94 MATERIALS AND METHODS ............................................................... 99 1. Metastatic breast cancer patient samples ....................... 101 1.1. Peripheral blood samples from metastatic BC patients 101 1.2. Diagnostic Leukapheresis (DLA) samples ................. 103 1.2.1. Standard separation protocol vs. Cluster separation protocol .......................................................... 104 1.2.2. Filtered: Before vs. After tumour cell spiking .. 105 2. Generation of individual ctc and ctc cluster in vitro models 106 2.1. Cell lines ................................................................... 106 2.2. Generation of single CTC and CTC cluster in vitro models 107 3. In vitro assays ................................................................... 109 3.1. Proliferation / Cell viability assays ........................... 109 3.2. Transwell migration assays ...................................... 109 3.3. Transwell invasion assay .......................................... 111 3.4. Fluid Shear Stress (FSS) assays ................................. 111 3.5. Endothelial adhesion assays ..................................... 113 3.6. Soft agar colony formation assays ........................... 114 4. In vivo assays .................................................................... 114 4.1. Zebrafish (Danio rerio) embryo xenografts .............. 114 4.2. Mouse (Mus musculus) experiments ....................... 117 4.2.1. Mouse lung colonization assay......................... 118 19 4.2.2. Mouse orthotopic BC xenografts ..................... 119 4.3. mCTC cell line generation ......................................... 121 5. Molecular characterisation .............................................. 121 5.1. Gene expression analysis ......................................... 121 5.2. RNA-sequencing (RNA-seq) análisis ......................... 124 5.3. Genomic analysis (WES, whole exome sequencing) 124 6. Data analysis ..................................................................... 125 RESULTS ............................................................................................ 128 1. CTC cluster detection in samples derived from metastatic breast cancer patients .............................................................. 130 1.1. CTC and CTC cluster enumeration in peripheral blood samples................................................................................. 130 1.1.1. Cohort description: patient characteristics ...... 130 1.1.2. CTC and CTC cluster enumeration and correlation with clinicopathological variables .................................... 133 1.1.3. CTC and CTC clusters as predictor factors of patient outcome ............................................................... 136 1.2. Optimization of ParsortixTM immuno-independent detection of CTCs and CTC clusters in BC liquid biopsy samples 148 1.2.1. Metastatic BC peripheral blood samples ......... 148 1.2.2. Metastatic BC Diagnostic Leukapheresis (DLA) products 151 2. CTC cluster models have a differential metastatic behaviour than individual CTCs ................................................................. 161 2.1. Functional characterization shows differences between individual CTC and CTC cluster experimental models 164 2.1.1. CTC clusters show a higher migration and invasion capacity than individual CTCs ........................................... 164 20 2.1.2. CTC clusters show a lower ability to adhere to the endothelium but a higher capacity to form colonies ....... 166 2.1.3. CTC clusters disseminate less within the circulation of the zebrafish embryo than individual CTCs 169 2.1.4. CTC clusters survive better in the circulation of the zebrafish embryo xenografts and have a proliferative advantage in comparison with single CTCs ...................... 178 2.1.5. Gene expression profiling supports the higher metastatic potential of the CTC cluster model ................ 184 2.1.6. CTC cluster has a higher colonization capacity in mice than individual CTCs ................................................ 187 3. Use of a metastatic orthotopic BC murine model as a source for CTCs and CTC clusters ............................................. 192 3.1. Establishment of an ex vivo CTC cluster surrogate model 195 3.2. Molecular analysis to study single CTC and CTC cluster genomic heterogeneity ........................................................ 203 DISCUSSION ...................................................................................... 217 CONCLUSIONS .................................................................................. 234 SUPPLEMENTARY MATERIAL ............................................................ 239 APPENDIX ......................................................................................... 249 List of publications: .................................................................. 251 Favourable report (Clinical Research Ethics Committee of Galicia, CEIC) ......................................................................................... 252 Material Transfer Agreement for the use of Diagnostic Leukapheresis products ........................................................... 254 Favourable report Animal Experimentation Ethical Committee of the University of Santiago de Compostela (CEEA) ................... 262 Animal experimentation training certificate. B function (euthanasia).............................................................................. 263 21 Animal experimentation training certificate. C function (animal procedures) .............................................................................. 265 AGRADECIMIENTOS.......................................................................... 268 REFERENCES ..................................................................................... 275 22 23 ABBREVIATIONS 24 25 ABBREVIATIONS AP Anatomopathological BC Breast Cancer bFGF basic Fibroblast Growth Factor cDNA complementary DNA CI Confidence Interval CK cytokeratin CTC Circulating Tumour Cell DEP Dielectrophoresis DLA Diagnostic Leukapheresis dpi days post-injection DoC Duct of Cuvier ECM Extracellular Matrix ECOG Eastern Cooperative Oncology Group EGF Epidermal Growth Factor eGFP enhanced Green Fluorescent Protein EGFR Epidermal Growth Factor Receptor EMT Epithelial to Mesenchymal transition EpCAM Epithelial Cell Adhesion Molecule ER Estrogen Receptor FBS Fetal Bovine Serum FDA Food and Drug Administration FOV Field Of View FPKM Fragments per Kilobase of transcript per Million mapped reads FSS Fluid Shear Stress GAPDH Glycerlaldehyde3-Phosphate Dehydrogenase GEP Gene Expression Profiling GFR Growth Factor Reduced He expected Heterozygosity HER2 Human Epidermal growth factor Receptor 2 hpf hours post-fertilization hpi hours post-injection HR Hazard Ratio HR Hormone Receptor IHC Immunohistochemistry INÉS MARTÍNEZ PENA 32 30% dos pacientes con cancro de mama acabará progresando cara un estado diseminado da enfermidade ó longo do tratamento. Por tanto, incrementar o coñecemento e comprensión sobre o proceso de metástase é esencial para mellorar o tratamento dos pacientes. Neste contexto, as células tumorais circulantes (CTC) desenvolven un papel esencial, xa que son as responsables da formación das metástases. As CTC correspóndense con aquelas células tumorais que son liberadas á circulación sanguínea dende o tumor primario, ou ben dende as lesións metastáticas preexistentes. As CTC poden circular no torrente sanguíneo como células individuais, ou ben como pequenas agrupacións oligoclonais de células, chamadas clústeres de CTC. As CTC, tanto individuais como en forma de clústeres de CTC, poden ser detectadas a partir de mostras de sangue periférica ou ‘biopsias líquidas’. A ‘biopsia líquida’, en contraposición á biopsia sólida tisular convencional, constitúe unha técnica minimamente invasiva que aporta información actualizada e a tempo real do estado de progresión do tumor. Na actualidade, é posible detectar e estudar as CTC presentes nas mostras de biopsia líquida, grazas o desenvolvemento e mellora das tecnoloxías de enriquecemento e illamento de CTC. As diversas metodoloxías de enriquecemento de CTC clasifícanse en dous grandes grupos: i) enriquecemento baseado en propiedades biolóxicas; ii) enriquecemento baseado en propiedades físicas. Os sistemas de enriquecemento fundamentados en propiedades biolóxicas baséanse no uso de anticorpos para a detección de diferentes marcadores. Estes marcadores poden ser expresados polas CTC (enriquecemento positivo), ou ben ser expresados polo resto de células da mostra pero non polas CTC (enriquecemento negativo). Entre os sistemas de enriquecemento baseados en características biolóxicas destaca o sistema CellSearch®, o cal se basea na detección do antíxeno EpCAM (Epithelial Cell Adhesion Molecule) para a identificación de CTC. CellSearch® é o único sistema de enriquecemento de CTC aprobado pola FDA (Food and Drug Administration) para o seu uso clínico en cancro de mama, próstata e en cancro colorrectal. Este tipo de sistemas de identificación de CTC posúe unha gran utilidade. Non obstante, a detección de CTC baseada en EpCAM Resumo 33 unicamente permite detectar as CTC de fenotipo epitelial, perdendo aquelas CTC de fenotipo máis mesenquimal. Ademais, CellSearch® non permite a recuperación das CTC identificadas en estado viable, restrinxindo a variedade de análises posteriores. Estas limitacións propiciaron o desenvolvemento dos sistemas de illamento baseados en propiedades físicas e por tanto, independentes da expresión de marcadores por parte das CTC. Estes sistemas inmuno-independentes teñen a capacidade potencial de identificar unha poboación de CTC fenotípicamente máis heteroxénea. Estas metodoloxías están baseadas en diferentes propiedades físicas, como densidade (gradiente de densidade), tamaño (microfiltración), tamaño e deformabilidade (microfluídica), propiedades eléctricas (dielectroforese). Así, por exemplo, a leucoaférese diagnóstica (DLA) é un proceso que emprega a densidade como parámetro para o illamento de CTC. A DLA é un procedemento estándar empregado no ámbito clínico para o illamento das células mononucleares mediante a centrifugación continua do sangue. Recentemente, demostrouse que as CTC poden ser co-illadas conxuntamente coas células mononucleares durante a centrifugación continua, xa que posúen densidades similares. Ademais, en comparación coas mostras de sangue periférica, os produtos de DLA teñen a vantaxe de que derivan de grandes volumes de sangue, maximizando a probabilidade de detección de CTC. Os sistemas microfluídicos permiten identificar as CTC presentes nunha mostra, fundamentándose no seu maior tamaño e menor deformabilidade en comparación co resto de células sanguíneas. Os dispositivos microfluídicos son uns dos máis empregados na actualidade, xa que realizan múltiples pasos dentro dun chip (‘lab on a chip’), permitindo un procesamento continuo que minimiza a perda de mostra durante o proceso. ParsortixTM é un sistema microfluídico semiautomatizado capaz de capturar CTC en diversos tipos de mostras (sangue periférico, ascite, medula ósea). Trátase dun sistema coste-efectivo, versátil e sinxelo que permite recuperar as CTC identificadas nun estado viable, o que posibilita o seu posterior cultivo in vitro, ou caracterización transcriptómica, entre outros. INÉS MARTÍNEZ PENA 34 A información máis elemental que se pode obter a partir dunha mostra de biopsia líquida (aproximadamente 7,5 mL de sangue) é o reconto de CTC. A pesar de ser un parámetro aparentemente sinxelo de determinar, o reconto de CTC posúe un gran valor prognóstico. De feito, estableceuse que a detección por CellSearch® de ≥5 CTC/7,5 mL de sangue periférica constitúe un factor de prognóstico negativo en pacientes con cancro de mama, mentres que aqueles pacientes que posúen <5 CTC/7,5 mL presentan un prognóstico favorable. Por tanto, as CTC constitúen unha ferramenta esencial e de gran potencial para impulsar o desenvolvemento dunha medicina personalizada, orientada á monitorización da progresión tumoral durante a terapia. Pola súa parte, os clústeres de CTC correspóndense cunha subpoboación minoritaria, xa que unicamente representan entre o 1% e o 30% de todas as CTC. A pesar da súa reducida frecuencia en sangue, pénsase que os clústeres de CTC posúen un potencial metastático entre 23 e 50 veces maior que as CTC individuais. De feito, estímase que os clústeres de CTC son os responsables da formación da maioría das metástases e a súa presenza en mostras de biopsia líquida correlaciónase con unha prognose adversa. A maior supervivencia e maior resistencia a anoikis (morte celular por ausencia de anclaxe) dos clústeres de CTC na circulación sanguínea, a súa menor incidencia de apoptose na localización secundaria e a expresión de marcadores de tipo ‘stem’ son algúns dos factores que explican parcialmente o maior potencial metastático dos clústeres de CTC. Ademais, a enumeración dos clústeres de CTC constitúe un factor prognóstico independente, que aporta información adicional o reconto de CTC por si só. A pesar da gran relevancia dos clústeres de CTC no proceso de progresión tumoral e formación de metástases, existe un coñecemento limitado sobre a súa orixe, a súa bioloxía, ou sobre as características fenotípicas que os fan máis metastáticos que as CTC individuais. Isto débese a súa reducida frecuencia no sangue, así como a limitada capacidade das tecnoloxías de illamento actuais para identificar e illar os clústeres de CTC sen alterar a súa integridade física, o que restrinxe de maneira significativa a cantidade de material Resumo 35 biolóxico para o seu estudo. Polo tanto, faise necesario o desenvolvemento de ferramentas alternativas que contribúan a superar estas limitacións. Neste senso, a xeración de modelos experimentais de clústeres de CTC permite incrementar a cantidade de material biolóxico dispoñible para o seu estudo. A presente tese de doutoramento ten por obxectivo principal a realización dun estudo comparativo entre as CTC individuais e os clústeres de CTC, coa fin de profundar na bioloxía dos clústeres de CTC, determinar as características diferenciais que lle aportan maior potencial metastático e, en última instancia contribuír a determinar o seu papel durante a progresión tumoral e a formación das metástases. Para acadar este obxectivo, realizamos o illamento de CTC individuais e clusters de CTC a partir de mostras biopsia líquida procedentes de pacientes con cancro de mama metastático, avaliamos a capacidade prognóstica das CTC e dos clústeres de CTC e optimizamos as condicións de illamento inmuno-independente de clústeres de CTC a partir de diferentes mostras biolóxicas de partida (biopsia líquida, produtos de DLA). Ademais, desenvolvemos modelos de CTC individuais e de clústeres de CTC empregando liñas celulares humanas de cancro de mama e, finalmente, caracterizamos, tanto funcional, como molecularmente os modelos para comprobar a súa idoneidade e o seu potencial como ferramenta de estudo da bioloxía dos clústeres de CTC. Para avaliar a capacidade prognóstica dos clústeres de CTC, recolléronse mostras de sangue periférico dunha cohorte de pacientes (N = 54) con distintos subtipos de cancro de mama metastático. As mostras foron tomadas en diferentes momentos da enfermidade, no momento basal, previo ó inicio do tratamento quimioterápico, e no momento de progresión a este. As mostras de biopsia líquida foron empregadas para o reconto de CTC individuais e de clústeres de CTC mediante o sistema CellSearch®. Esta enumeración permitiunos confirmar que no momento basal o reconto dos clústeres de CTC constitúe un factor prognóstico independente que aporta valor adicional o reconto de CTC individuais, particularmente en pacientes con elevados recontos de CTC individuais. Ademais, a presenza continuada o INÉS MARTÍNEZ PENA 36 longo do tratamento de clústeres de CTC no sangue periférico dos pacientes asóciase cun prognóstico adverso, xa que se correlaciona cunha redución da supervivencia. A detección unicamente de aquelas CTC de fenotipo epitelial (que expresan EpCAM), a imposibilidade de recuperar as CTC detectadas en estado viable e o feito de tratarse dun sistema que non permite a optimización de protocolos para o illamento de clústeres de CTC son algunhas das grandes limitacións do CellSearch®, así como doutros sistemas de enriquecemento inmuno-dependentes. Para superar estas limitacións, levamos a cabo a detección inmuno-independente de CTC e especialmente de clústeres de CTC a partir de mostras de sangue periférico obtidas de pacientes con cancro de mama metastático. O procesamento levouse a cabo no sistema microfluídico ParsortixTM, aplicando un protocolo de separación con menor velocidade de fluxo e menor presión, especialmente deseñado para preservar a integridade dos clústeres de CTC. Ademais, a optimización dunha inmuno-tinguidura sobre células vivas permitiu identificar tanto CTC individuais como clústeres de CTC nestas mostras. A optimización deste fluxo de traballo permite a recuperación de CTC en estado viable que poden ser posteriormente empregadas para diferentes análises que requiren de células vivas, como o cultivo in vitro ou estudos transcripcionais a nivel de célula individual (single-cell analysis). Esta metodoloxía posibilita un estudo máis detallado da bioloxía das CTC e dos clústeres de CTC, permitindo afondar na comprensión do proceso de metástase. Adicionalmente e co propósito de maximizar a frecuencia de detección de CTC e especialmente de clústeres de CTC, tamén se verificou se é factíbel o uso de produtos de DLA para o illamento de CTC e clústeres de CTC no ParsortixTM. A optimización dos parámetros máis relevantes do fluxo de traballo, como a filtración previa da mostra ou a selección do protocolo de separación máis axeitado, permitiunos identificar, illar e recuperar tanto CTC individuais, como clústeres de CTC. A pesar da súa elevada celularidade, demostramos que é factíbel usar mostras de DLA para o illamento de clústeres de CTC. Aínda que serán necesarias Resumo 37 probas adicionais, os datos indican que os produtos DLA constitúen un tipo de mostra con gran potencial para incrementar a capacidade de detección de clústeres de CTC, contribuíndo a superar a limitación da baixa frecuencia de clústeres de CTC detectados nas mostras comúns de sangue periférico (7,5 mL). Por outra banda, desenvolvemos modelos in vitro de CTC individuais e de clústeres de CTC, mediante o uso de liñas celulares humanas de cancro de mama. Utilizáronse dúas liñas celulares diferentes, cada unha pertencente a un subtipo de cancro de mama distinto e con diferentes características fenotípicas. Os modelos de CTC son recursos alternativos axeitados para vencer a limitación da baixa frecuencia de CTC e, sobre todo de clústeres de CTC no sangue dos pacientes. Tras a súa xeración, estes modelos foron caracterizados funcional e molecularmente, coa fin de comprobar a súa idoneidade. Os ensaios funcionais in vitro foron deseñados para simular diferentes etapas do proceso de metástase. Os datos obtidos cos ensaios in vitro foron confirmados polos ensaios funcionais in vivo, realizados sobre embrións de peixe cebra (Danio rerio) e sobre modelos murinos (Mus musculus). A caracterización funcional in vitro e no rato mostrou que os modelos de clústeres de CTC posúen maior capacidade de migración, invasión, supervivencia en circulación, formación de colonias e maior capacidade de colonización para xerar lesións metastáticas, en comparación co modelo de CTC individuais. Ademais, nos peixes inxectados co modelo de clústeres de CTC observouse menor diseminación celular pero maior capacidade de supervivencia e proliferación das células diseminadas en comparación cos inxectados con CTC individuais, suxerindo unha maior capacidade de supervivencia dos clústeres de CTC na circulación do peixe. Estes resultados están apoiados pola caracterización molecular, a cal mostrou que as células diseminadas nos peixes inxectados co modelo de clústeres de CTC posúen unha maior expresión de xenes tipo ‘stem’ (CD44, ITGA6), xenes relacionados coa regulación do ciclo celular e a proliferación (CDK4, E2F4) e unha maior expresión do xene de supervivencia PLAU, en comparación coas células diseminadas nos peixes inxectados con CTC individuais. Concordemente, tamén se observou unha menor expresión do INÉS MARTÍNEZ PENA 38 xene pro-apopótico BAX nas células diseminadas nos xenotrasplante de clústeres de CTC. Non só existe unha concordancia entre as observacións funcionais e moleculares, senón que estes resultados tamén están apoiados polas observacións publicadas na literatura, onde se suxire unha posible maior capacidade de supervivencia dos clústeres de CTC para explicar parte do seu maior potencial metastático. Consecuentemente, o modelo experimental de clústeres de CTC xerado a partir de liñas celulares de cancro de mama representa de xeito fidedigno os trazos fenotípicos dos clústeres de CTC illados do sangue de pacientes, sendo unha ferramenta útil para o estudo da súa bioloxía. O uso dun modelo ortotópico murino de cancro de mama metastático humano, de subtipo triple negativo, permitiunos monitorizar o proceso de progresión tumoral completo, dende a formación do tumor primario ata a aparición das metástases, nun sistema biolóxico complexo. Este modelo tamén posibilitou a captura e illamento de CTC e de clústeres de CTC a partir de mostras de sangue obtidas mediante punción cardíaca. O procesamento destas mostras no sistema microfluídico ParsortixTM e o posterior reconto do número de células illadas permitiunos corroborar que os clústeres de CTC son unha poboación minoritaria. Concretamente, neste modelo os clústeres de CTC representaron aproximadamente o 12,7% de todas as CTC. O procesamento de mostras de sangue derivadas deste modelo ortotópico empregando un kit de enriquecemento negativo posibilitou o illamento de CTC vivas que foron posteriormente postas en cultivo in vitro. Esta liña derivada das CTC do modelo murino, chamada mCTC, medra in vitro en condicións de baixa adherencia, en forma de pequenas agrupacións celulares, polo que podería constituír potencialmente un modelo de clústeres de CTC máis realista, dende o punto de vista fisiolóxico. A caracterización funcional preliminar da liña mCTC indica que posúe unha maior resistencia a estrés fluídico in vitro, así como unha maior capacidade de diseminación e supervivencia no peixe cebra, en comparación co modelo de clústeres de CTC. Os datos transcriptómicos (RNAseq) preliminares mostran a existencia de xenes diferencialmente Resumo 39 expresados entre a liña mCTC e as células control. Aínda que será necesaria unha validación posterior, algúns destes xenes poderían estar implicados nas características metastáticas da liña mCTC. Por tanto, a liña mCTC posúe un elevado potencial como modelo de clústeres de CTC fisioloxicamente máis realista, aínda que se necesitará unha caracterización máis exhaustiva para confirmar esta hipótese. Adicionalmente, as células tumorais illadas no ParsortixTM, tanto CTC individuais como clústeres de CTC foron recuperadas e empregadas para a realización de estudos xenómicos de célula individual coa fin de estudar o perfil mutacional das CTC e dos clústeres de CTC. Os datos xenómicos preliminares mostran que é posible inferir mutacións a partir de CTC individuais e de clústeres e que os clústeres de CTC parecen posuír unha maior variabilidade xenética que as CTC individuais, amosando SNV (Single Nucleotide Variants) específicos que non existen nas CTC individuais. Estes resultados poderían contribuír a explicar a policlonalidade dos clústeres de CTC e a entender o seu maior potencial metastático. Non obstante, requiriranse estudos adicionais para identificar os trazos xenómicos das CTC e dos clústeres de CTC responsables da xeración das metástases. Os clústeres de CTC desenrolan un papel fundamental na diseminación tumoral, xa que son os responsables da formación da maioría das metástases. En conxunto, o presente proxecto de tese confirma a relevancia clínica e o valor prognóstico dos clústeres de CTC e, por tanto, da biopsia líquida, nunha cohorte non seleccionada de pacientes con cancro de mama metastático. A pesar da súa reducida frecuencia en sangue, o reconto de clústeres de CTC aporta valor informativo adicional á enumeración das CTC individuais por si soas. Ademais, a optimización das condicións de enriquecemento de CTC en sistemas de illamento inmuno-independente, así como o uso de mostras de biopsia líquida alternativas a 7,5 mL de sangue periférico, como os produtos de DLA, non só poden permitir maximizar a capacidade de detección de CTC e clústeres de CTC, senón que posibilita potencialmente a captación dunha maior heteroxeneidade nas CTC, representando de xeito máis realista a poboación de CTC presente no sangue dos pacientes. Por outra banda, os modelos de clústeres de CTC desenvoltos no presente INÉS MARTÍNEZ PENA 40 proxecto recapitulan de maneira realista os trazos fenotípicos dos clústeres de CTC illados de paciente con cancro de mama. Por tanto, constitúen ferramentas con un gran potencial para o estudo da súa bioloxía. A combinación das distintas ferramentas e recursos mostrados neste proxecto de tese poden contribuír a superar as limitación impostas pola baixa frecuencia de clústeres de CTC no sangue dos pacientes e posibilita o aumento da dispoñibilidade de mostra inicial, o que permitirá afondar no coñecemento sobre o papel que desenvolven os clústeres de CTC durante a progresión tumoral e a formación de metástases. Este coñecemento resulta esencial para impulsar o avance da medicina personalizada e a oncoloxía de precisión. Resumo 41 INÉS MARTÍNEZ PENA 48 debe a su baja frecuencia en sangre, así como a la capacidad limitada de las tecnologías de aislamiento actuales para identificar y aislar los clústeres de CTC sin alterar su integridad física, lo que restringe significativamente la cantidad de material biológico inicial. Por tanto, se genera la necesidad de desarrollar herramientas alternativas que contribuyan a superar estas limitaciones. En este sentido, la generación in vitro de modelos de clústeres de CTC permite incrementar la cantidad de material biológico disponible para su estudio. El objetivo principal de la presente tesis doctoral es realizar un estudio comparativo entre las CTC individuales y los clústeres de CTC, con el fin de profundizar en la biología de los clústeres de CTC, determinar las características diferenciales que les aportan un mayor potencial metastásico y, en última instancia, contribuir a determinar su papel durante la progresión tumoral y la formación de metástasis. Para alcanzar este objetivo, se realizó el aislamiento de CTC individuales y clústeres de CTC a partir de muestras de biopsia líquida de pacientes con cáncer de mama metastásico, se evaluó la capacidad pronóstica de las CTC individuales y de los clústeres de CTC y se optimizaron las condiciones de aislamiento inmuno-independiente de los clústeres de CTC a partir de diferentes tipos de muestras biológicas (biopsia líquida, productos DLA). Además, se desarrollaron modelos in vitro de CTC individuales y clústeres de CTC, utilizando líneas celulares humanas de cáncer de mama y, finalmente, se llevó a cabo la caracterización, tanto funcional como molecular, de los modelos de CTC individuales y clústeres de CTC para comprobar su idoneidad y su potencial como herramienta para el estudio de la biología de los clústeres de CTC. Con el fin de evaluar la capacidad pronóstica de los clústeres de CTC, se recolectaron muestras de sangre periférica de una cohorte de pacientes (N = 54) con diferentes subtipos de cáncer de mama metastásico. Se tomaron muestras en diferentes momentos de la enfermedad, en el momento basal, previo al inicio del tratamiento quimioterápico, y en el momento de progresión a este. Las muestras de biopsia líquida se utilizaron para el recuento de CTC individuales y de clústeres de CTC, usando el sistema Resumen 49 CellSearch®. Este recuento nos permitió confirmar que en el momento basal la enumeración de los clústeres de CTC constituye un factor pronóstico independiente que aporta valor informativo adicional al recuento de CTC individuales, particularmente en pacientes con elevados recuentos de CTC individuales. Además, la presencia continuada a lo largo del tratamiento de clústeres de CTC en la sangre periférica de los pacientes se asocia con un pronóstico adverso, ya que se correlaciona con una reducción de la supervivencia. La detección de aquellas CTC de fenotipo epitelial que expresan EpCAM, la incapacidad de recuperar las CTCs detectadas en un estado viable, así como el hecho de tratarse de un sistema que no permite la optimización de protocolos para el aislamiento de clústeres de CTC, son algunas de las principales limitaciones del CellSearch®, así como de otros sistemas de enriquecimiento inmuno-dependientes. Para superar esta limitación, llevamos a cabo la detección inmuno-independiente de CTC y especialmente de clústeres de CTC, a partir de muestras de sangre periférica derivadas de pacientes con cáncer de mama metastásico. El procesamiento se realizó en el sistema microfluídico ParsortixTM, aplicando un protocolo de separación con menor flujo y menor presión, especialmente diseñado para preservar la integridad de los clústeres de CTC. Además, la optimización de una inmunotinción sobre célula viva hizo posible identificar tanto CTC individuales, como de clústeres de CTC en estas muestras. La optimización de este flujo de trabajo posibilita la recuperación de CTC viables que pueden ser utilizadas posteriormente para diferentes análisis que requieran de células vivas, como el cultivo in vitro o análisis transcipcionales a nivel de célula única. Esta metodología posibilita la realización de un estudio más detallado de la biología de las CTC y de los clústeres de CTC, lo que permitirá una comprensión más profunda del proceso de metástasis. Adicionalmente y con el propósito de maximizar la frecuencia de detección de CTC y de clústeres de CTC, también se testó el posible uso de muestras de DLA para el aislamiento de CTC y de clústeres de CTC en el ParsortixTM. La optimización de INÉS MARTÍNEZ PENA 50 los parámetros más relevantes del flujo de trabajo, como la filtración previa de la muestra o la selección del protocolo de separación más apropiado, nos permitió identificar, aislar y recuperar tanto CTC individuales, como clústeres de CTC. A pesar de su elevada celularidad, ha sido posible demostrar que es factible utilizar muestras de DLA para el aislamiento de clústeres de CTC. Aunque será necesaria la realización de pruebas adicionales, los datos indican que los productos de DLA constituyen un tipo de muestra con gran potencial para incrementar la capacidad de detectar clústeres de CTC, contribuyendo a superar la limitación de la baja frecuencia de clústeres de CTC en muestras de sangre periférica (7,5 mL). Por otra parte, se han desarrollado modelos in vitro de CTC individuales y clústeres de CTC, mediante el uso de líneas celulares humanas de cáncer de mama. Se utilizaron dos líneas celulares diferentes, cada una perteneciente a un subtipo de cáncer de mama distinto y con diferentes características fenotípicas. Los modelos de CTC son recursos alternativos adecuados para superar la limitación de la baja frecuencia de CTC y especialmente de clústeres de CTC en la sangre de los pacientes. Tras su generación, estos modelos fueron caracterizados funcional y molecularmente, con el fin de comprobar su idoneidad. Se diseñaron diferentes ensayos funcionales in vitro para simular las distintas etapas del proceso de metástasis. Los datos obtenidos con los ensayos in vitro fueron respaldados mediante ensayos funcionales in vivo, realizados en embriones de pez cebra (Danio rerio) y en modelos murinos (Mus musculus). La caracterización funcional in vitro y en ratón mostró que los modelos de clústeres de CTC tienen mayor capacidad de migración, invasión, supervivencia en circulación, formación de colonias y mayor capacidad de colonización para generar lesiones metastásicas, en comparación con el modelo de CTC individuales. Además, en los peces inyectados con el modelo de clústeres de CTC se observó una menor diseminación celular hacia la cola pero una mayor capacidad de supervivencia y proliferación de las células diseminadas en comparación con los inyectados con CTC individuales, lo que sugiere una mayor capacidad de supervivencia de los clústeres de CTC en la Resumen 51 circulación de los peces. Estos resultados están respaldados por la caracterización molecular, la cual mostró que las células diseminadas en peces inyectados con el modelo de clústeres de CTC poseen mayor expresión de genes de tipo ‘stem’ (CD44, ITGA6), genes relacionados con la regulación del ciclo celular y con proliferación (CDK4, E2F4) y mayor expresión del gen de supervivencia PLAU, en comparación con las células diseminadas en peces inyectados con CTC individuales. Asimismo, también se observó una menor expresión del gen proapoptótico BAX en células diseminadas en los xenotrasplantes de clústeres de CTC. No solo existe una concordancia entre las observaciones funcionales y moleculares, sino que estos resultados también están respaldados por observaciones publicadas en la literatura, donde se sugiere una mayor capacidad de supervivencia de los clústeres de CTC para explicar parcialmente su mayor potencial metastásico. Consecuentemente, el modelo de clústeres de CTC generado a partir de líneas celulares de cáncer de mama es representativo de las características fenotípicas de los clústeres de CTC aislados de la sangre de paciente, siendo una herramienta útil para el estudio de su biología. El uso de un modelo ortotópico murino de cáncer de mama metastásico humano, de subtipo triple negativo, permitió monitorizar el proceso de progresión tumoral completo, desde la formación del tumor primario hasta la aparición de metástasis en un sistema biológico complejo. Este modelo también permitió capturar y aislar CTC y clústeres de CTC, a partir de muestras de sangre obtenidas por punción cardíaca. El procesamiento de estas muestras en el sistema de microfluídica ParsortixTM y el posterior recuento del número de células aisladas nos permitió corroborar que los clústeres de CTC son una población minoritaria. Específicamente, los clústeres de CTC representaron aproximadamente el 12,7% de todas las CTC en este modelo. El procesamiento de muestras de sangre derivadas de este modelo ortotópico mediante un kit de enriquecimiento negativo permitió el aislamiento de CTC vivas que posteriormente fueron puestas en cultivo in vitro. Esta línea derivada de las CTC del modelo murino, denominada mCTC, crece in vitro en condiciones de baja INÉS MARTÍNEZ PENA 52 adherencia, en forma de pequeños agregados celulares, por lo que potencialmente podría constituir un modelo de clústeres de CTC más realista, desde un punto de vista fisiológico. La caracterización funcional de la línea mCTC indica que esta línea posee una mayor resistencia al estrés fluídico in vitro, así como una mayor capacidad de diseminación y supervivencia en el pez cebra, en comparación con el modelo in vitro de clústeres de CTC. Los datos transcriptómicos preliminares (RNAseq) muestran la existencia de genes diferencialmente expresados entre la línea mCTC y las células control. Aunque será necesaria una validación posterior, algunos de estos genes podrían estar implicados en las características metastásicas de la línea mCTC. Por lo tanto, la línea mCTC tiene un alto potencial como modelo de clústeres de CTC fisiológicamente más realista, aunque se necesitará una caracterización más completa para confirmar esta hipótesis. Además, las células detectadas, tanto CTC individuales como clústeres de CTC, fueron recuperadas y posteriormente empleadas para la realización de estudios genómicos de célula individual, con el fin de estudiar el perfil mutacional de las CTC y de los clústeres de CTC. Los datos genómicos preliminares mostraron que es posible inferir mutaciones a partir de CTC individuales y de clústeres de CTC y que los clústeres de CTC parecen poseer una mayor variabilidad genética que las CTC individuales, mostrando SNV (Single Nucleotide Variants) específicas que no existen en las CTC individuales. Estos resultados podrían contribuir a explicar la policlonalidad de los clústeres de CTC y a entender su mayor potencial metastásico. Sin embargo, serán necesarios estudios adicionales para identificar los rasgos genómicos de las CTC y los clústeres de CTC responsables de la generación de metástasis. Los clústeres de CTC desempeñan un papel clave en la diseminación tumoral, ya que son los responsables de la formación de la mayoría de las metástasis. En conjunto, el presente proyecto confirma la relevancia clínica y el valor pronóstico de los clústeres de CTC y, por lo tanto, de la biopsia líquida, en una cohorte no seleccionada de pacientes con cáncer de mama metastásico. A pesar de su baja frecuencia en sangre, el recuento de clústeres de CTC proporciona información adicional Resumen 53 a la enumeración de CTC individuales por sí sola. Además, la optimización de las condiciones de enriquecimiento de CTC en sistemas de aislamiento inmuno-independientes, así como el uso de muestras de biopsia líquida alternativas a 7,5 mL de sangre periférica, como los productos DLA, no solo pueden maximizar la capacidad de detección de CTC y clústeres de CTC, si no que permiten potencialmente la identificación de una mayor heterogeneidad en las CTC, representando de forma más realista la población de CTC presentes en la sangre de los pacientes. Por otro lado, los modelos de clústeres de CTC desarrollados en el presente proyecto recapitulan de manera fidedigna las características fenotípicas de los clústeres de CTC aislados de pacientes con cáncer de mama. Por tanto, constituyen herramientas con gran potencial para el estudio de su biología. La combinación de las diferentes herramientas y recursos desarrollados en este proyecto de tesis pueden contribuir a superar las limitaciones impuestas por la baja frecuencia de clústeres de CTC en la sangre de los pacientes e incrementan la disponibilidad de muestra inicial, lo que posibilita un mayor estudio y comprensión sobre el papel de los clústeres de CTC durante la progresión tumoral y la formación de metástasis. Este conocimiento es esencial para impulsar el avance de la medicina personalizada y la oncología de precisión. INÉS MARTÍNEZ PENA 54 55 SUMMARY INÉS MARTÍNEZ PENA 56 Summary 57 SUMMARY The term ‘cancer’ includes a group of genetically complex diseases. Breast cancer (BC) is the most common type of malignancy in women and is the second leading cause of cancerrelated deaths, after lung and bronchus cancer. Although there are different known genetic and environmental factors that increase the risk of developing BC, the aetiology of most BC cases is still unknown. BC is a heterogeneous group of diseases, which show differences in terms of incidence, prognosis, progression, or response to therapy. There are four different subtypes of BC, based on the molecular diagnosis, according to their expression of the progesterone receptor (PR), estrogen receptor (ER) and epidermal growth factor receptor 2 (HER2): luminal A (PR+/ER+/HER2-), luminal B (PR+/-/ER+/HER2+/-), HER2 overexpression (PR-/ER-/HER2+), and 'basal-like' or triplenegative (PR-/ER-/HER2-). The majority of cancer-related deaths are due to the development of a disseminated state or metastasis. Metastasis is a process by which tumour cells spread to secondary locations far away from the primary tumour, hence becoming a systemic disease. Metastasis is a complex, multi-step process that includes different steps: migration and invasion from the primary tumour, intravasation, survival into the bloodstream, extravasation in a secondary site, survival and proliferation generating a new tumour lesion or metastasis. Despite the great advances made in the prevention, diagnosis and treatment of BC, metastasis remains an incurable disease. In fact, between 5% and 10% of patients will show metastasis at the time of diagnosis, and approximately 30% of BC patients will eventually progress to a disseminated state of the disease throughout treatment. Therefore, increasing knowledge and understanding about the metastatic process is essential to improve patient care. In this context, circulating tumour cells (CTCs) play an essential role, as they are the responsible for the formation of metastases. CTCs are those tumour cells that are released into the bloodstream from the INÉS MARTÍNEZ PENA 64 level to evaluate the mutational profile of single CTCs and CTC clusters. Preliminary genomic data showed that it was possible to infer mutations from single CTCs and CTC clusters and that CTC clusters seemed to have a higher genetic variability than individual CTCs, showing specific SNVs (Single Nucleotide Variants) that were not found in individual CTCs. These results could help to explain the possible polyclonal origin of CTC clusters and to better understand their higher metastatic potential. However, further studies will be required to identify the genomic traits of those CTCs and CTC clusters responsible for the seeding of metastasis. CTCs clusters play a key role in tumour dissemination, as they are responsible for the formation of the majority of metastases. Taken together, this project confirmed the clinical relevance and the prognostic value of CTC clusters and of liquid biopsy in an unselected cohort of patients with metastatic BC. Despite their low frequency in blood, the enumeration of CTC clusters provided additional and valuable information to the enumeration of individual CTCs alone. Furthermore, the optimization of CTC and especially CTC cluster enrichment conditions in the ParsortixTM, as well as the use of liquid biopsy samples, such as DLA products, as alternatives to 7.5 mL of peripheral blood samples not only could maximize the detection capacity of CTCs and CTC clusters but also could increase the ability to detect a higher heterogeneity within isolated CTCs. This greater variability would provide a more realistic landscape of the CTC population existing in the blood of BC patients. On the other hand, the CTC cluster models developed in this project recapitulated the phenotypic features of CTC clusters isolated from BC patients. Therefore, these models are tools with enormous potential for the study of their biology. The combination of the different tools and resources presented in this thesis project can contribute to overcome the restrictions derived from the low frequency of CTC clusters in the blood of patients and would increase the availability of tumour material, which would allow a deeper understanding of the role of CTC clusters during tumour progression and metastasis formation. This Summary 65 knowledge is essential to prompt the advance of personalized medicine and precision oncology. INÉS MARTÍNEZ PENA 66 Introduction 67 INTRODUCTION INÉS MARTÍNEZ PENA 68 Introduction 69 INTRODUCTION 1. BREAST CANCER 1.1. Epidemiology and aetiology Breast cancer (BC) corresponds to the tumour with the highest incidence rate, and it is the second leading cause of cancer-related deaths in women, after lung and bronchus cancer 1. With more than 2 million new cases of BC diagnosed per year worldwide 2, it currently represents about 30% of all female cancers 2. The relevant advances in prevention, early diagnosis and treatment allowed the sustention of a declining trend in the BC death rate since 1989 1. Nevertheless, it is estimated that approximately 627,000 women died from BC during the year 2018, which is equivalent to 15% of all female cancer deaths (WHO) 3. Taking into consideration that breast carcinoma is a genetically complex disease, several environmental factors have been reported to be associated with an increased risk of developing BC. These risk factors comprise female gender, increased patient age, early menarche age, late menopause, late age at first childbirth, familial occurrence of BC at a young age, increased mammographic breast density, presence of benign breast disease, exposure to chest radiation, hormonal and alcohol intake as well as obesity 3–5. Furthermore, it was widely demonstrated that genetic mutations, such as mutations in BRCA1 and BRCA2 genes are associated with an increased risk of BC. Nevertheless, they only represent a minority of BC cases 6. This biological complexity contributes to explain the fact that the aetiology in the majority of BC cases is unknown. BC is a highly heterogeneous group of diseases, which involves various BC subtypes with phenotypic differences. Consequently, there are variations in terms of incidence, prognosis, progression and response to therapy between them 7–9. INÉS MARTÍNEZ PENA 70 Conventionally, BC subtypes were determined by using classical immunohistochemistry (IHC) based on different biomarkers, such as estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2), combined with classic clinical and histopathological parameters, which include cell proliferation index (Ki67), tumour size, tumour grade and nodal involvement. These criteria are usually used in the clinic for patient prognosis and management 10. 1.2. Subclassification of BC tumours Clinically, the tumour IHC status stratifies BC tumours into three main groups: luminal, HER2 overexpression subtype and triple-negative breast cancer (TNBC) 8. Nevertheless, with the development of the molecular techniques and the tissue microarray technology at the beginning of the XXI century, it was possible to apply gene expression profiling (GEP) to assess BC heterogeneity in a more accurate manner 9,10. One example of the combined approach of IHC status and GEP was the study performed by Sørlie and colleagues in 2001, in which they described a ‘molecular portrait’ of BC by using 456 cDNA clones. According to their research, BC tumours were classified into five subtypes with diverse clinical outcomes: luminal A, luminal B, HER2 overexpression, basal (or TNBC) and normallike tumours 11. Each subtype proposed by Sørlie was associated with a specific IHC status, except for normal-like, which shares IHC status with luminal A differing only in their expression pattern, with the normal-like resembling the normal breast profiling 10. For this reason, the results of Sørlie set the standard for BC tumour stratification. Nevertheless, there are alternative classifications, such as the one proposed by Fan and colleagues, who classified BC in 4 different categories, where the normal-like subtype was not identified according to Sørlie’s results 10. The luminal group is divided into two different subtypes, luminal A and luminal B. Luminal A is positive for both ER and PR but negative for HER2 marker. It also has a low proliferation index (Ki67 < 14%). On the other hand, luminal B is characterised by positive expression of ER, variable expression of PR and high Introduction 71 expression of Ki67 (≥ 14%). The luminal group is the most frequent subtype of BC and it is associated with a good prognosis 12,13. The HER2 overexpression subtype shows the negative expression of hormonal receptors, high expression of HER2 as well as high proliferation level (Ki67 ≥ 14%)7. Basal like or TNBC is characterised by the absence of hormonal receptors and HER2 expression and it shows an elevated expression of Ki67 (≥ 14%). TNBC is a highly aggressive BC subtype, and it lacks a specific treatment. Therefore, TNBC is linked to an adverse disease outcome 14 (table 1). Table 1. BC subtype classification according to IHC status. Luminal A Luminal B HER2 overexpression Basal-like (or TNBC) IHC status ER +++ + - - PR ++ (>20%) +/- - - HER2 - -/+ +++ - Ki67 (%) Low (<14%) High (>14%) High High Outcome Good Intermediate Poor Poor Abbreviations: IHC immunohistochemistry; ER Estrogen Receptor; PR Progesterone Receptor; HER2 Human Epidermal growth factor Receptor 2. 1.3. BC stages The Tumour-Node-Metastasis (TNM) staging system is the most effective method to divide tumour progression into different categories of clinical usefulness for patient treatment recommendations and prognosis, although there are several other cancer-staging approaches currently available 15. The TNM system was originally proposed by Pierre Denoix in 1943-1952 and, in 1958 the first recommendations of breast and larynx cancer staging were published 16. Afterwards, since the origin of the American Joined Committee on Cancer (AJCC) in 1977, it has been developed a staging system grounded in the TNM method. The staging approach suggested by AJCC has evolved throughout time, updating this system as new knowledge regarding cancer-staging and prognostic factors are published 17. INÉS MARTÍNEZ PENA 72 TNM cancer-staging system is based on anatomic traits, such as tumour size (T), nodal status (N) and metastases (M) 17. The TNM staging method classifies the primary tumour (T) in different categories (TX, Tis, T0-T4), according to the existence, size and extent of the primary tumour. The lymph node status (N) is divided into five groups (NX, N0-N3), determined by the presence and the regional extent of nodal invasion. The metastatic state (M) is subdivided into two categories, M0 and M1, which represent the presence or absence of metastatic lesions, respectively 18,19 (table 2). Table 2. TNM status (based on AJCC Cancer Staging Manual). Primary Tumour (P) TX Primary tumour cannot be assessed T0 No evidence of primary tumour Tis Carcinoma in situ (early cancer that has not spread to neighbouring tissue) T1 Tumour ≤ 20mm in greatest dimension T2 Tumour > 20mm but ≤ 50mm in greatest dimension T3 Tumour > 50mm in greatest dimension T4 Tumour of any size with direct extension to the chest wall and/or to the skin (ulceration or macroscopic nodules); invasion of the dermis alone does not qualify as T4 Lymph Nodes (N) NX Regional lymph nodes cannot be evaluated N0 No regional lymph node involvement N1 Metastases to movable ipsilateral level I (low-axilla) and II (midaxilla) axillary lymph node(s) N2 • Metastases in ipsilateral level I (low-axilla) and II (midaxilla) axillary lymph nodes that are clinically fixed or matted … or … • Metastases in ipsilateral internal mammary lymph nodes in the absence of axillary lymph node metastases N3 • Metastases in ipsilateral infraclavicular (level III axillary) lymph node(s) with or without level I (low-axilla) and II (mid-axilla) axillary lymph node involvement … or … • Metastases in ipsilateral internal mammary lymph node(s) with level I (low-axilla) and II (mid-axilla) axillary lymph node metastases … or … Introduction 73 • Metastases in ipsilateral supraclavicular lymph node(s) with or without axillary or internal mammary lymph node involvement Metastases (M) M0 No distant metastasis M1 Distant metastasis detected In 2017, it was published the 8th Edition of the AJCC TNM classification of malignant tumours, which maintains its basis on the traditional TNM system but it incorporates biomarkers into the anatomical cancer-staging method, such as ER, PR, HER2, histologic grade and commercial multigene assays 17,19. In the 8th edition of AJCC Staging Manual, patients are clinically classified by the classical TNM anatomic system as well as by the information provided by the biomarkers. This information generates a Clinical Prognostic Stage Group, which should be determined by initial assessment before any systemic therapy. 2. METASTASIS Metastasis is a complex and multi-step process by which new tumour lesions are generated in locations far from the primary tumour, hence becoming a systemic disease 20. Metastasis is responsible for 90% of all cancer-related deaths 21. Despite the relevant advances that have been made in prevention, diagnosis and treatment, BC metastasis continues to be an incurable disease. In fact, between 5% and 10% of BC patients show metastasis at the time of diagnosis and about 30% of all BC patients will develop metastasis throughout the time of treatment 22. The metastatic process can be divided into different stages of development: individual or collective migration and invasion; intravasation and survival into the bloodstream; extravasation; colonization, survival and growth in a distant tissue/organ (secondary site) 23 (figure 1). INÉS MARTÍNEZ PENA 80 cells against cytokeratins (CKs) 8, 18 and 19, the leukocyte marker CD45, and the nuclear staining DAPI (4’,6-diamidino-2phenylindole positivity). Hence, a CTC is defined as a cell EpCAM+/CKs+/DAPI+/CD45-, with a minimum size of 4x4 µm2 76,78. Sample processing in the CellSearch® is mainly used for CTC enumeration, and it allows downstream genomic analysis of the captured cells. However, CellSearch® CTC detection uses whole blood collected in tubes with preservative reagents, which do not allow downstream transcriptomic analysis, neither later recovery of cells in a viable state to perform functional characterization, among other downstream procedures 78. The CellSearch® also has the limitation of detecting only CTCs with an epithelial phenotype. Hence, it is not capable of detecting CTCs with a more undifferentiated (or stem) mesenchymal traits where low or no EpCAM expression is found. Despite its limitations, CellSearch® has great clinical utility, as a correlation between CTC enumeration and patient outcome was found. For instance, an enumeration of ≥5 CTC/7.5mL of blood correlates with a worse prognosis in metastatic breast and prostate cancer patients, while a cut-off of ≥3 CTC/7.5mL correlates with a negative prognosis in metastatic colorectal cancer patients 79–81. Unlike CellSearch®, other immune-dependent enrichment methods allow the recovery of CTCs in a viable state, permitting gene expression studies. For instance, CellCollector® (Gilupi) is a positive CTC enrichment methodology based on the detection of EpCAM. CellCollector® is a CE certified medical device that isolates CTCs in vivo, directly from the blood of the patient. It consists of a medical stainless steel wire functionalized with antibodies against EpCAM that is placed as a catheter. CellCollector® is able to detect CTCs even in patients with early BC and with no diagnosed metastasis. However, it requires a manual screening for the identification of CTCs 78,82. CELLectionTM Epithelial Enrich DynabeadsTM (ThermoFisher Scientific) is also a positive CTC enrichment kit based on the expression of EpCAM. It can be used with whole blood, bone marrow, or with peripheral blood mononuclear cell (PBMC) samples. DynabeadsTM are magnetic beads coated with Introduction 81 antibodies against EpCAM that are directly added to the sample to allow interaction with CTCs for later separation with a magnet. Moreover, DynabeadsTM can be customised for detecting different antigens, such as cell surface proteins or ECM components 83. There are many technologies based on EpCAM expression for CTC isolation. However, the suitability of EpCAM as a marker for CTC enrichment is currently under discussion, as the role of EpCAM in tumour progression is not fully understood. Besides, it has been reported that EpCAM can have a transient expression due to EMT 84. Consequently, alternative immunebased methods are being developed for detecting CTCs. The AdnaTest is a kit specifically designed for BC or prostate cancer. It is formed by magnetic beads functionalized with a cocktail of antibodies against different markers (EpCAM, HER2, CA 15-3 (MUC1), and optionally against PR and ER). This system uses whole blood for CTC detection and allows gene expression analyses by quantitative multiplexed PCR (qPCR). Thus, AdnaTest can provide useful information regarding the progression of the disease during therapy, as well as the genetic origin of the isolated CTCs 85–87. 3.2.1.2. Negative enrichment Moreover, negative immune-dependent enrichment techniques have been developed to overcome the limitations of those EpCAM-based technologies 88. RossetteSepTM (EasySepTM Direct Human CTC Enrichment Kit, Stemcell Technologies) is a good example and a widely used negative CTC enrichment system. This kit is used over whole blood samples and it contains a cocktail of antibodies that specifically recognise antigens present in blood cells. The density gradient centrifugation allows the separation of the blood components: contaminating cells (settle down to the bottom), plasma (the upper part), and the CTCs, which are located in the interface between the plasma and the lower part of contaminating cells. CTCs recovered with this system are viable and can be used for downstream assays, such as INÉS MARTÍNEZ PENA 82 functional or molecular phenotyping analyses, as it has been previously demonstrated 89,90. 3.2.2. CTC enrichment based on biophysical properties Biophysical-based CTC enrichment technologies are generating great interest as they do not rely on the identification of cell surface markers. These can potentially detect and isolate a more heterogeneous CTC population. These label-free enrichment technologies can be based on different physical properties. The main physical properties used by these systems are density (density gradient), size (microfiltration), different size and deformability (microfluidics), or electrical charges (dielectrophoresis) 91. 3.2.2.1. Density gradient Density-based methods were one of the first technologies applied for CTC isolation. These techniques have the advantage of fast sample processing time. However, they also have limitations, as they show low specificity that generates enrichment of CTCs with low purity 76,78. Diagnostic Leukapheresis (DLA) is a good example of a density-based method for CTC isolation that can overcome the limitation of low CTC frequency in peripheral blood samples. DLA is a standard procedure used in the clinic to collect Mononuclear Cells (MNCs) due to continuous centrifugation of a large volume of blood 92. It has been recently proposed that CTCs could be co-isolated along with MNCs during DLA centrifugation, as they have similar densities 93. Moreover, DLA screens large volumes of blood, thus increasing the likelihood of CTC detection in comparison with peripheral blood samples 92,94. It is estimated that the processing of DLA in the CellSearch® system increases between 0 and 32 fold the ability to detect CTCs in comparison with peripheral blood samples (7.5 mL) 95. It has been reported that DLA allows the enrichment of CTCs in a viable state that can be later used for downstream characterization 94. Besides, CTCs isolated from cryopreserved DLA products Introduction 83 maintain their viability, which makes sample processing more easily manageable 92. 3.2.2.2. Microfiltration Microfiltration technologies were developed in the 1960s, based on the higher size of CTCs in comparison with blood cells 91. Nowadays, there are different microfiltration-based devices for CTC isolation. For instance, ISET® (Isolation by Size of Epithelial Tumour cells) (Rarecell, Paris) was the first sizebased device for CTC detection. It uses a calibrated track-etched polycarbonate membrane with 8 µm cylindrical pores for CTC enrichment from diluted peripheral blood samples 91,96. 3.2.2.3. Microfluidics Microfluidic devices are based on the higher size and lower deformability of CTCs compared with blood cells. Currently, microfluidic CTC enrichment techniques are one of the most used immune-independent methods as they perform different steps of sample processing within a chip. Thus, they integrate the ‘lab on a chip’ concept and allow continuous sample processing to reduce cell loss. The ParsortixTM Cell Separation System (Angle plc) is a semiautomated device capable of isolating CTCs from different body fluids, such as blood, bone marrow or ascites 97. It uses a disposable cassette with serpentine channels that finish in a 6.5 µm gap to capture CTCs, while the rest of the blood cells flow through (figure 2). After blood enrichment, cells can be stained by in-cassette immunofluorescence for CTC identification, and/or recovered by an optional harvesting step. Harvested CTCs are viable, so they can be used to in vitro culture, or to perform functional and molecular characterization 98. Its costeffectiveness, versatility, simplicity, and label-free technology provide great potential to ParsortixTM and boost its use for a deeper understanding of CTC biology 99,100. INÉS MARTÍNEZ PENA 84 Figure 2. Overview of a ParsortixTM cassette. The sample flows through serpentine channels. These channels have a stair-like silhouette that finishes in a 6.5 µm gap in which CTCs are captured, while the rest of blood cells flow through. Red arrows indicate the direction of the flow inside the cassette. Image used with permission of Angle plc. 3.2.2.4. Dielectrophoresis Dielectrophoresis (DEP) separates CTCs based on their differential movement in the presence of an electric field, as they are charged particles. CTCs have a differential surface charge in comparison with other cells, thus an electric flow can separate them from the rest of the cells of the sample 101. DEPArrayTM (Menarini Silicon Biosystems) system combines microfluidics with DEP to detect and isolate individual CTCs based on their electrophoretic properties 76. CTCs are identified and selected by an image-based analysis. Then, selected cells are isolated inside dielectrophoretic cages and finally, cells are individually recovered for downstream analysis 102. DEPArrayTM is normally used in combination with other separation techniques, as a secondary isolation method to eliminate the background contaminating cells and increase the purity of the individually isolated CTCs 76,78. Introduction 85 4. CIRCULATING TUMOUR CELL CLUSTERS (CTC CLUSTERS) CTC clusters (a.k.a circulating tumour microemboli, circulating micrometastasis, circulating tumour aggregates, or tumour cell clumps) are defined as a small group of two or more (>100) tumour cells that travel together through the bloodstream 103. CTC clusters can be made of tumour cells exclusively (homotypic clusters), or be associated with other non-tumour cells, such as platelets, neutrophils, or PBMCs (heterotypic clusters) 104. It has been demonstrated that the presence of CTC clusters in the peripheral blood of cancer patients is associated with an adverse outcome. CTC clusters have a higher metastatic potential than individual CTCs 103,104. CTC clusters were firstly described in 1858 by Rudolph Virchow, a German physician known for being the founder of cellular pathology. Rudolf Virchow postulated the arrest of tumour microemboli in the vasculature as the origin of metastases 105. However, it was not until 1954 that the first studies regarding the relevance of CTC clusters in tumour progression were published. A work performed by Watanabe demonstrated that CTC clusters were more efficient in metastatic seeding than individual CTCs when intravenously injected in mice at equal numbers 106,107. Despite knowing the existence of CTC clusters for over a century, the origin of CTC clusters, or the traits that make them more metastatic are still in the process of being elucidated. 4.1. Origin of CTC clusters The origin of CTC clusters is still under discussion and two main hypotheses have been proposed. One possibility is that CTC clusters could be directly separated from the primary tumour and/or from metastasis. Another possibility is that CTC clusters originate from the aggregation and/or proliferation of individual CTCs within the bloodstream 107 (figure 3). INÉS MARTÍNEZ PENA 86 Figure 3. Overview of the possible origin of CTC clusters. CTC clusters can be generated by the direct detachment of a group of cells from the primary tumour or a metastatic lesion. Alternatively, CTC clusters could derive from the aggregation and/or proliferation of individual CTCs within the bloodstream. Although there is limited evidence about the origin of CTC clusters in BC, some studies point to a direct detachment of a group of cells from the tumour as the origin of CTC clusters and exclude the intravascular aggregation of cells. In fact, the injection of two different fluorescently-labelled tumour cells in the mammary fat pad of opposite flanks of mice demonstrated that BC metastases arise from a collective migration and invasion of a group of cells, and discarded the intravascular aggregation of individual tumour cells 45,108. Moreover, an in vitro assay performed in a platform that mimics the conditions of the bloodstream has shown that this environment does not favour the intravascular aggregation and/or proliferation of single CTCs 107. In line with this, molecular mechanisms associated with CTC Introduction 87 cluster formation have been reported. For instance, it has been described that plakoglobin (JUP) is essential for CTC cluster formation. In fact, the presence of high expression levels of plakoglobin correlated with an adverse outcome of BC patients and plakoglobin knockdown in murine models reduced CTC cluster formation and metastatic incidence 45. Plakoglobin is a relevant component of adherens junctions and desmosomes. Hence, plakoglobin can play a fundamental role in keeping the cell-to-cell adhesion of tumour cells within the CTC cluster 45. The intermediate filament keratin 14 (KRT14) was also found to be associated with CTC clusters. Like plakoglobin, keratin 14 is also a protein related to cell junctions that regulates cell-cell and cell-ECM adhesion. It has been demonstrated that keratin 14 is necessary for collective invasion and dissemination but not for the dissemination of individual cells 108. On the contrary, a recent work performed with murine Patient-Derived Xenograft (PDX) models of BC showed the feasibility of CTC cluster formation by adhesion of individual cells within the bloodstream. The authors reported that homophilic CD44-CD44 interactions allow cell aggregation to form CTC clusters in circulation, and that CD44 depletion prevented tumour cells from aggregation. Thus CD44 is required for CTC cluster formation by intravascular cell aggregation 109. In line with this, previous research showed that the primary attachment to endothelia allowed tumour cells to form cellular aggregates during the initial steps of metastasis development 110. Moreover, it has been proposed the ‘cell jamming’ as another alternative mechanism for CTC cluster formation during tumour spread. The ‘cell jamming’ model proposes that the density of the ECM determines the type of invasion (individual or collective) performed by mesenchymal tumour cells. According to this, high density ECM induces cell-cell interactions and collective migration, thus favouring CTC cluster formation 111. Consequently, the currently available data suggest a possible combination of all of the above-mentioned mechanisms to INÉS MARTÍNEZ PENA 88 explain the formation of CTC clusters. However, further studies are required to shed light on the origin of CTC clusters. 4.2. Metastatic traits of CTC clusters CTC clusters only represent 1-30% of all CTCs 104. Despite being a minor subpopulation of CTCs, it has been reported that CTC clusters have a 23-50-fold higher metastatic potential than individual CTCs. Preclinical studies have estimated that CTC clusters are the responsible for 50-97% of metastasis 45,108. However, the characteristics and mechanisms that make CTC clusters more metastatic than individual CTCs are still not fully understood. In this regard, a study in melanoma demonstrated that tumour cells with low metastatic potential can increase their metastatic ability when they interact and group with tumour cells with a high metastatic capacity 112. On the contrary, a previous study showed that combining tumour cells with different metastatic potentials within a cluster did not increase the ability to metastasize of the low-metastatic tumour cells. This study showed that the majority of the metastases were developed by those tumour cells with high metastatic potential 113. Therefore, further work is required to better understand the heterogeneity within CTC clusters and how tumour cells interact with each other in the cluster. Related to this, it should be highlighted that cell heterogeneity can play an important role in tumour spread. CTC clusters can simultaneously exhibit both mesenchymal and epithelial markers 54. This could be suggesting a possible cooperation between different cell clones within the CTC clusters, although so far it has not been demonstrated. Besides, genomic studies and the use of other OMIC techniques is essential to decipher the potential polyclonality of CTC clusters and the clonal cooperation of cells within CTC clusters, which will enlighten the origin of polyclonal metastases 114. Moreover, CTC clusters have a higher survival ability in circulation and a higher resistance to apoptosis at the site of metastasis than single CTCs 107. The survival advantage of CTC Introduction 89 clusters could be partially explained by the existence of strong cell-cell interactions between tumour cells that would make them more resistant to anoikis 56,115. Furthermore, the higher resistance to apoptosis observed in CTC clusters might also be due to the lower circulation half-life of CTC clusters (6-10 minutes) in comparison with individual CTCs (25-30 minutes) observed in BC murine models 45. Due to the larger size of CTC clusters, they can be physically entrapped in small capillaries more easily. This would favour CTC cluster survival and later growth to form metastatic lesions 104. This higher resistance to apoptosis was not only observed in animal models but also in BC patients. A study performed in TNBC by Paoletti and collaborators found that only 0.4% of cells were apoptotic in the CTC cluster population, while in single CTCs this percentage reached 20% 116. Changes in DNA methylation of cells within CTC clusters isolated from BC patients showed hypomethylation in the binding sites of transcription factors regulating stem-related genes, such as OCT4, NANOG, SOX2, and SIN3A, suggesting the presence of stem-like traits in CTC clusters. This methylation pattern correlated with a higher metastatic capacity of CTC clusters and a worse outcome of a subset of BC patients 117. The stem-like properties can increase cell plasticity and adaptation capacity to microenvironmental changes, which are characteristics that will favour the development of metastasis 109,117. CTC clusters have been mainly studied in mBC patients. However, it has been recently reported that it is also possible to detect CTC clusters in early-stage BC, suggesting that dissemination of CTC clusters may be an early event in BC 118,119. Recently published studies showed that CTC clusters isolated from early-stage BC patients and PDXs had mutations that were not present in the primary tumour and that can potentially provide information about the genes involved in tumour progression 114,120. Therefore, further studies are required in different BC stages of development to fully understand the role of CTC clusters in tumour spread. INÉS MARTÍNEZ PENA 96 OBJECTIVES The main objective of this thesis is to conduct a comparative study between individual CTCs and CTC clusters in order to go in depth in the biology of CTC clusters, to determine the biological traits of CTC clusters that confer their metastatic potential and that will help us to understand the role of CTC clusters in tumour progression. To accomplish this purpose, we have proposed different specific objectives: 1. To isolate CTCs, and specially CTC clusters from liquid biopsy samples from mBC patients. − To evaluate the potential of CTC clusters regarding their clinical prognostic utility in a cohort of mBC patients. − To evaluate the feasibility of CTC cluster isolation in the ParsortixTM from different types of liquid biopsy samples (peripheral blood samples, DLA products). 2. To establish in vitro experimental models of CTCs and CTC clusters by using human BC cell lines and to functionally and molecularly characterise them. − To evaluate the metastatic potential of the CTC and CTC cluster models by performing in vitro assays that mimic different steps of the metastatic cascade. − To assess the metastatic potential of the models in preclinical in vivo models (Danio rerio, Mus musculus). − To compare gene expression signatures between individual CTCs and CTC clusters. 3. To use a mouse model of mBC as a source for CTCs and CTC clusters to deepen into their biology − To generate an ex vivo CTC derived cell line and to functionally and molecularly characterize it as Objectives 97 a potential and more physiological realistic model of CTCs. − To identify genetic differences between individual CTCs and CTC clusters that could explain their differential contribution to tumour spread. INÉS MARTÍNEZ PENA 98 Materials and Methods 99 MATERIALS AND METHODS INÉS MARTÍNEZ PENA 100 Materials and Methods 101 MATERIALS AND METHODS 1. METASTATIC BREAST CANCER PATIENT SAMPLES 1.1. Peripheral blood samples from metastatic BC patients Peripheral blood samples from patients with mBC were provided by the Oncology Department of the University Hospital of Santiago de Compostela (CHUS). Before collecting the samples, patients were properly informed regarding the conditions of the biological sample cession by signing the corresponding informed consent (protocol code: RLL-BL2015_01), which was previously approved by the Clinical Research Ethics Committee (CEIC) of Galicia (code 2013/462 or 2015/772). Each sample consists of approximately 7.5 mL of peripheral whole blood. Samples were used for CTC and CTC cluster detection by using two different approaches for CTC enrichment. CellSearh® (Menarini Silicon Biosystems) is an immune-dependent device that isolates CTCs and CTC clusters based on their expression of EpCAM. CellSearh® is the only system approved by the FDA for clinical use. On the other hand, ParsortixTM (Angle plc) is an immune-independent system that is based on the differential physical properties of tumour cells for CTC and CTC cluster isolation. More specifically, ParsortixTM isolates CTCs and CTC clusters based on their larger size, and lower deformability, in comparison with the blood cells. For CTC and CTC cluster isolation in the ParsortixTM system, a specific protocol was used for maximising the integrity of CTC clusters and avoiding CTC cluster disaggregation. Thus, this protocol is characterised by a lower pressure, and a lower flow rate, in comparison with standard separation conditions (50 mbar and 99 mbar, respectively). Peripheral blood samples used for CTC and CTC cluster enumeration and correlation with patient outcome were collected in CellSave Preservative tubes (Menarini Silicon Biosystems) and INÉS MARTÍNEZ PENA 102 processed within the following 96 hours. Blood samples were collected at baseline (before starting the first line of systemic therapy, or before starting a new line of therapy), as well as at follow up, during the course of the treatment. Follow up samples were exclusively collected from those patients who were about to start the first line of systemic therapy. The follow up period varied from 3 to 5 weeks after the beginning of the therapy. Blood samples were processed at the Liquid Biopsy Analysis Unit (Health Research Institute of Santiago de Compostela, IDIS), using the CellSearch® System (Menarini Silicon Biosystems), and the CellSearch® Epithelial Circulating Tumour Cell Kit. Hence, biological samples were enriched for CTCs and CTC clusters using metallic beads coated with an antibody that specifically recognises EpCAM antigen. Moreover, after enrichment cells were stained with fluorescent antibodies that recognise CK 8, 18, and 19. Fluorescent antibody against the WBC marker CD45, and the double-stranded DNA staining DAPI were also used for cell staining. Thus, CTCs were identified as EpCAM+/CKs+/DAPI+/CD45-. CTC clusters were recognised as groups of ≥2 cells EpCAM+/CKs+/DAPI+/CD45-, with intact cytoplasm membranes and non-overlapping nuclei. CTC identification was manually evaluated by two trained technicians. Peripheral blood samples were also used to optimise a workflow for immune-independent CTC and CTC cluster isolation. These samples were collected into Vacuntainer K2 EDTA Blood Collection tubes (BD), when the blood extraction and the sample processing were performed within the same day, or into CellSave Preservative tubes (Menarini Silicon Biosystems), when the collection day and the processing day were different. Blood samples were processed in the size-based isolation system ParsortixTM (Angle plc) to proceed with single CTC and CTC cluster isolation. After separation, tumour cell identification was done by immunological characterisation using a cocktail of fluorescent conjugated antibodies against the epithelial markers EpCAM, the Epidermal Growth Factor Receptor (EGFR), and E-cadherin, as well as an antibody against CD45 for the identification of leukocytes (table 3). Moreover, NucBlueTM Live ReadyProbesTM Reagent (Molecular Probes), Materials and Methods 103 was used for nuclear staining. Therefore, it was considered as CTCs those cells with the following phenotype: EpCAM/EGFR/E-cadherin+, NucBlue+, CD45-. Table 3. Antibodies used for CTC identification in the ParsortixTM system. Epitope Antibody Proportion EpCAM Anti-EpCAM Alexa Fluor® 488 (Biolegend) 1:100 EGFR Anti-EGFR Alexa Fluor® 488 (Biolegend) 1:100 Ecadherin Anti-E-cadherin Alexa Fluor® 488 (Biolegend) 1:100 CD45 Monoclonal anti-CD45 PE (Exbio) 1:100 1.2. Diagnostic Leukapheresis (DLA) samples DLA is a density-based method that isolates cells due to continuous centrifugation of large volumes of blood. DLA is a standard procedure used in the clinic for the isolation of mononuclear cells. However, it has been proposed that CTCs could be co-isolated with the MNCs during DLA centrifugation as they have similar density. Given its origin, DLA samples can increase the likelihood of CTC and CTC cluster detection, in comparison with a 7.5mL of peripheral blood. DLA products obtained from mBC patients were provided by the research group of Prof. Dr. rer. nat. Hans Neubauer (University Hospital, Heinrich Heine University, Düsseldorf). DLA products were obtained after suitably signing the corresponding material transfer agreement (MTA) for sample cession by the HeinrichHeine University of Düsseldorf (Germany) (see appendix ‘Material Transfer Agreement for the use of Diagnostic Leukapheresis products’). In particular, it was used DLA samples negative for the presence of CTCs when analysed by CellSearch® system (Menarini Silicon Biosystems), and that were named as DLACTC-. DLACTCwere stored in liquid nitrogen upon arrival. DLACTCproducts were thawed quickly in the water bath (37 °C) INÉS MARTÍNEZ PENA 104 while opening the cryotube regularly to avoid cell damage due to overpressure. Then, DLACTCproducts were transferred to a 50 mL tube (Falcon) and diluted with cold 1x EDTA in PBS (100mM) until a total volume of 15 mL. Samples were kept in ice, and they were filtered twice using a 100 µm cell strainer (Fisher Scientific). DLACTCsamples were used to perform spiking tests in which a known number of fluorescently labelled BC cells as either single CTCs or as CTC clusters were added to the sample. Particularly, a 20 µL drop containing between 100 and 300 units, that is single cells and clusters, were added to each sample. Before spiking, tumour cells were stained with the nuclear staining NucBlueTM Live ReadyProbesTM Reagent (Molecular Probes), following the manufacturer’s instructions. Afterwards, spiked samples were processed in the ParsortixTM system (Angle plc) to optimise a workflow for DLA sample processing. After finishing the separation step with the corresponding protocol, captured CTCs and CTC clusters were counted based on enhanced Green Fluorescent Protein (eGFP) expression to calculate the isolation efficiency. Afterwards, isolated CTCs were harvested, and the CTCs that remained inside the cassette after harvesting were counted, to determine the predicted recovery efficiency. DLACTCsamples were used to perform the following assays: 1.2.1. Standard separation protocol vs. Cluster separation protocol DLA products are characterised by an elevated cellularity that could potentially block the ParsortixTM system during the separation step. Thus, two different separation protocols were evaluated for CTC detection in the ParsortixTM system to select the most suitable protocol to work with DLA samples. The same DLACTCsample was divided into two different subsamples after the dilution step with 1x EDTA-PBS. On the one hand, one of the DLACTCsubsamples was processed in the ParsortixTM system using the standard separation protocol (figure Materials and Methods 105 4, tube 1). On the other hand, the second DLACTCsubsample was processed with a protocol specifically designed to maximise CTC cluster capture and to avoid CTC cluster disaggregation (figure 4, tube 2). The cluster separation protocol has lower flow rate and lower pressure (50 mbar) than the standard separation protocol (90 mbar). Figure 4. Graphical representation of the experimental workflow followed for the optimization of the ParsortixTM separation protocol using DLA samples. Each DLA sample was divided into two halves. Each half was spiked with a known number of tumour cells as both, individual CTCs and CTC clusters. Afterwards, CTCs were isolated in the ParsortixTM using a standard separation protocol (tube 1), or a cluster separation protocol (tube 2). Isolated CTCs and CTC clusters were counted and harvested. Finally, the CTCs that remained inside the cassette were counted and a predicted recovery was determined. 1.2.2. Filtered: Before vs. After tumour cell spiking In this assay, the same DLACTCproduct was divided into two different tubes after the dilution with 1x EDTA-PBS (100 mM). One of the DLACTCsubsamples was passed through the 100 µm cell strainer twice before spiking (figure 5, tube 1), while the other DLACTCsubsample was filtered after performing the spiking step (figure 5, tube 2). Later on, samples were processed in the ParsortixTM system using the standard separation protocol, and CTC counting was INÉS MARTÍNEZ PENA 112 Table 4. Parameters of the fluid shear stress assay. Syringe diameter (mm) 12.5 Flow rate (mL/min) 4.25 Pressure (dyn/cm2) 1840 Volume (mL) 4 It was considered one passage (P) when the entire content of the syringes passed through the 30G needle, which was collected in a 15 mL clear polypropylene (PP) centrifuge tube (Corning). A total number of 10P were performed in each FSS assay, and a 2minute break between consecutive passages was allowed to permit cell resting (figure 9). Figure 9. Workflow of the fluid shear stress (FSS) assay. Cell suspensions were loaded into a 5 mL syringe that was linked to a 30G needle, as well as to a syringe pump. It was considered a passage (P) when the cell suspension completely flowed through the needle. Each FSS assay consisted of 10P with a 2-minute break between subsequent passages. Cell viability was assessed by bioluminescence after finishing the assay (0 h), as well as 24 hours later (24 h). Furthermore, 100 µL subsamples were taken before the first passage (P0), as well as after the even passages (P2, P4, P6, P8, and P10) and they were collected into a 96-well Flat Clear Bottom Black Polystyrene TC-treated Microplates (Corning) in order to evaluate cell viability right after finishing the assay (0 h). Besides, subsamples after even passages were collected to put in culture Materials and Methods 113 for 24 hours and measure the cell viability at the 24 hour timepoint (figure 9). Cell viability was assessed across the passages at 0 and 24 hours after the FSS assay by the bioluminescent signal emitted by the luciferase enzyme when added XenoLight D-Luciferin - K+ Salt Bioluminescent Substrate (PerkinElmer) as a substrate for the reaction. Bioluminiscence was measured using the system EnVision Multilabel Plate Reader (PerkinElmer), and the software Wallac EnVision Manager 1.12 (PerkinElmer). 3.5. Endothelial adhesion assays For endothelial adhesion assays EA.hy926 cells were seeded at a density of 5x104 cells per well, in 96-well plates (Corning) previously coated with gelatine. Endothelial cells were allowed to grow to generate a cell monolayer, by incubating them at 37 °C, 5% CO2, for 24-48 hours. Once the monolayer of endothelial cells was formed, 2.5x104 tumour cells per well were added under serum deprivation conditions. A minimum of 4 replicates of each CTC in vitro model were seeded in each experiment (figure 10). Tumour cells were incubated together with the endothelial monolayer for 45 minutes, at 37°C, and 5% CO2, to enable the interaction between both cell types. Non-attached tumour cells were washed away with PBS. Finally, tumour cells adhered to the endothelial monolayer were counted by microscopy, using the fluorescence microscope Leica DMi8 (Leica Microsystems). The whole surface area of each replicate was counted. Figure 10. Overview of the endothelial adhesion assay. INÉS MARTÍNEZ PENA 114 3.6. Soft agar colony formation assays Soft agar colony formation ability was assessed using 6-well and 12-well plates (Corning). Firstly, a lower layer of 0.5% agarose (Fisher Bioreagents) in the corresponding cell culture media (DMEM High Glucose, or MammocultTM, depending on the cell line) was generated. After the gelling of the bottom layer, 1x104 cells/well (6-well plate), or 5x103 cells/well (12-well plate) with 0.3% agarose in cell culture media were seeded. Duplicates of each CTC in vitro model were seeded in each experiment performed in 6-well plate, and three replicates of each condition were seeded in the case of 12-well plate assays. Cells were incubated at 37°C, 5% CO2 for 2-4 weeks, depending on the subtype of the cell lines used in each assay. Afterwards, the assays were stopped by incubating the wells with 0.005% crystal violet (VWR Chemicals) in methanol (SigmaAldrich) for 1 hour, at RT. Excess of staining was removed by washing the cells with PBS (Biowest) several times. Finally, the number of colonies was determined by microscopy using the fluorescence microscope Leica DMi8 (Leica Microsystems). In the CTC cluster models derived from BC cell lines only macroscopic colonies were counted, while for mCTC colony formation assays a ≥70 µm diameter cut-off was established for determining the number of colonies. 4. IN VIVO ASSAYS 4.1. Zebrafish (Danio rerio) embryo xenografts Wild-type zebrafish (ZF) embryos were obtained by natural mating of adult zebrafish. Twenty-four hours post-fertilization (hpf) embryos were maintained at 28 °C, in 0.003% w/v Nphenylthiourea (PTU) (Sigma) in E3 medium, to inhibit melanogenesis and obtain transparent embryos. ZF xenografts were generated using 48 hpf embryos. All procedures were performed according to current legislation (RD53/2013). In the RD53/2013, Danio rerio is included in the appendix I, as a Materials and Methods 115 species for experimental procedures that should be protected. However, this law only applies to embryo forms of mammals in the final third of their development, but does not apply to nonmammal embryos, such as zebrafish embryos. Firstly, 48 hpf embryos were dechorionated and anaesthetized with 0.003% w/v tricaine methanesulfonate (MS222) (Sigma) diluted in E3 medium with PTU. Once the embryos were immobilized by the anaesthetic, they were put in a 1.5% agarose-coated 100 mm ø Petri dish (VWR) to perform the microinjection of cells. Immediately preceding the microinjection, MDA-MB-231eGFP Luc and MCF-7eGFP Luc cell lines were used to generate in vitro models of single and cluster cell suspension, by differential disaggregation (see above ‘Generation of single CTC and CTC cluster in vitro models’). A million cells were washed once with PBS and resuspended in 6 µL of 2% polyvinylpyrrolidone (PVP) (Sigma), which is a polymer that prevents glass capillaries from blocking. Cells were maintained on ice during microinjection. Secondly, borosilicate glass capillaries (1mm O.D. ×0.58mm) (Harvard Apparatus) were used to make needles using the PC-10 puller (Narishige International). Cell suspensions were loaded into the needles and microinjection was performed by injecting approximately 300 cells per embryo. Microinjection was done by using an IM 300 microinjector (Narishige International), set with an outpout pressure of 68.95 kPa, and 0.03 ms of injection time, as well as the Nikon SMZ800 Stereo Zoom Microscope (Nikon). ZF microinjection was performed in the perivitelline space, next to the convergence of the Duct of Cuvier (DoC) (figure 11, a). After microinjection, ZF xenografts were examined for the presence of a fluorescent cell mass at the injection site, and the absence of tumour cells in the circulation was verified. Embryos showing tumour cells in circulation after injection were discarded and not considered for analysis. ZF xenografts were kept at 34 °C, a consensus temperature between ZF embryo development and human tumour cell viability. ZF xenografts were monitored at 0 INÉS MARTÍNEZ PENA 116 hours post injection (hpi), 24 hpi, 48 hpi and/or 72 hpi by using the inverted fluorescent microscope Leica DMi8 (Leica Microsystems). ZF xenograft monitoring was performed to evaluate cell dissemination, and cell survival. Moreover, caudal dissemination patterns were studied in each type of ZF xenograft, dividing the tail into three different regions, according to the vasculature: dorsal (includes dorsal longitudinal anastomotic vessel), ventral (involves the caudal vein, the posterior cardinal vein, and the dorsal aorta), and lateral (intersegmental vessels) (figure 11, b). Finally, ZF xenografts were euthanised at 120 hpf by tricain overdose, and visual verification of the absence of heart rate was used as the method for death confirmation. For molecular analysis of disseminated cells in the individual CTC and CTC cluster ZF xenografts, tails were physically separated from the rest of the body after the euthanasia for tumour cell isolation. Tails of each xenograft population were pooled in a 1.5 mL tube (Axygen). Each pool of tails was washed twice with PBS (Biowest), and later enzymatically and mechanically digested by using 100 µL of 0.25% Trypsin-EDTA (Lonza) in DMEM High Glucose (Biowest) at 37 °C, and by up-and-down pipetting for 10-15 minutes. After tissue digestion, trypsin was neutralised by 400 µL of DMEM High Glucose (with 10% FBS), and the digested product was centrifuged at 700 g, for 5 minutes. Cells were washed with PBS and pelleted for performing RNA extraction. ZF experiments were excluded when the survival rate was less than 70%. A minimum of 30 embryos per group were included in each assay. Materials and Methods 117 Figure 11. Zebrafish (ZF) embryo xenografts. Representation of the workflow followed for the generation of ZF xenografts. Firstly, embryo dechorionation was performed. Secondly, cells as either individual or as CTC clusters were injected in the perivitelline space, next to the convergence of the Duct of Cuvier (DoC). Finally, cell dissemination to the caudal region was assessed at 0, 24, 48 and/or 72 hours postinjection (hpi) (a). Tail division according to caudal vasculature. The caudal region was divided into three different sections, based on the vasculature: dorsal, lateral, and ventral (b). 4.2. Mouse (Mus musculus) experiments Murine experimental procedures were approved by the Animal Experimentation Ethical Committee of the University of Santiago de Compostela (CEEA) (project: 15010/2019/002), according to the RD 53/2013 (see appendix ‘Favourable report Animal Experimentation Ethical Committee of the University of Santiago de Compostela (CEEA)’, ‘Animal experimentation INÉS MARTÍNEZ PENA 118 training certificate. B function (euthanasia)’, and ‘Animal experimentation training certificate. C function (animal procedures)’). Mice were obtained from Barcelona Biomedical Research Park (PRBB, Barcelona), or Charles River and hosted in the animal facilities of the Centre for Research in Molecular Medicine and Chronic Diseases (CIMUS, University of Santiago de Compostela) (REGA: ES150780275701), and in the Experimental Biomedicine Centre (CEBEGA, Santiago de Compostela) (REGA: ES150780292901), where they were provided with food and water ad libitum, according to the guidelines of the recipient centre. 4.2.1. Mouse lung colonization assay To assess the lung colonization ability and tumourinitiating capacity of the in vitro models of single cells and clusters, suspensions of 5x105 cells in 100µL of PBS of the cell line MDA-MB-231eGFP Luc were injected into the lateral tail vein of SCID BEIGE mice, either as single cells, or clusters. Tumour cell injection was performed in mice immobilized using a mouse restrainer device, and 1 mL syringes linked to 27 gauge needles (BD Medical). Then, colonization and tumour-initiating abilities were evaluated by quantifying the pulmonary metastatic incidence over time. Mice were monitored immediately after injection (1 h), as well as every 3-4 days, for 18 days. Lung tumour incidence was tracked by the bioluminescence emitted by the luciferase activity, when 15 mg/mL XenoLight D-Luciferin - K+ Salt Bioluminescent (PerkinElmer) were intra-peritoneally injected as a substrate at a final concentration of 10 µL/g of body weight. At the endpoint, mice were euthanized by CO2 inhalation, and cervical dislocation was used to confirm the death. Afterwards, lungs were perfused by tracheal infusion with 1% penicillin-streptomycin (Lonza) in PBS. Lungs were surgically removed and fixated in formol (VWR) to perform later histopathological analysis (figure 12). Histological analyses were performed in collaboration with the Translational Molecular Materials and Methods 119 Pathology research group of the Vall d’Hebron Institute of Research (VHIR) (Barcelona). Figure 12. In vivo lung colonization assay. Tumour cell suspensions were injected in the lateral vein of the tail as either single CTCs or as CTC clusters. Metastatic incidence in the lungs was tracked over time by animal in vivo imaging. Finally, lungs were surgically removed for anatomopathological analysis. 4.2.2. Mouse orthotopic BC xenografts Tumour cells from the MDA-MB-231eGFP Luc cell line were used for this model. A total of 2x106 cells resuspended in 60 µL of a mixture 1:3 of DMEM and Matrigel Growth Factor Reduced (Corning) were injected into the mammary fat pad of 8week-old SCID BEIGE mice, using BD MicroFineTM 0.33 mm (29G) x 12.7 mm insulin syringes of 0.5 mL (BD Medical). Tumour progression was monitored weekly, during 9-10 weeks using the IVIS Spectrum in vivo imaging system (PerkinElmer), and the Living Image Software (PerkinElmer). Animal monitoring was performed measuring by bioluminescence, by intra-peritoneally injecting 15 mg/mL of XenoLight D-Luciferin - K+ Salt Bioluminescent (PerkinElmer), to generate a final concentration of 10 µL/g of body weight. Once primary tumour (PT) and lung metastases were developed, mice were euthanised by inhalation of CO2, and INÉS MARTÍNEZ PENA 120 cervical dislocation was used as the death confirmation method. Afterwards, blood samples were collected by cardiac puncture using insulin syringes and BD Vacuntainer K2 EDTA Blood Collection tubes (BD), or CellSave Preservative tubes (Menarini Silicon Biosystems). Murine blood samples were processed in the epitope-independent CTC isolation system ParsortixTM (Angle plc). Isolated CTC were fixed with PFA 2% for 30 minutes, washed with PBS, and pelleted in a final volume of 100 µL. Harvested cells from ParsortixTM were later individualised by the DEPArrayTM system (Menarini Silicon Biosystems). These cells were used to perform genomic analysis by Whole Exome Sequencing (WES). Moreover, primary tumours and lungs were also preserved in formol (VWR) and washed with ethanol (VWR) for performing anatomopathological (AP) examination (figure 13). Figure 13. Orthotopic BC model workflow. MDA-MB-231eGFP Luc cells were injected into the mammary fat pad of immunocompromised SCID BEIGE female mice. Tumour progression was monitored over time. Primary tumour (PT) and lungs were surgically removed for later gene expression and anathomopathological (AP) analyses. Besides, blood samples were collected for CTC isolation in the ParsortixTM system. Isolated CTCs and Materials and Methods 121 CTC clusters were used for genomic analyses, and for the obtention of an ex vivo CTC derived cell line. 4.3. mCTC cell line generation Moreover, a cell line was generated from CTCs isolated from the blood of one MDA-MB-231eGFP Luc orthotopically injected mouse. For this, blood was processed using the negative selection kit EasySepTM Mouse T Cell Isolation Kit (StemCell Technologies) for CTC isolation. Isolated cells were initially cultured in a low-attachment 96-well plate (Corning), using 200 µL MammocultTM (StemCell Technologies), supplemented with hydrocortisone (0.48 µg/mL), heparin (4 µg/mL), bFGF (20 ng/mL), EGF (20 ng/mL), B27 supplement (4% v/v), progesterone (0.4 µg/mL), β-estradiol (0.4 µg/mL), ultraGROTM (5% v/v), and P/S (1% v/v). Isolated cells were cultured under hypoxic conditions, at 37 °C for a week. Cell culture media was renewed every two days. Afterwards, cells were cultured under low-attachment normoxic conditions (5% CO2), at 37°C, and with supplemented MammocultTM. These cells grew indefinitely generating cellular aggregates and establishing a cell line called mCTC. 5. MOLECULAR CHARACTERISATION 5.1. Gene expression analysis To perform RNA extraction from the digested tissue of ZF tails, RNeasy Micro kit (Qiagen) was used, according to the manufacturer’s instruction. RNA quantification was done in the Nanodrop OneC Microvolume UV-Vis Spectrophotometer (FisherScientific). Additionally, RNA sample quality and integrity was evaluated using the 4200 TapeStation System (Agilent Technologies) and the TapeStation software (Agilent Technologies), in the case of RNA samples whose expression was studied by RNAsequencing. INÉS MARTÍNEZ PENA 128 RESULTS 129 INÉS MARTÍNEZ PENA 130 RESULTS 1. CTC CLUSTER DETECTION IN SAMPLES DERIVED FROM METASTATIC BREAST CANCER PATIENTS 1.1. CTC and CTC cluster enumeration in peripheral blood samples Peripheral blood samples (or liquid biopsy samples) were collected to evaluate the relevance of CTC and CTC cluster enumeration as a prognostic factor of the outcome of a cohort of unselected mBC patients. 1.1.1. Cohort description: patient characteristics A total number of 54 unselected female mBC patients were included in the cohort. Within this cohort, 27 patients (50.9%) were defined as having Hormone Receptor positive and Human Epidermal Growth Factor Receptor-2 negative (HR+/HER2-) tumours; 11 patients (20.8%) were determined as having HER2+ tumours; and 15 (28.3%) were defined as having triple-negative tumours (HR-/HER2-). Moreover, 1 patient showed undetermined levels of HER2 histological staining. The follow up period from baseline for alive patients varied from 22 to 603 days, and the median follow up time was 197 days. Besides, 40 patients (74.1%) had visceral metastases, while 14 patients (25.9%) showed non-visceral metastases. Visceral metastases included lung, liver, peritoneal and/or pleural location of metastatic lesions. Nonvisceral metastases referred to lymph node and/or bone involvement. Eastern Cooperative Oncology Group (ECOG) performance status was also determined at baseline for this cohort of patients. ECOG is a standardised method to measure the impact of the disease in the daily routine of patients. According to the ECOG performance status, 18 patients (33.3%) were fully active, and able to continue performing all the activities without restriction (grade 0); 31 patients (57.4%) showed restrictions on the physical activity of high intensity due to the disease Results 131 but were able to perform light or sedentary activities (grade 1); while 5 patients (9.3%) were able to do all selfcare activities but were unable to do work activities, and were up and about more than 50% of waking hours (grade 2). Moreover, NHG (Nottingham Histological Grade) was used to determine the characteristics of primary tumour, based on tubular/glandular formation, tumour cell mitotic activity, and tumour cell nuclear pleomorphism. NHG scale showed that 4 patients (7.4%) had a well differentiated primary tumour (grade I); 26 patients (48.1%) showed a moderately differentiated tumour, while 15 patients (27.8%) had a poorly differentiated primary tumour. Besides, 9 patients (16.7%) had an unknown primary tumour NHG status. Table 7 describe in detail patient characteristics. Table 7. Patient characteristics (n = 54). Variables Total, n (%) Mean age (years) 58.5 < 65 years 34 (63.0) ≥ 65 years 20 (37.0) Tumour stage IV 54 (100.0) Baseline ECOG 0 18 (33.3) 1 31 (57.4) 2 5 (9.3) Primary Tumour NHG I 4 (7.4) II 26 (48.1) III 15 (27.8) INÉS MARTÍNEZ PENA 132 Unknown 9 (16.7) Breast cancer subtypes HR+/HER227 (50.9) HER2+ 11 (20.8) HR-HER2- (TNBC) 15 (25.8) Unknown 1 (1.8) Site of metastases * Visceral 40 (74.1) Non-visceral 14 (25.9) Number of metastatic sites < 3 29 (53.7) ≥ 3 25 (46.3) First line of systemic therapy (n = 44) Chemotherapy 23 (52.3) Hormonal therapy 15 (34.1) Target therapy ** 6 (13.6) Other lines (n = 10) Chemotherapy 8 (80.0) Hormonal therapy 1 (10.0) Target therapy ** 1 (10.0) Abbreviations: ECOG Eastern Cooperative Oncology Group; NHG Nottingham Histologic Grade; HR Hormone Receptor; HER2 Human Epidermal growth factor Receptor-2; TNBC Triple-Negative Breast Cancer. *Visceral metastases include lung, liver, peritoneal and/or pleural locations. Nonvisceral metastases refer to lymph node and/or bone involvement. **In combination with chemotherapy Results 133 1.1.2. CTC and CTC cluster enumeration and correlation with clinicopathological variables A total number of 96 peripheral blood samples were used for individual CTC and CTC cluster detection. These liquid biopsy samples derived from 54 different patients that were recruited at baseline, 38 of whom had follow up sample collection. Baseline samples were collected before starting first line of systemic therapy, or before starting a new line of therapy (figure 14). The FDA-approved CellSearch® system allowed the detection and enumeration of CTCs and CTC clusters in these peripheral blood samples. Figure 14. CONSORT flow diagram of the cohort of metastatic breast cancer (mBC) patients. Patients are stratified according to the presence or absence of CTCs and CTC clusters at baseline and during follow-up. N refers to the number of patients. At baseline, 43 patients (79.7%) showed presence of CTCs, while 11 (20.4%) had no detectable CTCs in their blood. Besides, 12 (22.2%) of those patients positive for CTCs at baseline, showed less than 5 CTCs per 7.5 mL of blood; while 31 of them (57.4%) had 5 or more CTCs, and 14 patients (25.9%) showed at least one CTC cluster (a group of cells made up of ≥ 2 CTCs). During follow up, 16 (42.1%) patients showed no detectable CTCs, and 22 (57.9%) of them were positive for CTCs. Moreover, 13 (34.2%) INÉS MARTÍNEZ PENA 134 of the CTC positive patients had ≥ 5 CTC/7.5 mL, and 7 (18.4%) of them had at least one CTC cluster in their blood (table 8). Table 8. Analysis of CTC and CTC clusters in the study cohort. Baseline n = 54 Percent (%) Mean Min-Max CTCs 43 79.2 183.1 1-1970 CTCs ≥5 31 57.4 253.1 5-1970 CTC cluster ≥1 14 25.9 4.28 1-14 Follow up n = 38 CTCs 22 57.9 95.6 1-969 CTCs ≥5 13 34.2 160.6 5-969 CTC cluster ≥1 7 18.4 12.71 1-74 It should be highlighted that CTC clusters were exclusively detected in those patients with ≥ 5 CTCs per 7.5 mL of blood in both, at baseline (14/31), and during follow up (7/13). The presence of ≥ 1 CTC cluster was associated with the number of individual CTCs detected (figure 15). In addition, it was found a higher frequency of patients with ≥ 5 CTCs and CTC clusters in the HR+/HER2BC subtype. Nevertheless, no significant correlation was found between CTC and CTC cluster detection and BC subtype at baseline, neither during follow up (table 9). Besides, it was not found a correlation between the distribution of CTCs and CTC clusters regarding the number and location of metastatic lesions. Hence, CTC clusters are a subpopulation of low frequency within CTCs in mBC patients. The presence of CTC clusters is more likely in patients with high individual CTC shedding at baseline and/or during therapy. Results 135 Figure 15. Association between the number of CTCs and the presence or absence of CTC clusters. Box plots of the association between the number of CTCs and the presence/absence of CTC clusters at baseline (a), and during follow up (b). Representative images of CTC clusters isolated from the peripheral blood samples of mBC patients using the CellSearch® system (c). INÉS MARTÍNEZ PENA 136 Table 9. Enumeration of CTCs and CTC clusters distributed according to BC subtype. Variable Total, n (%) BC subtype (%) p value HR+/HER2HER2+ HR-/HER2Undet. Baseline n = 27 n = 11 n = 15 n = 1 CTCs/7.5 mL 0.861 < 5 CTCs 23 (100.0) 12 (52.2) 4 (17.4) 7 (30.4) 0 (0) ≥ 5 CTCs 31 (100.0) 15 (48.4) 7 (22.6) 8 (25.8) 1 (3.2) CTC clusters/7.5 mL 0.668 No 40 (100.0) 21 (52.5) 7 (17.5) 11 (27.5) 1 (2.5) Yes 14 (100.0) 6 (42.9) 4 (28.6) 4 (28.6) 0 (0) Follow up n = 19 n = 9 n = 9 n = 1 CTCs/7.5 mL 0.583 < 5 CTCs 25 (100.0) 11 (44.0) 7 (18.4) 6 (24.0) 1 (4.0) ≥ 5 CTCs 13 (100.0) 8 (61.5) 2 (15.4) 3 (23.1) 0 (0) CTC clusters/7.5 mL 0.184 No 31 (100.0) 15 (48.4) 9 (29.0) 6 (19.4) 1 (3.2) Yes 7 (100.0) 4 (57.1) 0 (0) 3 (42.9) 0 (0) Abbreviations: CTCs Circulating Tumour Cells; HR Hormone Receptors; HER2 Human Epidermal Growth Factor Receptor 2; Under. Undetermined 1.1.3. CTC and CTC clusters as predictor factors of patient outcome Regarding the prognostic value of CTCs and CTC clusters in relation to PFS and OS, it was observed that patients with ≥ 5 CTCs/7.5 mL of blood at baseline (31/54) had a tendency to show a lower PFS (Plog-rank = 0.059), and a significantly shorter OS (Plog-rank = 0.017) Results 137 (figure 16, a, b). Moreover, Cox regression analysis showed that there was a higher risk of death in those patients with ≥ 5 CTCs/7.5 mL (HROS = 3.15; 95% CI: 1.16-5.55; p = 0.024) but not an increased risk of progression (HRPFS = 2.11; 95% CI: 0.95-4.69; p = 0.065), when compared with patients with < 5 CTCs/7.5 mL of blood (table 10). Nevertheless, both observations became statistically significant when the analysis was adjusted for other clinicopathological variables (table 11). On the other hand, the CTC cluster-based analysis showed that those patients with ≥ 1 CTC cluster/7.5 mL of blood had a significantly higher risk of progression (HRPFS = 3.95; 95% CI: 1.80-8.68; p = 0.0006) and death (HROS = 4.23; 95% CI: 1.8-10.1; p = 0.0009) (figure 16, c, d). In fact, those patients showed a shorter OS and PFS (OS and PFS Plog-rank < 0.001). Besides, the significance of these observations was maintained when the analysis was adjusted for other clinicopathological parameters (table 11). During follow up, those patients with ≥ 5 CTCs/7.5 mL blood (13/38) did not show an increased risk of progression (HRPFS = 2.3; 95% CI: 0.76-6.7; p = 0.15), neither a shorter PFS (Plog-rank > 0.05). However, they showed a higher risk of death (HROS = 5.3; 95% CI: 1.421; p = 0.017), and a shorter OS (Plog-rank < 0.05), in comparison with patients who had < 5 CTCs/7.5 mL blood (table 10 and figure 16, e, f). These results were not significant when adjusted for other clinicopathological factors (table 11). The follow up analysis based on CTC cluster enumeration showed no differences in the PFS and OS between patients with or without CTC clusters in their blood samples. However, it was observed a tendency by which patients with ≥ 1 CTC cluster/7.5 mL blood had a higher risk of death (HROS = 3.0) (table 10 and figure 16, g, h). INÉS MARTÍNEZ PENA 144 Table 12. Study of the joint effect of CTCs and CTC clusters with patient survival (PFS and OS). Variable Total Events, n (%) HR (95% CI) p value HR (95% CI)* p value* Joint effect of CTC and CTC clusters: Baseline Associated with PFS < 5 CTCs and NO CTC clusters 23 9 (39.13) 1.00 1.00 ≥ 5 CTCs and NO CTC clusters 17 9 (52.84) 1.24 (0.47-3.24) 0.66 1.74 (0.57-5.30) 0.32 ≥ 5 CTCs, ≥ 1 CTC cluster 14 11 (78.57) 4.34 (1.76-10.6) 0.0013 5.16 (1.68-15.8) 0.0041 Associated with OS < 5 CTCs and NO CTC clusters 23 6 (26.1) 1.00 1.00 ≥ 5 CTCs and NO CTC clusters 17 7 (41.18) 1.88 (0.59-5.98) 0.28 1.84 (0.50-6.82) 0.36 ≥ 5 CTCs, ≥1 CTC cluster 14 10 (71.43) 5.79 (1.96-17.1) 0.0014 7.79 (1.93-31.4) 0.0038 Joint effect of CTC and CTC clusters: Follow up Associated with PFS < 5 CTCs and NO CTC clusters 25 8 (32.0) 1.00 1.00 ≥ 5 CTCs and NO 6 3 (50.0) 2.3 (0.57-9.27) 0.24 0.61 (0.08-4.53) 0.635 Results 145 CTC clusters ≥ 5 CTCs, ≥ 1 CTC cluster 7 4 (57.1) 2.2 (0.62-7.84) 0.22 4.45 (1.01-19.59) 0.0484 Associated with OS < 5 CTCs and NO CTC clusters 25 3 (12.0) 1.00 1.00 ≥ 5 CTCs and NO CTC clusters 6 3 (50.0) 5.71 (1.11-29.4) 0.037 1.2x1018 (0.0-Inf.) 0.996 ≥ 5 CTCs, ≥ 1 CTC cluster 7 4 (57.1) 5.00 (1.11-22.4) 0.0353 2.4x1018 (0.0-Inf.) 0.996 Abbreviations: CTC Circulating Tumour Cell; PFS Progression-Free Survival; OS Overall Survival; HR Hazard Ratio; CI Confidence Interval; Inf. Infinity. *Adjusted for age, ECOG, BC subtype, number of metastatic sites, metastases location, and treatment. Furthermore, mBC liquid biopsy samples were used to study the clinical prognostic value of CTC cluster enumeration over time. To this end, the continued presence of CTC clusters throughout time was studied, regarding survival variables, PFS and OS. In our cohort and among the patients with ≥1CTC (n = 43, only 2 patients (4.4%) had detectable CTC clusters at two different timepoints, 16 (35.6%) had CTC clusters in one timepoint, and 27 (60.0%) only had individual CTCs (figure 18, a). In terms of PFS, it was observed a shorter PFS time in patients with presence of CTC clusters in their blood, in comparison with those who only had individual CTCs (Plog-rank = 0.017) (figure 18, b). Nevertheless, it was only observed a significantly increased risk of progression in those patients with CTC clusters at one timepoint (HRPFS = 2.8; 95% CI: 1.22-6.34; p = 0.014). Moreover, those patients with detection of CTC clusters in one or two different timepoints had a significantly lower OS (Plog-rank = 0.007) (figure 18, c), and an increased risk of death (HROS = 3.3; 95% CI: 1.29-8.6; p = 0.012 for one timepoint, and HROS = 6.5; 95% CI: 1.31-32.7; p = 0.021). INÉS MARTÍNEZ PENA 146 Figure 18. Longitudinal analysis of CTCs and CTC clusters. Percentage of patients with individual CTCs only (dark bar), with CTC clusters at 1 timepoint (light grey) and with CTC clusters at 2 timepoints (dark grey) (a). Kaplan-Meier curves of progression-free survival (PFS) and overall survival (OS), according to the presence of single CTCs only, CTC clusters at 1 timepoint or CTC clusters at 2 timepoints (b, c). Therefore, these results denoted that CTC cluster enumeration at baseline is an independent prognostic marker, that provides additional value to CTC enumeration alone in this cohort of mBC patients in those patients with high CTC detection. Besides, the maintained presence of CTC clusters in the peripheral blood of mBC is associated with a worse disease outcome, as they are associated with a shorter survival. This work was extracted/adapted from our previous published work: Analysis of a Real-World Cohort of Metastatic Breast Cancer Patients Shows Circulating Tumor Cell Clusters (CTC-clusters) as Predictors of Patient Outcomes. Costa, C.; Muinelo-Romay, L.; Cebey-López, V.; Pereira-Veiga, T.; Martínez-Pena, I.; Abreu, M.; Abalo, A.; Lago-Lestón, R.M.; Abuín, C.; Palacios, P.; Cueva, J.; Piñeiro, R.; López-López, R. Cancers 2020, 12(5), 1111. ISSN 2072-6694 Results 147 Published 29/04/2020. Available on the website: https://doi.org/10.3390/cancers12051111 Author affiliation: Clotilde Costa Nogueira1,2,‡ , Laura Muinelo Romay2,3,‡, Víctor Cebey López4, Thais Pereira Veiga1, Inés Martínez Pena1, Manuel Abreu3, Alicia Abalo3, Ramón M. Lago Lestón3, Carmen Abuín1, Patricia Palacios4, Juan Cueva4, Roberto Piñeiro1,*, and Rafael López López1, 2,3,4,* 1Roche-Chus Joint Unit, Translational Medical Oncology Group, Oncomet, Health Research Institute of Santiago de Compostela (IDIS), Travesía da Choupana s/n, 15706 Santiago de Compostela, Spain 2CIBERONC, Centro de Investigación Biomédica en Red Cáncer, 28029 Madrid, Spain 3Liquid Biopsy Analysis Unit, Translational Medical Oncology Group, Health Research Institute of Santiago de Santiago de Compostela (IDIS), Travesía da Choupana s/n, 15706 Santiago de Compostela, Spain 4Department of Oncology, Complexo Hospitalario Universitario de Santiago de Compostela (SERGAS), 15706 Santiago de Compostela, Spain ‡These authors contributed equally *These authors jointly supervised this work Contribution to this work: I, Inés Martínez Pena, was involved in data analysis and reviewing of the manuscript. INÉS MARTÍNEZ PENA 148 1.2. Optimization of ParsortixTM immuno-independent detection of CTCs and CTC clusters in BC liquid biopsy samples 1.2.1. Metastatic BC peripheral blood samples As previously demonstrated, the FDA-approved CellSearch® system has highlighted the importance of CTC and CTC cluster enumeration for predicting mBC patient outcome. However, this system presents some limitations as it does not allow the recovery of CTCs and CTC clusters in a viable state, thus limiting the analysis that could be performed with the enriched CTCs and CTC clusters, and it does not allow the optimization of protocols for the isolation of CTC clusters. Given the limitations of this immune-dependent CTC isolation system, peripheral blood samples from mBC patients were also used to optimise a workflow for CTC and especially CTC cluster isolation in an immune-independent manner, using the ParsortixTM system. ParsortixTM is a microfluidic device that isolates CTCs based on their larger size and lower deformability, in comparison with blood cells. ParsortixTM allows the isolation of CTC clusters, as it was previously demonstrated 129 (figure 19). Results 149 Figure 19. Overview of a ParsortixTM cassette. The sample flows through serpentine channels. These channels have a stair-like silhouette that finishes in a 6.5 µm gap in which CTCs are capture, while the rest of blood cells flow through. Red arrows indicate the direction of the flow inside the cassette. Image used with permission of Angle plc. Firstly, we tried to isolate, and detect CTCs and especially CTC clusters from mBC patient liquid biopsy samples (n = 3) (table 13). For this purpose, a ParsortixTM separation protocol (named ‘cluster separation protocol’) with low pressure and reduced flow was used to minimise CTC cluster disaggregation during CTC capture. Secondly, a previously optimised live cell immunostaining was used to identify the captured CTCs and CTC clusters in the ParsortixTM. This in-cassette immunostaining included a conjugated cocktail of antibodies against the epithelial markers Epithelial Cell Adhesion Molecule (EpCAM), Epidermal Growth Factor Receptor (EGFR), and E-cadherin (E-cad). Besides, an antibody against the leukocyte marker CD45, as well as the double-stranded DNA nuclear dye NucBlueTM Live ReadyProbesTM Reagent was used. Thus, CTCs and CTC clusters were defined as isolated cells with the following phenotype: EpCAM/EGFR/E-cad+, NucBlue+ and CD45-. Moreover, a CTC cluster was defined by the presence of a group made up of ≥ 2 tumour cells. Hence, this workflow allowed the identification of both, putative CTCs and CTC clusters as a proof-of-concept experiment in a small number of patients (table 13 and figure 20). Table 13. mBC peripheral blood samples used for CTC and CTC cluster enumeration in the ParsortixTM system. BC subtype Parsortix CTCs CTC clusters UM176MV1 HER2+ 57 60 UM196MV1 HR+/HER2500 179 UM88MV3 HR+/HER29 7 Abbreviations: BC Breast Cancer; CTC Circulating Tumour Cell; HER2 Human Epidermal growth factor Receptor 2; HR Hormone Receptor INÉS MARTÍNEZ PENA 150 Figure 20. Representative images of single CTCs and CTC clusters isolated from the peripheral blood samples of mBC patients. Scale bars: 50 µm. Thus, it was possible to establish a workflow that allows CTC cluster isolation from peripheral blood samples of mBC patients. Unlike Results 151 the immune-based CTC capture, CTC isolation based on their differential physical properties allowed the recovery of CTCs in a viable state as they are positive for NucBlue Live Cell Stain, potentially extending the amount of downstream analysis. It also may increase the likelihood of detecting a more heterogeneous CTC population, although for this, antibodies against other epitopes should be used. 1.2.2. Metastatic BC Diagnostic Leukapheresis (DLA) products Despite the utility of peripheral blood samples for CTC and CTC cluster isolation, there is still a limitation regarding the low amount of biological material obtained for downstream analysis. This slows down the advancement of our knowledge in the biology of CTCs and CTC clusters. Therefore, alternative liquid biopsy samples, such as Diagnostic Leukapheresis (DLA) products, are being used to maximise the capture of CTCs. DLA is a standard density-based methodology used in the clinic for the isolation of mononuclear cells (MNCs) by continuous centrifugation of large volumes of blood (figure 21). It was recently proposed that CTCs could be co-isolated together with the MNCs during this centrifugation, as they have a similar density to MNCs. Hence, in comparison with peripheral blood samples, DLA products would increase the likelihood of CTC detection, as they come from the processing of large volumes of blood. INÉS MARTÍNEZ PENA 152 Figure 21. Graphical representation of a patient during DLA. In collaboration with the group of Translational Research in Gynaecology of the Heinrich-Heine University (Düsseldorf, Germany), headed by Prof. Dr. rer. nat. Hans Neubauer, we worked with DLA products from mBC patients. DLA samples have a cellularity of 5x107 MNCs/mL, while peripheral blood samples show a cellularity of 0.53x106 cells/mL of blood. Due to the high cellularity of DLA samples, a suitable optimization of the workflow is essential for the successful detection of CTCs and CTC clusters in DLA products. DLA products were used to optimise a workflow for DLA sample processing, and for testing the feasibility of CTC cluster isolation as there was no previous data on the presence of CTC clusters in these samples. We used mBC DLA samples which resulted negative for CTCs in the CellSearch® system, and that we called DLACTC-. These samples were initially utilised to select the most suitable CTC isolation protocol for DLA separation in the ParsortixTM system. To this end, each DLACTCsample was divided into two halves, and a known number of MDA-MB231eGFP Luc or MCF-7 eGFP Luc cells, either as individual cells (to mimic single CTCs) or as cellular aggregates (CTC clusters), were spiked into each of them. Two different separation protocols were compared: a standard protocol (pressure: 99 mbar), and a separation protocol specifically designed to preserve CTC cluster physical integrity Results 153 (‘cluster separation protocol’, previously mentioned. Pressure: 50 mbar). Afterwards, tumour cells were captured in the ParsortixTM system, and individual CTCs (figure 22) and CTC cluster (figure 23) were counted after separation, and after the harvesting steps. Moreover, it was determined a prediction of the harvesting recovery efficiency. Figure 22. Representative images of MDA-MB-231eGFP Luc and MCF-7eGFP Luc individual cells isolated from DLA spiking samples. Scale bars: 50 µm. INÉS MARTÍNEZ PENA 256 257 INÉS MARTÍNEZ PENA 258 259 INÉS MARTÍNEZ PENA 260 261 INÉS MARTÍNEZ PENA 262 FAVOURABLE REPORT ANIMAL EXPERIMENTATION ETHICAL COMMITTEE OF THE UNIVERSITY OF SANTIAGO DE COMPOSTELA (CEEA) 263 ANIMAL EXPERIMENTATION TRAINING CERTIFICATE. B FUNCTION (EUTHANASIA) INÉS MARTÍNEZ PENA 264 265 ANIMAL EXPERIMENTATION TRAINING CERTIFICATE. C FUNCTION (ANIMAL PROCEDURES) INÉS MARTÍNEZ PENA 272 Gracias a Lorena C., Cris V., y a Lucía A. (y a Dani) por los cafés, las cañas, las conversaciones, las risas y sobre todo, por ser tan comprensiv@s conmigo estos últimos meses. Muchas gracias a mi entrenador Norberto G. y a mis compañeros de karate por recordarme la importancia de la constancia y por estar siempre para ayudarme y apoyarme, tanto dentro como fuera del dojo. Después de estos 23 años de aprendizaje juntos, más que un grupo somos una gran familia. Muchísimas gracias a mi familia, tanto a mis padres, como a mis abuelos maternos, por vuestro apoyo incondicional y por enseñarme la importancia de pelear por aquello que me guste, aunque a veces no sea fácil. Sois un ejemplo de fortaleza, trabajo duro, constancia y lo más importante, de cómo son las buenas personas. Finalmente, gracias a Martín por estar conmigo, apoyándome cuando lo he necesitado y por tu paciencia infinita estos últimos meses. Has sido una fuente de calma siempre que ha hecho falta y has hecho llevaderos hasta los días más amargos. He elegido el mejor compañero de viaje que se puede tener. Espero no haberme dejado a nadie en el tintero. Si fuese así, perdonadme, ha sido un largo viaje de aprendizaje en el que me he cruzado con muchas personas. Para los que me haya podido dejar atrás en estos agradecimientos, gracias por haber enriquecido mi experiencia a lo largo de esto 4 años. Si tuviese que resumir estos 4 años, diría que sido un viaje muy inspirador, con sus altos y sus bajos pero que me ha hecho crecer mucho, tanto a nivel profesional, como a nivel personal. Se cierra una etapa importante para mí por todo lo que ha implicado, así que ahora toca disfrutarlo y sobre todo ¡celebrarlo juntos! Agradecimientos 273 INÉS MARTÍNEZ PENA 274 References 275 REFERENCES INÉS MARTÍNEZ PENA 276 References 277 REFERENCES 1. Siegel, R. L., Miller, K. D. & Jemal, A. Cancer statistics, 2020. CA. Cancer J. Clin. 70, 7–30 (2020). 2. Bray, F. et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA. Cancer J. Clin. 68, 394–424 (2018). 3. Ahmad, A. Breast Cancer Statistics: Recent Trends. in 1–7 (2019). doi:10.1007/978-3030-20301-6_1. 4. Barnard, M. E., Boeke, C. E. & Tamimi, R. M. Established breast cancer risk factors and risk of intrinsic tumor subtypes. Biochim. Biophys. Acta - Rev. Cancer 1856, 73–85 (2015). 5. ROJAS, K. & STUCKEY, A. Breast Cancer Epidemiology and Risk Factors. Clin. Obstet. Gynecol. 59, 651–672 (2016). 6. Tao, Z. Q. et al. Breast Cancer: Epidemiology and Etiology. Cell Biochem. Biophys. 72, 333–338 (2015). 7. Eroles, P., Bosch, A., Alejandro PérezFidalgo, J. & Lluch, A. Molecular biology in breast cancer: Intrinsic subtypes and signaling pathways. Cancer Treat. Rev. 38, 698–707 (2012). INÉS MARTÍNEZ PENA 278 8. Sonnenblick, A., Fumagalli, D., Sotiriou, C. & Piccart, M. Is the differentiation into molecular subtypes of breast cancer important for staging, local and systemic therapy, and follow up? Cancer Treat. Rev. 40, 1089–1095 (2014). 9. Prat, A. et al. Clinical implications of the intrinsic molecular subtypes of breast cancer. The Breast 24, S26–S35 (2015). 10. Dai, X. et al. Breast cancer intrinsic subtype classification, clinical use and future trends. Am. J. Cancer Res. 5, 2929–2943 (2015). 11. Sorlie, T. et al. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc. Natl. Acad. Sci. 98, 10869–10874 (2001). 12. Gao, J. J. & Swain, S. M. Luminal A Breast Cancer and Molecular Assays: A Review. Oncologist 23, 556–565 (2018). 13. Ignatiadis, M. & Sotiriou, C. Luminal breast cancer: from biology to treatment. Nat. Rev. Clin. Oncol. 10, 494–506 (2013). 14. Garrido-Castro, A. C., Lin, N. U. & Polyak, K. Insights into molecular classifications of triple-negative breast cancer: Improving patient selection for treatment. Cancer Discov. 9, 176–198 (2019). References 279 15. Plichta, J. K. et al. Implications for Breast Cancer Restaging Based on the 8th Edition AJCC Staging Manual. Ann. Surg. 271, 169– 176 (2020). 16. Sobin, L. H. TNM: Evolution and relation to other prognostic factors. Semin. Surg. Oncol. 21, 3–7 (2003). 17. Giuliano, A. E., Edge, S. B. & Hortobagyi, G. N. Eighth Edition of the AJCC Cancer Staging Manual: Breast Cancer. Ann. Surg. Oncol. 25, 1783–1785 (2018). 18. Hortobagyi, G. N. et al. Breast. in AJCC Cancer Staging Manual (ed. Amin, M.B., Edge, S., Greene, F., Byrd, D.R., Brookland, R.K., Washington, M.K., Gershenwald, J.E., Compton, C.C., Hess, K.R., Sullivan, D.C., Jessup, J.M., Brierley, J.D., Gaspar, L.E., Schilsky, R.L., Balch, C.M., Winchester, D.P., Asare, E.A., Madera, L. R. (Eds. .) 589– 636 (Springer International Publishing, 2017). doi:10.1007/978-3-319-40618-3_48. 19. Giuliano, A. E. et al. Breast Cancer-Major changes in the American Joint Committee on Cancer eighth edition cancer staging manual. CA. Cancer J. Clin. 67, 290–303 (2017). 20. Massagué, J. & Obenauf, A. C. Metastatic colonization by circulating tumour cells. Nature 529, 298–306 (2016). INÉS MARTÍNEZ PENA 280 21. Lambert, A. W., Pattabiraman, D. R. & Weinberg, R. A. Emerging Biological Principles of Metastasis. Cell 168, 670–691 (2017). 22. Jemal, A., Siegel, R., Xu, J. & Ward, E. Cancer Statistics, 2010. CA. Cancer J. Clin. 60, 277–300 (2010). 23. Cheung, K. J. & Ewald, A. J. A collective route to metastasis: Seeding by tumor cell clusters. Science (80-. ). 352, 167–169 (2016). 24. Lintz, M., Muñoz, A. & Reinhart-King, C. A. The Mechanics of Single Cell and Collective Migration of Tumor Cells. J. Biomech. Eng. 139, (2017). 25. Scully, O. J., Bay, B.-H., Yip, G. & Yu, Y. Breast cancer metastasis. Cancer Genomics Proteomics 9, 311–20 (2012). 26. Cheung, K. J. & Ewald, A. J. Illuminating breast cancer invasion: diverse roles for cell– cell interactions. Curr. Opin. Cell Biol. 30, 99–111 (2014). 27. Yu, W., Yang, L., Li, T. & Zhang, Y. Cadherin Signaling in Cancer: Its Functions and Role as a Therapeutic Target. Front. Oncol. 9, (2019). 28. Sousa, B., Pereira, J. & Paredes, J. The Crosstalk Between Cell Adhesion and Cancer References 281 Metabolism. Int. J. Mol. Sci. 20, 1933 (2019). 29. Kourtidis, A., Lu, R., Pence, L. J. & Anastasiadis, P. Z. A central role for cadherin signaling in cancer. Exp. Cell Res. 358, 78–85 (2017). 30. Li, D.-M. & Feng, Y.-M. Signaling mechanism of cell adhesion molecules in breast cancer metastasis: potential therapeutic targets. Breast Cancer Res. Treat. 128, 7–21 (2011). 31. Padmanaban, V. et al. E-cadherin is required for metastasis in multiple models of breast cancer. Nature 573, 439–444 (2019). 32. Blaschuk, O. W. N-cadherin antagonists as oncology therapeutics. Philos. Trans. R. Soc. B Biol. Sci. 370, 20140039 (2015). 33. Micalizzi, D. S., Haber, D. A. & Maheswaran, S. Cancer metastasis through the prism of epithelial-to-mesenchymal transition in circulating tumor cells. Mol. Oncol. 11, 770– 780 (2017). 34. Mendonsa, A. M., Na, T.-Y. & Gumbiner, B. M. E-cadherin in contact inhibition and cancer. Oncogene 37, 4769–4780 (2018). 35. Mittal, V. Epithelial Mesenchymal Transition in Tumor Metastasis. Annu. Rev. Pathol. Mech. Dis. 13, 395–412 (2018). INÉS MARTÍNEZ PENA 288 Heterogeneity in Circulating Tumor Cells: Sticky Friends on the Road to Metastasis. Cancers (Basel). 12, 1632 (2020). 78. Costa, C. & Dávila-Ibáñez, A. B. Methodology for the Isolation and Analysis of CTCs. in 45–59 (2020). doi:10.1007/978-3030-35805-1_4. 79. Cristofanilli, M. et al. Circulating Tumor Cells, Disease Progression, and Survival in Metastatic Breast Cancer. N. Engl. J. Med. 351, 781–791 (2004). 80. Danila, D. C. et al. Circulating Tumor Cell Number and Prognosis in Progressive Castration-Resistant Prostate Cancer. Clin. Cancer Res. 13, 7053–7058 (2007). 81. Abalde-Cela, S., Piairo, P. & Diéguez, L. The Significance of Circulating Tumour Cells in the Clinic. Acta Cytol. 63, 466–478 (2019). 82. N. Saucedo-Zeni, S. Mewes, R. Niestroj, L. Gasiorowski, D. Murawa, P. Nowaczyk, T. Tomasi, E. Weber, G.Dworacki, N.G. Morgenthaler, H. Jansen, C. Propping, K. Sterzynska, W. Dyszkiewicz, M. Zabel, M. Kiechle, U. Reuning, M. Schmitt, K. L. A novel method for the in vivo isolation of circulating tumor cells from peripheral blood of cancer patients using a functionalized and structured medical wire. Int. J. Oncol. (2012) References 289 doi:10.3892/ijo.2012.1557. 83. Schneck, H. et al. EpCAM-Independent Enrichment of Circulating Tumor Cells in Metastatic Breast Cancer. PLoS One 10, e0144535 (2015). 84. Eslami-S, Z., Cortés-Hernández, L. E. & Alix-Panabières, C. Epithelial Cell Adhesion Molecule: An Anchor to Isolate Clinically Relevant Circulating Tumor Cells. Cells 9, 1836 (2020). 85. Giordano, A. & Cristofanilli, M. CTCs in Metastatic Breast Cancer. in 193–201 (2012). doi:10.1007/978-3-642-28160-0_18. 86. Bredemeier, M. et al. Gene Expression Signatures in Circulating Tumor Cells Correlate with Response to Therapy in Metastatic Breast Cancer. Clin. Chem. 63, 1585–1593 (2017). 87. Keup, C. et al. RNA Profiles of Circulating Tumor Cells and Extracellular Vesicles for Therapy Stratification of Metastatic Breast Cancer Patients. Clin. Chem. 64, 1054–1062 (2018). 88. Labib, M. & Kelley, S. O. Circulating tumor cell profiling for precision oncology. Mol. Oncol. 15, 1622–1646 (2021). 89. Pereira‐Veiga, T. et al. CTCs‐derived INÉS MARTÍNEZ PENA 290 xenograft development in a triple negative breast cancer case. Int. J. Cancer ijc.32001 (2018) doi:10.1002/ijc.32001. 90. Ramirez, J.-M. et al. Prognostic Relevance of Viable Circulating Tumor Cells Detected by EPISPOT in Metastatic Breast Cancer Patients. Clin. Chem. 60, 214–221 (2014). 91. Boya, M., Chu, C.-H., Liu, R., OzkayaAhmadov, T. & Sarioglu, A. F. Circulating Tumor Cell Enrichment Technologies. in 25– 55 (2020). doi:10.1007/978-3-030-26439-0_2. 92. Franken, A. et al. Label-Free Enrichment and Molecular Characterization of Viable Circulating Tumor Cells from Diagnostic Leukapheresis Products. Clin. Chem. 65, 549– 558 (2019). 93. Stoecklein, N. H., Fischer, J. C., Niederacher, D. & Terstappen, L. W. M. M. Challenges for CTC-based liquid biopsies: low CTC frequency and diagnostic leukapheresis as a potential solution. Expert Rev. Mol. Diagn. 16, 147–164 (2016). 94. Reinhardt, F. et al. Diagnostic Leukapheresis Enables Reliable Transcriptomic Profiling of Single Circulating Tumor Cells to Characterize Inter-Cellular Heterogeneity in Terms of Endocrine Resistance. Cancers (Basel). 11, 903 (2019). References 291 95. Andree, K. C. et al. Toward a real liquid biopsy in metastatic breast and prostate cancer: Diagnostic LeukApheresis increases CTC yields in a European prospective multicenter study (CTCTrap). Int. J. Cancer 143, 2584–2591 (2018). 96. Vona, G. et al. Isolation by Size of Epithelial Tumor Cells. Am. J. Pathol. 156, 57–63 (2000). 97. Miller, M. C., Robinson, P. S., Wagner, C. & O’Shannessy, D. J. The ParsortixTM Cell Separation System—A versatile liquid biopsy platform. Cytom. Part A 93, 1234–1239 (2018). 98. Cohen, E. N. et al. Antigen-agnostic microfluidics-based circulating tumor cell enrichment and downstream molecular characterization. PLoS One 15, e0241123 (2020). 99. Koch, C. et al. Pre-Analytical and Analytical Variables of Label-Independent Enrichment and Automated Detection of Circulating Tumor Cells in Cancer Patients. Cancers (Basel). 12, 442 (2020). 100. Hvichia, G. E. et al. A novel microfluidic platform for size and deformability based separation and the subsequent molecular characterization of viable circulating tumor INÉS MARTÍNEZ PENA 292 cells. Int. J. Cancer 138, 2894–2904 (2016). 101. Alshareef, M. et al. Separation of tumor cells with dielectrophoresis-based microfluidic chip. Biomicrofluidics 7, 011803 (2013). 102. Fabbri, F. et al. Detection and recovery of circulating colon cancer cells using a dielectrophoresis-based device: KRAS mutation status in pure CTCs. Cancer Lett. 335, 225–231 (2013). 103. Castro-Giner, F. & Aceto, N. Tracking cancer progression: from circulating tumor cells to metastasis. Genome Med. 12, 31 (2020). 104. Aceto, N. Bring along your friends: Homotypic and heterotypic circulating tumor cell clustering to accelerate metastasis. Biomed. J. 43, 18–23 (2020). 105. VIRCHOW, R. As Based upon Physiological and Pathological Histology. Nutr. Rev. 47, 23–25 (1989). 106. Watanabe, S. The metastasizability of tumor cells. Cancer 7, 215–223 (1954). 107. Hong, Y., Fang, F. & Zhang, Q. Circulating tumor cell clusters: What we know and what we expect (Review). Int. J. Oncol. 49, 2206– 2216 (2016). 108. Cheung, K. J. et al. Polyclonal breast cancer metastases arise from collective dissemination References 293 of keratin 14-expressing tumor cell clusters. Proc. Natl. Acad. Sci. 113, E854–E863 (2016). 109. Liu, X. et al. Homophilic CD44 Interactions Mediate Tumor Cell Aggregation and Polyclonal Metastasis in Patient-Derived Breast Cancer Models. Cancer Discov. 9, 96– 113 (2019). 110. Glinsky, V. V et al. Intravascular metastatic cancer cell homotypic aggregation at the sites of primary attachment to the endothelium. Cancer Res. 63, 3805–11 (2003). 111. Haeger, A., Krause, M., Wolf, K. & Friedl, P. Cell jamming: Collective invasion of mesenchymal tumor cells imposed by tissue confinement. Biochim. Biophys. Acta - Gen. Subj. 1840, 2386–2395 (2014). 112. Küsters, B. et al. Micronodular transformation as a novel mechanism of VEGF-A-induced metastasis. Oncogene 26, 5808–5815 (2007). 113. Fidler, I. J. & Talmadge, J. E. Evidence that intravenously derived murine pulmonary melanoma metastases can originate from the expansion of a single tumor cell. Cancer Res. 46, 5167–71 (1986). 114. Thangavel, H. et al. A CTC-Cluster-Specific Signature Derived from OMICS Analysis of Patient-Derived Xenograft Tumors Predicts INÉS MARTÍNEZ PENA 294 Outcomes in Basal-Like Breast Cancer. J. Clin. Med. 8, 1772 (2019). 115. Paoli, P., Giannoni, E. & Chiarugi, P. Anoikis molecular pathways and its role in cancer progression. Biochim. Biophys. Acta - Mol. Cell Res. 1833, 3481–3498 (2013). 116. Paoletti, C. et al. Significance of Circulating Tumor Cells in Metastatic Triple-Negative Breast Cancer Patients within a Randomized, Phase II Trial: TBCRC 019. Clin. Cancer Res. 21, 2771–2779 (2015). 117. Gkountela, S. et al. Circulating Tumor Cell Clustering Shapes DNA Methylation to Enable Metastasis Seeding. Cell 176, 98112.e14 (2019). 118. Reduzzi, C. et al. Circulating Tumor Cell Clusters Are Frequently Detected in Women with Early-Stage Breast Cancer. Cancers (Basel). 13, 2356 (2021). 119. Krol, I. et al. Detection of clustered circulating tumour cells in early breast cancer. Br. J. Cancer 125, 23–27 (2021). 120. Silvestri, M. et al. Detection of Genomically Aberrant Cells within Circulating Tumor Microemboli (CTMs) Isolated from EarlyStage Breast Cancer Patients. Cancers (Basel). 13, 1409 (2021). References 295 121. Bidard, F.-C. et al. Clinical validity of circulating tumour cells in patients with metastatic breast cancer: a pooled analysis of individual patient data. Lancet Oncol. 15, 406–414 (2014). 122. Zhang, L. et al. Meta-Analysis of the Prognostic Value of Circulating Tumor Cells in Breast Cancer. Clin. Cancer Res. 18, 5701– 5710 (2012). 123. Smerage, J. B. et al. Circulating Tumor Cells and Response to Chemotherapy in Metastatic Breast Cancer: SWOG S0500. J. Clin. Oncol. 32, 3483–3489 (2014). 124. Amintas, S. et al. Circulating Tumor Cell Clusters: United We Stand Divided We Fall. Int. J. Mol. Sci. 21, 2653 (2020). 125. Wang, C. et al. Longitudinally collected CTCs and CTC-clusters and clinical outcomes of metastatic breast cancer. Breast Cancer Res. Treat. 161, 83–94 (2017). 126. Mu, Z. et al. Prospective assessment of the prognostic value of circulating tumor cells and their clusters in patients with advancedstage breast cancer. Breast Cancer Res. Treat. 154, 563–571 (2015). 127. Jansson, S., Bendahl, P.-O., Larsson, A.-M., Aaltonen, K. E. & Rydén, L. Prognostic impact of circulating tumor cell apoptosis and INÉS MARTÍNEZ PENA 296 clusters in serial blood samples from patients with metastatic breast cancer in a prospective observational cohort. BMC Cancer 16, 433 (2016). 128. Larsson, A.-M. et al. Longitudinal enumeration and cluster evaluation of circulating tumor cells improve prognostication for patients with newly diagnosed metastatic breast cancer in a prospective observational trial. Breast Cancer Res. 20, 48 (2018). 129. Szczerba, B. M. et al. Neutrophils escort circulating tumour cells to enable cell cycle progression. Nature 566, 553–557 (2019). 130. Costa, C. et al. Analysis of a Real-World Cohort of Metastatic Breast Cancer Patients Shows Circulating Tumor Cell Clusters (CTC-clusters) as Predictors of Patient Outcomes. Cancers (Basel). 12, 1111 (2020). 131. Hou, J.-M. et al. Clinical Significance and Molecular Characteristics of Circulating Tumor Cells and Circulating Tumor Microemboli in Patients With Small-Cell Lung Cancer. J. Clin. Oncol. 30, 525–532 (2012). 132. Hou, J.-M. et al. Circulating Tumor Cells as a Window on Metastasis Biology in Lung Cancer. Am. J. Pathol. 178, 989–996 (2011). References 297 133. Kats-Ugurlu, G. et al. Circulating tumour tissue fragments in patients with pulmonary metastasis of clear cell renal cell carcinoma. J. Pathol. 219, 287–293 (2009). 134. Topal, B., Roskams, T., Fevery, J. & Penninckx, F. Aggregated colon cancer cells have a higher metastatic efficiency in the liver compared with nonaggregated cells: an experimental study. J. Surg. Res. 112, 31–37 (2003). 135. Aceto, N., Toner, M., Maheswaran, S. & Haber, D. A. En Route to Metastasis: Circulating Tumor Cell Clusters and Epithelial-to-Mesenchymal Transition. Trends in Cancer 1, 44–52 (2015). 136. Wang, C. et al. Longitudinally collected CTCs and CTC-clusters and clinical outcomes of metastatic breast cancer. Breast Cancer Res. Treat. 161, 83–94 (2017). 137. Waks, A. G. & Winer, E. P. Breast Cancer Treatment. JAMA 321, 288 (2019). 138. Paoletti, C. et al. Significance of Circulating Tumor Cells in Metastatic Triple-Negative Breast Cancer Patients within a Randomized, Phase II Trial: TBCRC 019. Clin. Cancer Res. 21, 2771–2779 (2015). 139. Paoletti, C. et al. Circulating Tumor Cell Clusters in Patients with Metastatic Breast INÉS MARTÍNEZ PENA 304 (2018).