Supercritical technology applied to the development of drug delivery systems for bone regeneration

TESE DE DOUTORAMENTO SUPERCRITICAL TECHNOLOGY APPLIED TO THE DEVELOPMENT OF DRUG DELIVERY SYSTEMS FOR BONE REGENERATION Leticia Goimil García ESCOLA DE DOUTORAMENTO INTERNACIONAL PROGRAMA DE DOUTORAMENTO EN INVESTIGACIÓN E DESENVOLVEMENTO DE MEDICAMENTOS SANTIAGO DE COMPOSTELA ANO 2019 DECLARACIÓN DO AUTOR/A DA TESE Supercritical technology applied to the development of drug delivery systems for bone regeneration D./Dna. Leticia Goimil García 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 selo caso, na tese faise referencia ás colaboracións que tivo este traballo. 3) A tese é a versión definitiva presentada para a súa defensa e coincide coa versión enviada en formato electrónico. 4) Confirmo que a tese non incorre en ningún tipo de plaxio doutros autores nin de traballos presentados por min para a obtención doutros títulos. En Santiago de Compostela, 31 de Maio de 2019 Asdo. AUTORIZACIÓN DO DIRECTOR / TITOR DA TESE Supercritical technology applied to the development of drug delivery systems for bone regeneration Dna. Carmen Álvarez Lorenzo D. Carlos A. García González INFORMA/N: Que a presente tese, correspóndese co traballo realizado por Dna. Leticia Goimil García, baixo a nosa dirección, e a utorizamos 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. En Santiago de Compostela, 31 de Maio de 2019 Asdo. Asdo “...Human beings, little bags of thinking water held up briefly by fragile accumulations of calcium...” ― Terry Pratchett El “hacerse a uno mismo” es un concepto muy extendido — generalmente ligado al triunfo del esfuerzo (uni)personal— y, sin embargo, también muy erróneo. Son muy pocas las cosas que se pueden conseguir enteramente en solitario, si es que hay alguna. Tampoco debe ser el objetivo. Pedir ayuda no es una muestra de debilidad; una tarea que parecía inabarcable se vuelve menos intimidante al tener a alguien a nuestro lado. Especialmente en investigación, donde el fin último son las personas, no se debe minimizar la importancia de lo humano. A veces son nuestra inspiración, nuestra guía o nuestra motivación. Un empujón cuando crees que no puedes más. Una sonrisa en la que apuntalarse para seguir adelante. La realización de una tesis es uno de esos trabajos imposibles en apariencia, tan inmensos que a veces es difícil no desistir, no perder el rumbo. Por suerte, no estamos solos. El día que comencé la carrera de Farmacia, paseando con mi abuelo Manolo por la zona vieja, él señaló una tienda y dijo “Cuando termines la carrera te compro uno de esos anillos”. No pudo llegar a verme licenciarme. El día de la graduación llevé unos pendientes que me había regalado años atrás; aunque no significaba nada para el resto del mundo, era un detalle importante para mí. Una forma de tenerlo presente a mi manera. Tampoco ha podido acompañarme hasta aquí mi abuela Maru, a quien recuerdo en su cocina contándome la historia de El Cid y enseñándome que “arisco” es una de las peores cosas que se pueden ser. Tal vez ya no estén, pero este logro también es en parte suyo, y sencillamente no podía no dedicárselo. Gracias por tanto. Huelga decir que la realización de esta tesis hubiese sido irrealizable sin la dirección y la dedicación de Carmen Álvarez Lorenzo y Carlos García González, sin la inestimable colaboración del profesor José Luis Gómez Amoza o sin la oportunidad del profesor Hermínio C. de Sousa y la doctora Mara E.M. Braga para realizar una estancia de investigación en su laboratorio. Tampoco estaría aquí de no ser por la profesora Begoña Seijo y el amor por la Tecnología Farmacéutica que transmitía en sus clases. Por supuesto, también agradezco a los demás profesores del departamento, particularmente a Ángel Concheiro, Mariana Landín, Ramón Martínez I. RESUMEN La medicina regenerativa está experimentando un creciente auge con un enorme desarrollo en los últimos años. Esta tendencia se halla ligada a la creciente demanda de soluciones sociosanitarias para problemas derivados de los procesos naturales de envejecimiento y de las necesidades de reparación de lesiones asociadas no sólo a accidentes laborales y de tráfico sino también a nuevos hábitos de vida, como la práctica generalizada del deporte. Las aproximaciones clásicas para hacer frente al deterioro o la pérdida de un tejido se basan, principalmente, en incorporar un material con capacidad para suplir la pérdida física pero sin recuperar la función completa de ese tejido. Un buen ejemplo de ello son las prótesis metálicas que se implantan en sustitución de huesos o porciones de hueso en el caso de fracturas grandes o resecciones quirúrgicas asociadas a tumores. El principal problema de las prótesis inertes es que, aunque sirven de puente entre los extremos de la lesión, también actúan como obstáculo frente a la regeneración natural del tejido. En el caso de una persona en crecimiento, esto obliga incluso a intervenciones periódicas para sustituir la prótesis por otra de dimensiones mayores. En el caso de un adulto, la durabilidad de la prótesis puede resultar insuficiente dada la creciente esperanza de vida de la población. El estudio del proceso de regeneración natural de los tejidos en los seres vivos encierra un gran interés para solventar estas limitaciones ligadas a las prótesis inertes. Se pretende aprovechar el conocimiento sobre las cascadas de curación y reparación para plantear aproximaciones biomiméticas. La gran mayoría de los tejidos de nuestro organismo se encuentran en un proceso continuo de regeneración para suplir el desgaste natural que permite que las pequeñas lesiones se reparen espontáneamente. Cuando la lesión supera un cierto tamaño denominado crítico (que depende de cada tejido), los mecanismos de regeneración fallan y el organismo para LETICIA GOIMIL GARCÍA _____________________________________________________________ ii mantener la homeostasis desencadena procesos de cicatrización que dan lugar a tejidos de prestaciones inferiores a las del tejido no lesionado. Si la herida es muy grande, puede ocurrir que no se produzca la cicatrización, convirtiéndose en una lesión crónica. La regeneración natural de los tejidos se basa, a grandes rasgos, en la formación de un nicho o soporte adecuado en el que se puedan asentar, diferenciarse y proliferar las células. Para imitar este proceso resulta necesario, por lo tanto, contar con materiales que puedan servir de andamios (scaffolds, en la terminología inglesa), que alberguen las células adecuadas o las sustancias activas necesarias para atraerlas y regular su crecimiento. La utilización de porciones sanas del mismo tejido como scaffold en una zona lesionada ha sido ampliamente investigada. El injerto de hueso del propio paciente (autólogo) reúne las tres características deseables para la regeneración de lesiones de tamaño superior al crítico: osteoconducción, es decir, sirve de guía (andamio) en el proceso de reparación del hueso natural; osteoinducción, contiene las sustancias necesarias para fomentar que las células no diferenciadas se conviertan en osteoblastos activos; y osteogénesis, aporta células de hueso que deben contribuir a la remodelación ósea. A pesar de los buenos resultados clínicos que se obtienen, el injerto autólogo tiene los inconvenientes de requerir una intervención adicional para extraer tejido de la zona dadora, que en pacientes mayores o con alguna patología crónica puede ser un área muy limitada y de curación lenta, e implica el riesgo de infecciones. Las alternativas naturales con injertos procedentes de donantes (aloinjertos) o de animales (xenoinjertos) tienen riesgo de rechazo, de transmisión de enfermedades y su disponibilidad es limitada. Además hay que tener en cuenta que sólo el injerto autólogo es osteogénico, ya que en las alternativas mencionadas el tejido se debe descelularizar antes de la implantación. El tejido óseo es, después de la sangre, el tejido con mayor demanda lo que ha motivado un mayor esfuerzo investigador en el diseño de scaffolds sintéticos específicos para hueso. Los scaffolds sintéticos deben tener una estructura 3D porosa que guie y promueva el crecimiento de los tejidos, que facilite la difusión de nutrientes y oxígeno y la eliminación de productos de desecho de las células, y que Resumen _____________________________________________________________ iii cuente con propiedades mecánicas similares al tejido a reemplazar. La porosidad debe ser de entre el 65 y el 80 %, y debe combinar microporos (~ 10 µm) que faciliten la adhesión celular y macroporos (~ 350 µm) que permitan la colonización celular no sólo en superficie sino también en profundidad. Además de ser biocompatible ‒es decir, que ni los materiales ni los productos de degradación pueden resultar tóxicos para las células‒ el scaffold debe biodegradarse a la misma velocidad a la que se forma el nuevo tejido, de manera que proporcione soporte mecánico sin impedir la regeneración completa del tejido. El scaffold sintético no debe desempeñar un papel meramente pasivo en la regeneración sino que debe modular la respuesta del organismo al daño y también el comportamiento de las células que lo colonicen. En este sentido, se ha ensayado la incorporación de una amplia variedad de sustancias activas, principalmente factores de crecimiento con capacidad osteoinductiva. Desde un punto de vista tecnológico, los dos principales retos que plantea la preparación de scaffolds rígidos porosos son: a) conseguir que su arquitectura porosa interna se ajuste a las características de la estructura del hueso, con una población bimodal de poros interconectados; y b) mantener la cantidad y la actividad de los factores de crecimiento o de los fármacos necesarios para promover la adhesión, diferenciación y proliferación celular. Los procedimientos clásicos de preparación de materiales porosos implican, en su mayoría, la exposición de polímeros y porógenos a disolventes orgánicos volátiles, la aplicación de calor hasta alcanzar temperaturas relativamente elevadas para fundir el polímero y/o evaporar el porógeno, o la lixiviación con disolventes que arrastren el porógeno. Ninguna de estas técnicas permite un control preciso de la estructura porosa. Además, tienen el riesgo de que las sustancias activas se pierdan por degradación o por arrastre, lo que redunda en un bajo rendimiento de carga. En los últimos años, la tecnología de fluidos supercríticos y comprimidos se está revelando como una herramienta capaz de dar respuesta a los retos principales de la medicina regenerativa. Si bien esta tecnología se viene utilizando desde hace tiempo en campos muy variados, tan sólo recientemente ha empezado a ensayarse con LETICIA GOIMIL GARCÍA _____________________________________________________________ iv biomateriales. Los fluidos alcanzan condiciones supercríticas cuando la temperatura y la presión se sitúan por encima de las del punto crítico, caracterizados por su temperatura (Tc) y presión crítica (Pc), respectivamente. Los fluidos supercríticos están dotados de propiedades fisicoquímicas singulares que los hacen muy útiles para el procesado de materiales, destacando su elevada difusividad, que permite que penetren en materiales con morfologías y texturas complejas. El CO2 supercrítico (scCO2) es el fluido supercrítico más empleado debido a que las condiciones moderadas de su punto crítico (73,8 bar, 31,1 ºC) son adecuadas para materiales lábiles. Además, no es inflamable, es económico, relativamente inerte y está considerado como seguro (GRAS). El interés en el scCO2 para la preparación de scaffolds porosos se basa en su capacidad para disolverse en polímeros amorfos y semicristalinos. El scCO2 actúa como plastificante, reduciendo el punto de fusión y la temperatura de transición vítrea de los polímeros. Los poros se forman en el material durante la despresurización cuando la remoción del CO2 da lugar a una inestabilidad termodinámica que provoca la expansión de la matriz polimérica. La porosidad del producto resultante depende de la cantidad de CO2 que se adsorbe durante el espumado, y se puede ajustar regulando la temperatura, la presión y el tiempo de procesado. La interconectividad de los poros depende de la velocidad de despresurización y de la velocidad de enfriamiento como parámetros de procesado principales. La evaporación completa del CO2 resulta en un producto sólido poroso libre de disolventes y de impurezas. Resulta una técnica muy versátil que se puede aplicar al diseño de sistemas de diferente composición (orgánicos e híbridos), morfología (esponjas, micro y nanopartículas, monolitos), porosidad (meso y macroporosidad) y arquitectura interna (homogénea, multicapa, multicomponente). Esta Tesis Doctoral se planteó con el objetivo general de diseñar scaffolds porosos útiles en regeneración ósea. Para ello, se seleccionaron como componentes principales poliésteres biodegradables, tales como la poli(-caprolactona), PCL, y el ácido poli(láctico-co-glicólico), PLGA. Se implementaron procedimientos de procesado utilizando scCO2 o CO2 comprimido para obtener tres Resumen _____________________________________________________________ v tipos de scaffolds: (1) scaffolds de PCL y PLGA con almidón pregelificado para cesión controlada de dexametasona; (2) scaffolds de PCL conteniendo aerogeles de almidón para cesión sostenida de ketoprofeno para regeneración ósea; y (3) scaffolds de PCL y aerogeles de fibroína de seda para cesión controlada de dexametasona. De acuerdo con estos objetivos, el trabajo se llevó a cabo en tres etapas: 1. Preparación de scaffolds, mediante espumado con CO2 comprimido, de PCL y PLGA con almidón pregelificado y cargados con dexametasona para regeneración ósea, y evaluación de su estabilidad durante el almacenamiento. Esta etapa de la Tesis se centró en dos objetivos principales: evaluar las posibilidades que ofrece la incorporación de almidón pregelificado para modular la estructura porosa y la cesión de sustancias activas a partir de los scaffolds; y obtener información sobre la estabilidad de los scaffolds en condiciones de almacenamiento controladas que reflejan las condiciones climáticas de la zona II según el International Council for Harmonization (ICH). Este último es un aspecto crítico para determinar el periodo de vida útil del scaffold una vez preparado y establecer las características requeridas para el material de acondicionamiento. No obstante, hasta la fecha se ha prestado muy poca atención a la estabilidad de los scaffolds por lo que la información disponible es muy escasa. Para llevar a cabo este estudio se contó con la colaboración de Ph. Jaeger de la compañía Eurotechnica GmbH, que evaluó la capacidad de absorción de CO2 en las matrices poliméricas mediante balanza de suspensión magnética, y de I. Ardao del grupo BioFarma de la Universidade de Santiago de Compostela, que realizó los estudios de modelización de la cesión in vitro de dexametasona. En primer lugar, se prepararon mezclas de PCL y PLGA (50:50) de baja viscosidad inherente con y sin almidón pregelificado y dexametasona. Las mezclas se compactaron utilizando una máquina de comprimir excéntrica para obtener compactos de 400 mg de masa (14x10x2.6 mm). A continuación, se aplicó un protocolo de espumado asistido por CO2 comprimido previamente patentado en el grupo I+D LETICIA GOIMIL GARCÍA _____________________________________________________________ vi Farma, y que se adaptó a los compactos. Los compactos se expusieron a CO2 a 26 ºC y 60 bar durante 30 min. Posteriormente, la presión se redujo a 30 bar a una velocidad de 10 bar/min, y se volvió a incrementar de nuevo a 60 bar mediante la incorporación de CO2 líquido a 1 ºC. Este protocolo se repitió tres veces antes de proceder a la despresurización a 10 bar/min. Se eligieron estas condiciones de temperatura y presión suaves para hacerlas compatibles con la estabilidad de moléculas activas lábiles como la dexametasona, y también para evitar un espumado descontrolado del PLGA de baja viscosidad inherente contenido en el scaffold. A continuación, los scaffolds se caracterizaron en cuanto a características morfológicas, comportamiento frente a la temperatura, biodegradabilidad y propiedades mecánicas. El proceso de espumado hizo que el volumen de los scaffolds aumentara cuatro veces con respecto al volumen de los compactos no procesados, lo que supone valores de porosidad total de 68-75 %. Los poros presentaron dos poblaciones de macroporos, una de tamaño comprendido entre 50 y 200 µm, y otra de unas 10 µm de tamaño medio. El efecto de la composición de los scaffolds sobre la velocidad de erosión se evaluó en tampón fosfato salino de pH 7,4. La velocidad de erosión fue similar en todos los scaffolds durante los primeros 14 días, con una pérdida de peso de ~25 % y que está asociada a la degradación del PLGA. La incorporación tanto de dexametasona como de almidón pregelificado dio lugar a valores de pérdidas de masa más elevadas a partir de los 14 días, y que se justifican por la formación de poros mayores a medida que se disuelven/dispersan en el medio acuoso. Para evaluar la estabilidad, los scaffolds (no embalados) se mantuvieron en una cámara a 25 ºC y 65 % de humedad relativa, imitando las condiciones climáticas de la zona II según la ICH (Europa, Estados Unidos y Japón). Se monitorizaron los cambios en la estructura de los scaffolds, sus propiedades fisicoquímicas y mecánicas, y los perfiles de cesión de dexametasona al cabo de 1 y 3 meses. Se observaron disminuciones significativas del volumen de los scaffolds, debido a procesos de densificación, que fueron más marcadas en el caso de aquellos preparados con dexametasona y almidón. Al cabo de tres meses, los scaffolds preparados con almidón Resumen _____________________________________________________________ vii mostraron un descenso muy notable de la porosidad y el volumen total de los poros, lo que puede ser atribuido al hinchamiento de la fracción de amilopectina del almidón por efecto de la humedad. También se observaron cambios en las propiedades mecánicas y en los perfiles de cesión. En conjunto, los resultados obtenidos indican que la exposición directa de los scaffolds a 25 ºC y 65 % de humedad relativa causa efectos negativos sobre las propiedades de los scaffolds, de distinta magnitud dependiendo de la composición. La dexametasona y el almidón conducen a estructuras menos cristalinas y más sensibles a la humedad ambiental. La velocidad de cesión de dexametasona se produce mayoritariamente por la erosión de la matriz. El posible control del almidón pregelificado sobre la cesión de la dexametasona se reduce cuando los scaffolds se almacenan durante un período de tiempo prolongado. 2. Procesado supercrítico de aerogeles de almidón y de scaffolds de PCL conteniendo aerogeles para cesión sostenida de ketoprofeno para regeneración ósea. El objetivo principal de esta etapa de la Tesis fue preparar scaffolds porosos de poli(-caprolactona), PCL, conteniendo ketoprofeno y micropartículas de aerogel de almidón. El ketoprofeno es un anti-inflamatorio no esteroídico (AINE) que se administra por vía oral durante el período postoperatorio para aliviar el dolor y reducir la inflamación. El ketoprofeno no inhibe el proceso de curación del hueso, por lo que la cesión localizada podría permitir regular la respuesta inflamatoria que está normalmente asociada al proceso de implantación utilizando dosis bajas y minimizando efectos secundarios. Por su parte, las micropartículas de aerogel de almidón se caracterizan por una elevada mesoporosidad y pueden incorporar sustancias activas. La hipótesis del trabajo fue que la incorporación de micropartículas de aerogel de almidón debe permitir modular la estructura porosa del scaffold y la velocidad de cesión del ketoprofeno. Esta parte de la Tesis se llevó a cabo en colaboración con el grupo de H. C. de Sousa del Chemical Engineering Department, FCTUC, de la Universidade de Coimbra en el marco de una estancia predoctoral de 5 meses. LETICIA GOIMIL GARCÍA _____________________________________________________________ viii En primer lugar se puso a punto, por primera vez, un procedimiento de preparación de microesferas de aerogel de almidón del tamaño de una micra. Para ello se desarrolló un procedimiento de emulsión-gelificación en el que se partió de una emulsión de fase interna acuosa (dispersión de almidón al 15 %) y de fase externa oleosa (aceite de parafina) utilizando un agente emulsificante adecuado. La mezcla se llevó a autoclave, se aplicaron ultrasonidos manteniendo la mezcla a 95 ºC y después se enfrió para la completa gelificación del almidón. Las microsferas de hidrogel de almidón obtenidas se sumergieron en etanol y el alcogel resultante se secó en condiciones supercríticas para obtener las microesferas de aerogel. Parte de las microesferas se cargaron con ketoprofeno utilizando también scCO2 mediante impregnación supercrítica. Se prepararon lotes de scaffolds de PCL sin/con aerogel y sin/con ketoprofeno libre en el scaffold o previamente incorporado al aerogel. Las mezclas de los componentes se expusieron a scCO2 a 37 ºC y 140 bar, durante 60 min en condiciones estáticas. La despresurización se llevó a cabo a una velocidad de venteo de 1,8 g/min hasta presión atmosférica. Todos los scaffolds se caracterizaron en cuanto a sus propiedades morfológicas, físico-químicas y mecánicas mediante análisis de adsorción-desorción de N2, microscopía electrónica de barrido, porosimetría de intrusión de mercurio, modelado 3D, análisis mecánico dinámico y calorimetría diferencial de barrido. Los andamios preparados con aerogeles de almidón presentaron mayor porosidad e interconectividad de los poros, lo que debe facilitar la penetración de células madre mesenquimales, aunque a costa de un pequeño debilitamiento de las propiedades mecánicas. Los estudios de cesión de ketoprofeno se llevaron a cabo sumergiendo piezas de scaffold en tampón fosfato salino de pH 7,4, en condiciones sink, a 37 ºC y bajo agitación. La monitorización de la concentración de ketoprofeno en el medio de cesión reveló que todos los scaffolds podían sostener la cesión durante varios días. No obstante, la cesión fue más rápida desde los scaffolds preparados con aerogeles de almidón, probablemente debido a la más alta accesibilidad del agua a la estructura interna del scaffold. Resumen _____________________________________________________________ ix En su conjunto, los resultados de esta etapa ponen de manifiesto que la combinación de PCL con microesferas de aerogel de almidón y el procesado de las mezclas mediante espumado supercrítico permite obtener scaffolds cargados con ketoprofeno que cuentan con una estructura porosa, en términos de porosidad total, distribución de tamaños de poro e interconectividad, adecuada para el crecimiento celular y capaz de proporcionar perfiles de cesión sostenida de sustancias activas. 3. Preparación de scaffolds de PCL y aerogeles de fibroína de seda con dexametasona, mediante espumado con scCO2, para regeneración ósea. En estudios previos, se ha observado que la fibroína de seda promueve la adhesión y el crecimiento de células madre mesenquimales en scaffolds. Sin embargo, su utilización en forma de partículas de aerogel no ha sido investigada previamente. Por otra parte, es conocido que la dexametasona base y la dexametasona-21-fosfato de sodio presentan una solubilidad distinta en medios acuosos, lo que puede condicionar su velocidad de cesión. Tomando como base estas premisas, en la última etapa de la Tesis, se prepararon scaffolds a partir de mezclas de PCL, micropartículas de aerogel de fibroína de seda y dexametasona con el objetivo de contrastar las hipótesis siguientes: (i) las micropartículas de aerogel de fibroína de seda pueden mejorar la estructura porosa del scaffold y facilitar la infiltración celular y el transporte de fluido biológico; y (ii) la incorporación de dexametasona en forma de base o en forma de sal puede llevar a diferentes perfiles de cesión y, por lo tanto, a diferentes resultados de regeneración ósea. Para llevar a cabo este estudio se contó con la colaboración del grupo de J.L. Cenis del Instituto Murciano de Investigación y Desarrollo Agrario y Alimentario (IMIDA), que proporcionó fibroína de seda fresca obtenida en condiciones GMP, y del grupo de C. Evora de la Universidad de La Laguna, donde se llevaron a cabo los estudios in vivo. En primer lugar, se prepararon aerogeles de fibroína de seda a partir de una emulsión de fase interna disolución de fibroína de seda (8 %) y fase externa aceite de parafina, utilizando un agente emulsificante adecuado. Se ensayaron distintas condiciones de LETICIA GOIMIL GARCÍA ______________________________________________________________ xvi 1. INTRODUCTION 1.1. BONE FUNCTIONS, PROPERTIES AND REQUIREMENTS Bone is a connective tissue with an extracellular matrix composed by a mineral phase of hydroxyapatite (HA) as main component and responsible for its compressive strength, and an organic phase that mainly consists of collagen type I, which represent 65 and 35 % of dry weight, respectively (Fig. 1.1) (Henkel et al. 2013). This composition provides not only compressive strength and stiffness, but also enough ductility to withstand repetitive impacts without being fractured. Bone tissue has various functions in the organism: serves as point of attachment for skeletal muscles, provides support and protection for soft tissues and organs, contributes to mineral homeostasis by storing and releasing calcium and phosphate, and accommodates red bone marrow, which produces blood cells, and yellow bone marrow, a reservoir of triglycerides (Doblaré et al. 2004, Tortora and Derrickson 2012, Florencio-Silva et al. 2015). According to their dimensions and morphology, bones can be classified into long (e.g., femurs), short (carpals and tarsals), flat (cranial bones), sesamoid (kneecaps) and irregular (vertebrae). Bone shapes are interlinked to their mechanical functions: while long bones support body weight and help movement, flat and irregular bones protect internal organs, short bones provide stability and fine movement, and sesamoid bones protect tendons from stress (Tortora and Derrickson 2012). Bone structure can be divided into cancellous and cortical bone. Cortical bone is structured in osteons or haversian systems, composed by concentric lamellae surrounding a harvesian canal, which contains blood vessels, lymphatic vessels and nerves. There are spaces, lacunae, between the layers of lamellae containing the osteocytes, connected by canals with diameters ranging from 0.13 to 0.39 m denominated canaliculi (Fig. 1.1) (You et al. 2004). This wide net of LETICIA GOIMIL GARCÍA _____________________________________________________________ 2 Fig. 1.1: Structure of cortical bone: (a) 3D sketch of cortical bone, (b) cut of a Haversian system and (c) micrograph of a Haversian system. Reproduced from Sprio et al. (2011) with permission of Elsevier. canaliculi is filled with extracellular fluid, which provides osteocytes with nutrients and oxygen and removes waste products. Cancellous bone is built by the bone marrow and presents a lamellar architecture with a high porosity, up to 90 %. Cancellous bone is always surrounded by a layer of varying thickness of cortical bone, which possesses higher density and stiffness; compressive strength and Young modulus of cortical bone can reach up to 230 MPa and 30 GPa respectively, while the values for cancellous bone are 12 MPa and 0.5 GPa. This combination of cortical and cancellous bone confers enough resistance to bones, capable of absorbing impacts and resisting deformation, without adding too much weight to the structure (Velasco et al. 2015, Osterhoff et al. 2016). Despite its inert appearance, bone is a highly vascularized tissue with a constant remodeling that allows a high regenerative capacity of the tissue without scarring due to the action of bone cells: osteogenic cells, osteoblasts, osteocytes and osteoclasts (Fig. 1.2). Osteogenic cells, stem cells derived from mesenchymal stem cells, are the only bone cells that experiment cellular division, and differentiate into 1. Introduction _____________________________________________________________ 3 osteoblasts. Osteoblasts are located on the bone surface and are responsible for bone formation synthesizing and secreting the organic components of bone matrix, such as collagen fibers, and initiate calcification. Once they are surrounded by the growing bone matrix, osteoblasts turn into osteocytes, and acquire the function of regulating bone metabolism through the exchange of nutrients and metabolic wastes between blood and bone tissue. Lastly, osteoclasts are large macrophage-like cells that take part in bone resorption and calcium homeostasis by secreting enzymes and acids that decompose bone matrix. Fig. 1.2: Types and morphology of bone cells: osteocytes, osteoblasts, osteogenic cells and osteoclasts. OpenStax College - Anatomy & Physiology http://cnx.org/content/col11496/1.6/. The activity of osteoblasts and osteoclasts varies through life. Bone production is regulated by insulin like growth factor (IGF) and thyroid hormones, which promote cell division and stimulate the activity of the osteoblasts. This results in a predominance of bone formation over bone resorption in childhood. Androgens and estrogens increase the activity of osteoblasts in puberty and slow bone LETICIA GOIMIL GARCÍA _____________________________________________________________ 4 resorption during adulthood. In the later years of life, as the production of androgens and estrogens decays, bone resorption prevails. This bone resorption, and sequent decline of bone density, is more notable in women, since the decrease in estrogens is more pronounced (Marsell and Einhorn 2011, Tortora and Derrickson 2012, Florencio-Silva et al. 2015). In addition to hormones, mechanical stimulus also has a big impact on bone formation and maintenance. The added pressure when walking, jumping or weightlifting leads to an increment in bone production, enhancing bone strength. Insufficient mechanical stimulus, due to either a lack of activity (sedentary lifestyle) or a low gravity environment (astronauts), causes a loss of bone mass up to 1 % a week (Tortora and Derrickson 2012, Alghadir et al. 2015). Apart from declining bone density, modifications in bone ductility also increase its brittleness. These changes are not only due to the lower production of proteins in the later stages of life, but also because of modifications in the structure of the collagen matrix regarding the kind of cross-linking of the fibers: enzymatic cross-links, produced by lysyl hydroxylase and lysyloxidase, and non-enzymatic cross-links, formed for example by glycationor oxidation-induced non-enzymatic processes. This second mode of cross-linking increases with age, while enzymatic cross-linking decays, and has been found to be linked to lower toughness and resistance to crack propagation of bone tissue (Boskey and Coleman 2010). The healing of bone fractures can be naturally achieved when the size of the defect does not exceed the ability of self-repair. In these cases, the fracture can be resolved through either direct or indirect bone healing. Direct bone healing happens when the gap between the ends of the fracture is lower than 1 mm and requires anatomical stabilization so that lamellae, Harvesian canals and blood vessels can be repaired without the formation of callus. Indirect bone healing is the most common bone healing process and is favored by micromotion and weight-loading, so it does not need a rigid fixation. Indirect bone healing takes place following five steps: I) haematoma and inflammation, II) soft callus formation, III) fibrous tissue formation, IV) hard callus formation and V) callus resorption and 1. Introduction _____________________________________________________________ 5 remodeling (Fig. 1.3) (Marsell and Einhorn 2011, Einhorn and Gerstenfeld 2015). Haematoma is formed immediately with peripheral and intramedullary blood cells and is coagulated as a response to the inflammatory process, creating a support for callus formation. During the initial inflammation, which has its peak at 24 hours after the damage, the release of pro-inflammatory factors (tumor necrosis factor-α (TNF-α), interleukin-1 (IL-1), IL-6, IL-11 and IL-18) attracts inflammatory cells and promotes angiogenesis. Also, TNF-α is capable of inducing osteogenic differentiation of mesenchymal stem cells (MSCs). Nevertheless, despite the role of inflammation during bone regeneration, a sustained or chronic process has a negative impact since it is associated with a decreased angiogenesis and delayed osteoclast recruitment, apart from the direct tissue damage derived from inflammation (Abou-Khalil et al. 2014, Loi et al. 2016). Fig. 1.3: Steps of indirect fracture healing. Reproduced from Einhorn and Gesterfield (2015) with permission of Springer Nature. LETICIA GOIMIL GARCÍA _____________________________________________________________ 6 The reconstruction of bone tissue starts with the growth of a cartilaginous callus that helps stabilize the structure and its following ossification, starting at the ends of the fracture, until a hard callus is formed. As the mineralized cartilage is replaced with woven bone, the callus becomes more rigid and capable of bearing weight. However, this structure does not completely fulfill the original properties of the tissue. Remodelling starts 3-4 weeks after the injury, and it consists of the resorption of the hard callus and the deposition of lamellar bone carried by osteoclasts and osteoblast, respectively. This process usually takes years to complete until mature bone tissue is obtained (Marsell and Einhorn 2011). Critical size is the limit upon which a bone injury cannot spontaneously heal during a lifetime. Thus, this kind of damages requires the implantation of grafts that bind the ends of the fracture and enable bone healing to take place. The size for a bone defect to be considered critical sized in humans is generally considered to be above 1-3 cm and comprising at least 50 % loss of bone circumference. However, this value can differ since successful fracture healing not only depends on the amount of tissue lost, but also on bone regeneration capacity, which at the same time varies with the condition of the patient and the location of the defect, since regions such as the tibial diaphysis present low blood supply and soft tissue covering (Schemitsch 2017). The variability in the prognosis regarding full recovery of bone tissue implies that there is a need for grafts with different properties that suit the prospected healing timeline. 1.2. STRATEGIES FOR BONE REGENERATION Bone is the second most transplanted tissue, only surpassed by blood transfusions, with an incidence of over 2 million of procedures annually worldwide (Campana et al. 2014). Life expectancy in Europe has greatly increased from 40 to 80 years through the last century due to the advances in medicine, hygiene and nutrition, along with a decrease in natality. In Spain the population above 65 years represented an 18.8 % of total in 2017, and it is prospected to increase up to a 34.6 % by 2066 (Abellán García et al. 2018). Given that age is one of the main risks for fractures and post-surgical complications, 1. Introduction _____________________________________________________________ 7 researches aimed at preventing and treating bone fractures are of paramount importance in order to favor healthy and active ageing. 1.2.1. Biological grafts: autograft, allograft, xenograft and decellularized matrix The history of prosthesis goes back thousands of years ago, with the use of materials like shells, metals or wood to replace lost or damaged tissues; 2500 years BC the Egyptians implemented the use of gold wires to stabilize teeth and around 500 BC the Etruscans combined gold bands and bovine bones to fix oral problems. Bone grafts aim at repairing bone defects by both providing a structure for tissue to grow and stimulating cell growth and differentiation. Biological grafts are the current gold-standard in the clinical practice and can be classified according to their origin as: autografts, pieces of bone harvested from a healthy bone of the patient; allografts, from donors different from the patient; and xenografts, from animals. The use of bone grafts to repair bone tissue was firstly reported in 1668, when the Dutch surgeon Job van Meek’ren filled a defect in the cranium of a soldier using a canine xenograft (Klosterhoff et al. 2017). This intervention was considered blasphemous and the soldier was excommunicated; but when he tried to get the graft removed it was already incorporated into his cranium. It was not until 1820 that the first autologous bone grafting was performed by Philips von Walter to repair a cranium after trepanation (Von Walter 1821, Pryor et al. 2009). A survey of patients that suffered fracture nonunions and were treated with either autografts or allografts revealed the use of autografts led to statistically significant shorter healing procedures (172–225 days) compared to allografts (290–543 days). The patients whose fractures were repaired with autografts also had the lowest rate of surgical revisions (17 %) and incidence of postoperative infections (12.4 %), compared to those treated with allografts (47 % and 26.3 %, respectively) (Flierl et al. 2013). These factors have an economic impact derived from the costs of longer treatments and rehabilitation programs, but also on the quality of life of patients. LETICIA GOIMIL GARCÍA _____________________________________________________________ 8 Even though autologous bone grafts are the preferred clinical option due to their superior performance, their use implies longer surgical procedures, damage to healthy bone structures and postsurgical pain, so their application is limited (Table 1.1). Allografts and xenografts are restricted by their availability in tissue banks not able to give response to the growing demand and by possible immunological responses from the host. Decellularization appears as an alternative strategy to obtain scaffolds from allogenic or xenogenic tissues without generating this response by removing cells through a combination of mechanical, chemical and enzymatic techniques. This approach aims at obtaining the protein structure of the tissue, but a complete removal of the cells of a tissue or organ is difficult to attain and the processing technique usually alters the architecture of the extracellular matrix, leading to a high variability in studies regarding their application (Gilbert et al. 2006, Crapo et al. 2011). Table 1.1: Advantages and disadvantages of biological bone grafts. Bone graft Advantages Disadvantages Autograft Osteoinduction Osteoconduction Compatibility Availability Second surgical site Post-surgical complications Allo/xenograft Availability Immunological reactions Disease transmission Lower osteoinduction and osteoconduction Decellularized matrix Availability Compatibility Complex preparation High variability 1.2.2. Synthetic grafts The design of synthetic bone grafts aims to overcome the problems of availability and immunogenic reactions of biological grafts (Table 1.1). These artificial scaffolds must fulfill some basic requirements to be suitable for bone tissue engineering. First and foremost, they shall be biocompatible, meaning neither the materials nor their degradation products can be toxic to the cells, and also biodegradable, with a degradation rate that matches the growth of the new tissue. They need a high porosity, from 65 to 80 %, and a proper pore size distribution, with a microporosity of about 10 m, that improves cell attachment, and a macroporosity ideally reaching up to 1. Introduction _____________________________________________________________ 9 350 m (Lee et al. 2010) to allow cell proliferation. Since mechanical stimulus is an important factor for bone formation and mechanical stresses are inherent to the physiological function of bones, bone grafts should have mechanical properties similar to those of bone tissue. Lastly, the synthetic matrix should not merely be an inert structure, but also be endowed with bioactive properties such as promoting cell growth (osteoconductivity) and differentiation (osteoinductivity). This can be attained combining materials with different properties or by loading bioactive materials into the matrix that are locally delivered after implantation. 1.3. MATERIALS FOR BONE GRAFTS A wide range of materials has been tested for the fabrication of synthetic bone grafts, aiming to replicate the properties of bone tissue regarding porosity, strength or even composition. 1.3.1. Ceramics The development of bone grafts using calcium phosphate ceramics, such as HA or tricalcium phosphate (TCP), in their composition is a common approach since bone matrix is mainly composed by HA (Ghassemi et al. 2018). HA has a high biocompatibility and poor resorbablility, but this feature can be tuned with the incorporation of salts of zinc and manganese. TCP is also biocompatible, with a faster biodegradation than HA by dissolution and osteoclastic resorption (Chai et al. 2011). Apart from osteoconduction and osteoinduction, ceramic scaffolds show high compression resistance (up to 160 MPa). Nevertheless, the stiffness of bioceramics alone leads to poor fracture resistance, so composites combining ceramics with ductile components like collagen, the main component of bone extracellular matrix, have been designed to overcome this drawback (Fernandez de Grado et al. 2018). 1.3.2. Natural polymers Natural polymers have the intrinsic capacity of promoting cell attachment and differentiation, as well as degrading into non-toxic products in vivo; however, the resulting scaffolds often lack in LETICIA GOIMIL GARCÍA _____________________________________________________________ 16 between the tip of the capillary and the collector and the voltage of the electric field are among the processing parameters to tune the size of the fibers (Walker and Santoro 2017). Electrospining allows the incorporation of bioactive molecules into the blends of polymers, and is suitable for obtaining fibrous wound dressings and scaffolds for ligaments and tendons, but the use of solvents or high temperatures to liquefy the polymer can be a limit to its use (Lu et al. 2013). Fig. 1.9: Diagram of an electrospinning set-up. 1.4.2. Solvent-free preparation techniques 1.4.2.1. Melt molding Melt molding is used for polymers, being a suitable processing alternative for those insoluble in organic solvents, such as poly(glycolic acid), and consists of submitting them to high temperatures and then pouring the melted material in a mold to obtain the desired morphology. This technique allows the incorporation of additives such as ceramics or bioactive compounds, which are homogeneously dispersed in the polymeric matrix, and has the advantage of enabling a complete control on the morphology of the scaffold by changing the shape of the mold. Melt molding is sometimes coupled with particle leaching by adding a porogen to the melt that will be removed once the polymer solidifies (Fig. 1.10). Nevertheless, the high temperatures can degrade the compounds, and mold residues may have to be removed from the graft before further use (Allaf 2018). In the case of melt molding combined with particle leaching methods, there is also the risk of partly removing bioactive 1. Introduction _____________________________________________________________ 17 compounds loaded into the matrix, as previously remarked (Walker and Santoro 2017). Fig. 1.10: Diagram of melt molding-salt leaching. 1.4.2.2. Fused deposition modeling (FDM) FDM is a 3D printing method applied to thermoresponsive polymers. This technique heats the polymer above the glass transition temperature and extrudes it through a nozzle to build computerdesigned patterns layer by layer (Fig. 1.11). FDM avoids the use of organic solvents that are employed in other 3D printing techniques to solubilize polymers, and allows the combination of different polymers or polymer-ceramic blends. Layer thickness, speed of deposition, xyz resolution of movement, or the nozzle design are parameters that can be used to tune surface roughness and resolution. Moreover, FDM usually requires building supports that have to be removed to obtain the desired architecture (Do et al. 2015, Mohamed et al. 2015). Fig. 1.11: Diagram of fused deposition modeling. LETICIA GOIMIL GARCÍA _____________________________________________________________ 18 1.4.2.3. Gas foaming The principle of this technique is the generation of pores in a polymeric matrix through a nucleation-growth mechanism of gas bubbles that results in a macroporous material. The solvent-free version of this technique consists of three steps: 1) Dispersion of a porogen in a polymeric matrix. The porogen can be either a chemical blowing agent, that will decompose into an inert gas by a chemical reaction (sodium bicarbonate) or by thermal degradation (ammonium carbonate), or a physical blowing agent, such as an inert gas (nitrogen, argon or carbon dioxide) insufflated, or a volatile liquid (pentane) absorbed in the polymer. 2) Pore formation through the porogen removal. Gas bubbles are generated in this step following a nucleation-growth mechanism that results in pore formation after gas release. 3) Rapid solidification of the polymeric matrix. The temperature is rapidly lowered to allow the vitrification of the material (Annabi et al. 2010, Dehghani and Annabi 2011). Fig. 1.12: Diagram of gas foaming. 1. Introduction _____________________________________________________________ 19 The structures obtained through this technique are expanded polymeric foams with closed-cell or open-cell frames commonly used in the building sector. However, gas foaming can be adapted for regenerative medicine purposes by choosing biocompatible materials with non-toxic degradation products to create open-cell structures with suitable pore sizes and interconnectivity for cell penetration and growth (Dehghani and Annabi 2011). The limitations of this technology lay in the control of the pore size, pore interconnectivity and spatial homogeneity of the material, depending on the rheological properties of the liquefied polymer during foaming. Moreover, long processing times for complete porogen removal may occur in the case of using chemical-blowing agents. 1.4.2.4. Compressed CO2 and supercritical CO2-assisted foaming and sintering A fluid is at supercritical conditions when submitted to pressures and temperatures above its critical point (Fig. 1.13). Supercritical carbon dioxide (scCO2) is the most widely used supercritical fluid due to the relatively mild conditions of its critical point of CO2 (31.1 ºC and 73.8 bar), low toxicity and price. Physicochemical properties (density, viscosity and diffusivity) of CO2 at near-critical or supercritical conditions are intermediate between those of gas and liquid state and can be tuned by changing the conditions of pressure and temperature. The use of near critical or supercritical CO2 (scCO2) for the fabrication of scaffolds arises as an alternative green-processing method to operating with organic solvents or under high temperatures. The principle behind this processing strategy is that CO2 can dissolve in high amounts into certain polymers (PCL, PLGA) under nearcritical and supercritical conditions, resulting in a plasticizing effect able to reduce the melting and/or glass transition points of the polymers (Fig. 1.14). Sintering is a solvent-free method that classically employs high temperatures to melt the surface of particles to merge them by formation of necks between particles to obtain a porous matrix (Fig. 1.15a). This technique can be modified by exploiting the plasticizing LETICIA GOIMIL GARCÍA _____________________________________________________________ 20 effect of compressed CO2 to achieve the same results under mild conditions that do not cause degradation of materials (GarcíaGonzález et al. 2015). On the other hand, supercritical foaming method consists of putting a foamable polymeric material in contact Fig. 1.13: Phase diagram for CO2. Adapted from Witkowski et al. (2013) with permission of Elsevier. Fig. 1.14: Solid−liquid transition for poly(ε-caprolactone)−CO2 system. Comparison between experimental and literature data. Reproduced from Markocic et al. (2013) with permission of ACS Publications. 1. Introduction _____________________________________________________________ 21 with supercritical CO2 so that the CO2 dissolves into it —soaking stage—, obtaining a polymer in a molten or in a rubbery state, and then decreasing the pressure to induce a phase separation — depressurization stage— that will originate bubbles within the matrix (1.15b). The pore sizes and porosities of the resulting scaffolds are dependent of the processing conditions (mainly pressure, temperature and depressurization rate). Namely, the depressurization rate is related to the formation of nucleation sites. A fast pressure drop will cause a higher supersaturation and lead to higher nucleation rates leading to small pores and less homogenous structures. Conversely, reducing the venting rate allows the growth of the pores and even their coalescence, so the final structure has larger pores and higher pore interconnectivity. Fig. 1.15: Diagram of near-critical and supercritical CO2 (a) sintering and (b) foaming. In the supercritical foaming method, the avoidance of organic solvents or high temperatures to obtain a polymer with a porous structure suppresses the need of posterior purification processes and thermal degradation, leading to drug-incorporation yields of ca. 100 %. Besides, the morphological properties of the resulting scaffolds regarding porosity and pore size can be easily tuned controlling the conditions of soaking and depressurization conditions and durations. LETICIA GOIMIL GARCÍA _____________________________________________________________ 22 The most suitable materials for supercritical foaming are synthetic polymers, such as PCL and PLGA, which are highly biocompatible and have good mechanical properties for bone regeneration. However, the pores obtained usually show poor interconnection rates and smooth surfaces that hinder cell penetration and attachment (Ge et al. 2013). The use of organic solvents (acetone), biofunctional plasticizers (eugenol) or mesoporous silica were proposed as admixtures during the supercritical foaming of polyesters to improve these properties (Kiran 2010, de Matos et al. 2013, Salerno et al. 2017). However, those options pose problems of cytotoxicity, limited range of processability and reduction in mechanical properties. Alternative processing options are thus prospected to improve the pore interconnectivity and surface roughness of scaffolds obtained by supercritical foaming. Aerogels are defined as: ―Solid, lightweight and coherent open porous networks of loosely packed, bonded particles or nanoscale fibers, obtained from a gel following the removal of the pore fluid without significant structural modification‖ (García-González et al. 2019). In this PhD Thesis, the use of aerogel particles from natural polymers was evaluated as alternative ingredients to be incorporated in the scaffold formulations to be obtained by supercritical foaming. 1.5. REFERENCES A. Abellán García, A. Ayala García, J. Pérez Díaz and R. Pujol Rodríguez (2018). Un perfil de las personas mayores en España, 2018. Indicadores estadísticos básicos 17. Madrid, Spain. CSIC - Instituto de Economía, Geografía y Demografía (IEGD). R. Abou-Khalil, F. Yang, M. Mortreux, S. Lieu, Y.-Y. Yu, M. Wurmser, C. Pereira, F. Relaix, T. Miclau, R. S. Marcucio and C. Colnot (2014). "Delayed bone regeneration is linked to chronic inflammation in murine muscular dystrophy." Journal of Bone and Mineral Research 29(2): 304315. 1. Introduction _____________________________________________________________ 23 A. H. Alghadir, S. A. Gabr and E. Al-Eisa (2015). "Physical activity and lifestyle effects on bone mineral density among young adults: sociodemographic and biochemical analysis." Journal of Physical Therapy Science 27(7): 2261-2270. R. M. Allaf (2018). 4 - Melt-molding technologies for 3D scaffold engineering. Functional 3D Tissue Engineering Scaffolds. Y. Deng and J. Kuiper. Duxford, UK, Woodhead Publishing: 75-100. N. Annabi, J. W. Nichol, X. Zhong, C. Ji, S. Koshy, A. Khademhosseini and F. Dehghani (2010). "Controlling the porosity and microarchitecture of hydrogels for tissue engineering." Tissue engineering. Part B, Reviews 16(4): 371-383. A. Autissier, C. L. Visage, C. Pouzet, F. Chaubet and D. Letourneur (2010). "Fabrication of porous polysaccharide-based scaffolds using a combined freeze-drying/cross-linking process." Acta Biomaterialia 6(9): 3640-3648. A. L. Boskey and R. Coleman (2010). "Aging and bone." Journal of Dental Research 89(12): 1333-1348. V. Campana, G. Milano, E. Pagano, M. Barba, C. Cicione, G. Salonna, W. Lattanzi and G. Logroscino (2014). "Bone substitutes in orthopaedic surgery: from basic science to clinical practice." Journal of Materials Science. Materials in Medicine 25(10): 2445-2461. P. M. Crapo, T. W. Gilbert and S. F. Badylak (2011). "An overview of tissue and whole organ decellularization processes." Biomaterials 32(12): 32333243. F. Chai, G. Raoul, A. Wiss, J. Ferri and H. F. Hildebrand (2011). "Les biomatériaux de substitution osseuse : classification et intérêt." Revue de Stomatologie et de Chirurgie Maxillo-faciale 112(4): 212-221. M. B. C. de Matos, A. P. Piedade, C. Alvarez-Lorenzo, A. Concheiro, M. E. M. Braga and H. C. de Sousa (2013). "Dexamethasone-loaded poly(ɛcaprolactone)/silica nanoparticles composites prepared by supercritical CO2 foaming/mixing and deposition." International Journal of Pharmaceutics 456(2): 269-281. F. Dehghani and N. Annabi (2011). "Engineering porous scaffolds using gas-based techniques." Current Opinion in Biotechnology 22(5): 661-666. LETICIA GOIMIL GARCÍA _____________________________________________________________ 24 A.-V. Do, B. Khorsand, S. M. Geary and A. K. Salem (2015). "3D Printing of Scaffolds for Tissue Regeneration Applications." Advanced Healthcare Materials 4(12): 1742-1762. M o lar , M arc a and M me ( 00 ) Modelling one tissue fracture and healing: a review." Engineering Fracture Mechanics 71(13): 1809-1840. T. A. Einhorn and L. C. Gerstenfeld (2015). "Fracture healing: mechanisms and interventions." Nature Reviews. Rheumatology 11(1): 45-54. I. El Bialy, W. Jiskoot and M. Reza Nejadnik (2017). "Formulation, Delivery and Stability of Bone Morphogenetic Proteins for Effective Bone Regeneration." Pharmaceutical Research 34(6): 1152-1170. G. Fernandez de Grado, L. Keller, Y. Idoux-Gillet, Q. Wagner, A.-M. Musset, N. Benkirane-Jessel, F. Bornert and D. Offner (2018). "Bone substitutes: a review of their characteristics, clinical use, and perspectives for large bone defects management." Journal of Tissue Engineering 9: 2041731418776819-2041731418776819. M. A. Flierl, W. R. Smith, C. Mauffrey, K. Irgit, A. E. Williams, E. Ross, G. Peacher, D. J. Hak and P. F. Stahel (2013). "Outcomes and complication rates of different bone grafting modalities in long bone fracture nonunions: a retrospective cohort study in 182 patients." Journal of Orthopaedic Surgery and Research 8(1): 33. R. Florencio-Silva, G. R. d. S. Sasso, E. Sasso-Cerri, M. J. Simões and P. S. Cerri (2015). "Biology of Bone Tissue: Structure, Function, and Factors That Influence Bone Cells." BioMed Research International 2015: 17. C. A. García-González, T. Budtova, L. Durães, C. Erkey, P. Del Gaudio, P. Gurikov, M. Koebel, F. Liebner, M. Neagu and I. Smirnova (2019). "An Opinion Paper on Aerogels for Biomedical and Environmental Applications." Molecules 24(9): 1815. C. A. García-González, A. Concheiro and C. Alvarez-Lorenzo (2015). "Processing of Materials for Regenerative Medicine Using Supercritical Fluid Technology." Bioconjugate Chemistry 26(7): 1159-1171. J. Ge, M. Li, Q. Zhang, C. Z Yang, P. Wooley, X. Chen and S.-Y. Yang (2013). "Silica Aerogel Improves the Biocompatibility in a Polyε - Caprolactone Composite Used as a Tissue Engineering Scaffold." International Journal of Polymer Science 2013: 7. 1. Introduction _____________________________________________________________ 25 P. Gentile, V. Chiono, I. Carmagnola and P. V. Hatton (2014). "An overview of poly(lactic-co-glycolic) acid (PLGA)-based biomaterials for bone tissue engineering." International Journal of Molecular Sciences 15(3): 36403659. T. Ghassemi, A. Shahroodi, M. H. Ebrahimzadeh, A. Mousavian, J. Movaffagh and A. Moradi (2018). "Current Concepts in Scaffolding for Bone Tissue Engineering." The Archives of Bone and Joint Surgery 6(2): 9099. T. W. Gilbert, T. L. Sellaro and S. F. Badylak (2006). "Decellularization of tissues and organs." Biomaterials 27(19): 3675-3683. J. Henkel, M. A. Woodruff, D. R. Epari, R. Steck, V. Glatt, I. C. Dickinson, P. F. M. Choong, M. A. Schuetz and D. W. Hutmacher (2013). "Bone Regeneration Based on Tissue Engineering Conceptions — A 21st Century Perspective." Bone Research 1: 216. A. W. James, G. LaChaud, J. Shen, G. Asatrian, V. Nguyen, X. Zhang, K. Ting and C. Soo (2016). "A Review of the Clinical Side Effects of Bone Morphogenetic Protein-2." Tissue engineering. Part B, Reviews 22(4): 284297. E. Kiran (2010). "Foaming strategies for bioabsorbable polymers in supercritical fluid mixtures. Part II. Foaming of poly(ɛ-caprolactone-colactide) in carbon dioxide and carbon dioxide+acetone fluid mixtures and formation of tubular foams via solution extrusion." The Journal of Supercritical Fluids 54(3): 308-319. B. S. Klosterhoff, S. Nagaraja, J. J. Dedania, R. E. Guldberg and N. J. Willett (2017). Chapter 5 - Material and Mechanobiological Considerations for Bone Regeneration. Materials and Devices for Bone Disorders. S. Bose and A. Bandyopadhyay, Academic Press: 197-264. C. Koski, B. Onuike, A. Bandyopadhyay and S. Bose (2018). "Starchhydroxyapatite composite bone scaffold fabrication utilizing a slurry extrusion-based solid freeform fabricator." Additive Manufacturing 24: 4759. J. W. Lee, G. Ahn, J. Y. Kim and D.-W. Cho (2010). "Evaluating cell proliferation based on internal pore size and 3D scaffold architecture fabricated using solid freeform fabrication technology." Journal of Materials Science: Materials in Medicine 21(12): 3195-3205. LETICIA GOIMIL GARCÍA _____________________________________________________________ 32 3. PREPARATION AND STABILITY OF DEXAMETHASONE-LOADED SCAFFOLDS FOR BONE REGENERATION PROCESSED BY COMPRESSED CO2 FOAMING§ §The work described in this chapter has been published in Goimil et al. J. CO2 Utiliz., 24 (2018) 89–98 and carried out in collaboration with Prof. Philipp Jaeger, from the Hamburg University of Technology, and Dr. Inés Ardao, from BioFarma Research Group at USC. LETICIA GOIMIL GARCÍA ________________________________________________________ 34 3.1. INTRODUCTION Bone tissue usually heals completely after damage without scarring (Jahagirdar and Scammell 2009). Nevertheless, the overall rate of nonunion or delayed union represents 5-10 % of the total fractures and depends on the type of bone injury and the patient condition. In Europe, around one million patients per year undergo a surgical bone reconstruction procedure, with expectations of increasing numbers owing to population ageing, a parameter that compromises bone regeneration (Gómez-Barrena et al. 2015). Regenerative medicine is a rising discipline aiming to repair, replace or regenerate damaged tissues and organs. A suitable structure should be accordingly provided to cells for their attachment and proliferation, allowing the formation of a functional tissue similar to the one to be reconstructed (Mao and Mooney 2015). In case of bone tissue regeneration, autografts (from the patient), allografts (from donors or corpses) and xenografts (from animals) are natural grafts that can act as mechanical support for tissue growth as well as a source of growth factors that promotes tissue regeneration (Shibuya and Jupiter 2015). Synthetic porous grafts (scaffolds) are being developed to overcome the problems of availability and post-implantation risks of natural grafts. Compressed CO2 foaming is regarded as a key green technology for the processing of scaffolds (García-González et al. 2015). This technology is based on the role of CO2 as a porogen by putting in contact this compressed fluid with the matrix of the scaffold for CO2 sorption and the subsequent expansion of the scaffold (i.e. foaming) upon CO2 depressurization through a pore nucleation and growth mechanism (Fanovich and Jaeger 2012, Salerno et al. 2017). This technology is unique for preparing drug-loaded scaffolds in a solvent-free approach whilst avoiding leaching downstream steps (García-González et al. 2015). Problems of cytotoxicity of the scaffolds, thermal degradation of their components and low incorporation yields are thus precluded by using this technique. Moreover, compressed CO2 can also act as sterilizing agent for the scaffolds (Spilimbergo and Bertucco 2003). Nevertheless, compressed CO2 foaming was not effective so far in the processing of certain 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 35 biopolymers, e.g., PLGA, of low inherent viscosities (<0.5 dL·g-1) having attractive degradation profiles for scaffolding purposes (Sheridan et al. 2000, García González et al. 2015). Dramatic foam expansions have been obtained so far with this kind of matrices and a proper process design and optimization is urged to circumvent this limitation of the compressed CO2 foaming technology. The in vivo performance of the processed scaffolds depends primarily on their structure and composition (Ho and Hutmacher 2006). Porosity must be at least 65 %, ideally around 90 % (Zhang et al. 2015), while pore interconnectivity should be high to enable diffusion of oxygen and nutrients to the cells and removal of metabolites (Ho and Hutmacher 2006). The presence of micro and mesopores in the scaffolds allows the circulation of biological fluids and cell adhesion, while small (5-10 m) and large macropores (50200m) are required to facilitate vessel formation and tissue growth, respectively (García-González et al. 2015). At the same time, the mechanical properties of the scaffold should ideally be suitable to act as a temporary support while the growing tissue is still not able to withstand by itself the intrinsic mechanical demands of bone. The composition of the scaffold must be precisely selected with a matrix able to provide a degradation rate that suits the growth rate of the bone tissue, and with growth and differentiation agents to promote the proper bone tissue growth. Various biodegradable synthetic polymers, e.g., poly(lactic-co-glycolic) acid –PLGA– and poly(- caprolactone) –PCL– (Zhang et al. 2015, Diaz-Gomez et al. 2016, Diaz-Gomez et al. 2017), and natural polymers, e.g., starch, collagen, gelatin, silk, alginate or chitosan (Duarte et al. 2009, Chimenti et al. 2011, Cuadros et al. 2015, Zhang et al. 2015, Melke et al. 2016), are common materials employed as scaffold matrices. Among the growth and differentiation agents, dexamethasone (DX), a drug belonging to the group of glucocorticoids, can induce the differentiation of mesenchymal stem cells (MSC) to osteoblasts (Yoon et al. 2003). DXloaded scaffolds providing local delivery of the bioactive agent may have the advantage of inducing bone formation whilst avoiding or at least mitigating the undesired collateral effects derived from a systemic administration of glucocorticoids. LETICIA GOIMIL GARCÍA ________________________________________________________ 36 Stability under storage is a critical quality attribute to define the performance and the availability of scaffolds. A short shelf-life would hamper the supply of these grafts and raise the cost of the treatment (Sun and Gouk 2008). Nitrogen atmospheres or vacuum conditions are the most common packaging options of the commercial scaffolds. In these cases, the direct exposition to climatic conditions is initially avoided, although it might occur if the packaging is accidentally damaged (e.g., with a microhole) resulting in loss of the inert atmosphere or breaking of vacuum. Systems like knitted polylactide scaffolds and knitted silk fibroin (SERI©) surgical scaffolds have been tested for up to 3 years under room conditions, maintaining their mechanical properties for that period (Ellä et al. 2011, Jewell et al. 2015). Nevertheless, there is still a paucity of information on the effect of storage on the durability of scaffolds with more complex porous structures or loaded with drugs, which might be more sensitive to environmental factors. In this work, synthetic scaffolds made of mixtures of PCL and PLGA of low inherent viscosity (50:50 w/w) loaded with and without DX were prepared using a novel compressed CO2-assisted foaming method. The processing method was herein designed and optimized for DX-loaded scaffolds from a previously implemented one (García González et al. 2015) so that it could operate in one-pot, during short processing times, without use of organic solvents and under mild temperatures. The role of the incorporation of starch (St) in controlling the DX release profile from the scaffolds was evaluated. Then, the scaffolds (without packaging) were stored in chambers (25 ºC, 65% relative humidity) mimicking the zone II ICH-climatic conditions (Europe, USA and Japan) (de Matos et al. 2015) and the stability of the scaffolds in terms of structural, physicochemical and mechanical properties as well as DX release profiles were monitored for 3 months. 3.2. MATERIALS AND METHODS 3.2.1. Materials PLGA (50:50 lactic:glycolic ratio; Mw 16 kDa; amorphous; Tg=41.4 ºC) was purchased from Purac (Gorinchem, The 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 37 Netherlands). PCL (Mw 50 kDa; semicrystalline; Tm=61.6 ºC; ΔHm=95.9 J/g) was supplied by Polysciences (Warrington, PA, USA). Corn starch (Amylo N-460, 52.6 % amylose content) was bought from Roquette (Lestrem, France). Dexamethasone (DX, 97 % purity) was obtained from Sigma-Aldrich (Saint Louis, MO, USA). Carbon dioxide (99.8 % purity) from Praxair (Madrid, Spain) was used to process the scaffolds. Milli-Q water (resistivity > 18 MΩ·cm; MilliQ, Millipore®, Madrid, Spain), sulfuric acid (95-97 % purity, Merck, Darmstadt, Germany) and acetonitrile (99.9 % purity, Merck, Darmstadt, Germany) were also used. 3.2.2. Preparation of oven-dried starch gels (Stp) 10 g of corn starch were dispersed in 210 g of water under magnetic stirring and then heated for 20 min at 121 ºC (Autoclave Raypa, model AES-12, Terrassa, Spain). Then, the starch dispersion was transferred to a fridge at 4 ºC and maintained for 60 h for retrogradation, and subsequently poured on Petri dishes and put in an oven at 80 ºC for water evaporation. The resulting dry solid films were ground in a ball mill (Mix MM 400, Retsch Inc., Newton, PA, USA) to obtain powdered starch particles of ca. 20 m size. 3.2.3. CO2 sorption in the scaffold polymers A gravimetric method was used to monitor carbon dioxide sorption in PLGA and in PCL:PLGA (50:50 w/w) mixtures at elevated pressure. A high-pressure magnetic suspension balance (Rubotherm GmbH, Bochum, Germany) was used to measure the CO2 uptake by PLGA as a function of time and simultaneously detecting the volume change by a high pressure view cell (Eurotechnica GmbH, Hamburg, Germany), at the operating conditions of 66 bar and 26 ºC (compressed liquid) and 51 bar and 26 ºC (compressed gas) for PLGA and of 60 bar and 26 ºC (compressed gas) for the PCL:PLGA mixture. The polymer was previously filled in the powdered form into a glass vessel, molten at 70 °C, cooled down for solidification and then placed in the magnetic suspension balance (Fig. 3.1). The force resulting from the mass and the buoyancy of the polymer is transmitted to a microbalance at the outside of the autoclave via a LETICIA GOIMIL GARCÍA ________________________________________________________ 38 Fig. 3.1: Sketch of the high–pressure magnetic suspension balance used for CO2 sorption tests in polymers. Inset: Glass vessel containing the polymeric mixture (bottom) inside the balance before the start of the test. magnetic coupling. From the evolution of the weight as a function of time, the diffusion process could be followed and the CO2 solubility determined (<5 % deviation) once saturation was achieved (Crank 1975). 3.2.4. Compressed CO2 foaming of synthetic scaffolds Polyester-based scaffolds were prepared using a multi-step process sketched in Fig. 3.2. The components of the scaffolds (Table 3.1) were homogenized in a mortar and mixed for 5 min (Mixer Wab, model T2C, Switzerland) and then compressed in an eccentric tableting machine (FE236FC, Korsch, Berlin, Germany) to obtain compacts of 400 mg with prismatic dimensions (14x10x2.6 mm). Fig. 3.2: Schematic workflow for the preparation of polyester-based scaffolds using compressed CO2 technology. 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 39 Table 3.1: Composition (in weight percentage) of compressed CO2-processed scaffolds. Scaffolds PLGA:PCL 50:50 DX Stp St0-DX0 100 - - St0-DX 94.5 5.5 - St-DX 85 5 10 The compacted materials were processed in a high pressure equipment (Thar Technologies, Pittsburgh, PA, USA) adapted for compressed CO2 foaming. Briefly, the compacts were placed in a rotating basket (700 rpm) inside the foaming vessel and put in contact with CO2 at 26 ºC and 60 bar for 30 min. After this soaking period, pressure was lowered from 60 to 30 bar at a flow rate of 10 bar/min, and then pressure was increased again until 60 bar by adding liquid CO2 at 1 ºC. This process was repeated three times, before proceeding to complete the depressurization period at a flow rate of 10 bar/min. Finally, the external layer of the scaffolds was removed with a scalpel one day after processing and before any further use. The foaming process was also monitored with a view cell (Eurotechnica, Hamburg, Germany) and recorded with a CCDcamera. Visual behavior of the foaming process of St0-DX0 scaffold was compared to a similar experiment with direct depressurization from 60 bar to atmospheric pressure at 10 bar/min without addition of liquid CO2, to unveil the relevance of the foaming process strategy used in this work. 3.2.5. Scaffolds characterization 3.2.5.1. Structural, physicochemical and mechanical characterization of scaffolds Images from the cross-sections of the scaffolds were obtained by scanning electron microscopy (SEM, EVO LS15, Zeiss, Oberkochen, Germany). Overall porosity (  ) was calculated using the following equation: ( ⁄ ) (3.1) LETICIA GOIMIL GARCÍA ________________________________________________________ 40 where ρbulk is the bulk density estimated from the dimensions and the weight of the scaffolds (after the foaming process), and ρskel is the skeletal density of the scaffolds measured by helium pycnometry (Quantachrome; Boynton Beach, FL, USA) at 25 ºC and 1.03 bar. Textural properties of the scaffolds (total pore volume –Vp–, pore size distribution and open porosity –  MIP–) before and after storage were measured through mercury intrusion porosimetry (MIP) (Autopore IV 9500 model, Micromeritics, Norcross, GA, USA). MIP was operated with a 3 mL-penetrometer for solids and at working pressures ranging from 0.07 to 1724 bar. Thermal properties were analyzed through differential scanning calorimetry (DSC) (TA Instruments Q100; New Castle, DE, USA) using two heating cycles up to 125 ºC with a cooling cycle down to -10 ºC inbetween, at a rate of 10 ºC/min and under a nitrogen atmosphere. Mechanical properties were analyzed at 37 ºC using a Rheolyst AR1000N rheometer (TA Instruments, New Castle, DE, USA) fitted with an environmental test chamber, a solid torsion kit and a data analyser (AR2500). The scaffolds were located between two clamps (8 mm gap) and then subjected to an angular frequency sweep in the 0.05-20 rad/s range 3.2.5.2. Stability under storage Scaffolds of defined dimensions and weights of the three compositions (Table 3.1) were placed on top of a platform inside sterile hermetic glass vessels, which contained sulfuric acid solution (37 % v/v) in the bottom to maintain the required relative humidity (65 %) at 25 ºC (Wilson 1921). The glass vessels were stored for 1 and 3 months. Once the storage period was over, the scaffolds were collected for further characterization (Section 3.2.5.1). 3.2.5.3. DX release tests Prismatic pieces (10 mg) of scaffolds were individually transferred to flasks containing 50 mL of PBS pH 7.4. The flasks were kept in an oscillating shaker (JP Selecta, Unitronic 320 OR model, Abrera, Spain) at 37 ºC and 60 rpm for 3 weeks (Duarte et al. 2009). Aliquots of 1 mL were sampled through the release period at selected times and the volume was replaced with fresh PBS medium. DX 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 41 release tests were performed in triplicate. The dissolution profile of 1 mg of DX powder in 100 mL of PBS was also recorded in triplicate for the sake of comparison. DX concentration was quantified by HPLC (Waters; Milford, MA, USA) consisting on a 717 plus Autosampler, a 600 Controller and a 996 Photodiode Array Detector. A C18 column (Symmetry© C18, particle size 5 m, 3.9 mm diameter x 150 mm length) from Waters (Milford, MA, USA) was used at 30 ºC. The mobile phase consisted on a mixture of water:acetonitrile 67:33 (v/v), operating in the isocratic mode, and the flow rate was set at 1 mL/min (MoyaOrtega et al. 2010). The volume injected in the column was 20 L. DX was quantified using UV/Vis diode-array detector at 242 nm with a retention time of 5.5 min. Samples were filtered prior to analysis using 0.2 m nylon filters to avoid potential interference due to the presence of particles from the scaffolds. The calibration curve of DX in PBS was done in triplicate in the concentration range 0.025-50 g/mL (Fig. 3.3). Fig. 3.3: Calibration curve and equation for dissolved DX in PBS pH 7.4 by HPLC at 242 nm. 3.2.5.4. Erosion assay Fifteen prismatic scaffold pieces (15 mg) of each formulation were suspended in separated Eppendorf tubes containing 1 mL of PBS pH 7.4 and kept in an oscillating bath at 37 ºC and 60 rpm. The LETICIA GOIMIL GARCÍA ________________________________________________________ 48 fill the latter ones, and the computed size will be the same for all the pores (Giesche 2006). Accordingly, the large pores from St0-DX0 scaffolds might have narrow throat sizes and explain the high contribution of the small pores unveiled by MIP. The presence of DX and Stp in the scaffold formulations (St0-DX and St-DX) changed the pore size distribution. The population of mid- (10-100 µm) and largesized (>100 µm) pores increased with DX and Stp likely due to the secondary pore nucleation site provided by these admixtures and the observed formation of macrochannels, respectively. Thermal properties of the polymers were altered when blended and processed under compressed CO2 conditions (Table 3.3). Under the experimental conditions used for foaming (26 ºC, 60 bar), both PCL and PLGA were above their glass transition values (Tg) according to the view cell experiments (Section 3.1) and, therefore, in the glassy state. This information is crucial since it ensures a certain volume expansion after CO2 depressurization. After processing, St0DX0 scaffolds showed two melting thermal events (Tm,1 and Tm,2) one Table 3.2: Porosity and pore size distribution of the scaffolds before and after storage at 25 ºC and 65% relative humidity. Pores were separated according to their pore size into small (<10 µm), mid-sized (10-100 µm) and large pores (>100 µm). Volume contribution (%) Scaffolds  (%) MIP (%) Vp (cm3/g) <10 m 10-100 m >100 m St0-DX0 Non stored 69.1±2.7 65.5 1.78 71.20 24.50 4.30 1 month 67.6±1.6 54.0 1.55 49.31 43.60 7.09 3 months 68.2±3.7 67.2 1.64 50.00 42.34 7.66 St0-DX Non stored 71.8±14.2 63.9 2.04 50.08 42.85 7.07 1 month 71.1±9.8 63.6 1.78 52.46 37.82 9.72 3 months 65.3±16.0 61.8 1.49 49.97 43.53 6.50 St-DX Non stored 74.7±10.4 65.3 1.99 61.82 31.91 6.27 1 month 68.9±11.1 61.9 1.72 58.10 35.98 5.92 3 months 71.9±10.4 46.3 0.80 50.00 40.48 9.52 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 49 shifted towards lower values than the melting point of pure PCL (61.57 ºC) and the other towards higher values and with decreased crystallinity of PCL (ΔHm=80.44 J/gPCL) because of PCL-PLGA chemical interactions (Diaz-Gomez et al. 2016). Melting of PCL was assumed to take place only partially under the experimental conditions of foaming and PCL fractionation should have occurred. The molten fraction of PCL, corresponding to the lighter fraction of PCL, was responsible for the first thermal event observed (Tm,1). The Tg of PLGA was overlapped with the first melting point of PCL and could not be individually evaluated. The reduction in the melting point value Tm,1 with respect to the normal melting point of PCL was ascribed to PCL-PLGA interactions and to the intrinsic lower melting point values of the light fraction from PCL. The increase in the Tm,2 value was attributed to the average melting point of the heavier fraction of PCL, not molten upon the compressed CO2 foaming process. The identification of the thermal events and the hypothesis of the partial melting of PCL during the foaming process were confirmed with the second heating cycle of the DSC analysis. PCL completely melted during the first heating cycle (end cycle at 125 ºC) and favored the interaction between PCL and PLGA. A single melting event (T’m,1) was then observed during the second heating cycle shifted towards lower values than the normal melting point of PCL, thus indicating a stronger PCL-PLGA interaction and without PCL fractionation. The effect of the incorporation of admixtures (DX and St) in the scaffolds on the thermal transitions of the synthetic scaffolds was evaluated. DX had a minimum effect on the thermal properties of the scaffolds (St0-DX) with a limited decrease in the polymer crystallinity (ΔHm=72.19 J/gPCL). Differently, a dramatic depletion in the melting point values (Tm,1 and Tm,2) was observed with the incorporation of starch to the formulation (St-DX). The pregelification of the starch used in this work (Stp) favored the PCL-St interaction by hydrogen bonding between the hydroxyl groups of starch and the ester carbonyl groups of PCL and was responsible for these shifts in the melting point values (Averous et al. 2000, Ali Akbari Ghavimi et al. 2015, Diaz-Gomez et al. 2016). LETICIA GOIMIL GARCÍA ________________________________________________________ 50 Table 3.3: Thermal properties of the processed scaffolds before and after storage tests at 25 ºC and 65 % relative humidity. Scaffolds 1st heating cycle 2nd heating cycle Tm,1 (ºC) Tm,2 (ºC) H(J/g) T’m,1 (ºC) H(J/g) St0-DX0 Non stored 48.3 62.1 40.2 56.1 36.4 1 month 42.5 61.0 64.3 54.7 53.6 3 months 38.6 62.8 50.0 54.6 32.9 St0-DX Non stored 48.2 61.9 34.1 56.3 37.1 1 month 42.1 60.2 40.7 54.4 34.5 3 months 38.1 62.2 47.8 54.5 31.2 St-DX Non stored 43.2 57.4 36.4 53.2 32.0 1 month 42.2 61.2 54.0 54.6 46.4 3 months 34.5 61.7 39.1 54.2 25.0 Biodegradation is a critical design parameter of the scaffolds to suit the bone tissue growth rate (Woodruff et al. 2012). The mixture of PCL with a low molecular weight PLGA was accordingly chosen in this work to tune the erosion rate, since it combines the slow degradation of PCL in the order of months (Diaz-Gomez et al. 2016) with the much faster degradation of PLGA in the order of weeks (Lu et al. 2000). The effect of scaffolds composition on the erosion rate was assessed in PBS pH 7.4 medium (Fig. 3.6). The erosion rate was similar for the first 14 days for all the scaffolds formulations with a weight loss of ca. 25 wt.% due to the degradation of the polyesters, mainly PLGA. This degradation might be initially promoted owing to the acidic environment provided by the soluble degradation products with limited diffusion through the porous matrix of the scaffolds (Berkland et al. 2007). Upon degradation, the pore sizes of the scaffolds increased and the diffusion of the degradation products to the PBS medium was enhanced, resulting in a less acidic pore microenvironment and a subsequent decrease in the degradation rate of the polymers. The transition towards slower degradation rates took longer for the scaffolds St0-DX and St-DX (ca. 21 days). The 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 51 incorporation of DX in the formulation accelerated the degradation rate since it contributes to the formation of less crystalline polymeric matrices, to the scaffold weight loss due to the release of DX in the degradation medium (see Section 3.3.4) and to a more open structure as DX progressively dissolves. The presence of starch in the scaffolds (St-DX) gave rise to a morphology of higher open porosity and larger pores favoring the intimate contact of the degradation medium with the scaffold and a less crystalline polymeric matrix (Cai et al. 2014). Additionally, starch would degrade in PBS solution due to its partial dissolution (i.e. leaching) through the breakage of hydrogen bonds of starch molecules. Moreover, the starch used in this work is particularly susceptible to dissolution due to its high amylose content (52.6 % amylose) (Han and Lim 2004). Subsequently, the erosion rate of the scaffolds increased with the presence of starch, as already reported for PCL-starch blends (Ali Akbari Ghavimi et al. 2015). Fig. 3.6: Erosion tests of the processed scaffolds in PBS pH 7.4 medium (37 ºC, 60 rpm): St0-DX0 (in diamonds), St0-DX (squares) and St-DX (triangles). Bone fractures usually heal within 6 and 8 weeks, but the period to repair bone defects will vary depending on the defect size, and can take from 6 to 12 months to form a mature tissue (Woodruff et al. 2012). Nevertheless, these scaffolds incorporate an osteogenic factor -90 -80 -70 -60 -50 -40 -30 -20 -10 0 0 7 14 21 28 35 42 49 56 63 Weight loss (%) Time (days) LETICIA GOIMIL GARCÍA ________________________________________________________ 52 (DX), which is expected to accelerate the healing process and thus fit the needed degradation rate. In addition, some polymeric scaffolds have shown a faster degradation rate in vitro than in vivo due to the formation of connective tissue around the structure (Ellä et al. 2011). Regarding the mechanical properties, viscoelastic behavior was observed for all the scaffolds formulations (Fig. 3.7). The storage (G’) and loss (G’’) moduli were close to the values of 0.1 and 0.01 GPa, respectively, in the tested range of angular frequencies. These mechanical parameters fall slightly below the values reported in the literature for trabecular and cortical bones of human and animal origin: G’=0.3-10 GPa; G’’=0.01-0.60 GPa (Linde et al. 1988, Yamashita et al. 2001, Buechner and Lakes 2003). 3.3.3. Effect of storage The volume of all scaffolds decreased after the storage period at 25 ºC and 65 % relative humidity (Table 3.4). Loss in weight was negligible. Therefore, densification of the scaffolds and reduction in the overall porosity occurred during the storage period. The densification was more pronounced for the scaffolds containing DX and Stp and had a direct impact on the micromeritic properties as evaluated by MIP analysis (Table 3.2). For St0-DX0 scaffolds, a reduction in the population of small pores (<10 µm) took place upon storage, which was mainly related to the broadening of the pore throat sizes. Moreover, a reduction in the open porosity (MIP) was observed due to a certain pore collapse upon storage. The pore size distribution was almost unaltered in the case of St0-DX scaffolds, although a certain shift towards the lowest limit of each pore range took place with longer storage time periods, since a reduction in the overall pore volume (Vp in Table 3.2) and a reduction in the scaffold volume (Table 3.4) were observed. The most significant change in porous morphology upon storage was observed for the scaffolds containing starch (St-DX). After one month of storage under controlled temperature and humidity conditions, densification occurred with a certain reduction in the overall and open porosities as well as in the total pore volume (Tables 3.2 and 3.4). A dramatic change in the porous structure of St-DX took 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 53 Angular frequency (rad/s) 0.01 0.1 1 10 100 G' and G'' (Pa) 106 107 108 Angular frequency (rad/s) 0.01 0.1 1 10 100 G' and G'' (Pa) 106 107 108 Angular frequency (rad/s) 0.01 0.1 1 10 100 G' and G'' (Pa) 106 107 108 (a) (b) (c) Fig. 3.7: Storage (G’; black symbols) and loss (G’’; white symbols) moduli of the scaffolds (a) St0DX0, (b) St0-DX and (c) St-DX before and after storage at 25 ºC and 65 % relative humidity. Circles, triangles and squares stand for non-stored scaffolds, scaffolds stored for 1 month and scaffolds stored for 3 months, respectively. LETICIA GOIMIL GARCÍA ________________________________________________________ 54 Table 3.4: Changes (in percentage) of weight, volume, density and porosity of PCLPLGA-based scaffolds after being stored for 1 and 3 months at 25 ºC and 65 % relative humidity. Scaffolds St0-DX0 St0-DX St-DX Storage period 1 month 3 months 1 month 3months 1 month 3 months Change in weight (%) -0.480.18 -0.620.43 -0.380.47 -0.780.35 -0.500.27 -0.520.93 Change in volume (%) -6.197.53 -2.354.47 -12.5910.31 -15.819.45 -18.6912.19 -12.605.14 Change in bulk density (%) 6.688.94 1.894.85 15.1512.49 19.1112.68 24.8119.55 14.196.62 Change in overall porosity (%) -2.693.73 -0.901.92 -3.763.53 -5.671.90 -6.113.89 -4.873.23 place in a more prolonged storage period (three months) with a marked decrease in the open porosity and in the overall pore volume (Vp). The swelling capacity of the amylopectin fraction of the starch in the presence of moisture might explain the observed phenomena of reduction of porosity (Alcázar-Alay and Meireles 2015). Regarding the thermal transitions of the scaffolds, mid-term storage periods resulted in a noticeable effect on the first melting point (Tm,1) region of the scaffolds (Table 3.3). This depletion on Tm,1 values was more significant with longer time periods and was attributed to the overlapping of the first melting of PCL with the glass transition of PLGA coupled to the plasticizing effect of water on PLGA (Blasi et al. 2005). Water vapor would progressively adsorb on PLGA resulting in a decrease in the Tg inferring a reduction in Tm,1 (Bouissou et al. 2006). This depletion effect due to adsorbed water was also observed for T’m,1 during the second DSC-heating cycle for St0-DX0 and St0-DX scaffolds. A reverse trend was observed for StDX scaffolds where T’m,1 increased with time likely due to a higher starch-PCL interaction with the swollen starch. The erosion rate of the scaffolds in PBS pH 7.4 at 37 ºC varied depending on the previous storage time under 25 ºC and 65 % relative 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 55 humidity (Table 3.5). For St0-DX0 scaffolds, the increasing plasticizing effect of the adsorbed moisture on PLGA with time (Blasi et al. 2005) resulted in a severe weight loss, notably after 3 months of storage. The presence of DX in the scaffolds (St0-DX) and its release seemed to have a minor effect on the erosion rates when compared to the unloaded scaffolds (St0-DX0), although higher variability in the results was observed. The slower degradation of PCL due to its increased crystallinity over the storage time (Table 3.3) mitigated the higher weight loss that could be expected due to DX release to the PBS medium (Table 3.5). A reverse trend was observed with the incorporation of starch in the formulation (St-DX), with the most pronounced weight loss for non-stored scaffolds, and the lowest value obtained after 3 months of storage. This effect of starch in the scaffold erosion was attributed to the increased starch-PCL interaction of the St-DX scaffold after 3 months of storage if compared to the other two scaffold formulations (Table 3.3). Table 3.5: Weight loss (in percentage) after 3 weeks in PBS pH 7.4 medium at 37 ºC and 60 rpm of scaffolds previously stored at 25 ºC and 65% relative humidity during different time periods (0, 1 and 3 months). Scaffolds Non-stored 1 month-storage 3 month-storage St0-DX0 -13.381.79 -15.280.38 -34.401.61 St0-DX -13.765.51 -18.532.83 -34.708.15 St-DX -22.674.17 -18.6110.24 -23.525.58 The viscoelastic properties of the St0-DX0 scaffolds remained unaltered after long storage periods (Fig. 3.7a). Nevertheless, a different mechanical profile after 3 month-storage was observed, since at low angular frequencies G’’ predominates over G’ in a fluid-like (viscous) behaviour that indicated loss of mechanical stability. Scaffolds containing DX and Stp did not only follow this behaviour but also underwent other noticeable changes. The increase in the DXpolymeric matrix interactions upon storage of St0-DX scaffolds, as previously observed by DSC analysis, might explain the decrease in both the loss and storage moduli (Fig. 3.7b). The use of starch in the scaffold formulations resulted in the largest reduction in the values of LETICIA GOIMIL GARCÍA ________________________________________________________ 56 the viscoelastic moduli (Fig. 3.7c). The low interfacial adhesion between PCL and starch had a strong effect on the viscoelastic properties of the scaffolds (Mehr et al. 2015). The loss and storage moduli decreased upon storage likely due to the plasticizing effect of adsorbed water on starch and the increased PCL-St interaction as unveiled by DSC analysis. 3.3.4. Dexamethasone release The observed release profiles of DX from the scaffolds were composed of three steps (Fig. 3.8): (i) burst of DX during the first 2-3 hours; (ii) slow release of DMXT in the following days up to 2 weeks; (iii) faster release of the remaining DX payload during the following days. The dissolution profile of pure DX in PBS pH 7.4 at 37 ºC showed a DX concentration loss as the dissolution time progressed (Fig. 3.9). This decrease was attributed to the degradation of DX in the release medium and followed a zero-order kinetics with a kinetic constant of kd=2.2 10-4 mg/(mL·day). This degradation kinetic profile was similar to that reported in the literature for DX in PBS pH (7-7.4) containing sodium azide at 37 ºC (Hickey et al. 2002). The slightly faster degradation kinetics compared to the literature (kd=10-5 mg/(mL·day)) might be due to a different crystalline form or size of DX or the partial degradation of the bioactive compound by microbiological action (Hickey et al. 2002, Zhu et al. 2015). The observed release profiles were recalculated taking into account the degradation rate of DX to get the actual release profiles, which led to DX release percentages close to 90-100 wt.% after 3 weeks in PBS. After a certain lag time (td in Table 3.6) of less than 10 min needed for the wetting of the scaffold with PBS, a burst release was observed in the first 4 h. Thereafter, a sustained and constant DX release was observed for all of the scaffolds for at least three weeks according to the actual DX release profiles (cf. the actual DX release profile from non-stored DX-St scaffolds in Fig. 3.10) and regardless of the storage time period used (not shown). The burst release was related to the release of DX weakly bound to the scaffolds or in the surface of the pores, whereas the erosion of the polymeric matrix controlled the sustained release of DX embedded within the scaffolds. 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 57 Fig. 3.8: Release profiles of DX in PBS pH 7.4 medium at 37 ºC from (a) non-stored scaffolds, and scaffolds stored for (b) 1 month and (c) 3 months during a release period of 21 days (left) and the first 6 hours (right). Black circles and white triangles represent St0-DX and St-DX scaffolds, respectively. Black lines represent the proposed DX release kinetics. The actual DX release was hardly affected by the presence of starch. Both scaffold formulations before the storage tests showed similar weight loss rates during the first three weeks (Fig. 3.6). Accordingly the kinetic constants from both formulations were similar (Table 3.6). LETICIA GOIMIL GARCÍA ________________________________________________________ 64 C. A. García-González, A. Concheiro and C. Alvarez-Lorenzo (2015). "Processing of Materials for Regenerative Medicine Using Supercritical Fluid Technology." Bioconjugate Chemistry 26(7): 1159-1171. C. A. García-González, A. Vega-González, A. M. López-Periago, P. Subra-Paternault and C. Domingo (2009). "Composite fibrous biomaterials for tissue engineering obtained using a supercritical CO2 antisolvent process." Acta Biomaterialia 5(4): 1094-1103. C. A. García González, L. Diaz-Gomez, C. Alvarez Lorenzo and A. Concheiro Nine (2015). System for administering biologically active substances produced by foaming techniques using compressed gases or supercritical fluids. Spanish. WO 2017013288 A1. H. Giesche (2006). "Mercury Porosimetry: A General (Practical) Overview." Particle & Particle Systems Characterization 23(1): 9-19. L. Goimil, M. E. M. Braga, A. M. A. Dias, J. L. Gómez-Amoza, A. Concheiro, C. Alvarez-Lorenzo, H. C. de Sousa and C. A. GarcíaGonzález (2017). "Supercritical processing of starch aerogels and aerogel-loaded poly(ε-caprolactone) scaffolds for sustained release of ketoprofen for bone regeneration." Journal of CO2 Utilization 18: 237249. E. Gómez-Barrena, P. Rosset, D. Lozano, J. Stanovici, C. Ermthaller and F. Gerbhard (2015). "Bone fracture healing: Cell therapy in delayed unions and nonunions." Bone 70: 93-101. J. A. Han and S. T. Lim (2004). "Structural changes in corn starches during alkaline dissolution by vortexing." Carbohydrate Polymers 55(2): 193-199. T. Hickey, D. Kreutzer, D. J. Burgess and F. Moussy (2002). "Dexamethasone/PLGA microspheres for continuous delivery of an anti-inflammatory drug for implantable medical devices." Biomaterials 23(7): 1649-1656. S. T. Ho and D. W. Hutmacher (2006). "A comparison of micro CT with other techniques used in the characterization of scaffolds." Biomaterials 27(8): 1362-1376. 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 65 R. Jahagirdar and B. E. Scammell (2009). "Principles of fracture healing and disorders of bone union." Surgery (Oxford) 27(2): 63-69. M. Jewell, W. Daunch, B. Bengtson and E. Mortarino (2015). "The development of SERI® Surgical Scaffold, an engineered biological scaffold." Annals of the New York Academy of Sciences 1358(1): 4455. T. E. A. H. Küpper, S. Bettina, R. Burkhard, A.-V. Hemmerling, S. Volker and S. Juergen (2006). "Drugs and Drug Administration in Extreme Environments." Journal of Travel Medicine 13(1): 35-47. F. Linde, C. B. Gothgen, I. Hvid and B. Pongsoipetch (1988). "Mechanical properties of trabecular bone by a non-destructive compression testing approach." Engineering in Medicine 17(1): 23-29. L. Lu, S. J. Peter, M. D. Lyman, H.-L. Lai, S. M. Leite, J. A. Tamada, S. Uyama, J. P. Vacanti, L. Robert and A. G. Mikos (2000). "In vitro and in vivo degradation of porous poly(dl-lactic-co-glycolic acid) foams." Biomaterials 21(18): 1837-1845. A. S. Mao and D. J. Mooney (2015). "Regenerative medicine: Current therapies and future directions." Proceedings of the National Academy of Sciences of the United States of America 112(47): 14452-14459. E. Markočič, M. Škerget and Ž. Knez (2011). "Solubility and diffusivity of CO2 in poly(l-lactide)–hydroxyapatite and poly(d,llactide-co-glycolide)–hydroxyapatite composite biomaterials." The Journal of Supercritical Fluids 55(3): 1046-1051. N. G. Mehr, X. Li, G. Chen, B. D. Favis and C. D. Hoemann (2015). "Pore size and LbL chitosan coating influence mesenchymal stem cell in vitro fibrosis and biomineralization in 3D porous poly(epsiloncaprolactone) scaffolds." Journal of Biomedical Materials Research Part A 103(7): 2449-2459. J. Melke, S. Midha, S. Ghosh, K. Ito and S. Hofmann (2016). "Silk fibroin as biomaterial for bone tissue engineering." Acta Biomaterialia 31: 1-16. M. D. Moya-Ortega, C. Alvarez-Lorenzo, H. H. Sigurdsson, A. Concheiro and T. Loftsson (2010). "γ-Cyclodextrin hydrogels and LETICIA GOIMIL GARCÍA ________________________________________________________ 66 semi-interpenetrating networks for sustained delivery of dexamethasone." Carbohydrate Polymers 80(3): 900-907. R. A. Quirk, R. M. France, K. M. Shakesheff and S. M. Howdle (2004). "Supercritical fluid technologies and tissue engineering scaffolds." Current Opinion in Solid State and Materials Science 8(3): 313-321. A. Salerno, S. Dieguez, L. Diaz-Gomez, J. L. Gómez-Amoza, B. Magarinos, A. Concheiro, C. Domingo, C. Alvarez-Lorenzo and C. A. García-González (2017). "Synthetic scaffolds with full pore interconnectivity for bone regeneration prepared by supercritical foaming using advanced biofunctional plasticizers." Biofabrication 9(3): 035002. A. Salerno, S. Zeppetelli, E. Di Maio, S. Iannace and P. A. Netti (2012). "Architecture and properties of bi-modal porous scaffolds for bone regeneration prepared via supercritical CO 2 foaming and porogen leaching combined process." Journal of Supercritical Fluids 67: 114-122. M. H. Sheridan, L. D. Shea, M. C. Peters and D. J. Mooney (2000). "Bioabsorbable polymer scaffolds for tissue engineering capable of sustained growth factor delivery." Journal of Controlled Release 64(1–3): 91-102. N. Shibuya and D. C. Jupiter (2015). "Bone Graft Substitute: Allograft and Xenograft." Clinics in Podiatric Medicine and Surgery 32(1): 2134. M. M. C. G. Silva, L. A. Cyster, J. J. A. Barry, X. B. Yang, R. O. C. Oreffo, D. M. Grant, C. A. Scotchford, S. M. Howdle, K. M. Shakesheff and F. R. A. J. Rose (2006). "The effect of anisotropic architecture on cell and tissue infiltration into tissue engineering scaffolds." Biomaterials 27(35): 5909-5917. S. Spilimbergo and A. Bertucco (2003). "Non-thermal bacterial inactivation with dense CO2." Biotechnology and Bioengineering 84(6): 627-638. W. Q. Sun and S.-S. Gouk (2008). "Aging of a Regenerative Biologic Scaffold (AlloDerm Native Tissue Matrix) During Storage at Elevated 3. Preparation and stability of dexamethasone-loaded scaffolds for bone regeneration processed by compressed CO2 foaming ___________________________________________________________ 67 Humidity and Temperature." Tissue Engineering Part C: Methods 15(1): 23-31. J. L. Walker and M. Santoro (2017). 9 - Processing and production of bioresorbable polymer scaffolds for tissue engineering. Bioresorbable Polymers for Biomedical Applications, Woodhead Publishing: 181203. R. E. Wilson (1921). "Humidity Control by Means of Sulfuric Acid Solutions, with Critical Compilation of Vapor Pressure Data." Journal of Industrial & Engineering Chemistry 13(4): 326-331. M. A. Woodruff, C. Lange, J. Reichert, A. Berner, F. Chen, P. Fratzl, J.-T. Schantz and D. W. Hutmacher (2012). "Bone tissue engineering: from bench to bedside." Materials Today 15(10): 430-435. J. Yamashita, B. R. Furman, H. R. Rawls, X. Wang and C. M. Agrawal (2001). "The use of dynamic mechanical analysis to assess the viscoelastic properties of human cortical bone." Journal of Biomedical Materials Research 58(1): 47-53. J. J. Yoon, J. H. Kim and T. G. Park (2003). "Dexamethasonereleasing biodegradable polymer scaffolds fabricated by a gasfoaming/salt-leaching method." Biomaterials 24(13): 2323-2329. J. Zhang, A. Zhou, A. Deng, Y. Yang, L. Gao, Z. Zhong and S. Yang (2015). "Pore architecture and cell viability on freeze dried 3D recombinant human collagen-peptide (RHC)-chitosan scaffolds." Materials Science and Engineering C: Materials for Biological Applications 49: 174-182. L. Zhu, Z. Yang, Q. Yang, Z. Tu, L. Ma, Z. Shi and X. Li (2015). "Degradation of dexamethasone by acclimated strain of Pseudomonas Alcaligenes." International Journal of Clinical and Experimental Medicine 8(7): 10971-10978. LETICIA GOIMIL GARCÍA ________________________________________________________ 68 4. SUPERCRITICAL PROCESSING OF STARCH AEROGELS AND AEROGEL-LOADED PCL SCAFFOLDS FOR SUSTAINED RELEASE OF KETOPROFEN FOR BONE REGENERATION§ §The work described in this chapter has been published in Goimil et al. J. CO2 Utiliz. 18 (2017) 237-249 and carried out in collaboration with the CIEPQPF, Chemical Engineering Department, FCTUC, University of Coimbra and within the frame of a 4-month PhD research stay. LETICIA GOIMIL GARCÍA _____________________________________________________________ 70 4.1. INTRODUCTION Synthetic scaffolds are a promising alternative to biological grafts to promote bone repair in those situations where the natural selfregeneration is compromised. These synthetic constructs should give response to the typical problems associated with safety issues and to the scarcity of biological grafts with respect to the current increasing demand. Moreover, advanced synthetic scaffolds should have a superior performance to provide an accurate porous 3D-structure that guides and promotes tissue growth, enables the diffusion of nutrients and oxygen supplies and cell waste disposal, and acts as a provisional mechanical support (García-González et al. 2015). Finally, bone repair materials should also have a performance aligned with the current social changes associated with the increase in life expectancy of population and with the world commitment of keeping older people autonomy and enhancing their quality of life in the so-called “active ageing” (Wold Health Organization (WHO) 2002). Bioactive compounds can be incorporated in scaffolds for their local administration to promote bone healing efficiency by tackling several types of post-surgery complications due to infections and other biological processes that can impair the proper tissue regeneration (Romagnoli et al. 2013). Namely, it is necessary to reduce the time taken to solve severe foreign-body inflammatory responses (Vacanti et al. 2012). Otherwise, poor osteointegration of the scaffold due to fibrous encapsulation and granuloma formation can occur (Sridharan et al. 2015, Przekora and Ginalska 2016). Ketoprofen is a nonsteroideal anti-inflammatory drug (NSAID) that is orally administered during the post-operative period to relief pain and to reduce inflammation. The mode of action of ketoprofen is the nonselective inhibition of cyclooxygenases (COX-1 and COX-2), enzymes involved in the production of prostaglandins (Wibberley et al. 2006, Pountos et al. 2012). Ketoprofen does not inhibit bone healing process according to in vivo animal studies with goats and rabbits where no effects on MSCs proliferation and osteogenesis were observed after subcutaneous administration of daily doses of ketoprofen (ca. 2 mg/kg) at least for several weeks (van der Heide et al. 2008, Nyangoga et al. 2010). 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 71 Several processing strategies are reported in the literature to incorporate bioactive compounds in bone scaffolds and to tune their release kinetics patterns to provide a local administration (Tezcaner and Keskin 2011). However, common scaffold processing techniques like solvent casting/particle leaching, phase separation or rapid prototyping have limited versatility to prepare drug-loaded scaffolds. These techniques usually need the use of either organic solvents, or high operating temperatures, or multi-step processing with leaching or purification steps, which may originate problems of cytotoxicity, premature drug degradation, low drug loading yields, reproducibility concerns or long processing times among others. Supercritical foaming of polymeric scaffolds emerges as the only one-step straightforward alternative able to overcome most of the abovementioned drawbacks. This technique exploits the plasticizing and melting effects of supercritical carbon dioxide (scCO2) on some biodegradable thermoplastic polymers such as poly-α-hydroxyesters (e.g. PLA, PLGA, PCL) to produce highly porous synthetic bone scaffolds in a simple, economical and reproducible procedure (DiazGomez et al. 2016). Nevertheless, the supercritical CO2 foaming process has some limited control on the production of materials presenting well-defined pore size distributions and pore interconnectivities and, consequently, on the release of drug-loaded scaffolds. Mesoporous silica particles incorporated in foamed polyesterbased scaffolds have been shown to favour heterogeneous pore nucleation, to improve the mechanical properties and to refine the pore size distribution (de Matos et al. 2013). Also, inorganic aerogel particles tested for similar purposes were able to preserve the porous structure of the wet gel in the dry form (García-González et al. 2012). Silica aerogel particles were observed to promote the survival and growth of fibroblastic and osteoblastic cells if incorporated in PCLscaffolds obtained by solvent casting (Ge et al. 2013). Nevertheless, the incorporation of silica in scaffold formulations is still a concern since rigorous downstream processes are required to remove remnants of toxic silica precursors and the full biodegradability, bioerosion and excretion profiles of silica aerogels are still not fully confirmed. The LETICIA GOIMIL GARCÍA _____________________________________________________________ 72 use of polysaccharide aerogels as a safer, biodegradable and even cheaper alternative mesoporous material to be incorporated in synthetic grafts seems promising (García-González et al. 2011) and it is prospected in this work. Purely organic aerogel scaffolds can present high mesoporosity and good cytocompatibility, (Pircher et al. 2014, Salerno and Domingo 2014, Martins et al. 2015) but weak mechanical properties for hard-tissue repair and limited control of macroporosity. Mechanical reinforcement of aerogels implies post-processing steps (Pircher et al. 2014). Techniques explored to confer macroporosity to aerogel scaffolds give restricted results concerning the overall macroporosity values and the control of the macropores sizes (Salerno and Domingo 2014, Martins et al. 2015, Stergar and Maver 2016). Differently, the combination of micron-sized polymeric aerogel particles with biodegradable polyesters using the supercritical CO2 foaming technique to prepare scaffolds would assemble and combine the high porosity and mesopore volume of aerogels, the good mechanical properties of polyesters and the capacity of tuning macropores sizes and densities of the scCO2 foaming method. Starch aerogel powders emerge as an interesting admixture of polymeric-based scaffolds due to the intrinsic biocompatibility and biodegradability of starch and to the excellent textural properties of aerogels (Maleki et al. 2016). Moreover, starch is an abundant and relatively cheap polysaccharide of particular interest for regenerative medicine purposes (Elvira et al. 2002, Silva et al. 2007). Starch is able to interact with polyesters changing its crystallinity depending on the starch gelification method used (Diaz-Gomez et al. 2016). Starch aerogels have been prepared from different sources (corn, pea, tapioca, potato), formats (monoliths, beads, particles) and using different gelation mechanisms (thermal, inclusion complexing, microwaves) (Mehling et al. 2009, García-González et al. 2011, García-González et al. 2012, García-González and Smirnova 2013, Kenar et al. 2014, Milovanovic et al. 2015, Muñoz García et al. 2015). Namely, aerogel particles reported in the literature have particle sizes in the order of hundreds of microns or larger for starch and of tens of microns or higher for other polysaccharides (Alnaief et 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 73 al. 2011, García-González et al. 2015). Nevertheless, starch aerogels in the size range of few microns would be ambitioned for several pharmaceutical and biomedical applications and, namely, for their incorporation in composite scaffolds presenting uniform physicochemical and mechanical properties. The aim of this work was to prepare starch aerogel microspheres and to evaluate their potential roles once incorporated in ketoprofenloaded PCL-based scaffolds on the control of the pore structure of the material and of the ketoprofen release over time. One-micron sized starch aerogel microspheres were developed by emulsion-gelation under ultrasound sonication followed by supercritical drying, for the first time. Ketoprofen was loaded in the starch aerogel particles by a supercritical CO2 impregnation method or directly incorporated in the scaffold by a supercritical CO2 foaming process for the sake of comparison. The resulting scaffolds were evaluated regarding their morphological, mechanical and physicochemical properties, as well as to their ketoprofen-release profiles. 4.2. MATERIALS AND METHODS 4.2.1. Materials PCL (PCLraw, 50 kDa, Tm=61.4 ºC, 66.7 % crystallinity) was purchased from Polysciences (Warrington, PA, USA) in the powdered form. Corn starch in the native form (Straw, amylo N-460; 52.6 % amylose content, 12.9 % loss on drying, conform to USP and EP Pharmacopeias) was obtained from Roquette (Lestrem, France), and carbon dioxide (99.9 % purity) was supplied by Praxair, Inc. (Madrid, Spain). Ketoprofen (K, Tm= 95.8 ºC, 99.7 % purity) was provided by Acofarma (Terrassa, Spain). Polyglycerol polyricinoleate (PGPR) was obtained from Palsgaard (Juelsminde, Denmark). Paraffin oil was provided by Panreac (Castellar del Vallès, Spain). Ethanol (99.8 % purity) was obtained from Omnilab (Bremen, Germany) and diethyl ether (98 % purity) from Labkem (Mataró, Spain). 4.2.2. Starch aerogel preparation Starch gels in the form of microspheres were obtained by the emulsion-gelation method (García-González et al. 2012). Briefly, a LETICIA GOIMIL GARCÍA _____________________________________________________________ 80 4.2.6. Ketoprofen release studies Scaffold samples of 10 mg were suspended in 50 mL of PBS pH 7.4. The dissolution profile of 1 mg of ketoprofen dissolved in 100 mL of PBS was determined as reference for the release profiles. Experiments were performed under sink conditions (solubility of ketoprofen in PBS pH 7.4 = 2.2 mg/mL) (Mukae et al. 1990). Flasks were put in an oscillating bath (Unitronic 320 OR, JP Selecta, Barcelona, Spain) at 37 ºC and 60 rpm for 3 weeks. Aliquots of 1 mL were sampled through that period at selected times, and withdrawn volumes were replaced with fresh medium. Ketoprofen concentration was measured by UV-Vis spectrophotometry (8453, Agilent, Santa Clara, CA, USA) at λ = 260 nm. Samples were filtered prior to analysis using 0.2 m nylon filters to avoid potential interference due to the presence of particles from the scaffolds. Calibration curve was obtained from triplicate dilution series of ketoprofen in PBS, ranging from 0.001 to 0.025 mg/mL, attaining a R2=0.9994. Modeling of the ketoprofen release was carried out by fitting the obtained release data to the Korsmeyer-Peppas equation (Eq. (4.6)) (Siepmann and Peppas 2011) and using GraphPad Prism version 6.04 for Windows (GraphPad Software, La Jolla, CA, USA) software: (4.6) where F is the fraction of drug released at a time t, k is a constant related to the macromolecular polymeric network structure, and n is the diffusional exponent (Peppas 1985, Siepmann and Peppas 2011). 4.2.7. Statistical analysis All results were expressed as mean ± standard deviation. 1-way ANOVA and Tukey’s multiple comparisons tests were used to evaluate the effect of starch aerogel content on the scaffolds on the bulk density and the overall porosity of the scaffolds (Statgraphics Centurion XVI, StatPoint Technologies Inc., Warrenton, VA, USA). 4.3. RESULTS AND DISCUSSION 4.3.1. Starch aerogel microspheres processing Starch aerogel microspheres with a unimodal particle size distribution of mean size of ca. 1.2 µm were obtained by applying the n tkF  4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 81 emulsion-gelation method followed by supercritical CO2 drying (Fig. 4.1). After supercritical drying, the particle size of these aerogel particles was considered as suitable for their incorporation in the PCLbased scaffolds without interfering in the isocratic properties of the resulting material. Remnants of the native starch granules were not observed by SEM, confirming the effectiveness of the gelation process. Previous attempts to get polysaccharide aerogel spheres using different processing techniques (prilling, microfluidics) resulted in particle diameters from hundreds of microns to some milimeters (De Cicco et al. 2016, Zhou et al. 2016). Particularly, previous results on polysaccharide aerogel spheres obtained by emulsion-gelation methods resulted in particle sizes in the range of tens to hundreds of microns (Alnaief et al. 2011, García-González et al. 2012, GarcíaGonzález et al. 2015). In this work, the aerogel particle size threshold of ten microns has been trespassed by combining the proper choice of the emulsifier source (polyglycerol polyricinoleate) and content (3% (w/w)) and a vigorous agitation method (ultrasound probe). In addition, a relatively low polydispersity index (PdI) was also achieved. SEM pictures with higher magnifications confirmed the good sphericity of the particles as well as the presence of porosity in the Fig. 4.1: Particle size evaluation of starch aerogel microspheres (StA): (a) SEM image of starch aerogels and (b) and particle size distribution of the starch gel particles. Inset: representative parameters from the dynamic light scattering analysis. 364 364 0 5 10 15 20 10 100 1000 10000 Intensity (%) Size (nm) Zav (d.nm) PdI 1210 0.127 (a) (b) (a) LETICIA GOIMIL GARCÍA _____________________________________________________________ 82 material (Fig. 4.2). StA sample showed a porous and wrinkle surface partially coated with fused particles. This coating is likely due to the presence of the emulsifier (PGPR) in the samples. To confirm this hypothesis, starch gels were washed with diethylether (a solvent for PGPR and a non-solvent for starch) prior to drying. A porous structure formed by a mesh of strands arising from the starch components rearrangement upon gelation was unveiled after this treatment (StADEE sample). FIB-SEM technique allowed obtaining a cross-section of the aerogel microspheres (Fig. 4.3). The mesoporous inner structure of both aerogel particles was thus confirmed. The analysis of the textural properties of starch aerogels (StA) confirmed the presence of a mesoporous structure having average pore diameters between 24 and 25 nm, and with relatively high BETsurface areas (93 m2/g), BJH-pore volume (0.69 cm3/g) and porosities (87.7±0.2 %). These values were similar to those reported in literature Fig. 4.2: Morphology (left) and textural appearance (right) of starch aerogel microspheres obtained without (StA, top) and with (StA-DEE, bottom) emulsifier removal treatment. 200 nm 200 nm 100 nm 100 nm 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 83 Fig. 4.3: FIB-SEM images of the inner structure of starch aerogel microspheres obtained without (StA, left) and with (StA-DEE, right) emulsifier removal treatment. for larger starch aerogel particles (García-González et al. 2012). The surfactant removal treatment resulted in an aerogel (StA-DEE) with one-fold increase in BET-surface area (188 m2/g) and BJH-pore volume (1.25 cm3/g) with respect to the untreated sample. These higher values are comparable to those reported for starch aerogels in the form of monoliths obtained by standard gelation methods (GarcíaGonzález et al. 2012, García-González and Smirnova 2013). BJHmean pore diameters of the starch aerogels was of ca. 24 nm for both treatments showing that the diethylether was successful in removing the emulsifier whilst preserving the mesoporous structure. C-constant values obtained from the BET equation, which are related with the adsorbate/adsorbent interaction energies and thus associated with material hydrophilicity, (Marcinko et al. 2003, García-González et al. 2009) showed that untreated starch aerogel had an intermediate Cvalue of 36. This C-value is likely due to the combined effects of the relative contributions of the outer surface presence of fatty acids (from the PGPR) and of the inner surface presence of hydroxyl groups (from starch). The absence of PGPR for StA-DEE aerogel microspheres significantly increased the C-value (C=102) due to the improved interactions that can be established between the hydroxyl groups in both inner and outer surfaces of the aerogel and the adsorbed N2. However, in this work the intermediate hydrophilicity of StA aerogels was preferred for the subsequent incorporation in the scaffolds since it would be expected to favor the interactions of aerogel microspheres Table 4.2a: Composition, morphological and textural properties of supercritically processed aerogels and scaffolds. Mean values and standard deviation (n=3). Scaffold ρbulk, g/cm3 ρskel, g/cm3 ε, % ABET, m2/g Vp,BJH, cm3/g dp,BJH, nm AMIP, m2/g Vp,MIP, cm3/g dp,MIP, μm εMIP, % PCL 0.482± 0.027 1.106± 0.012 56.4± 2.4 - - - 5.61 0.64 0.46 42.05 PCL-Straw 0.517± 0.111 1.146± 0.008 54.9± 9.6 - - - 3.66 0.49 0.46 35.22 PCL-StA 0.386± 0.078 1.142± 0.012 66.2± 6.8 1.3 0.0017 15.1 5.76 0.55 0.38 38.98 PCL-K 0.511± 0.082 1.098± 0.004 53.5± 7.4 1.2 0.0006 3.3 5.98 0.40 0.27 30.93 PCL-Straw-K 0.461± 0.091 1.155± 0.013 60.1± 7.9 1.4 0.0009 3.8 3.99 0.46 0.46 33.93 PCL-StA-K 0.345± 0.049 1.147± 0.005 69.9± 4.3 1.6 0.0040 14.1 5.48 0.41 0.30 31.68 PCL-StAK-K 0.329± 0.079 1.111± 0.012 70.4± 7.1 1.7 0.0054 14.9 1.55 0.72 1.87 44.81 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 85 Table 4.2b: Composition, morphological and textural properties of supercritically processed aerogels and scaffolds. Mean values and standard deviation (n=3). Scaffold Pore median, μm Interconnectivity, % Water uptake (<1s)a, % vol. Water uptake (50ms)b, % vol. kw, mD Cell infiltration, % PCL 74.1 92.5 8.33± 1.41 94.35 616.66 83 PCL-Straw 75.0 92.2 3.71± 2.36 94.21 421.77 77 PCL-StA 85.1 90.7 9.17± 3.75 94.62 4902.43 87 PCL-K 84.2 65.2 5.82± 2.45 86.62 2439.63 94 PCL-Straw-K 67.2 82.2 17.03± 8.97 81.51 67.54 83 PCL-StA-K 94.8 74.8 46.95± 9.59 91.85 7469.82 92 PCL-StAK-K 99.0 93.0 25.52± 2.78 98.72 5866.79 84 aIn vitro value; bIn silico value LETICIA GOIMIL GARCÍA _____________________________________________________________ 86 with both the hydrophobic PCL-based matrix scaffold and the aqueous body fluids. The obtained favourable flow properties of starch aerogel powders allow their volumetric dosing and mixing with the other components to prepare scaffolds by supercritical foaming. Starch aerogels presented an angle of repose of 29.0±1.9º, which is comparable to those reported for reference materials with good flowability like spray-dried lactose (31-33º) (DFE pharma 2013, Huang et al. 2013). In addition, the angle of repose was improved with respect to native starch Straw (36.9±2.1º). Finally, the reduced particle sizes and high porosities/surface areas of the prepared starch aerogel powders tend to be more cohesive than for larger and denser particles. Accordingly, the compressibility obtained for starch aerogel powder was 32.2±0.2 %, while for native starch was 30.5±0.5 %. 4.3.2. Scaffolds development and morphological characterization Lightweight PCL-based cylindrical scaffolds were obtained using the supercritical foaming process (Fig. 4.4 and 4.5 and Table 4.2). Obtained pure PCL porous materials had an external nonporous layer of some few microns and a macroporous inner structure of 200-300 μm pore size (Fig. 4.5a) partially interconnected through throats of two different families, one below 1 μm and other above 50 μm (Fig. 4.6a). Fig. 4.4: Visual appearance of scaffolds after supercritical foaming treatment and removal from the cylindrical mould. From right to left, PCL scaffolds with increasing content in starch aerogel (StA). 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 87 Starch was added in the formulation of PCL-based scaffolds to tune the porous morphology and the release profile of bioactive agents. The incorporation of corn starch in its native form (PCL-Straw) was not successful since the resulting porous material presented a poor PCL-starch miscibility and spatial distribution, as observed in the inset of Fig. 4.4b, as well as a higher densification and reduced porosity with respect to pure PCL scaffolds (Table 4.2). A similar densification of PCL scaffolds was also reported with the use of oven-dried starch gels (i.e., xerogels)(Diaz-Gomez et al. 2016). The incorporation of starch in the form of aerogel microparticles to the formulation gave rise to scaffolds having increased porosities and decreased bulk densities proportional to the starch aerogel content in the scaffold (Fig. 4.7). Namely, an increase in porosity and a decrease of bulk density of up to ca. 20% was observed for PCL-StA scaffolds when compared with the purely PCL scaffold. A visual change of the scaffold consisting on the presence of a dome-like top ending of the PCL-StA scaffolds was also noticed, instead of a flat surface as observed for PCL scaffolds (Fig. 4.4). Additionally, an increase in the pore size (300-600 μm) and in pore interconnectivity through larger pores seems to take place for PCL-StA scaffolds (Fig. 4.5c,left). The aerogel particles seem to be mainly located nearby pore surfaces giving rise to an evident increase in surface roughness of the pores (Fig. 4.5c right). This preferential location of aerogel particles nearby pore surface seems to indicate that porous microspheres may be playing a decisive role in the generation of pores in PCL. In the supercritical foaming process, pores are essentially generated upon depressurization due to the supersaturation in CO2 (nucleation), and grow due to the diffusion of CO2 out from the matrix (GarcíaGonzález et al. 2015). The role of nanometric and micrometric inclusions as secondary pore generation sites in scaffolds obtained by supercritical foaming has been already reported for other matrices incorporating highly porous inorganic particles inducing an increased and dual porosity (Collins et al. 2008, de Matos et al. 2013). However, the role of aerogels in the scaffolds on pore generation goes beyond a secondary pore nucleation site behaviour as reported before, since a high interconnectivity of pores through large throats was also LETICIA GOIMIL GARCÍA _____________________________________________________________ 88 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 89 Fig. 4.5: Optical (left) and SEM (right) micrographs of PCL-based scaffolds obtained using supercritical: (a) PCL, (b) PCL-Straw, (c) PCL-StA, (d) PCL-K, (e) PCL-K-Straw, (f) PCL-K-StA and (g) PCL-K-StAK. Scale bars: (left) 1mm, (right) 100 µm and (inset) 10 µm. observed here for the first time (Fig. 4.6a). The inherent porosity of aerogels would allow secondary CO2 release pathways out from the scaffold matrix during the depressurization process and might likely favour the connection between pores (i.e. interconnectivity). LETICIA GOIMIL GARCÍA _____________________________________________________________ 96 point value (Lozano and Martínez 2006, García-González et al. 2012, García-González et al. 2015). The increased crystallinity was related to the presence of the starch in the formulation as explained before. Both behaviours are refrained by the DSC results from PCL-StAK-K scaffold formed by PCL, ketoprofen and ketoprofen-saturated starch aerogel microspheres. In this case, crystallinity is high due to the presence of the starch aerogels, whereas Tm value is similar to that of PCL-K scaffold because ketoprofen is in excess with respect to starch in the formulation thus promoting the interaction with PCL. 4.3.4. Permeability and human MSCs infiltration modeling of the scaffolds Scaffolds used for bone regeneration need porous structures with different pore ranges that are able to allow the transport of fluids containing nutrients, metabolites and wastes as well as the accessibility of cells within the material for bone ingrowth and neovascularization (García-González et al. 2015). The study of the water uptake and permeability through the scaffold and the percolation capacity of particles of defined sizes (namely similar to those of MSCs) give essential information for the in silico screening of the performance of scaffolds. Results are presented in Table 4.2. According to the in silico values, all prepared PCL-based scaffolds presented a fast water uptake with more than 80% of the open pore volume filled with the liquid in less than 50 ms after being immersed in pure water. Experimental values for water uptake differed from the modelled values. Values were in general lower than the modelled values due to the experimental procedure, where samples were weighed just after whipping with filter paper. The in silico water uptake values were obtained considering the same hydrophilicity for all the samples and this neglect might explain the differences obtained between in vitro and in silico values. Two distinct groups of scaffold were unveiled regarding in vitro values of instantaneous water uptake: those containing ketoprofen and starch aerogel (PCL-StA-K and PCLStAK-K) and the rest of the studied scaffolds. The interaction of ketoprofen with starch (cf. Section 4.3.3) seems to influence the amount of polysaccharide particles in the graft directly exposed to the 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 97 surface of the pore walls, thus improving the hydrophilicity of the scaffolds. The differences in composition, pore morphology (pore and throat sizes) and interconnectivity observed for the different PCL-based scaffolds studied (cf. Sections 4.3.2 and 4.3.3) strongly influenced the permeability of aqueous solutions and the MSCs percolation through the 3D-porous network of the scaffolds. Despite all scaffolds may present high pore interconnectivity values, the throat size distribution and the throat-to-pore ratio in scaffolds are additional key parameters varying the fluid flow properties and the MSCs percolation through the material. The high contribution of throats in the 1-10 μm region for the PCL and PCL-Straw scaffolds resulted in a low water permeability and a limited infiltration capacity of simulated MSCs within the scaffold. The loading of the scaffolds with ketoprofen slightly improved both properties. The use of starch aerogel microspheres in the formulations significantly altered the throat size distribution leading to enhanced water permeability (5-10 times higher) and MSCs infiltration capacity (up to 13 %). Water permeability values of starch aerogel containing scaffolds fell close to the lower limit of the permeability range reported for cancellous bone (104-108 mD) (Syahrom et al. 2013). Hence, the incorporation of starch aerogel micropheres to the PCLbased scaffolds inferred the most promising 3D-porous architecture for bone tissue ingrowth. 4.3.5. Ketoprofen release from PCL scaffolds The choice of the scaffold formulation, the modulation of the scaffold porosity, the drug loading strategy and the drug loading amount are factors significantly influencing the drug release profile from bone scaffolds (Tezcaner and Keskin 2011). In this work, a porous scaffolds consisting on a hydrophobic porous matrix (PCL) with highly porous hydrophilic components (starch particles) and having high chemical affinity through hydrogen bonding (Lozano and Martínez 2006, Messner et al. 2012) for ketoprofen was accordingly chosen to sustain the release during days (Fig. 4.9). As a result, all prepared scaffolds presented slower ketoprofen release profiles than LETICIA GOIMIL GARCÍA _____________________________________________________________ 98 that corresponding to the dissolution profile of the crystalline drug. An initial burst occurred during the first hours, and then sustained release was recorded for several days. This time-dependent ketoprofen release profile seems suitable to provide a prompt anti-inflammatory response in the initial stages (burst step) and also long-term action (sustained release step) during the bone healing process (Prabaharan et al. 2007, Raafat and Abd-Allah 2018). The use of starch aerogels as an admixture of the scaffold composition had a relevant role in the tuning of the ketoprofen release profile. Scaffolds containing starch aerogels (PCL-StA-K and PCLStAK-K) had a faster initial release (within 30 min) if compared with scaffolds not containing starch (PCL-K, 1 h) or with unprocessed starch (PCL-Straw-K, 2 h). Then, PCL-StA-K and PCL-StAK-K have sustained releases with more than 90% of the payload within 3 days, whereas the ketoprofen release took longer times for the rest of the scaffolds. The higher accessibility of water to the scaffolds containing starch aerogel microspheres unveiled by the water permeability study (Table 4.4) is likely responsible for the observed accelerated release profile, especially during the burst stage. The preferential location of starch aerogel microspheres in the pore surface coupled to the affinity of ketoprofen for the hydroxyl groups from starch would explain the accelerated but sustained release profile of ketoprofen from the scaffolds. Drug release profiles from the scaffolds were fitted to the Korsmeyer-Peppas equation (Table 4.4) showing n values below threshold of 0.45 in all cases. According to these results, the drug release falls into a complex diffusion-controlled mechanism comprising a combination of diffusion through the matrix and through the pores filled with the solution from the release medium (Peppas 1985, Diaz-Gomez et al. 2016). Table 4.4: Kinetic fitting parameters of the ketoprofen release profiles from drugloaded scaffolds in PBS solution (pH 7.4) according to Eq. (4.6). Scaffold k, h-n n R2 PCL-K 0.224±0.038 0.304±0.049 0.884 PCL-Straw-K 0.294±0.016 0.261±0.016 0.969 PCL-StA-K 0.244±0.035 0.320±0.041 0.923 PCL-StAK-K 0.299±0.041 0.304±0.040 0.920 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 99 Fig. 4.9: Ketoprofen released from PCL-based scaffolds in PBS solution (pH 7.4) during 7 days. Ketoprofen dissolution profile (black diamonds) is provided for the sake of comparison. Legend: PCL-K (black squares), PCL-Straw-K (white circles), PCL-StA-K (black triangles) and PCL-StAK-K (white triangles). 4.4. CONCLUSIONS Biodegradable mesoporous particles from polysaccharides of a size as small as one micron are processed in this work for the first time. This format encompasses the outstanding textural properties of aerogels (high specific surface areas and pore volumes) and the excellent solid physical mixing capacity of microparticles. The incorporation of starch aerogel microspheres in highly porous PCL scaffolds processed by supercritical CO2 foaming significantly improved the potential uses of these biomaterials for bone regeneration applications, notably regarding pore throat size, water permeability and MSCs infiltration capacity. These PCL-based scaffolds also provided good mechanical properties and an in vitro sustained release of an anti-inflammatory drug (ketoprofen) during several days. In vivo ketoprofen release tests need to be carried out to confirm this two-stage release pattern under physiological conditions. The mild operating conditions used for the processing of the scaffolds (40 ºC, 150 bar) open up the possibility to incorporate other bioactive compounds (e.g., growth factors) for sustained release to facilitate other bone healing biological processes or to be used as subcutaneous implants or medical devices for drug release. LETICIA GOIMIL GARCÍA _____________________________________________________________ 100 4.5 REFERENCES R. K. Ahuja, M. Kodialam, A. K. Mishra and J. B. Orlin (1997). "Computational investigations of maximum flow algorithms." European Journal of Operational Research 97(3): 509-542. M. Alnaief, M. A. Alzaitoun, C. A. García-González and I. Smirnova (2011). "Preparation of biodegradable nanoporous microspherical aerogel based on alginate." Carbohydrate Polymers 84(3): 1011-1018. P. M. Buechner and R. S. Lakes (2003). "Size effects in the elasticity and viscoelasticity of bone." Biomech Model Mechanobiol 1(4): 295-301. N. J. Collins, G. A. Leeke, R. H. Bridson, F. Hassan and L. M. Grover (2008). "The influence of silica on pore diameter and distribution in PLA scaffolds produced using supercritical CO2." Journal of Materials Science: Materials in Medicine 19(4): 1497-1502. F. De Cicco, P. Russo, E. Reverchon, C. A. García-González, R. P. Aquino and P. Del Gaudio (2016). "Prilling and supercritical drying: A successful duo to produce core-shell polysaccharide aerogel beads for wound healing." Carbohydrate Polymers 147: 482-489. M. B. C. de Matos, A. P. Piedade, C. Alvarez-Lorenzo, A. Concheiro, M. E. M. Braga and H. C. de Sousa (2013). "Dexamethasone-loaded poly(ɛcaprolactone)/silica nanoparticles composites prepared by supercritical CO2 foaming/mixing and deposition." International Journal of Pharmaceutics 456(2): 269-281. E. De Paz, A. Martín, S. Rodríguez-Rojo, J. Herreras and M. J. Cocero (2010). "Determination of phase equilibrium (solid-liquid-gas) in poly-(εcaprolactone)-carbon dioxide systems." Journal of Chemical and Engineering Data 55(8): 2781-2785. DFE pharma. (2013). "Directly compressible lactose." Retrieved 06/07/2016, 2016, from https://www.dfepharma.com. L. Diaz-Gomez, A. Concheiro, C. Alvarez-Lorenzo and C. A. GarcíaGonzález (2016). "Growth factors delivery from hybrid PCL-starch scaffolds processed using supercritical fluid technology." Carbohydrate Polymers 142: 282-292. 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 101 C. Elvira, J. F. Mano, J. San Román and R. L. Reis (2002). "Starch-based biodegradable hydrogels with potential biomedical applications as drug delivery systems." Biomaterials 23(9): 1955-1966. C. A. García-González, M. Alnaief and I. Smirnova (2011). "Polysaccharidebased aerogels - Promising biodegradable carriers for drug delivery systems." Carbohydrate Polymers 86(4): 1425-1438. C. A. García-González, M. C. Camino-Rey, M. Alnaief, C. Zetzl and I. Smirnova (2012). "Supercritical drying of aerogels using CO2: Effect of extraction time on the end material textural properties." Journal of Supercritical Fluids 66: 297-306. C. A. García-González, E. Carenza, M. Zeng, I. Smirnova and A. Roig (2012). "Design of biocompatible magnetic pectin aerogel monoliths and microspheres." RSC Advances 2(26): 9816-9823. C. A. García-González, A. Concheiro and C. Alvarez-Lorenzo (2015). "Processing of materials for regenerative medicine using supercritical fluid technology." Bioconjugate Chemistry 26(7): 1159-1171. C. A. García-González, J. Fraile, A. López-Periago, J. Saurina and C. Domingo (2009). "Measurements and Correlation of Octyltriethoxysilane Solubility in Supercritical CO2 and Assembly of Functional Silane Monolayers on the Surface of Nanometric Particles." Industrial & Engineering Chemistry Research 48(22): 9952-9960. C. A. García-González, M. Jin, J. Gerth, C. Alvarez-Lorenzo and I. Smirnova (2015). "Polysaccharide-based aerogel microspheres for oral drug delivery." Carbohydrate Polymers 117: 797-806. C. A. García-González and I. Smirnova (2013). "Use of supercritical fluid technology for the production of tailor-made aerogel particles for delivery systems." Journal of Supercritical Fluids 79: 152-158. C. A. García-González, J. J. Uy, M. Alnaief and I. Smirnova (2012). "Preparation of tailor-made starch-based aerogel microspheres by the emulsion-gelation method." Carbohydrate Polymers 88(4): 1378-1386. J. Ge, L. Guo, S. Wang, Y. Zhang, T. Cai, R. C. H. Zhao and Y. Wu (2014). "The Size of Mesenchymal Stem Cells is a Significant Cause of Vascular Obstructions and Stroke." Stem Cell Reviews and Reports 10(2): 295-303. LETICIA GOIMIL GARCÍA _____________________________________________________________ 102 J. Ge, M. Li, Q. Zhang, C. Z. Yang, P. H. Wooley, X. Chen and S. Y. Yang (2013). "Silica aerogel improves the biocompatibility in a poly-εcaprolactone composite used as a tissue engineering scaffold." International Journal of Polymer Science 2013. A. Gomez-Carracedo, R. Martinez-Pacheco, A. Concheiro and J. L. GomezAmoza (2010). "Modelling of porosity and waterfronts in cellulosic pellets for understanding drug release behavior." International Journal of Pharmaceutics 388(1–2): 101-106. M. Guzmán, J. Molpeceres, F. García and M. R. Aberturas (1996). "Preparation, characterization and in vitro drug release of poly-εcaprolactone and hydroxypropyl methylcellulose phthalate ketoprofen loaded microspheres." Journal of Microencapsulation 13(1): 25-39. W. Huang, Y. Shi, C. Wang, K. Yu, F. Sun and Y. Li (2013). "Using spraydried lactose monohydrate in wet granulation method for a low-dose oral formulation of a paliperidone derivative." Powder Technology 246: 379-394. A. Johnson, I. M. Roy, G. P. Matthews and D. Patel (2003). "An improved simulation of void structure, water retention and hydraulic conductivity in soil with the Pore-Cor three-dimensional network." European Journal of Soil Science 54(3): 477-490. J. A. Kenar, F. J. Eller, F. C. Felker, M. A. Jackson and G. F. Fanta (2014). "Starch aerogel beads obtained from inclusion complexes prepared from high amylose starch and sodium palmitate." Green Chemistry 16(4): 19211930. F. Linde, C. B. Gothgen, I. Hvid and B. Pongsoipetch (1988). "Mechanical properties of trabecular bone by a non-destructive compression testing approach." Engineering in Medicine 17(1): 23-29. H. R. Lozano and F. Martínez (2006). "Thermodynamics of partitioning and solvation of ketoprofen in some organic solvent/buffer and liposome systems." Brazilian Journal of Pharmaceutical Sciences 42(4): 601-613. H. Maleki, L. Durães, C. A. García-González, P. del Gaudio, A. Portugal and M. Mahmoudi (2016). "Synthesis and biomedical applications of aerogels: Possibilities and challenges." Advances in Colloid and Interface Science 236: 1-27. S. Marcinko, R. Helmy and A. Y. Fadeev (2003). "Adsorption properties of SAMs supported on TiO2 and ZrO2." Langmuir 19(7): 2752-2755. 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 103 Ł. Marcinkowski, A. Kloskowski, A. Spietelun and J. Namieśnik (2015). "Evaluation of polycaprolactone as a new sorbent coating for determination of polar organic compounds in water samples using membrane–SPME." Analytical and Bioanalytical Chemistry 407(4): 1205-1215. E. Markočič, M. Škerget and Z. Knez (2013). "Effect of temperature and pressure on the behavior of poly(εcaprolactone) in the presence of supercritical carbon dioxide." Industrial and Engineering Chemistry Research 52(44): 15594-15601. M. Martins, A. A. Barros, S. Quraishi, P. Gurikov, S. P. Raman, I. Smirnova, A. R. C. Duarte and R. L. Reis (2015). "Preparation of macroporous alginate-based aerogels for biomedical applications." Journal of Supercritical Fluids 106: 152-159. T. Mehling, I. Smirnova, U. Guenther and R. H. H. Neubert (2009). "Polysaccharide-based aerogels as drug carriers." Journal of Non-Crystalline Solids 355(50-51): 2472-2479. M. Messner, O. Häusler and T. Loftsson (2012). Solution enhancement of drug substances using soluble amylose. Proceedings of PBP 8th World Meeting on Pharmaceutics, Biopharmaceutics and Pharmaceutical Technology, Istanbul (Turkey). S. Milovanovic, I. Jankovic-Castvan, J. Ivanovic and I. Zizovic (2015). "Effect of starch xeroand aerogels preparation on the supercritical CO2 impregnation of thymol." Starch/Staerke 67(1-2): 174-182. K. Mukae, Y. H. Bae, T. Okano and S. W. Kim (1990). "A New ThermoSensitive Hydrogel: Poly(ethylene oxide-dimethyl siloxane-ethylene oxide)/Poly(N-isopropyl acrylamide) Interpenetrating Polymer Networks I. Synthesis and Characterization." Polymer Journal 22(3): 206-217. A. Muñoz García, A. J. Hunt, V. L. Budarin, H. L. Parker, P. S. Shuttleworth, G. J. Ellis and J. H. Clark (2015). "Starch-derived carbonaceous mesoporous materials (Starbon®) for the selective adsorption and recovery of critical metals." Green Chemistry 17(4): 2146-2149. L. Nogueiras-Nieto, J. L. Gómez-Amoza, M. B. Delgado-Charro and F. J. Otero-Espinar (2011). "Hydration and N-acetyl-l-cysteine alter the microstructure of human nail and bovine hoof: Implications for drug delivery." Journal of Controlled Release 156(3): 337-344. LETICIA GOIMIL GARCÍA _____________________________________________________________ 104 H. Nyangoga, E. Aguado, E. Goyenvalle, M. F. Baslé and D. Chappard (2010). "A non-steroidal anti-inflammatory drug (ketoprofen) does not delay β-TCP bone graft healing." Acta Biomaterialia 6(8): 3310-3317. N. A. Peppas (1985). "Analysis of Fickian and non-Fickian drug release from polymers." Pharmaceutica Acta Helvetiae 60(4): 110-111. N. Pircher, S. Veigel, N. Aigner, J. M. Nedelec, T. Rosenau and F. Liebner (2014). "Reinforcement of bacterial cellulose aerogels with biocompatible polymers." Carbohydrate Polymers 111: 505-513. I. Pountos, T. Georgouli, G. M. Calori and P. V. Giannoudis (2012). "Do Nonsteroidal Anti-Inflammatory Drugs Affect Bone Healing? A Critical Analysis." The Scientific World Journal 2012: 606404. M. Prabaharan, M. A. Rodriguez-Perez, J. A. de Saja and J. F. Mano (2007). "Preparation and characterization of poly(L-lactic acid)-chitosan hybrid scaffolds with drug release capability." Journal of Biomedical Materials Research, Part B: Applied Biomaterials 81B(2): 427-434. A. Przekora and G. Ginalska (2016). "In vitro evaluation of the risk of inflammatory response after chitosan/HA and chitosan/β-1,3-glucan/HA bone scaffold implantation." Materials Science and Engineering: C 61: 355361. A. I. Raafat and W. M. Abd-Allah (2018). "In vitro apatite forming ability and ketoprofen release of radiation synthesized (gelatin-polyvinyl alcohol)/bioglass composite scaffolds for bone tissue regeneration." Polymer Composites 39(3): 606-615. C. Romagnoli, F. D’Asta and M. L. Brandi (2013). "Drug delivery using composite scaffolds in the context of bone tissue engineering." Clinical Cases in Mineral and Bone Metabolism 10(3): 155-161. R. A. Ruseckaite and A. Jiménez (2003). "Thermal degradation of mixtures of polycaprolactone with cellulose derivatives." Polymer Degradation and Stability 81(2): 353-358. A. Salerno and C. Domingo (2014). "Making microporous nanometre-scale fibrous PLA aerogels with clean and reliable supercritical CO2 based approaches." Microporous and Mesoporous Materials 184: 162-168. 4. Supercritical processing of starch aerogels and aerogel-loaded PCL scaffolds for sustained release of ketoprofen for bone regeneration _____________________________________________________________ 105 J. Siepmann and N. A. Peppas (2011). "Higuchi equation: Derivation, applications, use and misuse." International Journal of Pharmaceutics 418(1): 6-12. G. A. Silva, O. P. Coutinho, P. Ducheyne, I. M. Shapiro and R. L. Reis (2007). "The effect of starch and starch-bioactive glass composite microparticles on the adhesion and expression of the osteoblastic phenotype of a bone cell line." Biomaterials 28(2): 326-334. R. Sridharan, A. R. Cameron, D. J. Kelly, C. J. Kearney and F. J. O’Brien (2015). "Biomaterial based modulation of macrophage polarization: a review and suggested design principles." Materials Today 18(6): 313-325. J. Stergar and U. Maver (2016). "Review of aerogel-based materials in biomedical applications." Journal of Sol-Gel Science and Technology 77(3): 738-752. A. Syahrom, M. R. Abdul Kadir, J. Abdullah and A. Öchsner (2013). "Permeability studies of artificial and natural cancellous bone structures." Medical Engineering and Physics 35(6): 792-799. A. Tezcaner and D. Keskin (2011). Bioactive Agent Delivery in Bone Tissue Regeneration. Active Implants and Scaffolds for Tissue Regeneration. M. Zilberman. Berlin, Heidelberg, Springer Berlin Heidelberg: 193-223. H. Tsuji and Y. Ikada (1998). "Blends of aliphatic polyesters. II. Hydrolysis of solution-cast blends from poly(L-lactide) and poly (ε-caprolactone) in phosphate-buffered solution." Journal of Applied Polymer Science 67(3): 405-415. N. M. Vacanti, H. Cheng, P. S. Hill, J. D. T. Guerreiro, T. T. Dang, M. Ma, S. Watson, N. S. Hwang, R. Langer and D. G. Anderson (2012). "Localized delivery of dexamethasone from electrospun fibers reduces the foreign body response." Biomacromolecules 13(10): 3031-3038. H. J. L. van der Heide, G. Hannink, P. Buma and B. W. Schreurs (2008). "No effect of ketoprofen and meloxicam on bone graft ingrowth: A bone chamber study in goats." Acta Orthopaedica 79(4): 548-554. A. Wibberley, G. P. McCafferty, C. Evans, R. M. Edwards and J. P. Hieble (2006). "Dual, but not selective, COX-1 and COX-2 inhibitors, attenuate acetic acid-evoked bladder irritation in the anaesthetised female cat." British Journal of Pharmacology 148(2): 154-161. LETICIA GOIMIL GARCÍA _____________________________________________________________ 112 preparation of bone scaffolds with advanced properties (Duarte et al. 2009, Bhamidipati et al. 2013, García-González et al. 2015, Salerno and Pascual 2015). Supercritical foaming allows reducing the temperature required for the thermal events (melting point, glass transition) of synthetic polymers up to conditions compatible with temperature-sensitive bioactive agents (Duarte et al. 2009, Fanovich and Jaeger 2012, de Matos et al. 2013, Salerno and Pascual 2015). This solvent-free technology exploits the plasticizing effect of compressed CO2 on polymers to form a porous structure during the controlled removal of CO2 (i.e. depressurization) (Bhamidipati et al. 2013, García-González et al. 2015). Moreover, supercritical foaming is an auspicious strategy to process scaffolds with virtually 100 % yield of incorporation of bioactive agents with retained activity (Diaz-Gomez et al. 2016). The mixing of the polymer with bioactive agents is also promoted during the CO2 soaking period of the foaming process due to the rubbery state of the polymer under these conditions (de Matos et al. 2013, Fanovich et al. 2013). However, the supercritical foaming process usually results in scaffolds with high macroporosity but low pore interconnectivity that hinders the access of cells and the transport of nutrients and metabolic wastes throughout the graft (García-González et al. 2015, Goimil et al. 2017). The preparation of bioactive agentloaded scaffolds, by using the supercritical foaming, combined with the incorporation of aerogel microparticles, obtained by supercritical drying, is proposed in the present work to obtain 3D-scaffolds with advanced architectures and performance. In the regenerative medicine context, silk fibroin in the form of aerogel microparticles, i.e. solid nanostructured materials that preserve the mesoporous structure of the wet gel precursor (Maleki et al. 2016), and obtained by supercritical drying of gels might be an advantageous admixture for the processing of scaffolds, which has not been explored so far. The presence of other aerogels in synthetic scaffolds processed by supercritical foaming has been shown to increase the pore sizes and their interconnectivity resulting in enhanced cell infiltration capacity (Goimil et al. 2017). Silk aerogels may allow combining the intrinsic bioactive properties of silk fibroin and the 5. scCO2-foamed silk fibroin aerogel/PCL scaffolds containing dexamethasone for bone regeneration _____________________________________________________________ 113 textural properties of aerogels. Silk fibroin aerogel monoliths, prepared by a combination of a sol-gel mechanism followed by supercritical drying (Mallepally et al. 2014, Marin et al. 2014, Mallepally et al. 2015), have shown textural properties dependent on the fibroin molecular weight and the concentration in the aqueous precursor solution (Mallepally et al. 2015). These previously obtained silk fibroin aerogel monoliths were cytocompatible with human foreskin fibroblasts but lacked the macroporosity needed to favor the cell penetration and growth. On the other hand, the supercritical processing of mixtures of polyesters and silk fibroin (as raw material, not as aerogels) have been shown to be feasible at various conditions (140-200 bar, 37-155 ºC) (Kang et al. 2009, Diaz-Gomez et al. 2017). The innovation in the preparation of silk aerogel in the form of particles and their use in the processing of synthetic macroporous scaffolds by supercritical foaming is a feasible solvent-free processing approach with an advanced performance. The particular behavior of these scaffolds stems on the synergistic combination of two CO2based technologies that overcomes the individual processing limitations of poor pore interconnectivity (for supercritical foaming) and of lack of macroporosity (for silk aerogel scaffolds) of the scaffolds obtained for both processes separately. Moreover, the integration of the use of CO2 to promote silk fibroin gelation (Mallepally et al. 2014) and to facilitate solid state dispersion (de Matos et al. 2013) of the scaffold admixtures was also evaluated. Thus, in the present work, scaffold formulations containing PCL, silk fibroin aerogel microparticles and dexamethasone were prepared using the supercritical foaming process to contrast two main hypotheses: (i) silk fibroin aerogel microparticles may improve the porous structure of the scaffold facilitating cell infiltration and biological fluid transport; and (ii) the incorporation of dexamethasone as base or as a salt may lead to different drug release profiles and thus to different bone regeneration outcomes. First, SF aerogels were herein developed in the form of submicron-sized particles using supercritical drying. Dexamethasone was added in two different forms: dexamethasone base (DX) and dexamethasone-21-sodium phosphate (DS), which present distinct solubility in aqueous media. LETICIA GOIMIL GARCÍA _____________________________________________________________ 114 The scaffolds were prepared with various combinations of SF aerogels and DX and DS, and then evaluated in terms of morphology, physicochemical properties, in silico capacity for cell infiltration, and tuning of the release profile and subsequent promotion of cell osteodifferentiation. The bone regenerative performance of the scaffolds was assessed in in vivo tests using a murine model of critical-size calvarial defect (diameter: 8 mm). 5.2. MATERIALS AND METHODS 5.2.1. Materials Poly (-caprolactone) (PCLraw; 50 kDa) was purchased from Polysciences (Warrington, PA, USA) in the powdered form. Dexamethasone (DX; 99.9 % purity) and dexamethasone-21phosphate (DS; 99.95 % purity) were from Fagron (Nazareth, Belgium) and Guinama (Valencia, Spain), respectively. Carbon dioxide (99.9 % purity) was supplied by Praxair Inc. (Madrid, Spain). Span 80 (HLB 4.3) was purchased from Sigma-Aldrich (Saint Louis, MO, USA). Paraffin oil and absolute ethanol (Emparta®, 99.5 % purity) were from Panreac (Castellar del Vallès, Spain) and Merck (Darmstadt, Germany), respectively. 5.2.2. Preparation of silk fibroin solution For the preparation of silk fibroin, cocoons of Bombyx mori, obtained from silkworms reared in the sericulture facilities of the IMIDA (Murcia, Spain), were divided into 4 pieces and boiled in 0.02 M Na2CO3 for 30 min to remove the glue-like sericin proteins. Then, raw fibroin was rinsed thoroughly with water and dried at room temperature for 3 days. The extracted fibroin was dissolved in 9.3 M LiBr (Acros Organics, Spain) for 3 h at 60 °C to generate a 20 % w/v solution that was dialyzed against distilled water for 3 days (Snakeskin Dialysis Tubing 3.5 kDa MWCO, Thermo Scientific, USA) with 8 water exchanges. The resulting 8% w/v fibroin solution was recovered and filtered. 5. scCO2-foamed silk fibroin aerogel/PCL scaffolds containing dexamethasone for bone regeneration _____________________________________________________________ 115 5.2.3. Preparation of silk aerogel microparticles (SA) A water-in-oil emulsion was prepared from a mixture of paraffin oil and an aqueous SF solution (8 wt.%) in a 4:1 oil-to-water weight ratio. Span 80 was used as surfactant (3 and 6 wt.% with respect to the aqueous phase for emulsions #1 and #4, respectively) and added to the oil phase. The liquid-liquid mixture was homogenized using an ultrasound probe (450D, Branson Digital Sonifier, Nuevo Laredo, Mexico) with an amplitude of 50 % and for 2 min. The resulting emulsion was then magnetically stirred at 600 rpm for 5 h. As an alternative to the direct stirring of the emulsion, other gelation methods for SF were studied: portions of the 3 wt.% Span 80 emulsion were put either in contact with scCO2 (80 bar, 35 ºC) bubbling (1 g/min) for 30 min (emulsion #2), or ethanol (1:10 ethanolto-water content) was added prior to the stirring (emulsion #3). After SF gelation and ageing for 5 h, the dispersed phase of the emulsion turn a gel and a dispersion of gel microparticles in the oil phase was obtained. These oil dispersions (dispersions #1 to #4) were then centrifuged (5804 R, Eppendorf, Hamburg, Germany) at 5,000 rpm for 5 periods of 15 min with the addition of ethanol after each centrifugation period to remove the oil phase and the water of the hydrogel particles to obtain alcogels. For dispersion #4, a fraction (denoted as #4a) was firstly washed with diethyl ether before the ethanol washing steps to ensure a complete removal of the emulsifier. SF aerogel microspheres (aerogels #1 to #4) were obtained by supercritical drying of the alcogels. Briefly, SF alcogels soaked in ethanol were dried under a scCO2 (40 ºC, 120 bar) flow of 6 g/min during 3.5 h in a 100 mL-autoclave (Thar Technologies, Pittsburgh, PA, USA). A depressurization time of 70 min at a constant mass flow rate of 1 g CO2/min was used until atmospheric pressure was reached. SF powders were collected from the autoclave for further use and analysis. 5.2.4. Preparation of scaffolds by supercritical foaming The scaffold components in the form of dry powders were weighed in the proportions indicated in Table 5.1, physically mixed using a spatula and placed in Teflon cylindrical molds (internal LETICIA GOIMIL GARCÍA _____________________________________________________________ 116 diameter: 17.0 mm; height: 24.6 mm) (Brand GmbH, Wertheim, Germany). The mixtures were manually compacted using a 17.0 mmdiameter flat aluminum plunger and then placed in a 100 mL-highpressure stainless steel autoclave (Thar Technologies, Pittsburgh, PA, USA). The supercritical foaming was carried out following a pressurization-soaking-single depressurization stepwise protocol. The autoclave at 37 ºC was pressurized at 5 g/min of CO2 up to 140 bar (i.e. supercritical conditions) and left under these conditions in the static mode for 1 hour (soaking period). Then, the autoclave was depressurized at a venting rate of 1.8 g/min until atmospheric pressure was reached. The scaffolds were stored overnight at room conditions to allow complete CO2 desorption, before being weighed and the dimensions measured. The initial and final mass of the scaffolds were compared as a gravimetric method to determine the dexamethasone incorporation yields. The non-porous external layer of ca. 10 µm in the scaffolds, which is formed due to the faster diffusion of the CO2 from the surface of the polymer matrix, was removed with a scalpel prior to further analysis. Table 5.1: Notation and composition of the scaffolds (in weight percentage) containing poly(-caprolactone) (PCL), silk fibroin aerogel #4 (SA) and dexamethasone in the base (DX) and 21-phosphate (DS) forms, and processed by supercritical foaming (37 ºC, 140 bar). Scaffold PCL, wt.% SA, wt.% DX, wt.% DS, wt.% PCL 100 - - - PCL-SA 90 10 - - PCL-SA-DX 89.9 10 0.1 - PCL-SA-DS 89.9 10 - 0.1 PCL-SA-DXDS 89.9 10 0.05 0.05 PCL-SA-5DX 85 10 5 - PCL-SA-5DS 85 10 - 5 PCL-SA-5DXDS 85 10 2.5 2.5 5. scCO2-foamed silk fibroin aerogel/PCL scaffolds containing dexamethasone for bone regeneration _____________________________________________________________ 117 5.2.5. Structural and physicochemical characterization of the aerogel particles and the scaffolds Textural properties of the silk aerogel particles were studied by low-temperature N2 adsorption–desorption analysis (ASAP 2000; Micromeritics Inc.; Norcross, GA, USA). The aerogel powder was previously degassed at 70 ºC and under vacuum (<1 mPa) for 20 h. BET (Brunauer–Emmett–Teller) method was used to estimate the specific surface area (ABET), whereas BJH (Barrett–Joyner–Halenda) method was chosen to calculate the mean pore diameter (dp,BJH) and the specific pore volume (Vp,BJH). Attenuated Total Reflectance-Fourier Transformed Infrared Spectroscopy (ATR-FTIR) was used to analyze the structural conformation of SF after processing the aerogel particles. Spectra were acquired on a Nicolet iS5 spectrometer, equipped with an iD5 ATR accessory (Thermo Scientific, Madison, WI, USA) controlled with OMNIC v9.3.30 software. Dried aerogels were placed directly onto the diamond window (ca. 2 mg) without further manipulation. Samples of silk fibroin nanoparticles (Lozano ‐ Pérez et al. 2015) and a lyophilized film of regenerated SF (after dissolution in LiBr 9.3M and further dialysis) (Aznar-Cervantes et al. 2013) were measured as references for the β-sheet structure and for a predominantly random coil and α-helix conformation of SF, respectively. Measurements were made in absorbance mode, in the 550-1800 cm−1 spectral range using 64 scans and at a resolution of 4 cm−1. Spectra were focused on the 1800-800 cm-1 as the most significant range for SF conformation and N-B strong apodization and mertz phase correction were applied. The particle size distribution and mean hydrodynamic diameter (Z-average) of the microgel precursor of the optimized silk fibroin aerogel were measured using a Zetasizer Nano ZSP instrument (Malvern Instruments Ltd., Worcestershire, UK) by dynamic light scattering. After high-power ultrasound treatment of the samples (3 min, 30% amplitude) using a Sonifier Branson 450D (Emmerson Ultrasonic Corporation, Danbury, CT, USA), all measurements were performed in purified water at 25 ºC, at 173º angle relative to the source and with a gel concentration of 0.5 mg/mL. Values were LETICIA GOIMIL GARCÍA _____________________________________________________________ 118 calculated from the measurements performed in triplicate (12 runs per measurement). Images of the outer morphology of the SF-aerogels and cross sections of the scaffolds were obtained by scanning electron microscopy (SEM, EVO LS15, Zeiss, Oberkochen, Germany). Aerogels were sputtered with a 10 nm-iridium layer prior to imaging with the aim of improving the contrast. The scaffolds were measured and weighed to calculate their bulk density (ρbulk), and then their skeletal density (ρskel) was measured from five replicates by helium pycnometry (Quantachrome; Boynton Beach, FL, USA), operating at 25 ºC and 1.03 bar. The overall porosity (  ) of the scaffolds was calculated as follows: ( ⁄ ) (5.1) Mercury-intrusion porosimetry (MIP) analyses were used to determine the pore size distribution, open porosity (MIP) and MIPpore median of the scaffolds using a Micromeritics 9305 pore sizer (Norcross, GA, USA). MIP was operated with a 3 mL penetrometer for solids and using working pressures ranging from 0.07 to 1724 bar. MIP-cumulative curves were used to generate a 3D network model with similar percolation properties as those of the macrostructure of the scaffold using PoreXpert v.1.6.567 (PoreXpert Ltd, Plymouth, UK) software (García-González et al. 2018). The estimation of the pore interconnectivity of the scaffolds from the generated 3D-models was performed using a Boltzmann-annealed simplex algorithm where values are proportional to the number of interconnections (i.e., throats) between pores. The infiltration capacity of mesenchymal stem cells (MSCs) in the scaffolds was simulated using the filtration module from PoreXpert software and assuming an average MSCs size of 26.5±5.0 µm (Salerno et al. 2017). Water permeability (kw at 25ºC, 1.03 bar) was modeled assuming a Poiseuille flow (water) in the zdirection as follows: ( ) (5.2) being ωcell(Farcs) an averaging operator over the whole unit cell operating on the flow capacities of the pore throat-pore arcs parallels to the z-axis and calculated using the Dinic network analysis 5. scCO2-foamed silk fibroin aerogel/PCL scaffolds containing dexamethasone for bone regeneration _____________________________________________________________ 119 algorithm; and lcell and Acell the length and the cross-sectional area of the unit cell, respectively. Thermal properties of the scaffolds were evaluated by differential scanning calorimetry (DSC-Q100, TA Instruments; New Castle, DE, USA) performing two heating cycles under a nitrogen atmosphere with temperatures ranging from room temperature to 300 ºC and a cooling cycle down to -10 ºC in-between, using heating and cooling rates of 10 ºC/min. Crystallinity degree of PCL contained in the scaffolds (expressed in percentage) was calculated for each heating cycle using Eq. the following equation: (5.3) where x stands for the fraction (in weight) of PCL in the sample; Hm, the melting enthalpy of the sample in the heating cycle in J/g; and H0 m,PCL, the melting enthalpy of 100 % crystalline PCL (142 J/g) (Tsuji and Ikada 1998). 5.2.6. Dexamethasone release tests Scaffold pieces (250 mg) of PCL-SA-DX, PCL-SA-DS and PCLSA-DXDS were immersed in 25 mL of phosphate-buffered solution (PBS) pH 7.4 medium. The flasks were placed in an oscillating shaker (Unitronic 320 OR, JP Selecta, Barcelona, Spain) at 37 ºC and 60 rpm for 3 weeks. Aliquots of 3 mL were sampled through that period at selected times, and the withdrawn volumes were replaced with fresh medium. Concentrations of dexamethasone base were measured by HPLC (Waters, Milford, MA, USA) consisting on a 717 plus Autosampler, a 600 Controller, and a 996 Photodiode Array Detector, equipped with a C18 column (Symmetry© C18, particle size 5 m, 3.9 mm diameter x 150 mm length) from Waters (Milford, MA, USA), at 30 ºC and  = 242 nm. Samples were filtered prior to analysis using 0.2 µm nylon filters to avoid potential interferences due to the presence of particles from the scaffolds. Calibration curve was obtained from triplicate dilution series of DX in PBS pH 7.4 in the 0.00005 to 0.0125 mg/mL concentration range (R2 > 0.999). Concentrations of DS and total dexamethasone were measured by UV–Vis spectrophotometry (8453, Agilent, Santa Clara, CA, USA) at LETICIA GOIMIL GARCÍA _____________________________________________________________ 120 λ = 242 nm. Calibration curves were obtained from triplicate dilution series in PBS pH 7.4 in the 0.00125 to 0.01 mg/mL concentrations range (R2 > 0.99). The Higuchi model (Eq. 5.4) (5.4) and a modified Higuchi model with an initial burst stage (Eq. 5.5) (5.5) were fitted to the drug release profiles, D being the amount of dexamethasone released (in percentage with respect to the loaded amount) at time t, A the amount of dexamethasone released (in percentage) during the burst release period, and B the release rate kinetic coefficient (in dose % h-0.5). The Akaike criterion was applied to select the best fitting model (Akaike 1974). Kinetics fitting was performed using GraphPad Prism version 6.04 for WindowsGraphPad Software (La Jolla, CA, USA). 5.2.7. In vivo studies Male Sprague-Dawley rats, with weights ranging from 250 to 300 g, were used for the in vivo studies. Experiments were previously approved by the local committee for animal studies of the University of La Laguna (Spain) and were carried out in conformity with the European regulation on care and use of animals in experimental procedures. Surgery was performed under aseptic conditions using isoflurane for general anesthesia. A standardized, critical, circular defect (8 mm diameter) was created on the cranium of the rats to test the bone repair capacity of the scaffolds in a critical defect (Rodríguez-Évora et al. 2013). The scaffolds were cut using a hollow punch with an internal diameter of 8 mm, and the resulting cylinders were sliced (1.5 mm height) to create the implants for in vivo tests and sterilized by UVradiation for 1 h. The bone retrieved was replaced with the implants, the skin was repositioned over the area and stapled. Ketoprofen (3 mg/kg) and buprenorphine (0.05 mg/kg) were subcutaneously administered to lessen postsurgical pain. Upon recovery from the surgery, animals were allowed free activity, food and water uptake. 5. scCO2-foamed silk fibroin aerogel/PCL scaffolds containing dexamethasone for bone regeneration _____________________________________________________________ 121 Five scaffold formulations (PCL, PCL-SA, PCL-SA-DX, PCLSA-DS and PCL-SA-DXDS) were assessed regarding their bone inductive/regenerative effect, employing 5 groups of 6 rats each that were examined at 7 and 14 weeks after implantation. An additional group of 6 rats with an empty defect was used as surgical control and examined at the same time periods. After each period, 3 rats per group were sacrificed by CO2 inhalation. The implants were retrieved, fixed (10 % formalin solution) and decalcified in Histofix® decalcifier (Panreac, Spain) and prepared for histological analysis (Hernández et al. 2012). New bone formation and mineralization were identified by hematoxylin-erythrosin and VOF trichrome staining. In VOF staining sections, advanced mineralization areas appear colored in red, while less mineralized areas result in a blue coloration (Diaz-Gomez et al. 2017). All sections per specimen were histomorphometrically evaluated by light microscopy (LEICA DM 4000B) followed by image analysis using the Leica QWin v3Pro software. The area of the tissue within the defect was defined as the region of interest (ROI) to quantify the new bone formation. The original defect area within the ROI was circular (50 mm2 of area) and with the center being coincident with that of the defect site. New bone formation was accordingly calculated by using Eq. (5.6) and expressed as percentage of repair. (5.6) The number of ossification foci within the defect area was determined by counting the number of individual regions exhibiting bone extracellular matrix. Radiography images of the extracted crania were performed with a Philips Optimus X-ray equipment, applying the parameters established in the protocol for the thumb of the human hand (5.2 kV, 1.5 mA/s, and 9.7 mSv and 1 m of distance). 5.2.8. Statistical analysis An analysis of variance (one-way ANOVA, Tukey’s multiple comparison) using a level of significance (p) of 0.05 was used to compare the data of the in vivo tests from different experimental