Engineering a metastasis-on-a-chip system towards studying cell invasion and drug efficacy in lung cancer
Abstract
O cancro é uma das patologias líder em morbilidade e mortalidade, sendo o cancro do pulmão o tipo de cancro com mais incidência e mortalidade, mundialmente. A causa subjacente a esta mortalidade e morbilidade dos pacientes é a progressão cancerígena ou as metástases. De modo a alcançar diagnósticos clínicos precoces e precisos, é necessário desenvolver estratégias de deteção de biomarcadores cancerígenos bem como terapias direcionadas. Todavia, grande parte das terapias anticancerígenas desenvolvidas tendem a falhar aquando dos ensaios clínicos devido à falta de modelos in vitro replicativos do cenário in vivo, devido à carência de complexidade, de estímulos, relevância e equivalência fisiológica. Assim, tornam-se inadequados na previsão da eficácia terapêutica de compostos anticancerígenos. Portanto, é necessário desenvolver modelos biomimético tridimensionais (3D) das metástases de modo que repliquem verdadeiramente o microambiente tumoral bem como a vasculatura. Deste modo, microfluídica foi selecionada como a tecnologia ideal para atingir este fim visto que ultrapassa as limitações dos modelos tradicionais estáticos e possui características interessantes como a miniaturização, baixo-custo e controlabilidade sobre os estímulos impostos, num microdispositivo. O objetivo da presente tese foi desenvolver, otimizar e validar um dispositivo de microfluídica projetado para mimetizar a arquitetura tumoral bem como a vasculatura, um lung metastasis-on-a-chip. O sistema desenvolvido consiste em dois microcanais, um com células de cancro do pulmão dispostas tridimensionalmente, e outro canal adjacente endotelial de modo a mimetizar o vaso sanguíneo. A validação do dispositivo final foi feita através de estudos de atividade metabólica, de proliferação celular e de monitorização do sistema bem como a integridade da barreira endotelial, ao longo do tempo. Posteriormente, nanossistemas lipídicos foram introduzidos na plataforma e a sua capacidade de penetrar a cultura tumoral foi avaliada. Intravasamento e agressividade cancerígena também foi avaliada através de estudos fenotípicos e de monitorização da mobilidade celular. É esperado que o resultado final da presente tese assista no desenvolvimento terapêutico de nanoformulações que interrompam o processo metastático no cancro do pulmão. Os resultados também irão providenciar uma plataforma de testagem robusta de eficácia terapêutica e de estudos relevantes sobre o processo metastático e dos seus mecanismos de ação.
Full text
Ana Sofia Oliveira Martins Engineering a metastasis-on-a-chip system towards studying cell invasion and drug efficacy in lung cancer December 2021 UMinho | 2021 Ana Martins Engineering a metastasis-on-a-chip system towards studying cell invasion and drug efficacy in lung cancer Universidade do Minho Escola de Ciências
Ana Sofia Oliveira Martins Engineering a metastasis-on-a-chip system towards studying cell invasion and drug efficacy in lung cancer Master Thesis Master in Biophysics and Bionanosystems Work developed under the supervision of Dr. Lorena Diéguez and Prof. Dr. Ana Arminda Lopes Preto Almeida Universidade do Minho Escola de Ciências December 2021
II DIREITOS DE AUTOR E CONDIÇÕES DE UTILIZAÇÃO DO TRABALHO POR TERCEIROS Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Atribuição-NãoComercial-SemDerivações CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/
III Acknowledgements A presente dissertação representa um enorme marco a nível profissional e pessoal. Foi uma das experiências que mais me realizou sendo uma confirmação de que rumo profissional que optei é o mais indicado para mim. À minha orientadora, Professora Doutora Ana Arminda Lopes Preto Almeida, quero expressar a minha gratidão pela disponibilidade e por me ter aceitado como aluna. To my supervisor, Dr Lorena Diéguez, I am very grateful for the opportunity that she has granted me. She received me in this project with open arms and gave me every tool I needed to succeed. Her support, trust, and encouragement made my experience a lot better. Thank you! To the MESTASTARG group, Dr Catarina Gonçalves, Dr Miguel Xavier, Dr Alar Ainla and Dr Sara AbaldeCela I can only say that without any doubt I chose the right project to be in. I could always count on each of you and all of you were a huge part of my professional growth in the past year, and for that I am thankful. To the Medical Devices group, I want to express my gratitude, absolutely the best group in INL. A special thanks to Micaela Oliveira, Cláudia Lopes and María Cascallar for encouraging me every step of the way and helping me when I needed it the most. A todos os amigos que percorreram este caminho ao meu lado, mas com especial agradecimento ao Alexandre Mendonça, à Diana Sousa, ao Jorge Cunha, ao Fábio Lopes e à Leonor Pinto. Obrigada pelo constante apoio, amizade e gargalhadas. Por fim, à minha família por ter feito todos os possíveis e impossíveis por mim, os sacrifícios que fizeram e as vezes que me colocavam como prioridade para que eu tenha uma vida melhor. Mãe, Pai, Marco, Vitor, Lúcia, Sara, Bayle e Mia. Esta tese é dedicada a vocês, obrigado. “However bad life may seem, There is always something you can do and succeed at. While there’s life, there is hope” (Stephen W. Hawking)
IV STATEMENT OF INTEGRITY I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho.
V Desenvolvimento de um sistema metastasis-on-a-chip para a análise da invasão celular e eficácia farmacológica no cancro do pulmão Resumo O cancro é uma das patologias líder em morbilidade e mortalidade, sendo o cancro do pulmão o tipo de cancro com mais incidência e mortalidade, mundialmente. A causa subjacente a esta mortalidade e morbilidade dos pacientes é a progressão cancerígena ou as metástases. De modo a alcançar diagnósticos clínicos precoces e precisos, é necessário desenvolver estratégias de deteção de biomarcadores cancerígenos bem como terapias direcionadas. Todavia, grande parte das terapias anticancerígenas desenvolvidas tendem a falhar aquando dos ensaios clínicos devido à falta de modelos in vitro replicativos do cenário in vivo , devido à carência de complexidade, de estímulos, relevância e equivalência fisiológica. Assim, tornam-se inadequados na previsão da eficácia terapêutica de compostos anticancerígenos. Portanto, é necessário desenvolver modelos biomimético tridimensionais (3D) das metástases de modo que repliquem verdadeiramente o microambiente tumoral bem como a vasculatura. Deste modo, microfluídica foi selecionada como a tecnologia ideal para atingir este fim visto que ultrapassa as limitações dos modelos tradicionais estáticos e possui características interessantes como a miniaturização, baixo-custo e controlabilidade sobre os estímulos impostos, num microdispositivo. O objetivo da presente tese foi desenvolver, otimizar e validar um dispositivo de microfluídica projetado para mimetizar a arquitetura tumoral bem como a vasculatura, um lung metastasis-on-a-chip . O sistema desenvolvido consiste em dois microcanais, um com células de cancro do pulmão dispostas tridimensionalmente, e outro canal adjacente endotelial de modo a mimetizar o vaso sanguíneo. A validação do dispositivo final foi feita através de estudos de atividade metabólica, de proliferação celular e de monitorização do sistema bem como a integridade da barreira endotelial, ao longo do tempo. Posteriormente, nanossistemas lipídicos foram introduzidos na plataforma e a sua capacidade de penetrar a cultura tumoral foi avaliada. Intravasamento e agressividade cancerígena também foi avaliada através de estudos fenotípicos e de monitorização da mobilidade celular. É esperado que o resultado final da presente tese assista no desenvolvimento terapêutico de nanoformulações que interrompam o processo metastático no cancro do pulmão. Os resultados também irão providenciar uma plataforma de testagem robusta de eficácia terapêutica e de estudos relevantes sobre o processo metastático e dos seus mecanismos de ação. Palavras-chave: Cancro do pulmão; Metástases, Microfluídica; Organ-on-a-chip; Nanotecnologia
VI Engineering a metastasis-on-a-chip system towards studying cell invasion and drug efficacy in lung cancer Abstract Cancer is one of the leading pathologies in morbidity and mortality worldwide, being lung cancer the type with highest incidence and mortality. The underlying cause of mortality and morbidity in patients is cancer progression or metastasis. To achieve earlier and more accurate diagnosis in the clinic and to make better therapeutic decisions, it is necessary to engineer new tools for biomarker detection as well as to develop new targeted treatments. However, most newly developed anticancer therapies fail during clinical trials due to the lack of in vitro models that faithfully replicate the in vivo scenario, given their lack of complexity, stimuli, relevance, and physiologic equivalence. Hence, these in vitro models become inadequate to predict the therapeutic efficacy of anticancer compounds. As such, it is necessary to develop biomimetic three-dimensional (3D) models of metastasis that truthfully replicate the tumoral microenvironment as well as its vasculature. Thus, microfluidics was selected as the ideal technology to accomplish this end, given that it overcomes the limitations of traditional static models and has interesting characteristics including miniaturization, cost-effectiveness, and stimuli control, all in one microdevice. The main aim of this thesis was to develop, optimize and validate a microfluidic device to mimic the tumoral architecture and its vasculature in lung cancer, a lung metastasis-on-a-chip. The system developed consists of two microchannels, one displaying a 3D lung cancer cell culture, and an adjacent endothelial channel mimicking a blood vessel. The validation of the device was accomplished through metabolic activity studies, proliferation assessments, and monitoring the platform and endothelial barrier integrity over time. Afterwards, lipidic nanosystems were introduced in the platform and their capacity to penetrate into the tumour culture was assessed. Intravasation and cancer cell aggressiveness was also evaluated through phenotypic studies as well as cell motility monitorization. It is expected that the final result from this thesis assists in the development of therapeutic nanoformulations tailored to interrupt the metastatic process in lung cancer. The results will also provide a robust testing platform for the assessment of therapeutic efficacy, and to perform relevant studies to better understand the metastatic process and its mechanisms of action. Key-words: Lung cancer; Metastasis; Microfluidics; Organ-on-a-chip; Nanotechnology
VII Table of Contents Acknowledgements ....................................................................................................................... III Resumo ............................................................................................................................................ V Abstract .......................................................................................................................................... VI Table of Contents ..........................................................................................................................VII List of abbreviations ..................................................................................................................... IX List of Figures ................................................................................................................................ XI List of Supplementary Figures .................................................................................................. XVII List of Tables .............................................................................................................................. XVIII 1. Introduction ......................................................................................................................... 2 1.1. Cancer and lung cancer ............................................................................................................ 2 1.1.1. Non-small cell lung cancer (NSCLC) and metastatic NSCLC ....................................... 3 1.1.2. Circulating Tumour Cells (CTCs) ................................................................................ 5 1.2. Standard testing and screening models .................................................................................... 9 1.2.1. Traditional 2D cell culture in cancer modelling ........................................................... 9 1.2.2. Animals in cancer modelling ..................................................................................... 10 1.2.3. Limitations of current testing models ........................................................................ 11 1.2.4. 3D models and their improvements over 2D models .................................................. 12 1.3. Organ-on-a-chip (OOC).............................................................................................................. 13 1.3.1. Microfluidic technology ............................................................................................ 13 1.3.2. Microfluidics for cell culture.................................................................................... 16 1.3.3. Potential of OOC models of cancer ........................................................................... 20 2. Objectives .......................................................................................................................... 22 3. Techniques ........................................................................................................................ 24 3.1. Microfabrication techniques .................................................................................................... 24 3.1.1. Photolithography ...................................................................................................... 24 3.1.2. Soft lithography ........................................................................................................ 25 3.2. Microscopy ................................................................................................................................ 26 3.2.1. Fluorescence microscopy ......................................................................................... 26 3.2.2. Confocal laser scanning microscopy ......................................................................... 27 3.3. Cell culture techniques ............................................................................................................ 28 3.4. Immunocytochemistry .............................................................................................................. 28 4. Materials and Methods .................................................................................................... 31 4.1. Design of microfluidic devices ................................................................................................. 31 4.1.1. Design of the first generation ................................................................................... 31 4.2. Fabrication of microfluidic devices ......................................................................................... 33
XIV Figure 27-Human primary microvascular endothelial cells (HPMEC) metabolic activity (%) after 24h of culture. (A) Cell metabolic activity grown in tissue culture plates (TCPs) and in modified PDMS with O2 Plasma Treatment for 30 min (A) and 24h (B) prior to ECM thin coating with Collagen IV ( 30 µg.mL-1), Fibronectin ( 30 µg.mL-1) or Collagen IV/Fibronectin (1:1) mixture ( 30 µg.mL-1). The cell metabolic activity was determined by the resazurin reduction assay measuring the fluorescence of resorufin (λex= 560 nm, λem= 590 nm). Values show mean and standard deviation of four independent experiments: ****p < 0.0001. P-values were obtained using a one-way analysis of variance (ANOVA). 57 Figure 28Lung adenocarcinoma cell line A549 metabolic activity (%) after 2 days of culture. (A) Cell metabolic activity grown in tissue culture plates (TCPs), in the modified PDMS with 2% APTMS (A) or Sulfo-SANPAHTM (B) and ECM embedment of A549 cell suspension with Collagen IV (3 mg. mL-1), Matrigel (4 mg.mL-1) or Fibronectin (1 mg.mL-1). The cell metabolic activity was determined by the resazurin reduction assay measuring the fluorescence intensities of resorufin (λex= 560 nm, λem= 590 nm). Values show mean and standard deviation of four independent experiments. **p < 0.001; ****p < 0.0001. Pvalues were obtained using a one-way analysis of variance (ANOVA). ................................................. 59 Figure 29Lung adenocarcinoma cell line A549 metabolic activity (%) after 24h of culture. (A) Cell metabolic activity grown in tissue culture plates (TCPs) and in modified PDMS with O2 Plasma Treatment for 30 min (A) and 24h (B) prior to ECM embedment with A549 cell suspension functionalization with Collagen IV (3 mg. mL-1), Matrigel (4 mg.mL-1) or Fibronectin ( 1 mg.mL-1). The cell metabolic activity was determined by the resazurin reduction assay measuring the fluorescence of resorufin (λex= 560 nm, λem= 590 nm). Values show mean and standard deviation of four independent experiments. **p < 0.01***p < 0.001;****p < 0.0001. P-values were obtained using a one-way analysis of variance (ANOVA). ........................................................................................................................ 60 Figure 30Live-on-screen images of cell culture in the microfluidic device. (A) Day of cancer cell culture embedded in Matrigel (1:1) with cells still rounded. (B) One day after cell culture in chip demonstrating major cell loss and morphological changes in cancer cells. ......................................... 61 Figure 31Phase contrast images 24h after A549 lung cancer cells were seeded in the microfluidic chip displaying good adaptation and morphology to the environment provided. (A) (4x) (B) (10x). ............................................................................................................................................... 61 Figure 32Confocal micrographs 10x magnification of A549 cancer cells and HPMEC cells grown on the microfluidic device and stained for nuclei (DAPI, blue) and cytoplasm (GFP, green). Images were acquired after 5 days with continuous flow (15 µL∙h-1.) in the metastasis-on-a-chip with automated perfusion. ......................................................................................................................................... 62
XV Figure 33Confocal micrographs (10x magnification) of A549 cancer cells embedded in Matrigel ( 4 mg.mL-1) and grown on the microfluidic device. Cells were stained for nuclei (DAPI, blue) and FITClabelled antibody against cytoplasmatic CK (Green) expression. Images were acquired after 2, 5 and 7 days after cell culture under a semi-static environment. ..................................................................... 64 Figure 34Confocal images and orthogonal cross-section views of A549 cells grown on the microfluidic metastasis-on-a-chip after 2 days of culture, stained for CK expression (green) and counterstained nuclei (DAPI, blue). ................................................................................................... 65 Figure 35Immunocytochemistry in device of HPMEC cells growth, proliferation, and degradation on the Sulfo-SANPAH modified, fibronectin functionalized microfluidic device. Cells were stained for nuclei (DAPI, blue) and actin filaments (phalloidin, red).Images were acquired after 2, 5 and 7 days under periodic flow rate of 40 µL.min -1 by laser confocal microscopy with 10x magnification. ....................... 66 Figure 36Confocal micrograph and correspondent orthogonal cross-section views of HPMEC cells grown on the modified and functionalized microfluidic metastasis-on-a-chip after 2 days of culture. Orthogonal views are set in the micropillar area (endothelial barrier site). Cells are stained for actin filaments (phalloidin, red) and counterstained nuclei (DAPI, blue). ..................................................... 67 Figure 37Laser confocal microscopy images and correspondent orthogonal cross-section views of micropillar region (endothelial barrier site) of HPMEC cells grown on the Sulfo-SANPAH modified, Fibronectin functionalized microfluidic metastasis-on-a-chip after 5 days in culture under periodic flow rate ( 40 µL.mL-1). Cells are marked for actin filaments (phalloidin, red) and for nuclei (DAPI, blue). ... 68 Figure 38Immunocytochemistry confocal image and micropillar orthogonal cross-section views of HPMEC cells grown on microfluidic metastasis-on-a-chip after 7 days in culture and periodic flow rate ( 40 µL.mL-1). Cells are stained for actin filaments (phalloidin, red) and counterstained for nuclei (DAPI, blue). ............................................................................................................................................... 69 Figure 39Estimated number of the tumour channel and the endothelial channel over time. Values were obtained by nuclei counting in Image J software. Values show mean and standard deviation of five independent experiments. ....................................................................................................... 70 Figure 40Metabolic activity (%) of HPMEC cells and A549 cells on-chip. Measurements were acquired at days 2, 5, and 7-days post-flow. The cell metabolic activity was determined by the resazurin reduction assay measuring the fluorescence of resorufin (λex= 560 nm, λem= 590 nm). Values show mean and standard deviation of five independent experiments. ................................................................... 71 Figure 41Endothelial Barrier integrity over time. Fluorescence intensity profiles obtained from confocal micrographs after following 30 µM Fluorescein on seeded and unseeded microfluidic devices.
XVI Measurements were performed 2-, 5and 7-days post-flow. These line profiles were acquired after 30 min of flow across the endothelial barrier. ......................................................................................... 72 Figure 42Phase contrast images of HPMEC and H1650 cancer cells after 5 days under flow conditions displaying a robust chip structure to perform nanoemulsions assay. (A) 4X magnification; (B) 10X magnification............................................................................................................................. 73 Figure 43Sphingomyelin and Vitamin E nanoemulsions incubated 6 hours at a humified environment in the metastasis-on-a-chip system through the endothelial channel. Cells of both channels (H1650 and HPMEC) were stained for nuclei with DAPI (Blue). Nanoemulsions are labeled with Cy5 (Red). ............................................................................................................................................... 74 Figure 44Immunocytochemistry confocal images of intravasation events (red arrows) on the microfluidic metastasis-on-a-chip platform after 6 days in culture and periodic flow rate ( 40 µL.mL-1. HPMEC cells are stained for DAPI (nuclei, blue). Cancer A549 cells are stain for CK expression ( cytokeratin, green) and counterstained with DAPI. ............................................................................. 75 Figure 45Immunocytochemistry confocal image and r orthogonal cross-section views of A549 cell cluster cells grown on endothelial microchannel after 6 days in culture and periodic flow rate ( 40 µL.mL-1). Cells are stained for CK expression and DAPI. .................................................................. 76
XVII List of Supplementary Figures Figure A 1Confocal image ( magnification 10x) of the microfluidic device with cultured cell in both channels with visible system disruption due to bubble formation. Nucleus stained for DAPI (blue). ........................................................................................................................................................ 96 Figure A 2Confocal micrographs of the disrupted system in the automated perfusion microfluidic device. (A) Brightfield; (B) Nuclei stained with DAPI ( blue); (C) GFP labeled lung cancer cells (green); (D) Merged............................................................................................................................................. 96 Figure A 3Confocal images (10X magnification) of the full tumour channel (A549 cells) over time. Immunocytochemistry in chip, with cells expressing cytokeratin (Green) and nuclei (DAPI, blue).97 Figure A 4Confocal images (10x magnification) of the full endothelial channel over time. Images were acquired in days 2, 5, 7 post-flow. Immunocytochemistry in chip displays blue marked nuclei (DAPI) and red F-actin filaments (TRICT). ..................................................................................................... 97 Figure A 5Endothelial barrier integrity studies fluorescence studies profile ( Red line). Confocal micrograph displaying both microchannels with fluorescein. .............................................................. 98 Figure A 6Graphic representation of flow change over time through the 5 µm gaps for the endothelial barrier testing. ................................................................................................................ 98 Figure A 7Confocal micrographs (10X magnification) of an intravasation event. Immunocytochemistry in device displays the lung cancer cells with cytokeratin expression (CK, green) and nuclei counterstained (DAPI, blue). Endothelial channel is only marked for nuclei with DAPI (lower channel). .......................................................................................................................................... 99
XVIII List of Tables Table 1Overview of PDMS characteristics and some possible applications. ........................... 15 Table 2Second-generation microfluidic design parameters and respective descriptions. ........ 32 Table 3Modular experiment solutions description and concentrations. .................................. 39 Table 4Parameters calculated for the first-generation device with a 1 mbar pressure input. .. 49 Table 5Second-generation device’s calculated fluidic parameters for 1mbar pressure input. . 50 Table 6Evaluation of the first-generation device performance parameters. The presence (✓) or the absence (✘ ) of the capacity of laminar flow, the deformability of PDMS pillars and the overall outcome in function of different matrigel concentrations. ................................................................................. 54 Table 7Evaluation of the second-generation device with 300 µm wide channels performance parameters. The presence (✓) or the absence (✘ ) of the capacity of laminar flow, the deformability of PDMS pillars and the overall outcome in function of different matrigel concentrations. ....................... 54 Table 8Evaluation of the second-generation device with 600 µm wide channels performance parameters. The presence (✓) or the absence (✘ ) of the capacity of laminar flow, the deformability of PDMS pillars and the overall outcome in function of different matrigel concentrations. ....................... 55
1 CHAPTER 1 Introduction
2 1. Introduction 1.1. Cancer and lung cancer Cancer is believed to be a major barrier for life’s expectancy. Despite an extraordinary amount of effort and although there is a greater knowledge of cancer biology, statistics demonstrate that it is not only one of the main causes of death but also expected to continue to grow rapidly over the next years [1], [2]. Cancer is an abnormal and rapid cellular division that can occur in different parts of the body. It can be caused by genetic predisposition as well as environmental factors. Furthermore, cancer can occur when individuals have an unhealthy lifestyle. For instance, obesity, smoking, alcohol, and physical inactivity are behaviours that have been linked with cancer [3]. Although there is still no definitive cure, many breakthroughs have been made that helped to understand better this complex disease. Alongside the establishment of cancer hallmarks, an important characteristic was found – tumours are not solely composed of one type of cells but instead have shown to be very heterogeneous structures containing tumour cells with different phenotypic as well as genetic characteristics [2], [3]. This view of cancer reveals that therapeutic intervention may destroy cancer cells and their environment, but it can also ignite selective pressure for the expansion of resistant variants which relates to therapeutic failure. Therefore, early detection of the disease is key especially in cancer types that have high incidence and lower survival rates, which is the case of lung cancer. In 2020, lung cancer was the second most diagnosed cancer (11.7%) and accounted for 1.8 million (18%) deaths, remaining the leading cause of cancer-related death [1]. Every year, the number of deaths due to lung cancer is far greater than in other types of cancer, namely breast, colorectal, or prostate cancer, to name a few. Chronic smoking is the main contributor, however, urbanization and environmental pollution are also accountable for the increased number of patients with respiratory diseases [4], [5]. The large-scale incidence and mortality of lung cancer proves that there is an immediate need for more advanced therapies and tools to improve the clinical outcomes of patients. Signs of progress in this area have been substantial and show great promise yet challenges remain. Understanding lung cancer biology and mechanisms of tumour progression is key to develop promising (target) therapies, circumvent or overcome resistance to therapies, and built sharper diagnostic tools that detect the disease in early stages paving the way for a better prognosis. Consequently, decreasing lung cancers fatalities and building a new scenario in which the disease is less threatening and more manageable with personalized medicine.
3 1.1.1. Non-small cell lung cancer (NSCLC) and metastatic NSCLC Lung cancer can be categorized into two major groups — Small-cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) [4], [6], [7]. The first, with more aggressive behaviour, grows and spreads faster, but results in a better response to chemotherapy and radiotherapy. In contrast, NSCLC develops very slowly, has a late onset of clinical symptoms, and as such most patients are diagnosed with advanced-stage disease resulting in a very poor prognosis. Approximately 85% of patients diagnosed with lung cancer suffer from NSCLC, which can be further divided into lung adenocarcinoma, squamous cell carcinoma, and large cell carcinoma, based on their histological features [6]–[9]. Additionally, there are also several variants and combinations of clinical subtypes. Adenocarcinoma is the most diagnosed type and arises from the alveolar cells located in the smaller airway epithelium. Squamous cell carcinomas account for up to 30% of lung cancers and arise from cells located in the airway epithelium [7], [8]. Even though these subtypes start from different types of lung cells, they are grouped in the NSCLC category because their treatment and prognoses are often similar. The clinical outcome of NSCLC is directly related to the stage of disease at the time of the diagnosis. Hence, regular screening can be useful in early detection and may improve the patient’s survival. Unfortunately, NSCLC is detected when the patient starts to experience symptoms which often is related to the advancement of the disease – when cancer has begun to metastasise. Metastasis is the development of secondary tumours in different parts of the organism, distant from the original primary tumour. Cancer spreading or metastasis is the main cause of cancer therapy failure and the underlying cause of cancer-related mortality. Despite all efforts and research, it remains complex and poorly understood, being accountable for 90% of cancer fatalities [10]–[14]. The onset of the metastatic dissemination begins when cancer cells detach from the primary tumour, induce some level of degradation to the membrane basement and enter the underlying interstitial matrix – invasion. As mentioned, cancer is complex and gains various characteristics to ensure survival, being described as cancer hallmarks [3]. One of the hallmarks, inducing angiogenesis, leads to a subsequent step on the cascade where the detached cells use tumour vasculature to gain access to the body’s circulatory system – intravasation (Figure 1). In circulation, these cells are called circulating tumour cells (CTCs) and it is believed that they have a better chance to survive the dissemination process if they form clusters of cells [15]. If they endure pressure in blood vessels and other adverse conditions, they can further infiltrate the stroma – extravasation – upon arresting in the microcirculatory system of the target tissue by disrupting the tight junctions of the endothelial lining (Figure 1) of blood vessels [16]. Furthermore, they can remain quiescent for a long time, as occult micrometastases (OM), a reason why cancer relapse can arise years
4 after the first tumour. Once at the suited environment, the tumour niche restarts its activity – Undergo proliferation and colonize new sites further evolving from microto macro-metastasis (Figure 1). Although lung cancer starts as a local disease it can metastasise to distant sites, such as brain, liver and/or bone. Specifically in NSCLC, highly metastatic locally and in distal organs, the 5-year mean survival is lower than 5% in the metastatic setting [17], [18]. This is mainly due to the formation of OM [19], which can only be detected by highly invasive molecular techniques, not approved for clinical practices. Therefore, these OMs remain occult and nontreated leading to their “stealth” evolution to fully grown metastasis. Early detection of these micrometastases and immediate treatment to prevent progression to macroscopic metastasis can ultimately improve survival. Hence, careful analyses of metastatic cells behaviour and characteristics are of utmost importance. Figure 1Representation of the metastatic cascade, its main components, and events: invasion, intravasation, dissemination, extravasation, and colonization. (Adapted from Biorender)
5 1.1.2. Circulating Tumour Cells (CTCs) As mentioned, metastasis is the underlying cause of cancer-associated mortality. Still, our understanding of this process is far from complete. Since dissemination occurs mostly through the haematogenous route, CTCs that have been shed into the vasculature may potentially reach metastatic sites and are therefore of high interest. They seem to play a vital and active role in cancer spreading. As such, possess potential as a real-time marker for disease progression and are interesting for metastatic mechanisms studies, namely intraand extravasation studies. 1.1.2.1. CTCs, ECM, hemodynamic forces, and their metastatic role For most intravasating cancer cells the circulatory route is harsh and difficult to survive. In fact, melanoma studies performed in animal models suggest that a small percentage of CTCs (<0.1%) metastasize [20]. Their interaction with the microenvironmental components encountered determine not only their survival but also the ability of extravasation in distant sites. The extracellular matrix (ECM) is a scaffold of macromolecules that form networks and comprise cells present in tissues, having the capacity of altering the phenotypic properties of cells. Consequently, it affects the cell’s tendency to proliferate, migrate and survive. In normal conditions, ECM remodelling is crucial to maintain tissue homeostasis [21],[22]. Unfortunately, ECM deregulation is correlated with the development and progression of several pathologic conditions [22], as metastasis. In this scenario ECM remodelling is usurped and leads to tumorigenesis of the stroma, and as such its components are involved in the invasive and metastatic processes of CTCs. For instance, the dismantlement of ECM through enzymes like matrix metalloproteinases (MMPs), which are crucial in cell proliferation, immune response, and angiogenesis, leads to the invasion step of the metastatic cascade. They are associated with a poor prognosis and demonstrate the impact that the ECM has in the multistep process [22]. The bi-directional interaction between CTCs and the ECM is crucial. CTCs originate brief to lasting tissue alteration given their capability to sense and act in response to tissue mechanics (as is the case of ECM stiffening, compression, deformation, and remodelling). Most CTCs circulate as single cells, but as stated some disseminate through clusters, increasing their likelihood to form metastases. However, these clusters are not merely composed of CTCs. Their enhanced survival over single CTCs relies on heterogeneity ranging from stromal cells to immune components from the original microenvironment, neutrophils and even platelets [3], [15], [23]. All of which pay their contribution to the survival of the malignant cluster. For instance, the formation of a coating shield formed
12 1.2.4. 3D models and their improvements over 2D models Bidimensional cell cultures are a standard approach but due to their relevant disadvantages, current strategies have been developed to improve and elevate pre-clinical testing to a new level, being the case of 3D cell culture. In physiological conditions, cells reside in a 3D environment and interact with other cells as well as their ECM, as said. These are vital to the proper functionality of cells, and even more to precisely mimic physiological pathological conditions such as tumour metastasis. There are various types of 3D cell culture models that have been used in cancer research, being divided into scaffold-based and non-scaffold based. As the name suggests, the first is based on the use of synthetic matrix while the second relies on the self-assembling capability of cells. Among these groups, that differ in their culture technique, two 3D models stand out – spheroids (non-Scaffold based) and organoids ( Scaffold-based). Spheroids entail mainly cell-cell interactions in a 3D format that greatly improves their complexity over their 2D counterpart, while organoids incorporate 3D displayed cells in a scaffold or matrix environment elevating their complexity further, providing cell-cell interactions as well as cell-extracellular matrix interactions. 3D cell culture techniques have been greatly improved and confer many advantages. They demonstrate improved cell morphology, differentiation, mechanical properties, and viability than those in 2D culture. Furthermore, its environment promotes enhanced stimuli response, gene, and protein expression as well as cell function [47]. As an example, a lot of attention has been brought to organoids and their higher degree of complexity as they are self-organized 3D organ-like cell clusters that can be derived from embryonic stem cells (ESCs), induced pluripotent stem cells (iPSCs), primary cells, as well as tumour cells [48], [49]. Their high relevance as a study model over their 2D counterpart is mainly due to their substantial capacity to mimic the organ from which it derived in organization as in function [34]. Environmental cues are of great importance in many cellular aspects that are essential to drug development, an expensive process upon which first experimental tests regarding pharmacokinetic and pharmacodynamic profiles are done using 2D cell culture before further experiments. With this, by using a more complex cell model such as 3D cell culture, that exhibit environmental cues drug discovery can be improved through better drug candidate selection and costs can be reduced. Still, the major disadvantage of 3D cultures is that they reside in a static environment, and physical stimuli that are crucial, especially for metastasis, are not present as shear stress or continuous flow regimes.
13 1.3. Organ-on-a-chip (OOC) Albeit the effort made in oncology research as well as in anticancer pharmacological or therapeutical development research, the number of diagnosed and untreatable cases of NSCLC is growing and represents the leading cause of death, among cancer-related fatalities, worldwide. This expresses the necessity for new, more effective, and more advanced technologies to be used either for screening purposes or to better understand the tumour microenvironment, its cues and cancer disease. In this regard, organ-on-chip , born from the convergence of tissue engineering and microfluidic technology, is a game-changing technology that has flourished over the past years and is expected to give solutions to problems present in traditional testing models as well as in the more recent 3D models [43], [44], [48]. Dynamic OOC models that allow stimuli manipulation, compartmentalization due to their acquired miniaturization characteristics and cell-cell/cell-extracellular matrix interactions can ultimately improve therapy research and improve pre-clinical trials outcome, possibly reducing the economic and life burden of failed pharmacological drugs. Their cost-effective characteristic can also considerably accelerate scientific research. With special emphasis on tumour pathophysiology, OOC systems represent an ideal platform to accomplish the goals above mentioned. 1.3.1. Microfluidic technology The fact that 3D systems have more in vivo relevance as opposed to 2D systems is well established. That said, it is important to note that even though their increased relevance and complexity, they still lack some key points to mimic truthfully the in vivo scenario, notably continuous perfusion, and shear forces. In this context, microfluidic technology arises to overcome said limitations. Originally intended for physics and engineering fields, microfluidic technology has been used for the development of platforms referred to as miniaturised total analysis systems (µ-TAS) or as Lab-on-a-chip (LOC) [50], [51]. As a science that manipulates small amounts of fluid within channels of microto nanometre dimensions [52], [53],[54] microfluidics carries remarkable characteristics that transform it into an elegant technology to implement cell culture in a dynamic environment, that can be engineered to closely resemble the human organism and its pathologies.
14 1.3.1.1. The physics principles of microfluidics Microfluidics manipulates liquids in channel networks of small dimensions. Such low volumes of fluid behave differently and have different properties that would not be present on macro-scale [55]. For instance, microfluidic devices present an almost laminar type of flow that on larger channel dimensions would not occur – fluids would mix convectively [52], [53]. In fact, the dimensionless Reynolds number (Re) of a fluid flow allows determining its flow regime. That is if it is laminar or turbulent. The Reynolds number can be calculated as follows: 𝑅𝑒 =𝜌𝜐𝐷ℎ 𝜇 (Equation 1) Where 𝜌 is the fluids density, 𝜐 the average flow velocity, 𝐷ℎ the characteristic length scale of the system, and 𝜇 the fluid dynamic viscosity. If the Re number is inferior to 2.000, typically indicates a laminar flow. Around 2.300, the fluid begins to experience signs of turbulent flow and from that point forward, the fluid is in a turbulent regime [52]. To acknowledge the importance of laminar flow in microfluidic systems, because here, the particle’s velocity in a fluid stream is not randomized in time. As such two or more flow streams in contact will not mix and will remain parallel to each other, contacting only by diffusion [52]. Hence, without complete physical separation, particles can still diffuse freely from one stream to the other. This can be described by the tendency of a concentrated group of particles to move to regions of lower concentration due to Brownian motion [52], [55]. The equation can be described as follows: 𝑑2= 2𝐷𝑡 (Equation 2) Where 𝑑 is the distance a particle moves at a specific time, 𝑡, and D is the diffusion coefficient of the material. However, diffusion is often neglected because diffusion time is usually greater than the time it takes for a particle to cross the whole microfluidic channel [55]. Moreover, gravity force at these dimensions is largely reduced but surface tension and capillary forces are dominant [53], [55] and therefore allows a pumpless fluid movement. Inside the microchannels, the fluid is submitted to resistance. Hydraulic resistance dictates the flow rate within the channel given by the following equation: 𝑄 = ∆𝑃 𝑅 (Equation 3) Where 𝑄 is the flow rate, ∆𝑃 is the drop of pressure across the channel and 𝑅 is the channel resistance. By extension, the hydraulic resistance of a rectangular microchannel (the most common geometry), is given by equation 4.
15 𝑅 = 12𝜇𝐿 𝜔ℎ3 (Equation 4) Where 𝜇 is the fluidic viscosity, 𝐿 is the channel length, 𝜔 and ℎ are the width and height of the channel. 1.3.1.2. Poly(dimethylsiloxane) for microfabrication Microfluidic device technology emerged as a branch of MEMS (Micro electro mechanical systems) [56], [57]. In this field, the most commonly used material is silicon. As such, the first fabricated microfluidic devices were silicon-based [56]. With excellent surface stability, solvent compatibility, and good chemical resistance due to its well-developed surface chemistry, silicon was in fact an attractive material to use [58]. For cellular attachment and growth, its surface can be easily modified with silanes to reduce nonspecific adsorption [59]. Nevertheless, its optical properties are not ideal since it is not a transparent to visible light material [56] and silicon-based fabrication of microfluidic devices is an expensive process that requires specialized installations. Adding to silicon, glass microfluidic devices were also used [56], but equally expensive and time-consuming, although leading to a new generation of microfluidic biosensors. Over the years, silicon and glass inorganic materials were gradually replaced by polymeric materials [56]– [58]. They gained a lot of attention in the microfluidic field due to their low-cost production, easy manufacturing, and attractive properties, such as mechanical strength, optical transparency, and chemical stability [57]. Among various polymers, a thermoplastic material named poly(dimethylsiloxane) (PDMS) is the most common material used since its discovery in the late 1990s [59], [60]. PDMS is widely used because it has wonderful characteristics (Table 1). Table 1Overview of PDMS characteristics and some possible applications. Property Description Application References Biocompatibility It does not cause unwanted consequences on biological material Cell culture, suitable for biomedical implants and organ-on-chip [51], [57], [61] Deformation High compliance or low stiffness. With an elastic modulus of ~1-3MPa, it facilitates demoulding and bonding Cell response to stimuli [61], [62], [63], [64] Gas permeability Gas permeable, allowing for CO2 and O2 diffusion Cell culture, organ-on-achip [56],[63] Inert material Chemically inert surface allows for wanted patterning of subtract Tissue engineering, biosensing [61],[63] Optical transparency Optical transparent within the 2401100 nm range, visible light Optical applications, cell monitorization [65], [66], [67]
16 Its inorganic siloxane backbone coupled with organic methyl groups attached to silicon (O-Si(CH3)2) [68] leads to a unique combination of interesting properties such as absorption, biocompatibility, deformation, gas permeability, chemical inert surface, and optical transparency [63]. The first, absorption, is due to PDMS hydrophobic nature and its likeliness to absorb into the polymer matrix nonpolar solvents of organic origin, and small hydrophobic molecules [63], [69]. This can negatively affect the experimental procedure, especially in cell culture since cells will not adhere to the channels. However, this can be circumvented, by reducing absorption through coatings or chemical modifications of the inner surfaces of the microchannels [64], [69]. Secondly, PDMS is biocompatible which means it is cell-friendly and can endure organic coatings for ECM formation and allow cell attachment [54], [57], [59], [61], [63]. The third characteristic, deformation, facilitates unmoulding of the device in fabrication and an unfailing bonding to other surfaces. However, the channel’s deformation under pressure-driven flow can affect the actual flow rate due to modifications in the cross-section area. Gas permeability of PDMS can permit O2 and CO2 diffusion, greatly impacting cells behaviour. However, due to the small volume reagent, evaporation is enhanced, and can significantly affect concentrations, chemical gradients, and balances as well as other key factors in the experiment [56],[63]. As such, continuous perfusion is extremely important to avoid this. Moreover, PDMS is optical transparent [67] and as such, it is highly suitable for optical applications and allows for cell movement monitoring, and image data collection, which is a major improvement from the first used microfluidic device materials [61]. PDMS chemical structure transforms it to an excellent candidate for fabrication of microfluidic devices meant for cell culture [50]–[53], [57], specifically, development of OOC platforms. 1.3.2. Microfluidics for cell culture All the physical properties of microfluidic technology and PDMS mentioned can be taken advantage of and implemented particularly to perform dynamic cell culture on a chip. For instance, since laminar flow is present inside the chip, mass transport is dominated by local diffusion rates. Consequently, nutrient, gas and drug supply to cells can be manipulated and carefully analysed. Due to the non-considered inertia forces, precise delivery of compounds is facilitated in both time and space, meaning that they can be delivered in certain areas at a controlled time point. Importantly, besides providing these properties the channel’s size and the overall environment of the microfluidic platforms are comparable to the intrinsic dimensions of blood vessels, cells, or cell mass [44]. Hence, gas and drug diffusion rates, shear stress and even cellular niches can be artificially recreated on-chip, representing the physiological environment
17 of the human body, with continuous perfusion and waste removal, making them suitable for emulating artificial tumours and the surrounding microenvironment [70]–[72]. It is no surprise that, since its first appearance, microfluidic devices for OOC purposes have been gaining a lot of attention. This branched field seeks to replicate the function of in vivo organs and pathologies in one simple device, integrating various components into a single platform, allowing for a closer view of the true scenario [44], [50], [51], [54], [63], [70]–[72] [73]. To note, that it is not intended to build an intact artificial living organ but instead to develop a stable in vitro screening platform, that could be useful to give answers to the still concealed questions of various diseases, for example. It could be used as specialized in vitro testing platforms for simulations, functional studies, and pharmacological testing [43], [44]. Thus, the emerging application of the OOC system aims to mimic the mechanics, function and physiological responses of an organ or pathology in a 3D microfluidic culture model. This means that they must have the capacity of mimicking environment interaction as well as all the criteria to exemplify the complex network of interactions occurring in, for example, the metastatic process [46], [54], [59]. This accomplishment brings a good alternative to preclinical studies, lowering the clinical trial huge drug candidate’s failure, by enhancing their predictability power. Also, lowering the drug development economic burden, since the manufacture of new drugs is a long and expensive process, and often, fails in human studies, for the reasons already mentioned. Hence, OOC can reduce experimental error making them a great “go-to” model for pre-clinical trials. Its spatial control ability also allows to recreate physiological barriers that exist in the human body. Its permits the recreation of the interconnected set of biochemical and mechanical cues in the cellular microenvironment and recapitulate the complicated interactions between different cells in vivo , which can more accurately simulate and control drug delivery and infiltration. Spatial control allows OCC to be vascularized. Vascularization is an important factor to keep in mind when developing OCC since micro vessels and capillaries cover organs almost entirely [74],[75]. The (micro)vascular system not only supplies oxygen and nutrients to tissue but also is related to pathophysiologic conditions [74]. This means that it defines the biological and physical characteristics of the microenvironment within tissues and plays a role in the initiation and progression of some pathologies such as cancer. For instance, a metastasis-on-chip system should be vascularized since the cascade of events − invasion, intravasation, extravasation – takes place in the circulatory route. Since the vascular system is composed of an endothelial cell lining, microfluidic devices designed to mimic vessels, contain endothelial barriers [74]–[77]. ECs are responsible for maintaining vascular homeostasis in response to circulating cytokines, hemodynamic shear stress and inflammatory cells. The
18 endothelium is of interest in the comprehension of tumour metastasis due to the tumour cell-endothelial barrier interactions during organ invasion, where endothelial junctions’ integrity highly affects the likelihood of pathology progression [76]–[79]. To conclude, personalized medicine is emerging in OOC platforms, allowing for the collection of patient samples and implementing them on-chip [43], [57]. This is a revolutionary area of discovery. It is known that each patient is unique and responds differently to therapies. Despite this, the therapies to which patients are submitted are generic and broadly used. As such, a personalized OCC is helpful and opens doors to personalized drug development (Figure 5). 1.3.2.1. Existing microfluidic-based in vitro models 1.3.2.1.1. Mimicking biological barriers The endothelial cell barrier has a crucial role in metabolic activity regulation, wound healing, immune response, and disease progression, highlighting the importance of understanding the endothelium functions to further comprehend disease mechanisms. To meet this purpose, a microvascular network was formed in an in vitro microfluidic platform. This fully enclosed, perfusable endothelialized microvessel system was built by microfabrication techniques that enabled the construction of vascular geometries and controllable stimuli. The microvascular networks allowed for long-term experiment cultures of ECs (HUECs, Human Umbilical Endothelial Cells). The reported approach enabled studies of basic vascular biology, also providing a model for diseases such as tumour angiogenesis. HUECs were successfully Figure 5Personalized medicine with organ-on-a-chip platforms cultured with patient samples [43].
19 grown in these enclosed channels and demonstrated appropriate morphology, barrier function and cellcell junction interaction [80]. In another study, Verbridge et al. developed one device with 3 parallel channels coated by collagen I. The middle channel contained endothelial channels and the neighbouring channels served as a source of biochemical gradients. This promoted endothelial sprouting and invasion into the surrounding matrix of collagen [81], demonstrating the role of ECs in cancer spreading. Further, a multi-step microfluidic model composed of ECs was created for metastatic studies. The device enabled a sequential analysis and measurement of individual events of metastasis. The microfluidic device included microchannels coated with matrigel matrix to mimic the basement membrane. The channels were lined with ECs (HMECs, Human Microvascular Endothelial Cells) to replicate endothelial lining [82]. Besides endothelial models, other barrier models have been fabricated in microfluidics. For example, the “lung-on-a-chip” system reported by Huh et al. contained an epithelial model to mimic the alveolarcapillary interface composed of two PDMS layers that sandwiched a PDMS porous membrane. In this case, an EC monolayer was created on the bottom layer of the membrane for vascularization purposes, and also ECs were cultured in the upper side of the membrane to mimic the alveolar-capillary interface [83]. 1.3.2.1.2. Metastasis-on-a-chip Given the impact of cancer metastasis and overall therapeutical failure, OOC platforms to assess each cascade event, metastasis-on-a-chip (MOC), can give relevant insights into the poorly understood process. A MOC system was designed by Skardal et al. The platform consisted of two independent chambers, cultured with gut and liver constructs. Although independent, they were connected in series by perfusion flow. This system allowed real-time track of colon cancer cells migrating from the gut construct to the liver construct, upon environmental stimuli. Moreover, the authors were able to observe tumour regions with lost membrane-bound adhesion markers and mesenchymal as well as proliferative markers, demonstrating a metastatic phenotype [84]. Another study conducted in a MOC system by Shin et al. investigated the inhibition of cancer cell invasion upon SW480 and LOVOs were treated with inhibitors and the binding capacity of metastatic cells to endothelial cells (HUVEC, Human Umbilical Vein Endothelial Cells). The system allows for intraand extravasation studies so that the cascade can be studied in just one chip. Composed of two separate
20 chambers, one of 3D displayed colon cancer cells in matrigel – the intravasation chamber – and one for detection of metastasized cells upon their interaction with adhesion molecules present in HUVECs [85]. 1.3.3. Potential of OOC models of cancer Microfluidic OOC models are an almost-physiological technology. These testing /screening systems that accurately mimic cancer and its vasculature hold great promise in understanding the biophysical cues and mechanisms of cancer cell spreading and have the potential to revolutionize pre-clinical trial testing and predictions, lowering the failure of anticancer drugs in clinical trials. These dynamic cancer models have remarkable characteristics holding therefore great promise for personalized medicine and potentially would help to reduce side effects and costs in drug development. Nevertheless, a stable platform is challenging, needing still improvements and further optimizations.
21 CHAPTER 2 Objectives
28 defines a sharp focus point using pinholes [94]. As such the obtained image is sharper and with higher quality and resolution. Moreover, the modern scanning mechanisms implemented in confocal laser scanning microscopy, based on deflected mirrors that move the focus point in the XY plane, enable the reconstruction of an image by information collected from each point, and if the focus point is also scanned in the Z-axis, a 3D image can be created with high resolution (Figure 7) [95]. 3.3. Cell culture techniques Cell culture is the standard in vitro testing model. In fact, the influence of cell culture technology in the scientific community has been immeasurable. Cell culture is a technique that enabled cells extracted from different tissues of plants or animals to grow in an artificial environment outside of the body. Once cells are collected and put in vitro culture, they become what is called a primary culture. When their confluence is reached and cells are subcultured, they become what is known as a cell line. The latter can be classified based on their life span. They can either have a finite life, which means that they can be subcultured a determinate number of times until they cease proliferation and enter a stage of senescence, or a continuous life span, that means that they can proliferate and be subcultured indefinitely and remain stable [96]. To maintain cells in culture, some crucial features must be always present as a cell growth medium, supplements, a treated and sterile container (flasks, Petri dishes, multi-well plates) and the appropriate cell incubation conditions (temperature and humified environment). Moreover, to maintain their viability, upon cells reach confluence they should be subcultured using a detachment solution as trypsin/EDTA to separate the cells from the substrate and each other [96]. As such the technique is simple and easy being a crucial asset to oncological studies. 3.4. Immunocytochemistry Immunocytochemistry (ICC) is a technique for the detection and visualization of proteins or antigens using antibodies that recognize and target the antigen of interest in cells [97], [98]. The antibody can be directly or indirectly linked to a fluorophore so that it can give qualitative or quantitative information about the cellular structure that is targeted using fluorescence microscopy or a plate reader. This technique can be performed on samples of cells grown in a monolayer or suspension. The procedure can be carried
29 out in different forms and depends on the wanted analysis, but usually, adherent cells are fixed to a microscope glass slide, cover slip or well plate [97]–[99]. This technique is crucial in research as well as in clinical laboratories since it can be used to give a more accurate prognosis, clinical staging, and subsequent therapeutic decisions, by identifying proliferation markers or oncoproteins. Moreover, these methods are also able to predict clinical response to therapy by analysing the expression level of these markers. As such, paired with fluorescence microscopy/confocal microscopy, ICC is an important asset to biological research since it enables the visualisation of specific cell structures and markers.
30 CHAPTER 4 Materials and Methods
31 4. Materials and Methods In this chapter, the methods, techniques and optimizations applied for the design, fabrication and functionalization of the multichannel microfluidic device are described in detail. The cell culture procedures are also described, as the optimizations implemented to build a dynamic organ-on-a-chip. 4.1. Design of microfluidic devices 4.1.1. Design of the first generation The “first generation” device had a heigh of ≈100 µm and consisted of two parallel 200 µm wide channels separated by an array of square-shaped micropillars ( Figure 8 A, B). These are spaced by 5 or 10 µm gaps. This previously designed microfluidic chip model was fabricated by standard SU8 lithography. Figure 8First-generation microfluidic device; A) Full CAD design consisting of two 200 µm lateral microchannels, inlets and outlets; B) magnification of the cell culturing area separated by square-shaped micropillars with 10 µm gaps. (A) (B)
32 4.1.2. Design of the second generation Devices were designed using CleWin4 from WieWeb Software (Hengelo, The Netherlands), consisting of two side-by-side parallel microchannels (Figure 9 A, B) separated by an array of square-shaped micropillars (Figure 9 B, C). Four different device layouts were designed with different features (Table 2) – two models with 600 µm wide channels and two models with 300 µm wide channels, differing in their gap sizes (5 µm or 10 µm). Circular pillars were added in the design (Figure 9 B, D) to avoid PDMS from collapsing since it has a high aspect ratio. Moreover, 5 mm features (Figure 9 E) were added to produce reservoirs that allow liquid storage. Thus, the device is versatile – allows perfusion by automated mechanisms through inlets or via gravity through liquid reservoirs. Table 2Second-generation microfluidic design parameters and respective descriptions. Parameters Size (µm) Description w 300 600 Channel’s width d1 6400 6800 Channel’s length d2 250 250 Circular pillars separation d3 5 10 5 10 Micropillars separation gaps d4 100 100 Micropillars length d5 75 75 Circular pillars diameter d6 5000 5000 Reservoir feature diameter 𝑑 5 𝑑 1 𝑑 2 𝑑 4 𝑑 3 𝑑 6 (A) (B) (C) (D) (E) Figure 9Second-generation microfluidic device CAD design; (A) Full CAD design; (B) Cell culturing area of both microchannels separated by micropillars; (C) magnified image of the square-shaped micropillar area; (D) circular micropillars; (E) Inlets/Outlets and reservoir features . 𝑊
33 4.2. Fabrication of microfluidic devices 4.2.1. Fabrication of first-generation devices Sylgard 184 PDMS pre-polymer and curing agent were weighed (10:1, w/w), degassed, poured over the master mould (Figure 10 A, B), and cured for 4h at 65 ºC. After curing, the PDMS replica was unmoulded from the wafer producing a device slab in which inlets and outlets were opened using a 1.5 mm diameter Kai Medical biopsy puncher. Microscope glass slides (size 25x75 mm, ThermoFisher Scientific) were cleaned with diH2O and 70% ethanol and dried with an air gun. Immediately after, PDMS replicas and cleaned glass slides were treated with oxygen plasma (Plasma Cleaner PDC-002-CE, HarrickPlasma) at medium power for 30 s and brought in contact to produce irreversible bonding. 4.2.2. Fabrication of second-generation devices Devices were fabricated by standard photolithography. Fabrication of the silicon master mould was performed using the SiO2 hard mask for the silicon dry etching process. As such, a 1µm think plasmaenhanced chemical vapour deposition (PECVD) SiO2 layer was first deposited on a single-side polished (1 0 0) 20 mm Si wafer using a CDV system (MPX from SPTS). The wafer was exposed to hexamethyldisilazane (HDMS, Sigma Aldrich, USA) vapour prime to improve the adhesion of the photoresist to the sample, obtained by spin coating of 1.2 µm of AZP4110 (Microchemicals GmbH, Germany) on SÜSS MicroTec optical track (SÜSS MicroTec AG, Germany). Direct write laser (DWL 2000 Heidelberg Instruments) was used to pattern the photoresist for the top geometry microfluidic channels. Following the post bake, the exposed photoresist was developed with AZ400K (Microchemicals GmbH, Germany), and the wafer was rinsed with deionized water and dried. Etching of SiO2 was completed on a reactive ion etching tool (APS from SPTS) with a C4F8 etching chemistry followed by removal of the photoresist, striped with an oxygen plasma (PVA GIGAbatch 360 M Figure 10- (A) Photolithography fabricated wafer of the first-generation microfluidic devices; (B) Microfluidic device wafer mould. (A) (B)
34 from Tepla). The silicon wafer was then etched by dry etching process performed on an inductively coupled plasma (ICP) — RIE tool (Pegasus from SPTS), using a SF6/C4F8 plasma, to transfer the SiO2 mask features to the bulk silicon. Trench depth was measured using a surface profilometer (KLA — Tencor P-16 Surface Profiler) until the desired depth of 100 µm was reached and lastly the top remaining SiO2 mask was removed on APS from SPTS. Prior to master replication, the produced wafer (Figure 11 A) was hydrophobized with a vapour-phase treatment by applying an anti-adhesion layer of trichloro(1H,1H,2H,2H-perfluorooctyl)silane (Sigma Aldrich) for one hour under vacuum followed by one-hour exposure to 65 ºC to promote the vapour phase and facilitate the detachment of solid PDMS from the wafer. For fast prototyping using soft lithography techniques, PDMS pre-polymer and curing agent were weighed (10:1, w/w), degassed, poured over the master mould, and cured at 65 ºC, overnight. After curing, the PDMS replica was peeled away from the master mould producing a device slab (Figure 11 B). Posteriorly, top and bottom PDMS sheets were fabricated with 0.5 mm and 1 mm in height, respectively, using the soft fabrication techniques described above. After degassing, the PDMS mixture was poured over a square box and allowed an even spreading of the mixture. Subsequently, it was cured at 65ºC, overnight and then peeled off. The top sheet was produced to create a bubble trap mechanism when paired with punched holes. The bottom sheet was developed to seal the device’s channels instead of using a microscope glass slide, in order to ensure an even environment for cell attachment. (A) (B) Figure 11- (A) Photolithography fabricated SiO2 wafer of the second-generation microfluidic device design; (B) Unmoulded PDMS device slab.
35 4.2.2.1. Incorporation of bubble traps In automated driven perfusion, the presence of gas bubbles in microchannels is a frequent challenge when culturing biological material as cells. The presence of a single bubble can negatively impair biological function and viability given that it increases the wall shear stress in a liquid-perfused microchannel by at least one order of magnitude. To avoid this problem, a simple bubble trap approach was implemented in the devices with perfusion driven by syringe pumps (Figure 12 A, B). In the device slab, bubble trap holes were produced (Figure 12 C) using a 1 mm diameter Kai Medical biopsy puncher. Then, a top PDMS sheet and the device slab (channel features down) were oxygen plasma-treated (Harrick Plasma Cleaner) for 30 s at high power and brought into contact producing an irreversible bond (Si-O-Si) by condensation. Afterwards, inlets and outlets were opened using a 1.5 mm diameter biopsy puncher and the device slab (channel features up) along with the bottom PDMS sheet were submitted to O2 treatment to produce irreversible bonding once again and seal the microchannels (Figure 12 D). Figure 12Syringe perfusion device with implemented bubble trap. (A) Schematic representation of the bubble trap mechanism. The air bubbles float upwards in the punched hole and diffuse across the top PDMS membrane; (B) representative image of the device with 1 mm circular holes (Red) and bonded top PDMS membrane (Blue) to build the bubble entrapment; (C) Top view of the implemented bubble trap mechanism in device; (D) Side view of the total assembled device. (A) (B) (C) (D)
36 4.2.1.2. Incorporation of gravity reservoirs In PDMS device slabs with 1 cm height, reservoirs were produced by a 5 mm diameter Kai Medical biopsy puncher, in both inlets and outlets (Figure 13 A). These reservoirs enable a ≈ 196 µL liquid load. Afterwards, the device and the bottom PDMS sheet were oxygen plasma treated for 30 s at high power, bonded, and placed in the oven at 65 ºC for 15 min to allow permanent bonding (Figure 13 B). 4.3. Cell Culture The human lung adenocarcinoma cell lines A549 (American Type Culture Collection, ATCC) and H1650 (American Type Culture Collection, ATCC) as well as the human pulmonary microvascular endothelial cells, HPMEC (supplied by i3S, Porto) were used in this project. The A549 cell line was transfected with Green Florescent protein (GFP) and TAS1R3 (Taste receptor 1 member 3) plasmid. TAS1R3 has been detected in metastatic cells. An additional A549 non-transfected cell line was used. Both lung adenocarcinoma cell lines were cultured in RPMI-1640 (Sigma, Aldrich), supplemented with 10 % fetal bovine serum (FBS, Gibco) and 1 % Penicillin/Streptomycin (Pen/Strep, Corning). HPMEC were cultured in M199 medium (Sigma, Aldrich) with 20 % fetal bovine serum, 1 % Penicillin/Streptomycin and supplemented with 25 µg. mL-1 endothelial cell growth supplement (Sigma, Aldrich) and 25 µg. mL-1 of heparin sodium (Sigma, Aldrich). Cells were routinely subcultured when reached a confluence of 70-80 %. The old cell medium was removed, and the cell monolayer was rinsed with Phosphate Buffer Saline (PBS, Sigma Aldrich) (1x) to remove residues and dead cells. Afterwards, the monolayer was submitted to 0.25 % (w/v) Trypsin-EDTA (BioConcept) and incubated for 5 minutes to dislodge and disperse cells. Posteriorly, a complete growth medium was added, the cell suspension aspirated carefully into a 15 mL Falcon tube and gently resuspended to avoid clumping, promote cell separation and non-less important neutralize the TrypsinEDTA solution. To know cell concentration (Equation 5) and number (Equation 6) the Neubauer hemacytometer procedure was used. A sample volume of 10 µL was collected from the resuspended Figure 13Perfusion via gravity device. (A) Top view of the PDMS fabricated device with 5 mm reservoirs; (B) Side view of the assemble device. (A) (B)
37 cell suspension and added to 10 µL of trypan blue solution (Corning), obtaining a dilution with a suitable concentration for cell counting. After resuspension of the cell solution, 10 µL were carefully introduced into the Neubauer chamber (Hirschmann® EM Techcolor), setting the pipette tip against the edge of the cover slip, and carefully expelling the solution. Finally, the Neubauer chamber was taken to the microscope (Nikon Eclipse TS100, Nikon), and viable cells were counted. 𝐶𝑒𝑙𝑙 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 (𝑐𝑒𝑙𝑙𝑠 𝑚𝐿 ) = 𝑎𝑣𝑒𝑟𝑎𝑔𝑒 𝑐𝑒𝑙𝑙𝑠 ×𝑑𝑖𝑙𝑢𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 × 104 (Equation 5) 𝑁º 𝑜𝑓 𝑐𝑒𝑙𝑙𝑠 = 𝐶𝑒𝑙𝑙 𝑐𝑜𝑛𝑐𝑒𝑛𝑡𝑟𝑎𝑡𝑖𝑜𝑛 × 𝑉𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑐𝑒𝑙𝑙 𝑠𝑢𝑠𝑝𝑒𝑛𝑠𝑖𝑜𝑛 (Equation 6) The remaining cell suspension was centrifuged, and the obtained cell pellet was resuspended in fresh complete growth medium. Lastly, appropriate volumes of cell suspension were added to new culture flasks with fresh medium and incubated at 3 7ºC in a 5 % CO2 humidified atmosphere. The medium renewal was performed 2 to 3 times per week to maintain good cell growth conditions. 4.4. System Optimisations The firstand second-generation models were carefully evaluated to understand their capacity to sustain an OOC system. Optimization studies entailed analytic calculations, modular experiments for device layout, automated mechanisms, and scaffold concentration selections. Moreover, in-plate PDMS modification and functionalization studies were performed and evaluated via metabolic activity assays. 4.4.1. Parameter calculations In order to calculate the basic fluidic properties of the device in question, and how this relates to flow and diffusion, fluid flow in microchannels was calculated by using the already mentioned fluidic resistance for a rectangular channel (Equation 4) and flow rate (Equation 3 ). To understand if the flow in the device remains laminar or turbulent, the Reynolds number (Equation 1) was calculated for each device. All the parameters were analytically calculated and confirmed using MatLab Software.
44 4.5.1. Syringe pump perfused device After complete cell seeding and attachment in chip, the devices were assembled on a home-built microscope and connected to a syringe pump using 1.6 mm OD PTFE tubing to perfuse previously degassed culture medium continuously through both microchannels at a constant flow rate (15 µL∙h-1). Daily monitorization was performed in the home-built microscope (Figure 19 A, B), to not cause unnecessary disturbance to the system. Medium filled syringes were replaced 2-3 times a week, and devices were assessed for up to 5 days. 4.5.2. Gravity perfused device In this format, after seeding (Figure 20 A), the devices were sited in their correct position in a 6-well plate to permit a sterile environment since the reservoirs are opened (Figure 20 B). Both reservoirs of the cancer channel were previously filled with 190 µL of complete growth medium. For the endothelial channel, however, 180 µL of medium were placed in one reservoir. This creates perfusion by pressure drop. Moreover, to circumvent the evaporation of the medium, PBS (1X) was added in between wells to Figure 20Gravity driven flow device's setup in 6-well plate (A) Six cell-seeded microdevices in the tilted position to allow endothelial cells preferential growth against the micropillars; (B) Microdevices in their correct position with media perfusion through the resevoirs. (A) (B) Figure 19Syringe pump driven device setup. (A) In the laminar flow hood, the syringe Pump is connected to the device placed on the homebuilt microscope with the microchannel features displayed in the live-on-screen monitor; (B) Device connected to tubing on the stage of the home built-microscope. (A) (B)
45 create a more humified environment. Plates were carefully taken to the incubator (37 ºC; 5 % CO2) and daily monitorization and image acquirement was performed using the facility microscope (Nikon Eclipse TS100, Nikon). Reservoirs were re-filled, and devices were assessed up to 7 days post complete cell culture. 4.6. Integrity studies on chip 4.6.1. Cell proliferation and metabolic viability in-chip As described in section 4.4.3.3., the viability of cells was assessed on-chip and for each microchannel. Control measures were obtained by previously seeded well-plates at a correct cell density. The 10 % Resazurin solution was prepared, syringes or reservoirs were filled, and the perfusion commenced. After the 4 h incubation period required, the sample was collected, and the fluorescence intensity measured, as described before. Given that resazurin is non-deteriorating to cells, the channels were subsequently washed with warm PBS, and other measurements were performed. To complement the data from the metabolic viability assays, a cell counting estimative was performed in ImageJ software (National Institutes of Health, USA) by counting DAPI stained nucleus across the whole cell culturing area. 4.6.2. Endothelial barrier integrity To ensure endothelial barrier integrity and test the permeability against small compounds before performing on-chip studies, a fluorescein solution (30 µM. mL-1) was prepared in a complete growth medium of HPMEC cells. The solution was pipetted into the inlet reservoir of the endothelial channel and the adjacent channel reservoirs (cancer channel reservoirs) were emptied to assess the diffusion across the barrier of the vascular channel. Time-course images of fluorescein permeation on the chip were collected and the absolute fluorescence intensities of the compound from a plot profile of the vertical section entailing both channels were quantified using Image J software. Measurements were obtained on days 2, 5 and 7 post-flow, of cultured devices. To evaluate and compare the permeability and integrity, fluorescein permeation was quantitated by calculating the ratio of fluorescence intensities on the vascular channel side to that of the tumour channel. The fluorescence intensities of the vascular channel were defined as 100 % of standardization.
46 4.6.3. Immunocytochemistry Microchannels were carefully washed with warm PBS (1X) to eliminate medium residues. Cells were subsequently fixated with 4% Paraformaldehyde (PFA, Sigma Aldrich) for 15 minutes at room temperature (RT), rinsed with cold PBS (1X) thoroughly and permeabilized with 0.2% Triton X-100. After washing three times with cold PBS, the fixed cells were incubated with a 2 % (w/v) Bovine Serum Albumin (BSA) solution in PBS (blocking buffer solution) to minimise non-specific antibody bonding, for 45 minutes. Subsequently, GFP non-transfected cancer cells were incubated with 5 µg∙mL-1 Monoclonal Anti-Cytokeratin (CK) pan-FITC antibody (Sigma Aldrich), while ECs were incubated with 0.1 µg∙mL-1 phalloidin-TRITC (eBioscience™) at 4 °C overnight. After extensive washing, cells were counterstained with 1 µg∙mL-1 DAPI (Sigma, St Louis, USA) solution for 15 min and preserved in PBS. Images were acquired using a confocal microscope equipped with 405nm and 561-nm diode lasers and a 488-nm argon laser and analysed using Zen 2010 software and ImageJ software. 4.7. On-Chip Studies 4.7.1. Nanoemulsions Nanosystems were formulated by the ethanol injection method, previously described by our partners at FIDIS (The Foundation of Health Research Institute of Santiago de Compostela) [100], [101]. These are composed of an organic and an aqueous phase. The organic phase components – Vitamin E (5 mg, Sigma Aldrich) and sphingomyelin (0.5 mg, Lipoid) – were injected in ultrapure water. To observe the migration and internalization of nanoemulsions by confocal microscopy, sphingomyelin-Cy5 (Avanti Polar Lipids) was added to the formulation. The organic phase, dissolved in 100 µL of ethanol, was injected with an insulin syringe in 1000 µL of the aqueous phase. The formulation was introduced into the endothelial channel and incubated for 6 h at 37ºC in a 5 % humidified environment. Cells were subsequently fixed and counterstained using DAPI (1 µg∙mL -1) to allow a clearer view of the nanoemulsions path. 4.7.2. Intravasation and cancer aggressiveness Intravasation studies were carried via monitorization of the system’s evolution after cell seeding or postendothelial degradation. Cells were fixed and ICC was performed as described in section 4.6.3. Confocal images were acquired in Z-stacks of 100 µm depth to verify if cells were displayed three-dimensionally.
47 4.8. Statistical analysis Results are shown as mean ± standard deviation. All sample data analysis was performed using Graph Pad Prism 8.3.1 (San Diego, CA, USA). P-values were determined by employing one-way analysis (ANOVA). Results were considered statistically significant if the p-value <0.05.
48 CHAPTER 5 Results and Discussion
49 5. Results and Discussion 5.1. Device calculations and simulations Using the previously mentioned equations in section 1.3.1.1., some important parameters for devices performance were calculated. Here the basic fluidic properties of the chip are assessed. For the automated perfusion, the used pressure input was 1 mbar for the calculations. Table 4Parameters calculated for the first-generation device with a 1 mbar pressure input. For the first-generation device, the obtained values provided valuable information (Table 4) for the upcoming experiments. First, the device resistance (without considering the tubing since it imposes more hydraulic resistance) gave the average flow rate and velocity in device of 12 µL.min-1 and 1.03 x 10-2 m.s-1, respectively, with a shear stress of 6.22 dyne.cm-2. In these conditions, the obtained Reynolds number demonstrates that the microfluidic chip remains in laminar flow conditions (Re<2.00) [52], which indicates that compartmentalization is possible (in Newtonian fluids). As for the second-generation device (Table 5), it was observed that the flow rate between these two layouts is ≈20 and ≈40 µL.min-1, for the 300 µm and the 600 µm wide channel devices, respectively. Both are expected to remain in laminar flow and there is a slight change in the wall shear stress. Although not very significant to an empty model only with liquid flow, this change may impose some differences to cell differentiation, display, and inflammation. Resistance (Pa.s.m3) Flow Rate (µL.min-1) Velocity (m.s-1) Reynolds number Shear Stress (dyne.cm-2) 1st Generation Device 4.86×1011 12.5 1.03×10−2 1.38 6.22
50 Table 5Second-generation device’s calculated fluidic parameters for 1mbar pressure input. The interpretation of results gave a clearer view of the fluid dynamics of each layout and of the flow rates to use in the first experimental trials. Wider channels impose less resistance and higher flow rates, yet this sort of calculation is for simple Newtonian liquids with the same input pressure as well as the same resistance in both channels (Figure 21). As such, differences in each channel may give different outcomes, for instance, the deformation of the micropillars or system collapse. For the second-generation layout, hydrostatic pressure can be used to drive the flow ( Reservoir’s option). Given that 1cm of water column (reservoirs height) is 1mbar of pressure, the flow rate and the velocities experienced in the devices are as calculated (Table 5). If all the reservoirs are empty except for one (Figure 22 A), the fluid first perfuses the whole channel until both reservoirs are at the same height, and only then begins the perfusion across the gaps. This is due to the less resistance of the main channel compared to the resistance of the microgaps. Resistance (Pa.s.m-3) Flow Rate (µL.min-1) Velocity (m.s-1) Reynolds number Shear Stress (dyne.cm-2) 2nd Generation Device 300 µm 3.04×1011 18.5 1.12×10−2 1.69 6.17 600 µm 1.52×1011 39.49 1.3×10−2 1.88 6.58 1 mbar 1 mbar Figure 21Schematic representation of the 600 µm wide channel entailing 5 µm gaps with a 1mbar input pressure applied in both microchannels.
51 Moreover, if the input volume is higher, the pressure input is higher (higher water column) as well and as a result the flow rate and velocity increase. Yet, since the velocity is higher, the flow rate decreases rapidly and reduces in less than half an hour (0.1 to 0.2 hours) (Figure 22 B). As such this can negatively impact the cells and their development as the shear stress would be variant and periodic. 5.2. Device performance in model experiments with microbeads Initial experiments were performed to assess the robustness of the chip to build a MOC system. For this, the microfluidic device had to go through a series of testing to ascertain a model that could sustain pressure caused by the scaffold for the 3D cancer cell culture and create an endothelial barrier against the micropillars in the adjacent microchannel. All this while still maintaining laminar perfusion without deformability of the micropillars or system collapse. The first key feature of the system is the perfusion mechanism. Hence, the two already mentioned mechanisms were tested using food colour first, since liquids of this sort are expected to behave as Newtonian fluids (water). As such the previous calculations gave the flow rate and pressure parameters to test in each mechanism. Figure 22Liquid flow dynamics of the second-generation device driven via gravity with pressure input in only one reservoir. (A) Schematic representation of the pressure input in the reservoir dependent on the height of the water column. (B) Graphic representation of the flow rate and solution exchange overtime as the water column in the input decreases and the water column in the output increases. Pressure h (A) (B)
52 For the syringe pump, at a defined flow rate of 12 µL.min-1 the fluids remain separate (Figure 23). Conversely, with the pressure controller at a defined input of 5mbar (accounting for tubing resistance) the device would not be perfused, and only start to experience flow when the pressure was increased to 12 mbar. In these conditions, quickly both fluids would mix, as the micropillars would deform given Figure 23Live-on-screen image of the laminar flow maintenance in the 1º-generation device using a syringe pump with a defined volumetric flow rate of 12 µL.min-1. Figure 24Live-on-screen images of the non-laminar flow in the 1º-generation device using the pressure controller automated mechanism with a pressure input of 5mbar. (A) Red and blue food colour solutions remain separate as the channels are not yet in contact. (B) system collapse and fluidic mixture of the blue food colour with the red food colour. (A) (B)
53 the unsustainable pressure caused by the input (Figure 24). This was probably due to the range of the pressure controller (≈1 bar) and consequent present fluctuations, resulting in an unsteady and uneven pressure through the microchannels. As such this mechanism was discarded for future experimentations with higher internal pressures. Modular experiments, however, demonstrated that for the two different layouts neither of the studied concentrations would permit a laminar flow – the two solutions would mix upon a deformity of the micropillars (Figure 25). Initial calculations did demonstrate that with a Newtonian liquid the device would have a laminar flow using the syringe pump. Nevertheless, these calculations did not account for different resistances in parallel channels. When introducing matrigel, the internal pressure of the chip increased since the scaffold imposes resistance to flow, allowing perfusion through its porous structures only. This might have contributed to the observed deformation of the micropillars, as the liquid would preferentially flow across the microgaps instead of perfusing the scaffold. Figure 25Live-on-screen image of the model experiment using microbeads outcome of the 1º generation device demonstrating system collapse and fluidic mixture.
60 from the HPMEC cell line with this treatment probably due to the different culturing methods for the ECMs proteins. Hence, for the cancer channel, it was elected the Sulfo-SANPAH treatment and matrigel functionalization for the embedment of cells. Figure 29Lung adenocarcinoma cell line A549 metabolic activity (%) after 24h of culture. (A) Cell metabolic activity grown in tissue culture plates (TCPs) and in modified PDMS with O2 Plasma Treatment for 30 min (A) and 24h (B) prior to ECM embedment with A549 cell suspension functionalization with Collagen IV (3 mg. mL-1), Matrigel (4 mg.mL-1) or Fibronectin ( 1 mg.mL-1). The cell metabolic activity was determined by the resazurin reduction assay measuring the fluorescence of resorufin ( λex= 560 nm, λ em= 590 nm). Values show mean and standard deviation of four independent experiments. **p < 0.01***p < 0.001;****p < 0.0001. P-values were obtained using a one-way analysis of variance (ANOVA). (A) (B)
61 5.4. Cell culture in devices with syringe pump driven perfusion The in-plate assay results were translated to the microfluidic device. In the syringe pump driven device, initially, cell culture on-chip technique was executed as in microbeads assays. This meant that cell suspension was introduced in chip by tubing. Nevertheless, this technique enabled gas bubble introduction. By slightly pressing the tube, air enters the tubing, and it is transferred to the device’s microchannels once it is connected to the inlet. These “non-passive” air bubbles resulted in system disruption and cell morphology changes (Figure 30). Hence, the culturing method technique was shifted to directly pipetting the cells into the channel, significantly lowering bubble introduction, in the moment of seeding. Before introducing flow, cells appeared healthy and adjusted well to the environment created with the ECM proteins (Figure 31). Nevertheless, when connecting the tubing, a high percentage of cells could not endure the pressure caused and would detach from the surface. Figure 30Live-on-screen images of cell culture in the microfluidic device. (A) Day of cancer cell culture embedded in Matrigel (1:1) with cells still rounded. (B) One day after cell culture in chip demonstrating major cell loss and morphological changes in cancer cells. (A) (B) 600 µm 100 µm Figure 31Phase contrast images 24h after A549 lung cancer cells were seeded in the microfluidic chip displaying good adaptation and morphology to the environment provided. (A) (4x) (B) (10x). (A) (B)
62 Yet, channels were perfused to nurture the remaining cells envisioning their proliferation. Daily live-onscreen monitorization demonstrated that cells would either quickly be washed from the device or change their morphology using a flow rate of 15 µL.h-1. These changes resulted from the low flow rate that translated into non-proper shear stress of cells, which is essential for their maintenance. However, if a higher flow rate was used, cells would be washed away from the channel. This is probably owned to the initial pressure caused by tubing placement that would cause cells to detach from the coating, as mentioned. Additionally, due to the bubble trap mechanism saturation (2 to 3 days post-flow), gas bubbles would be introduced in the system once again (Figure A 1, Annex). Their formation and accumulation in microfluidic devices impose a critical barrier especially regarding long-term cell culture experiments [113]. This event can occur due to different reasons but one important factor, in this case, was the change in temperature (RT vs incubator temperature). The dissolved gas in the cell medium is emitted as a result of the changing of temperature during experiments in form of gas bubbles. To note that before degassing, the cell culture medium was warmed to 37 ºC. Nonetheless, the whole process of filing the syringes and removing all bubbles in the syringe itself allowed this warmed media to equilibrate to room temperature. Other factors – as hydrophobic properties of the PDMS – can also affect bubble generation. Once formed, they are extremely difficult to remove and result in abrupt changes in flow, being also capable of inducing membrane cell damage or even death [113]–[115]. Interestingly, the remaining cells in these disturbed conditions had a preference towards the circular pillars or the channel gaps, as they would become entrapped. Figure 32Confocal micrographs 10x magnification of A549 cancer cells and HPMEC cells grown on the microfluidic device and stained for nuclei (DAPI, blue) and cytoplasm (GFP, green). Images were acquired after 5 days with continuous flow (15 µL∙h-1.) in the metastasis-on-a-chip with automated perfusion. DAPI/Nucleus GFP/Cytoplasm PRotein MERGE
63 With all these impairments, at day 5 post flow, the endothelial barrier was not formed, and its channel would be almost empty. Aside from the generation of gas bubbles, this might have occurred as a result of non-proper shear stress given that fibronectin and cells need it to maintain integrity and function. Nonetheless, as shown, the integrity of the system would fail (Figure 32), and in this design, the automated mechanism pumping approach did not work, because of the overall sensitivity.
64 5.5. Cell culture in devices with gravity-driven perfusion With the same PDMS modification and functionalization methodology, the gravity-driven device was seeded as described. The shift from perfusion methodology rapidly changed the scenario in the same device layout. Here, gas bubbles rarely form inside the microchannels, unless manually introducing them inappropriate handling. For this, the device performance improved drastically, as cells rapidly adjust to the conditions provided and proliferate (Figure 33). 2 days DAPI/Nucleus FITC/ Cytokeratin MERGE 5 days 7 days Figure 33Confocal micrographs (10x magnification) of A549 cancer cells embedded in Matrigel ( 4 mg.mL-1) and grown on the microfluidic device. Cells were stained for nuclei (DAPI, blue) and FITC-labelled antibody against cytoplasmatic CK (Green) expression. Images were acquired after 2, 5 and 7 days after cell culture under a semi-static environment.
65 The microfluidic MOC system was designed to recapitulate the 3D architecture of human micrometastases and vasculature, to provide a more physiologically relevant model towards nanoformulation testing and to study the aggressiveness of lung cancer cells. For this, lung cancer cells (A549 or H1650) were grown in one microchannel and kept in non-perfusion conditions while endothelial cells, HPMEC, were grown in a monoculture over a thin fibronectin coating with perfusion and shear stress conditions. The morphology of the adenocarcinoma cells and HPMEC monolayer were evaluated by immunocytochemistry using a FITC-labelled antibody against cytoplasmatic CK expression, and rhodamine-labelled phalloidin to stain F-actin, respectively. DAPI was used to counterstain cells and to visualise the nucleus in both cases to make sure the signal obtained came from cells and not artefacts. Figure 33 demonstrates the system evolution in the tumour microchannel. Under the provided conditions, A549 cells were able to adapt and proliferate on the modified PDMS subtract. Incorporating A549 cells in matrigel matrix appeared to influence cellular display and differentiation. In fact, as the system evolves, tumour cells self-assemble with matrigel into cell clusters. In this tumour microchannel, cell proliferation is visible and reached an almost confluent state at day 7. To note that this channel is not under dynamic forces, since both reservoirs are filled just to keep culture hydrated, as such, it was hypothesized that if the channel was submitted to perfusion a higher proliferation could be achieved. Figure 34Confocal images and orthogonal cross-section views of A549 cells grown on the microfluidic metastasis-on-a-chip after 2 days of culture, stained for CK expression (green) and counterstained nuclei (DAPI, blue).
66 Nevertheless, 3D displayed cells entail a higher degree of complexity and mimic the in vivo condition more closely. Specifically, for metastasis, it is known that cells depend on their environment to achieve multiple ends as intravasation [50]. Hence, these cell clusters ensure a more complex environment on-chip. The 3D display of cancer cells was confirmed in image analysis of the obtained confocal scans in 100 stacks, that covered all the 100 µm of microchannels depth. Orthogonal views of the aggregate’s regions (Figure 34) suggest that these are 3D-displayed, and with an average size of ± 47 µm. Due to cancer cell aggregates, some areas of the tumour microchannel are empty, which was considered to have an impact on their progression since in a short span they self-aggregated and started to redirect themselves towards the micropillars – the blood DAPI/Nucleus TRITC/Actin MERGE 2 days post-flow 5 days post-flow 7 days post-flow Figure 35Immunocytochemistry in device of HPMEC cells growth, proliferation, and degradation on the Sulfo-SANPAH modified, fibronectin functionalized microfluidic device. Cells were stained for nuclei (DAPI, blue) and actin filaments (phalloidin, red).Images were acquired after 2, 5 and 7 days under periodic flow rate of 40 µL.min -1 by laser confocal microscopy with 10x magnification.
67 vessel region, highlighting their proliferative and invasive behaviour. Cells in both, the endothelial and the cancer channel, proliferate rapidly in the conditions provided. Differently from the tumour channel, the blood vessel channel is submitted to fluid dynamics forces since media perfusion is applied. ECs proliferation and overall display inside the microchannel is influenced highly by shear stress that ultimately contributes to the evolution of the vascular channel as reported in literature [116]. From day 2-post flow to day 5 post-flow, it is observed a considerable cellular proliferation but also actin filaments show enhanced expression and realignment (Figure 35). This may be due to ECs response to the patterns of imposed dynamic forces and consequent rearrangement of their morphology and cytoskeleton as found in literature [116], [117]. In fact, ECs not only display themselves against the micropillars but form an endothelial lumen, covering the top and bottom of the channel, the micropillars region and the opposite wall, while the middle remains hollow as supported by orthogonal views of the regions of interest. This was probably due to the FN thin coating method in the modified PDMS surface, but also by flow action. Orthogonal views from day 2-post flow (Figure 36) to day 5 post-flow (Figure 37) demonstrate the evolution of the endothelial tubing. After two days with the presence of shear forces, the endothelial cells begin to Figure 36Confocal micrograph and correspondent orthogonal cross-section views of HPMEC cells grown on the modified and functionalized microfluidic metastasis-on-a-chip after 2 days of culture. Orthogonal views are set in the micropillar area (endothelial barrier site). Cells are stained for actin filaments (phalloidin, red) and counterstained nuclei (DAPI, blue).
68 form an endothelial monolayer with, yet some “opened” spots as demonstrated in Figure 36. This suggests that the barrier is still in formation process. Yet, at 5 days post-perfusion, a complete endothelial barrier is formed against all the micropillar area covering all the 100 µm depth of the microchannel (Figure 37) and 6800 µm length, forming a tubularlike form. This outcome might have arisen because of the tilting step to promote micropillar coverage with ECs, combined with the PDMS modification and functionalization by thin FN coating given that the latter enhances early attachment of ECs to artificial surfaces [118]. This since, the FN coating excess was washed, meaning that only the ECM proteins that were connected to the surface remained, while the middle of the channel, that was not crosslinked by Sulfo-SAPAH, became hollow. Consequently, the cells would preferentially move towards the fibronectin-coated areas. Moreover, although it is visible an enhanced F-actin signal, these cells (at the bottom of the channel) do not present themselves with a uniformly oriented direction. This occurs when ECs experience laminar shear stress within the range of the present device (5-10 dynes/cm2) and their shape shifts from polygonal to ellipsoidal [116], [117]. Figure 37Laser confocal microscopy images and correspondent orthogonal cross-section views of micropillar region (endothelial barrier site) of HPMEC cells grown on the Sulfo-SANPAH modified, Fibronectin functionalized microfluidic metastasis-on-a-chip after 5 days in culture under periodic flow rate ( 40 µL.mL-1). Cells are marked for actin filaments (phalloidin, red) and for nuclei (DAPI, blue).
69 Here, ECs present both polygonal and ellipsoidal forms and their direction is not clear probably owing to the periodic shear stress present. Nevertheless, at day 7-post flow, the overall obtained result is of cell loss and barrier compromise (Figure 38), which can be due to the fact that fibronectin and ECs cells need continuous perfusion in order to maintain their integrity. Yet, with these reservoirs, only periodic perfusion would occur. Findings displayed in Figure 38 suggest that fluid mechanical forces have a direct impact on endothelial cell structure and function, overtime since it is clear the decrease in cell density and the irregular shape of ECs[116]–[120]. Yet, it was hypothesised that the cancer channel evolution might have a crucial role in the degraded HPMEC microchannel on day 7 but also surface stiffness could have contributed to it as well [119], [121]. Cell seeding is performed at the same density in both channels. Yet quickly the tumour channel cell number surpasses the endothelial channel (Figure 39). The culturing method of the different microchannels may influence this outcome but the nature of each cell line probably played an important part in this outcome. Tumour cells are known to surpass adversities and stressful environments of hypoxia and low nutrient [3],[122] while endothelial cells as HPMEC need certain important parameters as shear Figure 38Immunocytochemistry confocal image and micropillar orthogonal cross-section views of HPMEC cells grown on microfluidic metastasis-on-a-chip after 7 days in culture and periodic flow rate ( 40 µL.mL -1 ). Cells are stained for actin filaments (phalloidin, red) and counterstained for nuclei (DAPI, blue).
76 location and penetrate the 5 µm gaps between pillars highlighting their deformability and morphological adjustment (cytoplasmatic structure appear elongated) in order to intravasate. Consequently, can invade the endothelial microchannel. It is postulated as in intravasation, tumour cells induce some level of degradation to the membrane basement and are capable to enter blood circulation, and as such this same scenario seems to be detected in this dynamic platform. Interestingly, Epithelial-mesenchymal transition is a known complex process by which epithelial cells lose their cell polarity, adhesion, change their phenotypic characteristics and gain invasive and migratory properties meaning that they differentiate into mesenchymal cells. As such, EMT plays an important part in the metastatic process [125], [126]. Cells can be identified as exhibiting epithelial phenotypes, epithelial-mesenchymal transition phenotypes ( Both epithelial and mesenchymal phenotypes are present in one cell) or mesenchymal phenotypes. Cytokeratin is known as an epithelial cell marker [126], [127]. Here, in the tumour microchannel, CK expression is assertive which implies their epithelial state. Yet, the cells that are intravasating to the endothelial microchannel (Figure 44, red arrows) seem to decrease in their CK expression signal. This can probably be due to the EMT process. During the transition, epithelial markers as cytokeratin are known to be downregulated while mesenchymal phenotypes are upregulated [126], [128]–[130]. Consequently, the acquired results could potentially indicate an occurrence of an EMT process due to its Figure 45Immunocytochemistry confocal image and r orthogonal cross-section views of A549 cell cluster cells grown on endothelial microchannel after 6 days in culture and periodic flow rate ( 40 µL.mL-1). Cells are stained for CK expression and DAPI.
77 low CK expression, suggesting that these cells are shifting their phenotype and gaining invasive and migratory properties. Apart from multiple intravasation events (Figure A 7, Annex), migratory events of cluster cells were detected as demonstrated in Figure 45. Orthogonal views demonstrate that these are 3D cell aggregates. As such, these might also point to the detachment of cancer cells and their dissemination through the blood vessel channel. These sorts of intravasating and dissemination events were detected on days 6, 7 and 10 post-flow, which correlates with the previous mention impact of the cancer channel in the endothelial barrier. Hence, as the tumour mass evolves together with its microenvironment, it is observed a negative impact on the endothelial barrier and further enhancement of the invasiveness of lung cancer cells.
78 CHAPTER 6 Conclusions and Future Perspectives
79 6. Conclusions and future perspectives The main aim of this work was to develop, optimize and validate a vascularized microfluidic organ-on-achip system that recapitulates truthfully the in vivo micrometastasis. The results from this thesis highlight the crucial link between engineering and biology fields, in order to create a robust and rational microfluidic device and how the physical principles correlate to the outcome in cell culture in-chip. Here it was demonstrated that the 600 µm wide channels comprising 5 µm gaps between the pillars provide a solid platform to create a MOC system with gravity-driven perfusion. PDMS modification with Sulfo-SANPAH water reconstituted reagent was proven the most efficient and free-of-toxicity methodology for this organ-on-a-chip platform enabling peptide conjugation to PDMS through photochemical immobilization of functional -NHS groups on PDMS’s surface. Additionally, to have a replicative 3D tumour tissue architecture and an endothelial vessel, different culturing methods were tested and optimized. Fibronectin thin coating proved to be the best endothelial functionalization method with cell metabolic activity exceeding 115% (p<0.001) and the matrigel embedded method demonstrated the best cell metabolic activity (105%, p<0.001) for lung adenocarcinoma cells. The conditions were translated to the microfluidic device and quickly begin to display promising results of cell adaptation and proliferation. The lung cancer cells self-assembled into matrix embedded clusters with a higher degree of biologic and physiologic complexity, while the endothelial cells formed an endothelial-like tubular form. The integrity of the chip was assessed through multiple testing to ensure a robust platform to perform various studies. The chip metabolic activity displayed good results from day 2 to day 5 having only a degradation of the endothelial barrier at day 7, being supported by the endothelial barrier integrity results in which fluorescence intensity values are higher seven days-post flow when compared to the fifth day ( formed endothelial barrier). Having a defined window of testing in the in vitro model, lipidic nanoemulsions in their non-functionalized form were introduced in the endothelial microchannel simulating intravenous administration and incubated. The results show that even without any sort of functionalization these nanoemulsions can penetrate the endothelial barrier and reach the tumour site. Moreover, intravasation-in-chip was detected by confocal imaging and downregulation of the epithelial marker cytokeratin expression that highlighted the effect of a 3D cancer cluster on endothelial barriers, as cancer phenotype shifts in the EMT process. To conclude, future work will entail further phenotype testing in the in vitro model using both epithelial and mesenchymal phenotype markers to extend the study of the EMT process of intravasation. Moreover,
80 in the current platform, an automated mechanism will be associated with the reservoirs to enhance the dynamic environment and improve the system. Within the same present model cancer cellfibroblast coculture will be implemented to complement the complexity of the system, in which cell aggressiveness, phenotype and motility will also be assessed. At the same time, the functionalized nanoemulsions will be tested for their capacity to mediate a specific interaction between the overexpressed TAS1R3 masker and the functionalized aptamer and their consequent efficacy to interrupt micrometastasis.
81 CHAPTER 7 References
82 7. References [1] H. Sung et al. , “Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries,” CA: A Cancer Journal for Clinicians , vol. 71, no. 3, pp. 209–249, May 2021, doi: 10.3322/CAAC.21660. [2] S. Tran, P.-J. DeGiovanni, B. Piel, and P. Rai, “Cancer nanomedicine: a review of recent success in drug delivery,” Clinical and Translational Medicine , vol. 6, no. 1, Dec. 2017, doi: 10.1186/S40169-017-0175-0. [3] D. Hanahan and R. A. Weinberg, “Hallmarks of cancer: the next generation,” Cell , vol. 144, no. 5, pp. 646–674, 2011, doi: 10.1016/j.cell.2011.02.013. [4] C. Gridelli et al. , “Non-small-cell lung cancer,” Nature Reviews Disease Primers , vol. 1, May 2015, doi: 10.1038/nrdp.2015.9. [5] I. T. Lennes et al. , “Non-Small Cell Lung Cancer Clinical Practice Guidelines in Oncology NCCN Clinical Practice Guidelines in Oncology on Non-Small Cell Lung Cancer,” Journal of the National Comprehensive Cancer Network , vol. 8, no. 7, pp. 740–741, 2010, doi: 10.6004/jnccn.2010.0056. [6] R. S. Herbst, D. Morgensztern, and C. Boshoff, “The biology and management of non-small cell lung cancer,” Nature , vol. 553, no. 7689, pp. 446–454, Jan. 2018, doi: 10.1038/nature25183. [7] N. Duma, R. Santana-Davila, and J. R. Molina, “Non–Small Cell Lung Cancer: Epidemiology, Screening, Diagnosis, and Treatment,” Mayo Clinic Proceedings , vol. 94, no. 8, pp. 1623–1640, Aug. 2019, doi: 10.1016/j.mayocp.2019.01.013. [8] C. Zappa and S. A. Mousa, “Non-small cell lung cancer: Current treatment and future advances,” Translational Lung Cancer Research , vol. 5, no. 3, p. 288, Jun. 2016, doi: 10.21037/tlcr.2016.06.07. [9] M. Yuan, L. L. Huang, J. H. Chen, J. Wu, and Q. Xu, “The emerging treatment landscape of targeted therapy in non-small-cell lung cancer,” Signal Transduction and Targeted Therapy , vol. 4, no. 1, pp. 1–14, 2019, doi: 10.1038/s41392-019-0099-9. [10] J. Fares, M. Y. Fares, H. H. Khachfe, H. A. Salhab, and Y. Fares, “Molecular principles of metastasis: a hallmark of cancer revisited,” Signal Transduction and Targeted Therapy , vol. 5, no. 1, pp. 1–17, Dec. 2020, doi: 10.1038/s41392-020-0134-x.
83 [11] Y.-H. V. Ma, K. Middleton, L. You, and Y. Sun, “A review of microfluidic approaches for investigating cancer extravasation during metastasis,” Microsystems & Nanoengineering , vol. 4, no. 1, Apr. 2018, doi: 10.1038/micronano.2017.104. [12] J. J. F. Sleeboom, H. E. Amirabadi, P. Nair, C. M. Sahlgren, and J. M. J. den Toonder, “Metastasis in context: Modeling the tumor microenvironment with cancer-on-a-chip approaches,” DMM Disease Models and Mechanisms , vol. 11, no. 3, Mar. 2018, doi: 10.1242/dmm.033100. [13] J. D. O’Flaherty et al. , “Circulating tumour cells, their role in metastasis and their clinical utility in lung cancer,” Lung Cancer , vol. 76, no. 1, pp. 19–25, 2012, doi: 10.1016/j.lungcan.2011.10.018. [14] V. Plaks, C. D. Koopman, and Z. Werb, “Circulating tumor cells,” Science , vol. 341, no. 6151, American Association for the Advancement of Science, pp. 1186–1188, 2013. doi: 10.1126/science.1235226. [15] C. Dive and G. Brady, “SnapShot: Circulating Tumor Cells,” Cell , vol. 168, no. 4, Cell Press, pp. 742-742.e1, Feb. 09, 2017. doi: 10.1016/j.cell.2017.01.026. [16] A. F. Chambers, A. C. Groom, and I. C. MacDonald, “Dissemination and growth of cancer cells in metastatic sites,” Nature Reviews Cancer , vol. 2, no. 8, pp. 563–572, Aug. 2002, doi: 10.1038/nrc865. [17] S. RL, M. KD, and J. A, “Cancer Statistics, 2017,” CA: a cancer journal for clinicians , vol. 67, no. 1, pp. 7–30, Jan. 2017, doi: 10.3322/CAAC.21387. [18] H. FR et al. , “Lung cancer: current therapies and new targeted treatments,” Lancet (London, England) , vol. 389, no. 10066, pp. 299–311, Jan. 2017, doi: 10.1016/S01406736(16)30958-8. [19] J. X and K. MJ, “Clinical significance of micrometastasis in lung and esophageal cancer: a new paradigm in thoracic oncology,” The Annals of thoracic surgery , vol. 74, no. 1, pp. 278–284, 2002, doi: 10.1016/S0003-4975(01)03376-8. [20] K. J. Luzzi et al. , “Multistep Nature of Metastatic Inefficiency: Dormancy of Solitary Cells after Successful Extravasation and Limited Survival of Early Micrometastases,” The American Journal of Pathology , vol. 153, no. 3, pp. 865–873, Sep. 1998, doi: 10.1016/S0002-9440(10)65628-3.
84 [21] A. D. Theocharis, S. S. Skandalis, C. Gialeli, and N. K. Karamanos, “Extracellular matrix structure,” Advanced Drug Delivery Reviews , vol. 97, pp. 4–27, Feb. 2016, doi: 10.1016/J.ADDR.2015.11.001. [22] N. K. Karamanos, A. D. Theocharis, T. Neill, and R. v. Iozzo, “Matrix modeling and remodeling: A biological interplay regulating tissue homeostasis and diseases,” Matrix Biology , vol. 75–76, pp. 1–11, Jan. 2019, doi: 10.1016/J.MATBIO.2018.08.007. [23] J. D. O’Flaherty et al. , “Circulating tumour cells, their role in metastasis and their clinical utility in lung cancer,” Lung Cancer , vol. 76, no. 1, pp. 19–25, Apr. 2012, doi: 10.1016/j.lungcan.2011.10.018. [24] S. van Helvert, C. Storm, and P. Friedl, “Mechanoreciprocity in cell migration,” Nature cell biology , vol. 20, no. 1, pp. 8–20, 2018, doi: 10.1038/s41556-017-0012-0. [25] G. Follain, N. Osmani, A. Azevedo, G. Allio, L. M.-D. cell, and undefined 2018, “Hemodynamic forces tune the arrest, adhesion, and extravasation of circulating tumor cells,” Developmental Cell , vol. 45, no. 1, pp. 33–52, 2018, doi: 10.1016/j.devcel.2018.02.015. [26] V. M. Freitas, G. Hilfenhaus, and M. L. Iruela-Arispe, “Metastasis of Circulating Tumor Cells: Speed Matters,” Developmental Cell , vol. 45, no. 1, pp. 3–5, Apr. 2018, doi: 10.1016/J.DEVCEL.2018.03.005. [27] M. Reck, S. Popat, N. Reinmuth, D. de Ruysscher, K. M. Kerr, and S. Peters, “Metastatic non-small-cell lung cancer (NSCLC): ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up,” Annals of Oncology , vol. 25, pp. iii27–iii39, Sep. 2014, doi: 10.1093/annonc/mdu199. [28] K. C. Arbour and G. J. Riely, “Systemic therapy for locally advanced and metastatic nonsmall cell lung cancer: A review,” JAMA - Journal of the American Medical Association , vol. 322, no. 8, pp. 764–774, Aug. 2019, doi: 10.1001/jama.2019.11058. [29] M. Saxena and Gerhard Christofori, “Rebuilding cancer metastasis in the mouse,” Molecular Oncology , vol. 7, no. 2, pp. 283–296, 2013, doi: 10.1016/j.molonc.2013.02.009. [30] D. Laskowska, E. Ziokowska, and K. Mitura, “Nanotechnology in oncologycurrent state of knowledge,” Engineering of Biomaterials , vol. 22, no. 152, p. undefined, 2019.
85 [31] S. Li et al. , “Nanotechnology: Breaking the Current Treatment Limits of Lung Cancer,” Advanced Healthcare Materials , vol. 10, no. 12, p. 2100078, Jun. 2021, doi: 10.1002/ADHM.202100078. [32] C. Fornaguera and S. Borros, “Nanomedicine in NSCLC.” https://encyclopedia.pub/item/revision/5ff1a31c7817014f3878dfe5c4c08453 (accessed Oct. 21, 2021). [33] R. G. Harrison, M. J. Greenman, F. P. Mall, and C. M. Jackson, “Observations on the living developing nerve fiber,” The Anatomical Record , vol. 4, no. 1, pp. 116–128, Dec. 2016, doi: 10.1002/ar.1090010503. [34] D. Caballero, S. Kaushik, V. M. Correlo, J. M. Oliveira, R. L. Reis, and S. C. Kundu, “Organon-chip models of cancer metastasis for future personalized medicine: From chip to the patient,” Biomaterials , vol. 149, pp. 98–115, Dec. 2017, doi: 10.1016/j.biomaterials.2017.10.005. [35] M. Kapałczyńska et al. , “2D and 3D cell cultures – a comparison of different types of cancer cell cultures,” Archives of Medical Science : AMS , vol. 14, no. 4, p. 910, 2018, doi: 10.5114/AOMS.2016.63743. [36] S. Breslin and L. O’Driscoll, “Three-dimensional cell culture: the missing link in drug discovery,” Drug Discovery Today , vol. 18, no. 5, pp. 240–249, Mar. 2013, doi: 10.1016/J.DRUDIS.2012.10.003. [37] M. L. Coluccio et al. , “Microfluidic platforms for cell cultures and investigations,” Microelectronic Engineering , vol. 208, pp. 14–28, Mar. 2019, doi: 10.1016/j.mee.2019.01.004. [38] K. Kilian, B. Bugarija, B. T. Lahn, and M. Mrksich, “Geometric cues for directing the differentiation of mesenchymal stem cells,” Proceedings of the National Academy of Sciences , vol. 107, no. 11, pp. 4872–4877, 2010, doi: 10.1073/pnas.0903269107. [39] G. Mahmud et al. , “Directing cell motions on micropatterned ratchets,” Nature physics , vol. 5, no. 8, pp. 606–612, 2009, doi: 10.1038/nphys1306. [40] D. J and B. JS, “Modelling glandular epithelial cancers in three-dimensional cultures,” Nature reviews. Cancer , vol. 5, no. 9, pp. 675–688, Sep. 2005, doi: 10.1038/NRC1695. [41] K. Maddison and A. Clarke, “New approaches for modelling cancer mechanisms in the mouse,” The Journal of Pathology: A Journal of the Pathological Society of Great Britain and Ireland , vol. 205, no. 2, pp. 181–193, Jan. 2005, doi: 10.1002/path.1698.
92 [105] T. Ohashi, D. P. Kiehart, and H. P. Erickson, “Dynamics and elasticity of the fibronectin matrix in living cell culture visualized by fibronectin-green fluorescent protein,” Proceedings of the National Academy of Sciences , vol. 96, no. 5, pp. 2153–2158, Mar. 1999, doi: 10.1073/PNAS.96.5.2153. [106] M. Nakamura, M. Mie, H. Mihara, M. Nakamura, and E. Kobatake, “Construction of multifunctional extracellular matrix proteins that promote tube formation of endothelial cells,” Biomaterials , vol. 29, no. 20, pp. 2977–2986, Jul. 2008, doi: 10.1016/j.biomaterials.2008.04.006. [107] M. Shimizu, K. Minakuchi, M. Moon, and J. Koga, “Difference in interaction of fibronectin with type I collagen and type IV collagen,” Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology , vol. 1339, no. 1, pp. 53–61, 1997, doi: 10.1016/S0167-4838(96)00214-2. [108] M. Sgarioto, P. Vigneron, J. Patterson, F. Malherbe, M. D. Nagel, and C. Egles, “Collagen type I together with fibronectin provide a better support for endothelialization,” Comptes Rendus - Biologies , vol. 335, no. 8, pp. 520–528, Aug. 2012, doi: 10.1016/j.crvi.2012.07.003. [109] F. Akther, S. B. Yakob, N. T. Nguyen, and H. T. Ta, “Surface Modification Techniques for Endothelial Cell Seeding in PDMS Microfluidic Devices,” Biosensors , vol. 10, no. 11, Nov. 2020, doi: 10.3390/bios10110182. [110] C. C. Wu, C. Y. Yuan, and S. J. Ding, “Effect of polydimethylsiloxane surfaces silanized with different nitrogen-containing groups on the adhesion progress of epithelial cells,” Surface and Coatings Technology , vol. 205, no. 10, pp. 3182–3189, Feb. 2011, doi: 10.1016/j.surfcoat.2010.11.036. [111] A. Shakeri, S. Khan, and T. F. Didar, “Conventional and emerging strategies for the fabrication and functionalization of PDMS-based microfluidic devices,” Lab on a Chip , vol. 21, no. 16, pp. 3053–3075, Aug. 2021, doi: 10.1039/d1lc00288k. [112] B. Li, J. Chen, and J. H. C. Wang, “RGD peptide-conjugated poly(dimethylsiloxane) promotes adhesion, proliferation, and collagen secretion of human fibroblasts,” Journal of Biomedical Materials Research - Part A , vol. 79A, no. 4, pp. 989–998, Dec. 2006, doi: 10.1002/jbm.a.30847.
93 [113] J. H. Sung and M. L. Shuler, “Prevention of air bubble formation in a microfluidic perfusion cell culture system using a microscale bubble trap,” Biomedical Microdevices , vol. 11, no. 4, pp. 731–738, Feb. 2009, doi: 10.1007/s10544-009-9286-8. [114] E. W. Young and D. J. Beebe, “Fundamentals of microfluidic cell culture in controlled microenvironments,” Chemical Society Reviews , vol. 39, no. 3, pp. 1036–1048, Feb. 2010, doi: 10.1039/B909900J. [115] W. Zheng, Z. Wang, W. Zhang, and X. Jiang, “A simple PDMS-based microfluidic channel design that removes bubbles for long-term on-chip culture of mammalian cells,” Lab on a Chip , vol. 10, no. 21, pp. 2906–2910, Nov. 2010, doi: 10.1039/C005274D. [116] C. F. Dewey, S. R. Bussolari, M. A. Gimbrone, and P. F. Davies, “The Dynamic Response of Vascular Endothelial Cells to Fluid Shear Stress,” Journal of Biomechanical Engineering , vol. 103, no. 3, pp. 177–185, Aug. 1981, doi: 10.1115/1.3138276. [117] A. R. Wechezak, R. F. Viggers, and L. R. Sauvage, “Fibronectin and F-actin redistribution in cultured endothelial cells exposed to shear stress.,” Laboratory Investigation; a Journal of Technical Methods and Pathology , vol. 53, no. 6, pp. 639–647, Dec. 1985. [118] B. L. Ivarsson, R. P. Cambria, J. Megerman, and W. M. Abbott, “Fibronectin enhances early shear stress resistance of seeded adult human venous endothelial cells,” Journal of Surgical Research , vol. 47, no. 3, pp. 203–207, Sep. 1989, doi: 10.1016/00224804(89)90108-X. [119] O. Thoumine, R. M. Nerem, and P. R. Girard, “Changes in organization and composition of the extracellular matrix underlying cultured endothelial cells exposed to laminar steady shear stress.,” Laboratory Investigation; a Journal of Technical Methods and Pathology , vol. 73, no. 4, pp. 565–576, Oct. 1995. [120] F. Tovar-Lopez et al. , “A Microfluidic System for Studying the Effects of Disturbed Flow on Endothelial Cells,” Frontiers in Bioengineering and Biotechnology , vol. 7, p. 81, Apr. 2019, doi: 10.3389/fbioe.2019.00081. [121] J. D. Mih, A. S. Sharif, F. Liu, A. Marinkovic, M. M. Symer, and D. J. Tschumperlin, “A multiwell platform for studying stiffness-dependent cell biology,” PloS one , vol. 6, no. 5, pp. 1–10, 2011, doi: 10.1371/journal.pone.0019929. [122] M. Hockel and P. Vaupel, “Biological consequences of tumor hypoxia,” Seminars in Oncology , vol. 28, pp. 31–41, Apr. 2001, doi: 10.1016/S0093-7754(01)90211-8.
94 [123] J. Fang, H. Nakamura, and H. Maeda, “The EPR effect: Unique features of tumor blood vessels for drug delivery, factors involved, and limitations and augmentation of the effect,” Advanced Drug Delivery Reviews , vol. 63, no. 3, pp. 136–151, Mar. 2011, doi: 10.1016/J.ADDR.2010.04.009. [124] V. Torchilin, “Tumor delivery of macromolecular drugs based on the EPR effect,” Advanced Drug Delivery Reviews , vol. 63, no. 3, pp. 131–135, Mar. 2011, doi: 10.1016/J.ADDR.2010.03.011. [125] N. Gavert and A. Ben-Ze’ev, “Epithelial–mesenchymal transition and the invasive potential of tumors,” Trends in Molecular Medicine , vol. 14, no. 5, pp. 199–209, May 2008, doi: 10.1016/J.MOLMED.2008.03.004. [126] A. Nawshad, D. LaGamba, A. Polad, and E. D. Hay, “Transforming Growth Factor-β Signaling during Epithelial-Mesenchymal Transformation: Implications for Embryogenesis and Tumor Metastasis,” Cells Tissues Organs , vol. 179, no. 1–2, pp. 11–23, 2005, doi: 10.1159/000084505. [127] V. Tirino et al. , “TGF-β1 exposure induces epithelial to mesenchymal transition both in CSCs and non-CSCs of the A549 cell line, leading to an increase of migration ability in the CD133+ A549 cell fraction,” Cell Death & Disease , vol. 4, no. 5, pp. e620–e620, May 2013, doi: 10.1038/cddis.2013.144. [128] R. Strippoli, I. Benedicto, M. L. P. Lozano, A. Cerezo, M. López-Cabrera, and M. A. del Pozo, “Epithelial-to-mesenchymal transition of peritoneal mesothelial cells is regulated by an ERK/NF-κB/Snail1 pathway,” Disease Models & Mechanisms , vol. 1, no. 4–5, pp. 264– 274, Nov. 2008, doi: 10.1242/DMM.001321. [129] P. Savagner, “The epithelial–mesenchymal transition (EMT) phenomenon,” Annals of Oncology , vol. 21, no. 7, pp. 89–92, Oct. 2010, doi: 10.1093/ANNONC/MDQ292. [130] J. H. Kim et al. , “Transtorming growth factor β1 induces epithelial-to-mesenchymal transition of A549 cells,” Journal of Korean Medical Science , vol. 22, no. 5, pp. 898–904, Oct. 2007, doi: 10.3346/JKMS.2007.22.5.898.
95 CHAPTER 8 Annex
96 8. Annex (D) (C) (A) (B) Figure A 2Confocal micrographs of the disrupted system in the automated perfusion microfluidic device. (A) Brightfield; (B) Nuclei stained with DAPI ( blue); (C) GFP labeled lung cancer cells (green); (D) Merged. Figure A 1Confocal image ( magnification 10x) of the microfluidic device with cultured cell in both channels with visible system disruption due to bubble formation. Nucleus stained for DAPI (blue).
97 2 days post-flow 5 days post-flow 7 days post-flow Figure A 4Confocal images (10x magnification) of the full endothelial channel over time. Images were acquired in days 2, 5, 7 post-flow. Immunocytochemistry in chip displays blue marked nuclei (DAPI) and red F-actin filaments (TRICT). (B) 2 days post-flow 7 days post-flow 5 days post-flow Figure A 3Confocal images (10X magnification) of the full tumour channel (A549 cells) over time. Immunocytochemistry in chip, with cells expressing cytokeratin (Green) and nuclei (DAPI, blue).
98 Figure A 5Endothelial barrier integrity studies fluorescence studies profile ( Red line). Confocal micrograph displaying both microchannels with fluorescein. Figure A 6Graphic representation of flow change over time through the 5 µm gaps for the endothelial barrier testing.
99 Figure A 7Confocal micrographs (10X magnification) of an intravasation event. Immunocytochemistry in device displays the lung cancer cells with cytokeratin expression (CK, green) and nuclei counterstained (DAPI, blue). Endothelial channel is only marked for nuclei with DAPI (lower channel).