Дисертації з теми "3-dimensional cell culture"

Zebrafish (Danio rerio) have been extensively used over the past two decades to study muscle development, human myopathies and dystrophies, due to its higher degree of homology with human disease causing genes and genome. Despite its unique qualities, zebrafish have only been used as an in-vivo model for muscle development research, due to the limitations surrounding lack of a consistent isolation and culture protocol for zebrafish muscle progenitor cells in-vitro. Using different mammalian myoblast isolation protocols, a novel and robust protocol has been developed to successfully isolate and culture zebrafish skeletal muscle cells repeatedly and obtain differentiated long multi nucleated zebrafish myotubes. Commitment to myogenic lineage was confirmed by immuno-staining against muscle specific protein desmin, and expression pattern of different genetic markers regulating myogenesis. In order to recapitulate the in-vivo bio-physiological environment for zebrafish skeletal muscle cells in-vitro, these cells were successfully cultured in tissue engineered three dimensional (3D) constructs based on fibrin and collagen models. Maturation of tissue engineered collagen and fibrin based constructs was confirmed using the basic parameters described in the literature i.e. collagen three times greater contraction from the original width (Mudera, Smith et al. 2010) and fibrin constructs tightly coiled up to 4mm of diameter (Khodabukus, Paxton et al. 2007). In-vitro characterisation of zebrafish skeletal muscle cells showed hypertrophic growth of muscle mass compared to hyperplasic growth in-vivo as suggested for fish species in literature (Johnston 2006), which is different from human and other mammals. Comparative analysis of zebrafish muscle cells cultured in monolayer against cultured in 3D tissue engineered constructs showed significant increase in fusion index, nuclei per myotube (two-fold) and myotubes per microscopic frame (two-fold). Cells cultured in tissue engineered construct closely resembled in-vivo muscle in terms of their unidirectional orientation of myotubes. These tissue engineered 3D zebrafish skeletal muscle models could be used for various purposes such as drug screening, effect of different temperature extremes, studying underlined pathways involved in human diseases; and with further refinements it would potentially replace the need for studies on live fish in these areas.

Thesis (M.Phil.)--Hong Kong University of Science and Technology, 2004.
Includes bibliographical references (leaves 122-133). Also available in electronic version. Access restricted to campus users.

Abstract Desmoglein 3 (Dsg3) is an adhesion receptor in desmosomes, but relatively little is known about its role in cancer. In this study, the function of Dsg3 was investigated in oral squamous cell carcinoma (SCC) cell lines in vitro using locally established human leiomyoma tumor microenvironment (TME) matrices. Since Dsg3 has been identified as a key regulator in cell adhesion, we hypothesized that it may play a role in oral SCC cells adhesion and motility. Thus, one aim of the study was to explore this hypothesis by both gain and loss of function methods in four human buccal mucosa SCC SqCC/Y1 cell lines: transduction of vector control (Ct), full-length (FL) or two different C-terminally truncated Dsg3 mutants (Δ238 and Δ560). Live cell imaging was performed for 2D migration and 3D sandwich, alongside other assays. In 3D sandwich, we tested the effects of the monoclonal antibody, AK23, targeting the extracellular domain of Dsg3 in SqCC/Y1 cells. Our results showed that loss of Dsg3 disrupted cell adhesion and protein expression. In 2D assays, FL and Dsg3 mutants migrated faster with higher accumulated distances than Ct. In contrast with 2D, mutants showed accelerated invasion over the Ct in 3D models. The AK23 antibody inhibited only the invasion of FL cells. The TME in vivo consists of cellular and matrix elements playing a leading role in carcinoma progression. To study carcinoma cells invasion in vitro, mouse Matrigel® and rat type 1 collagen are the most commonly used matrices in 3D models. Since they are non-human in origin, they do not perfectly mimic human TME. To address this, we have developed a solid organotypic myoma disc model derived from human uterus leiomyoma tumor. Here, we introduce a novel Myogel, prepared from leiomyoma similar to Matrigel®. We validated Myogel for cell-TME interactions in 3D models, using SqCC/Y1 and HSC-3 cell lines. Compared with Matrigel® and type I collagen, oral SCC cell lines invaded more efficiently in Myogel containing matrices. This study describes promising 3D models using human TME mimicking Myogel which is suitable to analyze oral SCC cells both in carcinoma monocultures and in co-cultures, such as with TME fibroblasts. We also introduce a possible novel therapeutic target against Dsg3 to suppress cancer cell invasion
Tiivistelmä Desmogleiini 3 (Dsg3) on desmosomien adheesioreseptori, jonka merkityksestä syövässä tiedetään vähän. Koska Dsg3 on tärkeä epiteelisolujen välisissä liitoksissa, oletimme sillä olevan vaikutusta myös suun karsinoomasolujen tarttumisessa ja niiden liikkuvuudessa. Testasimme hypoteesiamme muuttamalla Dsg3:n toimintaa ihmisen posken karsinoomasolulinjassa SqCC/Y1, josta oli aiemmin valmistettu neljä erilaista muunnosta: tyhjän vektorin sisältävä kontrollisolulinja (Ct), kokopitkää Dsg3 tuottava solulinja (FL), sekä kaksi Dsg3 C-päästä lyhennettyä mutanttisolulinjaa (Δ238 ja Δ560). Immunofluoresenssi-menetelmää käyttäen analysoimme solulinjoissamme solujen välisiä liitoksia. Lisäksi mittasimme solujen liikkeitä 2D-migraatio- ja 3D-sandwich-kokeissa. Testasimme myös Dsg3:n solunulkoista osaa tunnistavan monoklonaalisen vasta-aineen (AK23) vaikutusta solujen invaasioon. Osoitimme, että Dsg3:n rakenteen muuttaminen ja toiminnan estyminen häiritsi solujen tarttumista. 2D-kokeissa sekä FL että mutanttilinjat (Δ238 ja Δ560) migroivat kontrollisoluja nopeammin ja pidemmälle, mutta 3D-kokeissa vain mutanttilinjat invasoituivat kontrollisoluja tehokkaammin. AK23-vasta-aine esti vain FL-solujen invaasiota. Syöpäsolujen 3D-invaasiota mittaavissa kokeissa käytetään yleensä hiiren kasvaimesta valmistettua kaupallista Matrigeeliä® tai rotan kudoksista eristettyä tyypin I kollageenia. Tutkimusryhmämme on jo aiemmin kehittänyt organotyyppisen myoomamallin, jossa valmistamme myoomakudosnapit ihmisen kohdun leiomyoomakasvaimista. Tässä työssä valmistimme leiomyoomasta Myogeelia, vertasimme sitä Matrigeeliin®, sekä tutkimme tarkemmin Myogeeli-valmisteen soveltuvuutta 3D-tutkimuksiin. Totesimme, että kielen (HSC-3) ja posken (SqCC/Y1) karsinoomasolut invasoituivat tehokkaimmin Myogeeli-pitoisissa matrikseissa kuin Matrigeeliä® tai kollageeniä sisältävissä kasvatusalustoissa. Tutkimustulostemme perusteella Myogeeli-pohjaiset 3D-mallit soveltuvat hyvin sekä syöpäsolulinjojen invaasiotutkimuksiin että yhteisviljelmiin, joissa syöpäsoluja viljellään yhdessä syöpäkasvaimen ympärillä olevien solujen, kuten fibroblastien, kanssa

Environ 90% des candidats-médicaments échouent en phase clinique, pour des raisons d’efficacité ou de toxicité qui impliquent souvent le système nerveux central (SNC). Ce fort taux d’échec souligne un manque de pertinence des modèles expérimentaux utilisés en amont, dont les modèles in vitro de cellules humaines. En effet, ces derniers ne prennent pas en compte toute la complexité du SNC, où les neurones organisés en 3 dimensions (3D) interagissent avec leur microenvironnement composé de cellules, de facteurs solubles et des molécules de la matrice extracellulaire (MEC). Les objectifs de ce travail étaient i) d’étudier l’influence de ces trois composantes du microenvironnement sur les cellules neuronales dans des modèles cérébraux in vitro par imagerie cellulaire automatisée, et ii) de développer des modèles cérébraux in vitro plus pertinents pour évaluer les effets neurotoxiques ou thérapeutiques de molécules par criblage phénotypique, notamment dans le cadre de la maladie de Parkinson (MP).Dans un premier temps, la technologie BIOMIMESYS® Brain a été développée. Cette matrice à base d’acide hyaluronique permet de mimer la MEC et de cultiver des cellules cérébrales en 3D dans des plaques 96 puits. La sensibilité des cellules Luhmes, une lignée de neurones dopaminergiques, aux inducteurs de la MP a été étudiée : les cellules ont montré une sensibilité plus faible dans BIOMIMESYS® Brain qu’en 2 dimensions (2D). Cette différence a pu être expliquée par deux phénomènes : une rétention partielle des molécules toxiques dans la matrice, et un phénotype de neurone dopaminergique moins mature qu’en 2D. L’importance du microenvironnement cellulaire a ensuite été étudiée au travers d’une co-culture de cellules Luhmes et d’astrocytes primaires humains en 2D. Cette co-culture a ensuite été transposée dans la matrice BIOMIMESYS®, formant ainsi un modèle complexe incluant à la fois le microenvironnement glial et le microenvironnement matriciel.En parallèle, l’influence du microenvironnement moléculaire a été étudiée sur les cellules SH-SY5Y, une lignée cellulaire issue d’un neuroblastome, couramment utilisée pour évaluer la neurotoxicité de molécules. Dans cette étude, les 24 principaux milieux décrits dans la littérature pour différencier ces cellules en neurones ont été criblés. Les 3 conditions les plus différenciantes en matière de ralentissement de la prolifération cellulaire et de croissance des neurites ont été sélectionnées : l’acide rétinoïque, la staurosporine, et l’Adénosine Monophosphate cyclique (AMPc) associée à du supplément B21. L’expression de marqueurs protéiques neuronaux et la sensibilité des cellules à des composés de toxicités connues ont été mesurées, en 2D et en 3D dans BIOMIMESYS® Brain. La maturité neuronale et la sensibilité aux composés neurotoxiques différaient selon le milieu, en étant les plus hautes en milieu B21+AMPc. La culture en 3D modifiait aussi la réponse des cellules, avec une sensibilité plus faible comparée aux cellules cultivées en 2D.Cette thèse a mis en évidence que le microenvironnement des neurones, qui inclut la MEC, les cellules gliales et les facteurs solubles, modifie la réponse neuronale in vitro et devrait par conséquent être considéré avec attention dans la recherche académique comme industrielle, dès les étapes de criblage de nouveaux médicaments
About 90% of drug candidates fail in clinical trials, for efficacy- and toxicity-related reasons, which often involve the Central Nervous System (CNS). This high failure rate highlights a lack of relevance in experimental models used upstream, including human in vitro models. Indeed, they do not take into account the complexity of the CNS, in which neurons are organized in 3 dimensions (3D) and interact with their microenvironment, composed of cells, soluble factors and extracellular matrix (ECM). The objectives of this PhD were i) to study the influence of these three microenvironment components on neuronal cells in cerebral in vitro models by automatized cellular imaging, and ii) to develop more relevant cerebral in vitro models for phenotypic screening, to assess neurotoxic or therapeutic effects, in the frame of Parkinson’s Disease (PD).First, the BIOMIMESYS® Brain technology has been developed. This acid hyaluronic based-matrix allows the simulation of the ECM and a 3D culture of cerebral cells in 96-well plates. The sensitivity of Luhmes cells, a dopaminergic neuronal cell line, to PD inducers has been studied: the cells displayed a lower sensitivity in BIOMIMESYS® Brain compared to cells cultured in 2 dimensions (2D). This difference was explained by two phenomena: a partial retention of toxic molecules in the matrix, and a lower neuronal maturity compared to cells cultured in 2D.The importance of the cellular microenvironment has been studied through a co-culture of Luhmes cells and primary human astrocytes in 2D. This co-culture has then been transposed in BIOMIMESYS® matrix, to form a complex model including both the glial and the matricial microenvironments.In parallel, the influence of the molecular microenvironment has been studied on the SH-SY5Y cells, a cell line derived from a neuroblastoma, commonly used for neurotoxicity assessment. In this study, the 24 major differentiation media described in the literature to differentiate these cells into neurons have been screened. The 3 most differentiating conditions in terms of proliferation slowdown and neurite elongation have been selected: retinoic acid, staurosporine, and cyclic Adenosine Monophosphate (cAMP) combined to B21 supplement. The neuronal protein marker expression and the cell sensitivity to compounds of known-toxicity have been measured, in 2D and in 3D in BIOMIMESYS® Brain. Both maturity and sensitivity of these neurons varied according to the differentiation medium, and were higher in B21+cAMP. The 3D cell culture modified also the cell response, with a lower sensitivity of cells cultured in 2D.This PhD highlighted that the microenvironment of neurons, including the ECM, the glial cells and the soluble factors, can modify the neuronal response in vitro, and should thus be considered carefully in academic research and as early as possible in the drug discovery industrial process

Traumatic brain injury (TBI) results from a physical insult to the head and often results in temporary or permanent brain dysfunction. However, the cellular pathology remains poorly understood and there are currently no clinically effective treatments. The overall goal of this work was to develop and characterize a novel three-dimensional (3-D) in vitro paradigm of neural trauma integrating a robust 3-D neural co-culture system and a well-defined biomechanical input representative of clinical TBI. Specifically, a novel 3-D neuronal-astrocytic co-culture system was characterized, establishing parameters resulting in the growth and vitality of mature 3-D networks, potentially providing enhanced physiological relevance and providing an experimental platform for the mechanistic study of neurobiological phenomena. Furthermore, an electromechanical device was developed that is capable of subjecting 3-D cell-containing matrices to a defined mechanical insult, with a predicted strain manifestation at the cellular level. Following independent development and validation, these novel 3-D neural cell and mechanical trauma paradigms were used in combination to develop a mechanically-induced model of neural degeneration and reactive astrogliosis. This in vitro surrogate model of neural degeneration and reactive astrogliosis was then exploited to assess factors influencing neural stem cell (NSC) survival and integration upon delivery to this environment, revealing that specific factors in an injured environment were detrimental to NSC survival. This work has developed enabling technologies for the in vitro study of neurobiological phenomena and responses to injury, and may aid in elucidating the complex biochemical cascades that occur after a traumatic insult. Furthermore, the novel paradigm developed here may provide a powerful experimental framework for improving treatment strategies following neural trauma, and therefore serve as a valid pre-animal test-bed.

The thesis describes the performance of the HepG2 cell line, a proliferating liver cell line, potentially to be used as the cellular component of a bioartificial liver (BAL). HepG2 cells in 3-Dimensional (3-D) spheroidal culture demonstrate dramatically improved function, compared to monolayer culture. The main aims of the thesis were to investigate the phenomenon in this culture system, whereby function in 3-D culture is optimal between Days 8-10, and thereafter there is a decline in function in spite of continued proliferation and viability. For optimal use in a BAL, it is necessary to increase function per cell either at later time points of 3-D HepG2 culture when cell numbers are greater, and/or increase cell number at times of peak function i.e. Days 8-10 of 3-D culture. The initial hypothesis was that the downregulation of activity observed from Day 11 occurred at the transcriptional level. Approaches to determine expression of liver-enriched transcription factors at times of peak and diminished function in 3-D culture and in monolayer culture were not conclusive. To understand the mechanisms underlying the temporal change in function in 3-D culture, the hypothesis that hypoxia and/or other forms of stress were responsible for the drop in function observed was explored. Microarray data and Western blot analysis highlighted the expression of genes and proteins known to be upregulated by hypoxia at Day 15. Ambient oxygen concentration was increased to attempt to increase cell performance but was ineffective. Genes and proteins implicated in oxidative stress were expressed. Results from assays to measure oxidative stress in HepG2 culture demonstrated an increase at Day 15 compared with Day 8 spheroidal cultures. Attempts to alleviate this stress by supplementing the culture medium with additional anti-oxidants at later times of 3-D culture did not enhance cell performance. Some stress-related proteins and genes were more strongly expressed in Day 8 cultures, as an adaptive response to increased metabolic activity during peak function, while others in Day 15 were turned off. These genes were investigated further at a functional level and results reflected the pattern observed. 3-D culture was manipulated with the addition of extracellular matrix (ECM) in order to enhance cell performance. Cell proliferation was measured by total nuclei quantification, incorporation of BrdU as a measure of DNA synthesis and Ki-67 labelling, as a measure of the total growth fraction. Positively labelled cells were seen throughout the spheroid indicating that even at the centre of the spheroids; there was maintenance not only of viability, but also the capacity to proliferate, indicating other microenvironmental changes may be responsible for the diminished function observed at Day 15. The thesis has highlighted the possible causes for the downregulation of function observed and modulation of the 3-D environment to overcome this, and emphasised the complex relationship between cell performance and the stress response, in order to improve a system which could provide the basis for the biological component of a BAL.

博士
國立成功大學
工程科學系碩博士班
98
Cell-based assays have been widely utilized in life science-related area to quantitatively investigate the link between the cellular responses and the tested conditions for decades. Conventional cell handling techniques mainly involve the cell isolation, separation, immobilization, liquid dispensing, and cell culture practice. These operations, however, might not be able to deal well with the biological sample with a small size. In addition, the commonly-used cell culture protocols might consume more experimental research resources (e.g. number of cells), and therefore the throughput of a cell-based assay might be compromised. More importantly, traditional cell cultures could not provide a stable, well-defined, and physiologically-meaningful culture conditions for cell-based assays due to the design of cell culture format. During the past decade, there have been tremendous advances in microfluidics. Due to the significant differences in several physical phenomena between microscale and macroscale devices, microfluidic technology provides unique functionality, which is not previously possible by using traditional techniques. This study reports several new microfluidic devices for high-performance cell handling and for high-throughput cell culture. All these devices fabricated based on a computer numerical controlled (CNC) milling or SU-8 lithography process for molds and polydimethylsiloxane (PDMS) replica molding processes. Firstly, to achieve cell isolation and separation, a new microfluidic-based filter was presented. The filtration separation mechanism is based on the pneumatically tunable deformation of PDMS membranes, which block the fluid channel with a varied degree. This defines the dimensions of the remaining passageway of fluid channel and thus the passage of the microbeads/cells with a specific size. Because of the miniaturization and tunable characteristics of separation performance, not only is the proposed device applicable to perform cell separation under the circumstance that either harvested specimen is limited to the cell content in a sample is sparse, but it also paves a new rout to separate/isolate cells in a simple, controllable and cell-friendly manner. To immobilize cells for 3-D cell culture purpose, a new microfluidic device for continuous generation of alginate microbeads was proposed. The working mechanism is based on the use of a pneumatically-driven vibrator to continuously spot tiny alginate microdroplets in a thin oil layer. The temporarily formed alginate microdroplets are soon sinking into a sterile calcium chloride solution to become gelled microbeads. By regulating the alginate suspension flow rate and the pulsation frequency of the integrated vibrator, the alginate microbeads can be produced in a size-controllable manner. Furthermore, a microfluidic-based pneumatically-driven micro-dispenser was demonstrated for precise pipetting of sub-microliter samples. The key feature of the micro-dispenser is the use of a suction membrane to provide a driving force for precise and quick aqueous liquid sampling and pipetting. The micro-dispenser features in the elegant control of the releasing time of the air pressure in the pneumatic chamber of the pressure-generating unit, contributing to precise pipetting of aqueous liquid volumes ranging from 0.05 μl to 0.45 μl (the minimum unit is 0.05 μl) achieving the multi-volume dispensing capability. By means of proper combinations, the liquid of various volumes would be easily sampled. In addition, a new perfusion-based, micro three-dimensional (3-D) cell culture platform was proposed for high-throughput bioassays using enabling microfluidic technologies. The main characteristics of the chip are the capability of multiple medium deliveries without any back-flow by using the new design pneumatic C-shape micropumps, and the function of efficient cells/hydrogel scaffold loading. Based on the inherent natures of miniaturized perfusion 3-D cell culture, the cell culture chip not only can provide stable, well-defined and more biologically-relevant culture environments, but also features in low consumption of research resource. All these traits are found particularly useful for high-precision and high-throughput 3-D cell culture-based assays. Finally, all the microfluidic devices proposed in the research were demonstrated to perform the process including separation, microencapsulation of the chondrocytes and investigation the effect of extracellular pH on chondrocyte functions. Experimental results showed that the chondrocytes from the limited enzymatically-digested tissue suspension can be successfully separated by using the microfluidic-based filter with an excellent cell separation efficiency of 93 % and a high cell viability of 96%. Moreover, the separated chondrocytes were encapsulated in alginate microbeads with high cell viability (94±2%) by using the microfluidic alginate microbead generator. Besides, a micro-scale perfusion 3-D cell culture-based assay to study the effect of extracellular pH on chondrocyte was successfully demonstrated using the proposed cell culture chip and the micro-dispenser was used to adjust the different pH value of the medium. The results were also compared with the same evaluation based on conventional static cell culture with larger culture scale. As a whole, these microfluidic systems proposed in the study provide a simple, automatic, controllable, uniform, cell friendly, less contaminated manner for cell manipulation and culturing and may facilitate a high-throughput cell culture based assay in the more in vivo-like environment.

碩士
長庚大學
生化與生醫工程研究所
101
Biocompatibility of ophthalmic biomaterial is known to play an important role in determining their success for application as medical devices and surgical implants. The development of in vitro biocompatibility testing model can potentially eliminate the need for animal use associated with in vivo testing model. This study aims to assess the biocompatibility of biomaterial by using a perfusion bioreactor cell culture system. Two- and Three-dimensional corneal keratocyte culture models were respectively established by cell seeding on fibronectin-coated polydimethylsiloxane or cell encapsulation within agarose gels. After a 5-day exposure to glutaraldehyde (GTA) treated gelatin discs, the cell from 3-D culture showed lower toxicity and apoptosis than those of 2-D counterparts, indicating that the cell culture model may affect biocompatibility of test materials. To further correlate the relationship between in vitro and in vivo tissue responses to the biomaterials, the GTA cross-linked gelatin discs were also implanted in an ocular anterior chamber. The results of Live/Dead and terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assays showed that similar levels of death percentage and apoptotic index were observed for the groups of in vitro 3-D cultured keratocytes and in vivo stromal cells. Our findings suggest that a 3-D perfusion bioreactor cell culture system may provide an appropriate testing environment for precise determination of ocular biocompatibility of newly developed materials.

yes
Protocols which permit the extraction of primary astrocytes from either embryonic or postnatal mice are well established however astrocytes in culture are different to those in the mature CNS. Three dimensional (3D) cultures, using a variety of scaffolds may enable better phenotypic properties to be developed in culture. We present data from embryonic (E15) and postnatal (P4) murine primary cortical astrocytes grown on coated coverslips or a 3D polystyrene scaffold, Alvetex. Growth of both embryonic and postnatal primary astrocytes in the 3D scaffold changed astrocyte morphology to a mature, protoplasmic phenotype. Embryonic-derived astrocytes in 3D expressed markers of mature astrocytes, namely the glutamate transporter GLT-1 with low levels of the chondroitin sulphate proteoglycans, NG2 and SMC3. Embroynic astrocytes derived in 3D show lower levels of markers of reactive astrocytes, namely GFAP and mRNA levels of LCN2, PTX3, Serpina3n and Cx43. Postnatal-derived astrocytes show few protein changes between 2D and 3D conditions. Our data shows that Alvetex is a suitable scaffold for growth of astrocytes, and with appropriate choice of cells allows the maintenance of astrocytes with the properties of mature cells and a non-reactive phenotype.
BBSRC

碩士
長庚大學
生化與生醫工程研究所
101
Microfluidic cell culture systems have been widely used in cell culture-based assays (e.g. drug testing). However, most of previous works can not provide a stable, well-defined, and more biologically-relevant culture environment for a high-throughput and high-precision cell-based assay. To tackle these issues, this research aims to develop a microfluidic cell culture system consisting of a microfluidic cell culture chip and a controller for micro-scale perfusion 3-D cell culture-based assays. Its advantages include the function for both efficient and high throughput micro-scale 3-D culture construct preparation and loading, the capability for multiplexed medium delivery, and the design of waste medium reservoir array facilitating the subsequent high throughput bioassay works. Furthermore, a chemosensitivity assay was successfully demonstrated using the proposed cell culture system. Comparnig with the chemosensitivity assays using other cell culture models, results showed that the different cell culture format could lead to different evaluation outcomes. In establishing in vitro cell-based assays, therefore, it might be necessary to investigate the fundamental physiological variations of the cultured cells in different culture models to avoid any misinterpretation of data. Overall, the proposed cell culture system not only can provide more stable, well-defined and biologically-relevant culture environments, but it also features in low consumption of research resources. All these features are found valuable for a high throughput and high precision cellular assays.

碩士
長庚大學
生化與生醫工程研究所
101
Traditional cell culture devices are widely used in cell culture-based testing (eg: drug testing). However, most of the cell culture tools may not be able to provide a stable, quantitative physiological significance possessed cell culture environment for high-throughput and high accuracy of the drug testing. Therefore, this study aims to develop a miniaturized three-dimensional cell culture perfusion system. The cell culture system main features are: (a) to gas-driven approach to multi-channel transmission medium and (2) the efficient conduct highly accurate small sample (example: three-dimensional cell culture samples) injection. Development section of the device, we polydimethyl siloxane as the material production of the cell culture system, in addition we also evaluated the gas-driven multi-channel medium conveying performance. Finally, we will demonstrate in the original cell culture. On the application in terms of the ultimate vision of this research is the use of cell culture systems developed by anticancer drug sensitivity testing in the hope that the future can provide patients with cancer chemotherapy drug choice of targets. In achieving this goal, some of the cell culture test results for such usage patterns affect basic research will be established. Experimental results show that the Institute is the development of the gas-drive system uses gas operation time and the intervals to control the perfusion flow rate at the action time is greater than one second when the coefficient of variation can reach 2.32% to 1.20% and the culture environment pH reaches stability, and the original cancer cells in different cell culture environment pH value below its physiological performance and growth capabilities were significantly different, so the system in providing stable homogeneous cell culture environment, with good performance. The use of this system for further experimental results also show that the original cancer cells in different cell culture models under its physiological performance and growth are significant differences, while affecting the original cancer cells to anticancer drugs chemical sensitivity, and therefore relevant in vitro experiments, which must be considered and the environment into the cell culture model is selected to obtain a physiological significance of the experimental results.