Influence of a three-dimensional, microarray environment on human Cell culture in drug screening systems

Abstract

We have used a modified 3D cellular microarray platform for the high-throughput analysis of growth, cytotoxicity, and protein expression profile of a human hepatocellular carcinoma cell line, HepG2, in alginate. The results obtained were compared to analogous studies in 2D and 3D environments at the microtiter scale. The antiproliferative effects of four drugs, tamoxifen, 5-fluorouracil, doxorubicin, and amitriptyline, were studied as a function of seeding density in the three different culture platforms. The chemosensitivity of HepG2 cells to all four compounds decreased substantially with increasing cell number in the 2D and 3D microtiter-based cultures, while no seeding density dependence was observed in the IC₅₀ values obtained in the 3D microarray culture platform. These results can be rationalized based on the development of confluence-dependent resistance in cultures where proliferation is restricted by cell–cell contacts and nutrient availability, as is the case for both of the microtiter-based cultures. Additionally, further development of an on-chip, in-cell immunofluorescence assay provided quantitative data on the levels of specific target proteins involved in proliferation, adhesion, angiogenesis and drug metabolism, and was used to compare expression profiles between 2D and 3D environments. The up-regulation of several CYP450 enzymes, β1-integrin and vascular endothelial growth factor (VEGF) in the 3D microarray cultures suggests that this platform provides a more in vivo-like environment allowing cells to approach their natural phenotype.

Introduction

A drug candidate in Phase I clinical testing often requires a decade of discovery followed by preclinical evaluation, yet still has only an 8% chance of reaching the bedside [1]. A leading cause for the failure of drug candidates during clinical trials, and even after a drug has been introduced into the market, is adverse toxicity that was not predicted by animal models [2]. Moreover, the ethical issues and financial constraints surrounding the use of animal-based models in drug screening and toxicity testing have placed increasing pressure to transition testing to in vitro, human cell-based assays that are inexpensive, faster, and potentially more predictive than the current animal testing paradigm [2], [3], [4].

Cell-based assays can facilitate evaluation of a target molecule in a cellular context at an early stage in drug discovery by simultaneously providing information on multiple biochemical and biological end-points, such as proliferation, chemoresistance, motility, differentiation, cell shape, drug absorption, metabolism, and protein expression and localization [5]. These assays are not only rich in information, but are often amenable to automated, high-throughput (HT) screening, reducing screening cost and time, and improving accuracy. However, there remain significant challenges in developing in vitro cell-based models that can recapitulate the in vivo tissue environment to evaluate biologically complex processes. One step toward developing more realistic culturing models is to constrain cells to a more in vivo-like, three-dimensional (3D) environment. This should facilitate cell–cell communication, and can be useful in determining how the cells perceive, interpret, and respond to cues from their microscale environment.

There is a wealth of information on how cells confined in conventional 2D monolayer cultures differ substantially in their properties from cells cultured in a 3D configuration. Lacking the physical and chemical cues defining their natural in vivo microenvironment, cells in 2D culture differ substantially in their shape and organization, in contacts with neighboring cells, and in their physiology and metabolism from cells observed in more physiologically relevant 3D environments [6], [7], [8]. For example, mammary epithelial cells grown in 2D and 3D environments exhibit dramatic differences in cell surface receptor expression, proliferation, cell morphology and organization, gene expression, signaling, and differentiation [7], [9], [10], [11], [12], [13], [14], [15], [16]. Important differences in cellular responses have also been observed in primary hepatocytes and human hepatoma cell lines cultured in 2D and 3D environments. For example, hepatocytes cultured in monolayers de-differentiate after only a few passages and lose liver-specific functions, most significantly their ability to express drug-metabolizing enzymes, which are essential for achieving more predictive toxicity assays [2].

Evidence suggests that the morphology and key functions of primary hepatocytes and human-derived liver cell lines, such as urea, fibrinogen, and albumin secretion, as well as expression and activity of phase I and phase II drug-metabolizing enzymes, can be at least partially regained in 3D cultures [17], [18], [19], [20], [21], [22], [23]. Importantly, a differential response to drugs of cells grown in 2D and 3D cultures has been observed, with a variety of studies showing an increased chemoresistance to anticancer drugs in 3D models [24], [25], [26], [27], [28]. This elevated chemoresistance, observed most commonly in multicellular spheroid models, has been attributed to several factors, including poor penetrability and diffusion of drugs, differences in metabolic state and cell cycle arrest at G₀/G₁ phase, up-regulation of genes conferring drug resistance, and increased pro-survival signaling.

Despite evidence of increased physiological relevance of 3D cultures for toxicity testing, little effort has been directed toward the miniaturization of standardized in vitro 3D models and assays that are compatible with large-scale automated approaches [29], [30], [31], [32]. Such steps, however, are necessary for incorporation of 3D culture techniques into commercial high-throughput (HT) screening. Previously, a miniaturized 3D cell-culture array for HT toxicity screening of drug candidates and their metabolites was described by Lee and co-workers [32], and immunofluorescence-based assays for target protein analysis have been adapted to the microarray platform [33], [34].

In the current study, this 3D platform was compared to more conventional 2D microwell plate assays to study differences in morphology, proliferation, cytoxicity, and protein expression of the human hepatoblastoma cell line, HepG2. The HepG2 cell line is a commonly used model to investigate liver cell function, since it shares many properties of hepatocytes, such as secretion of various lipoproteins, biosynthesis of multiple plasma proteins, and plasma membrane polarity [20], [22]. By performing comparative cytotoxicity studies in 3D nanoscale cultures and in 2D and 3D microtiter-scale cultures, the impact on cell function caused by scaling-down culture size was decoupled from the impact of changing culture dimensionality. Our results clearly demonstrate the importance of cell density when cell culture is used to model the response of tissue to anticancer drugs. Using an on-chip, in-cell immunofluorescence assay, we also demonstrate that the 3D microarray platform can be used in HT to study the key variables involved in 3D-dependent cell behavior and signaling. Specifically, a significant up-regulation was observed in the levels of proteins involved in proliferation, adhesion, angiogenesis and drug metabolism in 3D.