Ordered micropattern arrays fabricated by lung-derived dECM hydrogels for chemotherapeutic drug screening.

Aims: This study aims to evaluate ECM-coated micropattern arrays derived from decellularization of native porcine lungs as a novel three-dimensional cell culture platform.
Methods: ECM derived from decellularization of native porcine lungs was exploited to prepare hydrogels. Then, dECM-coated micropattern arrays were fabricated at four different diameters (50, 100, 150 and 200 ​μm) using polydimethylsiloxane (PDMS). Two lung cancer cell lines, A549 and H1299, were tested on a dECM-coated micropattern array as a novel culture platform for cell adhesion, distribution, proliferation, viability, phenotype expression, and drug screening to evaluate the cytotoxicity of paclitaxel, doxorubicin and cisplatin.
Results: The ECM derived from decellularization of native porcine lungs supported cell adhesion, distribution, viability and proliferation better than collagen I and Matrigel as the coated matrix on the surface. Moreover, the optimal diameter of the micropattern arrays was 100-150 ​μm, as determined by measuring the morphology, viability, proliferation and phenotype of the cancer cell spheroids. Cell spheroids of A549 and H1299 on dECM-coated micropattern arrays showed chemoresistance to anticancer drugs compared to that of the monolayer. The different distributions of HIF-1α, MCL-1 (in the center) and Ki-67 and MRP2 (in the periphery) of the spheroids demonstrated the good establishment of basal-lateral polarity and explained the chemoresistance phenomenon of spheroids. Conclusions: This novel three-dimensional cell culture platform is stable and reliable for anticancer drug testing. Drug screening in dECM-coated micropattern arrays provides a powerful alternative to existing methods for drug testing and metabolic profiling in the drug discovery process.

Introduction

Traditionally, anticancer drugs are evaluated in conventional two-dimensional cell culture platforms (monolayers), such as growing cells on polystyrene dishes. However, conventional two-dimensional cultured cancer cells cannot mimic the complexity of cell–cell and cell-matrix interactions or the heterogeneity of tumors, which usually grow in a three-dimensional conformation in vivo [1]. Thus, the results of efficacy and cytotoxicity assays often exhibit differences from those of animal studies and human trials, leading to the failure of new drugs in clinical trials. To overcome this, several techniques are currently being developed that can mimic the microenvironment as spheroids or organoids to rebuild the interaction signals for cell survival, proliferation, differentiation, and gene expression. Recent three-dimensional cell culture models narrow the gap between two-dimensional cell culture models and animal disease models by mimicking the realistic microenvironment of cell–cell and cell-matrix and recapitulating natural three-dimensional tumors [2,3]. They are also useful in drug discovery for determining the sensitivity of chemotherapeutics to tumor cells at both the cellular and molecular levels.

The presence of an intact native extracellular matrix (ECM) is of central importance, as it not only provides a platform for cell growth, proliferation, differentiation, polarization, migration and organization but is also thought to mediate biochemical and molecular signaling, such as neovascularization and cancer progression [4,5]. The matrix scaffold acquired from decellularization of donor organs holds great potential for maintaining the composition of ECM. Key proteins, such as collagen, fibronectin, and laminin, as well as activated growth factors and cytokines are well preserved after decellularization derived from native organs [6,7]. Recently, lung dECM hydrogels derived from healthy and diseased lungs have been further developed as a model to study ECM with similar mechanical properties [8]. Besides, dECM can also serve as coating [9] and embedding [10], bioink for three-dimensional culture [11,12], and injectable biological scaffolds to protect against radiation lung injury [13]. Decellularized ECM-based hydrogels are therefore particularly gaining attention due to their property and perfect homogeneity [6,14,15]. Moreover, natural-synthetic polymer hydrogels have also been proposed to construct ideal bionic tissues and sophisticated structures [16,17].

Another obstacle that must be overcome is to standardize and control the size and arrangement of spheroids or organoids, which could strongly affect drug efficacy and toxicity [18], as currently common three-dimensional cultures will not be able to be repeatedly and accurately measured [2]. Advances have been made on this front to standardize the concordance of spheroid morphological characteristics to reduce heterogeneity and increase reproducibility. Micropattern arrays form a patterned arrangement of defined shape and size on planar substrates by covalently or noncovalently coating carbohydrates, peptides and proteins, thus offering the advantage of culturing cells at high throughput, increasing functionality and improving uniformity at the same time [19,20]. For example, polydimethylsiloxane (PDMS) is widely used in the fields of cytology, drug screening and tissue engineering to create specific micropattern arrays of proteins on nonattachment plates [19,21].

In this study, we exploited ECM derived from decellularization of native porcine lungs to prepare hydrogels. The characteristics of dECM-based hydrogels were measured and compared with traditional coating substrates, including collagen I and Matrigel. Furthermore, we explored the fabrication of dECM-coated micropattern arrays at different diameters (50, 100, 150 and 200 ​μm) using PDMS. Two lung cancer cell lines, A549 and H1299, were tested on our dECM-coated micropattern array as a novel culture platform for drug screening to evaluate the cytotoxicity of paclitaxel, doxorubicin and cisplatin.

Materials and methods

Preparation and evaluation of decellularized lung scaffolds

Male Bama miniature pigs weighing 30–40 ​kg were purchased from Sainuo Biomedical (Chengdu, China). All animal protocols were performed in accordance with the guidelines of the Laboratory Animal Welfare Act and standard operating procedures at the Sichuan University Research Center with the approvals of the Animal Experiment Center of Sichuan University. Animals were anesthetized with Zoletil 50 (10 ​mg/kg body weight, Virbac, France) and maintained with propofol (6 ​mg/kg/h, Qingyuan Jiabo, China). The lungs were excised, the pulmonary artery and bronchus of the porcine lungs were cannulated, and the lungs were rinsed with phosphate-buffered saline (PBS) for 10 ​min and then stored at −20°C for further decellularization as reported previously [22]. In brief, the lungs were thawed and perfused with double-distilled water (ddH2O), 0.2% dextrose (Sigma–Aldrich, MO, USA), 1% Triton X-100 (Biofroxx, Einhausen, Germany), and 1% sodium lauryl ether sulfate (SLES) (Biofroxx) at a rate of 100 ​mL/min. This was followed by 1% Triton X-100 to remove residual SLES. Subsequently, the lungs were washed with ddH2O to remove residual detergent. As a control group, the lungs were perfused with distilled water only.

Native or decellularized lung scaffold samples were fixed with 4% neutral formalin, and then the sections were stained with hematoxylin and eosin (H&E), Masson's trichrome (MT, G1340, Solarbio, Beijing, China), and Gomori's aldehyde fuchsin (GAF, G1593, Solarbio) according to established protocols. Images were captured on a NanoZoomer Digital Pathology (NDP) scanning system (Hamamatsu, Hamamatsu City, Japan). In addition, primary antibodies against collagen-IV (ab6586, 1:500, Abcam, MA, USA), fibronectin (ab6328, 1:200, Abcam), and laminin (ab11575, 1:200, Abcam) were used to determine the retention of crucial proteins in the lung. The images were captured using an N-SIM-S Super Resolution Microscope (Nikon, Tokyo, Japan). The morphology of the ECM was also determined by scanning electron microscopy (SEM). Electron micrographs of lung cross-sections were obtained at 5.0×kV and 600 ​× ​magnification using a Hitachi S-4800 SEM (Hitachi, Tokyo, Japan).

Total DNA from native and decellularized tissues (10 ​mg dry weight) was extracted with a Tissue DNaeasy Kit (Tiangen Biotech, Beijing, China) and quantified using a NanoDrop One spectrophotometer (Thermo, MA, USA). Sulfated glycosaminoglycans (GAGs) were quantified using the Blyscan GAG assay kit (Biocolor, Belfast, UK). After incubation with the Blyscan dye reagent and centrifugation, the samples were quantified by measuring the absorbance at 650 ​nm.

Preparation of the dECM hydrogel

The dECM hydrogel was prepared as described with modifications [[23], [24], [25]]. After decellularization, the lungs were cut into 1 ​× ​1 ​× ​1 ​cm diameter disks for lyophilization. Lyophilized lung scaffolds were ground using a Wiley Mill (Retsch, MM400, Haan, Germany) and filtered through an 80 mesh screen. The 1 ​g ground lung scaffold was then enzymatically digested in 50 ​mL of 2 ​g/L porcine pepsin (Sigma–Aldrich) in 0.01 ​M HCl at a 60 ​rpm stirring rate for 48 ​h at 25°C. The lung dECM hydrogel was neutralized to a pH of 7.2–7.4 with 0.1 ​M NaOH to terminate the pepsin activity. Then, the neutralized lung dECM hydrogel was diluted with 10 ​× ​PBS at a ratio of 9:1 (v/v), and the concentration of dECM was adjusted to 10 ​g/L using 1 ​× ​PBS and stored at 4°C for further use. All procedures were performed on ice. The protein concentration of the dECM hydrogel was measured using a NanoDrop spectrophotometer at 280 ​nm (ND-2000c; Thermo, USA).

Cell culture

Human lung cancer A549 and H1299 ​cells were cultured in high-glucose Dulbecco's modified Eagle's medium (DMEM, HyClone, MA, USA) containing 10% fetal bovine serum (Gibco, NY, USA) and 1% penicillin–streptomycin solution (HyClone) in an incubator at 37°C under 5% CO2 with saturated humidity. The medium was changed three times per week.

Plate coating and characteristics evaluation

The dECM hydrogel was diluted to a concentration of 0.1 ​g/L with PBS, and collagen I (BD Biosciences, MA, USA) and Matrigel (Corning, MA, USA) at the same concentration were used as positive coating controls. Five hundred microliters of coating hydrogel was added to 24-well cell culture plates and then placed into a 37°C incubator for 1 ​h. The excess liquid was discarded, and the plates were washed with PBS. Uncoated plates were used as a negative control. The plates were seeded with A549 or H1299 ​cells, and nonadherent cells were counted after 4 ​h of culture using a Countess™ II FL Automated Cell Counter (Invitrogen, CA, USA). The plates were cultured in an incubator at 37°C under 5% CO2 with saturated humidity. The viability of cells at 1, 3 and 5 days was assessed using FluoroQuench fluorescent stain (One Lambda, CA, USA) according to the manufacturer's protocol. Representative samples were imaged using a fluorescence microscope (Axio Observer D1/AX10 cam HRC, Carl Zeiss, Gttingen, Germany). The Cell Counting Kit-8 (CCK-8, MCE, NJ, USA) assay was performed in a 96-well plate and measured using a NanoDrop spectrophotometer at 450 ​nm to quantify cell proliferation capacity.

Label-free LC–MS/MS proteomic analysis of the dECM hydrogel

dECM samples were lysed, and proteins were extracted and digested. The molecular weight of the dECM hydrogel was determined by sodium dodecyl sulfate (SDS) - polyacrylamide gel electrophoresis (PAGE) according to the manufacturer's instructions.

LC–MS/MS analysis was performed on a timsTOF Pro mass spectrometer (Bruker) coupled to Nanoelute (Bruker Daltonics). The raw MS data for each sample were combined and searched using MaxQuant 1.5.3.17 software for identification and quantitation analysis. The bioinformatic analysis was performed and visualized.

Micropattern array printing

Polydimethylsiloxane (PDMS) seals were made using a laser etching characteristic pattern silicon wafer as a template as reported previously [26]. A 20 ​× ​20 ​mm2 PDMS substrate consists of round patterns with a diameter of 50, 100, 150 or 200 ​μm and an interspace of 50 ​μm. The dECM hydrogel was filtered and sterilized, and the entire process was performed aseptically. The PDMS seals were coated with 0.1 ​g/L dECM hydrogel with 20 ​μg/mL fluorescein isothiocyanate isomer (FITC, Sigma) for 20 ​min at room temperature. The PDMS seals were removed, and the dECM hydrogel was dried at 37°C for 10 ​min. The coated PDMS seals were then placed on a 35-mm nontreated cell culture dish under 0.2 ​N pressure for 10 ​min. The morphology of the FITC-labeled microarray was observed using a fluorescence microscope. The printed dishes were stored at room temperature for further use. Before seeding the cells, the printed dishes were incubated with 10 ​g/L pluronic F-127 solution (Sigma–Aldrich) for 1 ​h to prevent nonspecific cellular adherence. The culture dishes were sterilized by ultraviolet irradiation for 1 ​h.

Spheroid culture

A549 or H1299 ​cells were suspended in 3 ​ml medium at a density of 1 ​× ​105 ​cells/ml and inoculated into dECM-printed dishes. The culture medium was removed after 6 ​h of incubation, and the printed dishes were washed with PBS to remove the unattached cells. In culture, 2.5 ​mL medium was added to the dish for spheroid culture and changed every other day. As a control group, monolayer cells were cultured under the same conditions. The spheroids and monolayer cells were all cultured in an incubator at 37°C under 5% CO2 with saturated humidity. The morphology of spheroids was observed using EVOS TM XL Core (Invitrogen) at 6 ​h and every day after seeding. The diameters of spheroids were analyzed via ImageJ software. The viability of cell spheroids was assessed using FluoroQuench fluorescence viability stain according to the manufacturer's protocol. Cytoplasmic actin microfilament systems were localized using rhodamine-phalloidin (PHDR1, 100 ​nM, Cytoskeleton, CA, USA) for 45 ​min at room temperature. Representative samples were imaged by confocal microscopy (Nikon).

Quantitative real-time polymerase chain reaction (qRT–PCR)

RNA was extracted from cell spheroids at each time point using an RNeasy Mini Kit (Qiagen, CA, USA) according to the manufacturer's instructions. Total RNA (1.0 ​μg) was reverse-transcribed to complementary DNA (cDNA) using an iScript cDNA Synthesis Kit (Bio–Rad, CA, USA). qRT-PCRs were carried out using iTaq Universal SYBR Green Supermix (Bio–Rad) according to the manufacturer's instructions. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) housekeeping gene was used as an endogenous internal control, and the results were normalized using the 2−ΔΔCt method (for promoter sequences, Table S1).

Chemotherapeutic drug cytotoxicity screening

To evaluate the sensitivity of cells to antitumor drugs, A549 or H1299 ​cells were inoculated into dECM-printed dishes as testing platforms for drug screening. After three days of culture, cell spheroids were incubated with paclitaxel (CSN19486, CSNpharm, IL, USA), doxorubicin (CSN16255, CSNpharm) and cisplatin (HY-17394, MCE) at concentrations of 5 ​mg/L, 10 ​mg/L, 50 ​mg/L and 100 ​mg/L for 24 ​h and 48 ​h, and DMSO (Sigma–Aldrich) and deionized water served as negative control. The morphology of the cell spheroids was observed and imaged. The viability of cell spheroids was assessed using FluoroQuench fluorescent stain and imaged.

Features of spheroids on dECM-coated micropattern arrays

Relative changes in proliferative ability, physioxia, anti-apoptosis, drug resistance and intracellular pH levels were measured.

Monolayer cells or spheroids were fixed with 4% neutral formalin and then stained with primary antibodies against Ki-67 (ab16667, 1:100, Abcam), HIF-1α (ab1, 1:100, Abcam), MCL-1 (ab32087, 1:100, Abcam), and MRP2 (ab172630, 1:100, Abcam) to determine their characteristics.

Cytosolic pH was measured using the fluorescent dye probe Protonex™ Red 600 (Cat# 21,207, 1 ​μM; AAT Bioquest) following the manufacturer's instructions. Monolayer cells or spheroids were incubated at 37°C for 30 ​min. Cells were then gently washed in Hanks solution with 20 ​mM HEPES and analyzed.

In addition, monolayer cells or spheroids were also stained with N-cadherin (66219-1-Ig, 1:100, Proteintech), β-catenin (ab32572, 1:250, Abcam), and cytokeratin 7 (ab181598, 1:100, Abcam) to determine cell adhesion and interaction.

Statistics

All data were analyzed with SPSS software version 25 and organized using GraphPad Prism 8. Statistical analysis was performed using one-way ANOVA and Dunnett's post-hoc test. A level of p ​< ​0.05 was accepted as significant. At least three parallel experiments were conducted using different samples. Data are presented as the mean ​± ​SEM.

Results

Characterization of decellularized lung scaffold

Visual inspection, as shown in Fig. 1A, revealed that the detergents produced white and translucent lung scaffolds after perfusion. SEM images of the decellularized tissue showed that the ultrastructure of the ECM was preserved and that the cells were successfully removed after decellularization (Fig. 1B). H&E staining confirmed that SLES removed all visible cell components in the proximal and distal lung and bronchus. No cell debris or genetic material was observed in any decellularized group (Fig. 1C). The architecture of the lung and bronchus was preserved after decellularization, as shown by MT, GAF and AB-PAS staining (Fig. 1D–E, Figure S1). Immunofluorescence staining confirmed the removal of cell nuclei and demonstrated the retention of crucial ECM proteins, including collagen IV, fibronectin and laminin (Fig. 1F–H). The level of residual DNA in tissue indicated the removal of cellular material (Fig. 1I). Meanwhile, there was a reduction in GAGs in the decellularized lung scaffolds, but it was not significant (Fig. 1J).

Fig. 1

Morphology of the decellularized lung scaffolds. (A) Macroscopic view of native and decellularized lung scaffolds. (B) SEM images of native lung and decellularized lung scaffolds. (C) H&E, (D) Masson's trichrome and (E) Gomori aldehyde fuchsin staining of native lung and decellularized lung scaffolds. (F) Collagen IV, (G) Fibronectin and (H) Laminin immunofluorescence staining of native lung and decellularized lung scaffolds. (I) Relative DNA content in native lung and decellularized lung scaffolds. (J) Sulfated glycosaminoglycan (GAG) content in native lung and decellularized lung scaffolds. ∗p ​< ​0.05 versus the native group. Scale bar ​= ​50 ​μm.

Characteristics evaluation of dECM-coated plates

The decellularized lungs were lyophilized and ground (Fig. 2A). The powder was then digested by pepsin to form the dECM hydrogel (Fig. 2B). To evaluate attachment capacity, the plates were seeded with A549 or H1299 ​cells. The number of adherent cells on the dECM-coated plates was greater than that on the uncoated, collagen I or Matrigel-coated plates in both A549 and H1299 ​cells after 4 ​h of culture (Fig. 2C–D). The proliferation of H1299 ​cells cultured on the dECM-coated plates was highest, whereas in A549 ​cells, the proliferation on the dECM-coated plates was higher than that on the uncoated and collagen I-coated plates, close to Matrigel without statistical significance (Fig. 2E–F). Cells were more evenly distributed and spread out without cavities on the dECM- and Matrigel-coated plates than uncoated and collagen I-coated plates, suggesting strong adhesion (Fig. 2G–H). In addition, the viabilities of both cell types were higher on the dECM-coated plates (Fig. 2G–J).

Fig. 2

Preparation and characteristics of extracellular matrix hydrogels derived from decellularized lung scaffolds. (A) Extracellular matrix powder. (B) Extracellular matrix hydrogel. The cell adhesion rate of (C) A549 and (D) H1299 ​cell lines on extracellular matrix hydrogel at 4 ​h after cell seeding. The proliferation rate of (E) A549 and (F) H1299 ​cell line extracellular matrix hydrogels at 24 ​h after cell seeding. Live/dead staining of (G) A549 and (H) H1299 ​cell lines cultured on extracellular matrix hydrogel on days 1, 3 and 5. Quantification of viability of (I) A549 and (J) H1299 ​cell lines cultured on extracellular matrix hydrogel. ∗p ​< ​0.05 versus the uncoated group. #p ​< ​0.05 versus the ECM group. Scale bar ​= ​100 ​μm.

Distribution and composition of the dECM

The molecular weights of the dECM hydrogel were evaluated using SDS-PAGE, and the results showed the consistency of components and repeatability of the dECM hydrogel (Fig. 3A). Label-free protein quantification by LC–MS/MS not only allows relative comparison of proteins across samples but also allows estimation of absolute amounts. Using this approach, we analyzed dECM hydrogels derived from decellularized porcine lungs and quantified and identified over 200 core structural components in porcine dECM. We ranked the relative abundance of core matrisome proteins, providing an in-depth dECM expression profile of healthy porcine lungs. Porcine dECM is enriched in glycoproteins, collagens and laminins such as fibrilin-1, collagen 1a1, 1a2 and 6a3, and laminin gamma 1. The dECM also holds receptors and serves as a reservoir for a large number of secreted proteins, such as growth factors. In addition, the components of dECM-modifying enzymes and other dECM-associated proteins, which, although they do not contribute to the structure of the dECM, affect their functionality as an instructing “niche” (Fig. 3B). The KEGG analyses revealed that significantly enriched pathways were “Focal adhesion”, “Regulation of actin cytoskeleton” and “Tight junction” in “Cell motility” and “Cellular community” (Fig. 3C).

Fig. 3

Composition of the ECM. (A) The molecular weights and distributions of the dECM hydrogel were evaluated using SDS-PAGE. (B) Porcine ECM is enriched in glycoproteins, collagens, laminins, proteoglycans, growth factors and receptors, and integrins. (C) KEGG analysis of porcine ECM.

Characteristics of the micropattern array

As shown in Fig. 4A–B, uniformly round micropattern arrays of 50, 100, 150, and 200 ​μm on dishes was created by the PDMS seals using dECM hydrogel in bright field and fluorescence light. A549 and H1299 ​cells were seeded on the dECM-coated micropattern array to assess their ability to support cell growth. After 6 ​h of attachment, the attached cells formed monolayer cells on the micropattern array (Fig. 4C and F). Thereafter, size-controlled and regularly arranged cell spheroids were formed and grew with prolonged culture time (Fig. 4D and G). On day 3, the cell spheroids remained separated. The diameters of the A549 spheroids were 73.6 ​± ​8.5, 109.2 ​± ​3.6, 162.4 ​± ​2.7, and 220.2 ​± ​5.0 ​μm (Fig. 4D), while those of the H1299 spheroids at 50, 100, 150 and 200 ​μm were 63.9 ​± ​7.2 ​μm, 97.4 ​± ​3.7 ​μm, 144.6 ​± ​8.2 ​μm and 194.7 ​± ​6.5 ​μm, respectively (Fig. 4G). However, the cell spheroids grew larger and connected each other afterwards (Figure S3). The initial seeding number varied greatly, and the arrays were fragile to support long-term culture on the 50 ​μm pattern, whereas the spheroids tended to crowd and connect early on the 200 ​μm pattern, suggesting that the optimal diameter of round patterns on PDMS was 100–150 ​μm (Fig. 4D and G). The viability of both cell spheroids on the 50, 100 and 150 ​μm micropattern arrays remained greater than 98% over the 3-day culture; however, dead cells were observed in the 200-μm pattern (Fig. 4E and H).

Fig. 4

Fabrication of micropattern arrays and the formation of cancer cell spheroids. Micropattern arrays (A) in bright field and (B) fluorescence light derived from PDMS seals of different diameters (50, 100, 150, 200 ​μm). (C) A549 ​cell line cultured on micropattern arrays at 6 ​h after cell seeding. (D) The A549 ​cell line formed cancer cell spheroids at day 3 after cell seeding. (E) Live/dead staining of A549 cancer cell spheroids. (F) H1299 ​cell line cultured on micropattern arrays at 6 ​h after cell seeding. (G) The H1299 ​cell line formed cancer cell spheroids at day 3 after cell seeding. (H) Live/dead staining of H1299 cancer cell spheroids. Scale bars ​= ​100 ​μm.

The height of the dECM-coated micropattern with a diameter of 100 ​μm was analyzed by confocal microscopy. The results showed that the height of the micropattern was 35.6 ​± ​0.6 ​μm (Fig. 5A). The 3D architecture of spheroids on the micropattern array showed bell-shaped cell distribution. A 70–80 ​μm upthrust was caused by the proliferation of cells in the 100-μm pattern (Fig. 5B, Figure S4). We next analyzed the expression of key genes in lung cancer cells and showed that the expression levels of key genes, including Vimentin, N-cadherin, E-cadherin, Cd133, Cd44, Aldh, and β-actin, were higher in the 100-μm pattern, followed by the 150-μm pattern (Fig. 5C, Figure S4). The abovementioned results confirmed that the optimal diameter of the round patterns was 100–150 ​μm.

Fig. 5

Characteristics of the micropattern array. (A) The height of the dECM-coated micropattern was analyzed by confocal microscopy. (B) 3D view of A549 ​cell spheroids in micropattern arrays derived from PDMS seals of different diameters (50, 100, 150, 200 ​μm). (C) qRT–PCR of A549 ​cell spheroids in micropattern arrays of different diameters.

Drug screening

As the consequence, we chose cell spheroids on a 100 ​μm micropattern array as our candidates for the following drug screening. The cancer cells in spheroids of both A549 and H1299 showed greater resistance to paclitaxel, doxorubicin and cisplatin than monolayer cancer cells. The anticancer drugs at lower concentrations killed more cells; however, in a three-dimensional environment, cancer cell spheroids shrank, and the interior of the spheroids was not completely killed even at high concentrations. The cytotoxic effect of paclitaxel was lower in A549 ​cells, with survival rates of 92.7%, 85.3%, 70.4%, and 60.2% at 5, 10, 50, and 100 ​mg/L, respectively (Fig. 6A), while the survival rates of doxorubicin were 58.2%, 50.1%, 1.2%, and 0.0%, respectively (Fig. 6B), and the survival rates of cisplatin were 53.0%, 46.8%, 38.3%, and 31.9% (Fig. 6C) at 48 ​h after drug administration. In H1299, the survival rates were 88.3%, 84.8%, 68.2%, and 57.6%, respectively (Fig. 6D), while the survival rates of doxorubicin were 4.7%, 3.1%, 0.0%, and 0.0%, respectively (Fig. 6E), and the survival rates of cisplatin were 91.1%, 80.3%, 16.5%, and 10.9% (Fig. 6F) at 48 ​h after drug administration. As a control group, 0 ​mg/L anticancer drugs were used to treat cancer spheroids.

Fig. 6

Drug screening. Drug screening of paclitaxel (A, D), doxorubicin (B, E) and cisplatin (C, F) in A549 and H1299 ​cells. Scale bars ​= ​100 ​μm.

Features of spheroids on micropattern arrays

Immunofluorescence images showed that spheroids had higher expression of Ki67 (proliferation marker), HIF-1α (hypoxia marker), MCL-1 (anti-apoptosis marker) and MRP2 (drug resistance) than monolayer cells (Fig. 7A–D), and the results were confirmed by qRT-PCR (Figure S5). However, the distributions of positive cells were different. The HIF-1α-positive and MCL-1-positive cells were located in the center of the spheroids, while the Ki-67-positive and MRP2-positive cells were spread in the periphery of the spheroids (Fig. 7A–D). The cytosolic pH was lower in the interior than in the periphery, as evidenced by the higher fluorescence intensity of the pH indicator Protonex red (Fig. 7E). In addition, the expression of β-catenin, N-cadherin and cytokeratin 7 was widely enriched in cell–cell interaction regions, indicating the establishment of basal-lateral polarity (Fig. 7F–G, Figure S6).

Fig. 7

Features of spheroids on micropattern arrays. Immunofluorescence staining of (A) Ki67 (proliferation marker), (B) HIF-1α (hypoxia marker), (C) MCL-1 (anti-apoptosis marker) and (D) MRP2 (drug resistance) in A549 ​cell spheroids and monolayers. (E) Cytosolic pH of A549 ​cell spheroids and monolayers. Immunofluorescence staining of (F) β-catenin and (G) N-cadherin in A549 ​cell spheroids and monolayers. Scale bars ​= ​100 ​μm.

Discussion

In this experiment, we established an ECM-based hydrogel derived from decellularized porcine lungs and then made micropattern arrays using PDMS. The results demonstrated that dECM supported cell adhesion, distribution, viability and proliferation better than collagen I and Matrigel as the coated matrix on the surface. Moreover, we proved that the optimal diameter of the micropattern arrays was 100–150 ​μm by measuring the morphology, viability, proliferation and phenotype of the cell spheroids. Cell spheroids of A549 and H1299 on dECM-coated micropattern arrays were evaluated for the cytotoxicity of paclitaxel, doxorubicin and cisplatin as a testing platform.

Three-dimensional culture has great potential for preclinical drug screening. Cancer cells grow in three dimensions and can mimic the complexity and heterogeneity of tumors. Cancer spheroids or organoids can maintain their viability, morphology, phenotype, metabolism and function. In addition, dECM can affect cell shape, permeability and rearrangement of the cytoskeleton, thus enhancing cell-matrix interactions and mimicking the ultrastructure and microenvironment in vivo. Previous reports revealed that three-dimensional spheroids exhibited more chemoresistance than cells in monolayers due to poor penetration of anticancer drugs into the interior of the sphere [27,28]. Anti-proliferative drugs aimed at dividing cells, such as paclitaxel and cisplatin, showed a reduced effect on anti-apoptotic and quiescent cancer cells in the interior of spheroids. Hypoxia and low pH could also contribute to drug resistance inside spheroids [28,29]. Anticancer drugs such as doxorubicin and epirubicin are designed to generate free radicals, which could be less harmful to DNA in the absence of oxygen. Cancer-associated fibroblasts and macrophages in vivo tumors can form a physical barrier to resist [[30], [31], [32]] or even increase drug efflux [33,34]. In contrast, drug resistance can be alleviated when the three-dimensional environment is removed, reflecting that chemoresistance is related to interactions. In our study, monolayer cancer cells were sensitive to lower concentrations of anticancer drugs, indicating that three-dimensional cultures were more resistant than monolayers of both A549 and H1299 ​cells. Three-dimensional cell culture models pave the way for mimicking the realistic microenvironment of cell–cell and cell-matrix and recapitulating natural three-dimensional tumors.

Collagen, gelatin, fibrin, hyaluronic acid, chitosan and alginate are widely used natural polymeric biomaterials because of their resemblance to native ECM components [35,36]. A well-defined matrix, such as collagen I utilized in our study, can produce limited variability in experimental results and thus is suitable for the study of the function of a single component. Matrigel is a commercialized gelatinous protein mixture extracted from Engelbreth-Holm-Swarm mouse sarcoma. It is rich in laminin, contains collagen IV, heparan sulfate, entactin and a variety of growth factors and is a commonly used system for three-dimensional cell culture research in vitro. However, the use of an immortalized mouse cell line has been limited by the large differences in xenogenicity and variability between benign and malignant tumors. For instance, the composition of the ECM can affect epithelial-to-mesenchymal transition, chemosensitivity and invasive progression in cancer cells. According to a report, cancer cells seeded in a laminin-rich matrix showed less invasion and dissemination [37]. In this way, Matrigel cannot be regarded as the first choice. The native ECM, on the other hand, comprises key functional proteins, such as collagens, glycoproteins, glycosaminoglycans, proteoglycans, growth factors, cytokines, proteases and integrins [38,39]. We proved that as the surface coating matrix, dECM-based hydrogels were more conducive to cell adhesion, survival, proliferation, and phenotype exhibition in both A549 and H1299 ​cells. The dECM-based hydrogel derived from decellularized porcine lungs in our study was found to be stable at room temperature for up to 30 days. In addition, extensive qRT–PCR testing performed on a number of porcine-specific pathogens was negative, including porcine endogenous retrovirus and α-1,3-Gal. The dECM-based hydrogel also showed a neutral pH, transparency, permeability and compatibility with both normal and cancer cell types.

By comparing micropattern arrays at different diameters, we found that the optimal diameter of round patterns was 100–150 ​μm in the field of viability of cell morphology, viability, proliferation and phenotype exhibition. In the 50-μm pattern, the initiating seeding cells were only 0–3, and the arrays were too fragile to support long-term culture, even though they could grow into spheroids of suitable size. Meanwhile, in the 200-μm pattern, the connection and death of spheroids could occur early. This could be explained by the fact that necrosis occurred in the spheroids because of a lack of oxygen and nutrient diffusion with increasing diameter. Furthermore, enough space should be provided for cell growth rapidly and steadily.

It is shown in the ATCC website (https://www.atcc.org/) that A549 was derived from adenocarcinoma alveolar epithelial cells, while H1299 was established from metastatic lymph nodes, and these two were both non-small-cell lung carcinomas. Furthermore, H1299 is more malignant, notably possessing the P53 deletion mutation [40]. To our knowledge, H1299 ​cells more easily form tumors in a mouse model than A549 ​cells. The results of our study showed that H1299 tended to aggregate, proliferate and form spheroids, indicating a more aggressive, more rapid doubling time of tumor volume and higher stemness in H1299 spheroids, consistent with its phenotype. Meanwhile, the results also demonstrated that drug screenings should be performed after spheroid formation rather than at a fixed time after seeding. As previous results confirmed, A549 spheroids had an efficacy of resistance to cisplatin, and H1299 ​cells showed more sensitivity to cytotoxic anticancer drugs such as cisplatin, etoposide, gemcitabine, vinblastine, taxol, and doxorubicin. In our study, we showed that H1299 ​cells display a relatively higher sensitivity to doxorubicin and cisplatin treatment and, to a lesser extent, to paclitaxel. Under identical conditions, a 2-day treatment with cisplatin concentrations over 50 ​mg/L was sufficient to induce extensive cell death of the H1299 ​cells. However, treatment with cisplatin concentrations as high as 100 ​mg/L was needed to achieve a similar effect on A549 ​cells. These results demonstrated that the responses of different sources of cells to a variety of drugs can be compared directly and illustrated the necessity and usefulness of drug screening before therapeutic interventions as genetically distinct individuals.

According to a previous study, spheroids or organoids of cancer cells embedded in matrix showed higher chemoresistance than those cultured on three-dimensional culture or matrix-based culture alone [29]. The advantage of our study lies in the combination of three-dimensional and natural dECM-based culture in the platform. The most immediate application of this study will take place in the field of pharmaceutical development and individualization. Standardization is necessary before immortalized cancer cell lines can be applied in routine pharmacotoxicology. From organ sources and animal age, decellularized reagents and exposure times, to the shape, size and arrangement of cancer spheroids could strongly affect drug efficacy and toxicity [41,42]. The emergence of micropattern arrays will enhance the efficiency and affordability of the novel drug development process. Patient-derived xenografts (PDXs) can be tested in this platform, and their chemosensitivity profiles can be observed. More advanced, primary cancer cells from human patients can be directly seeded to establish patient-derived organoids (PDOs) on our platform and allow for the prediction of interindividual differences in drug sensitivity and metabolism mediated by genetic polymorphisms, including efficacy and adverse reactions. Drug screening in micropattern arrays provides a powerful alternative to existing methods for drug testing and metabolic profiling. The patient-derived in vitro cell model will be able to reduce unresponsiveness or undesired outcomes.

Credit author statement

Xinglong Zhu: Investigation Writing – original draft preparation. Yi Li: Investigation Writing- Reviewing and Editing. Ying Yang: Visualization. Yuting He: Investigation, Validation. Mengyu Gao: Investigation, Validation. Wanliu Peng: Validation. Qiong Wu: Resources. Guangyue Zhang: Validation. Yanyan Zhou: Validation. Fei Chen: Visualization. Ji Bao: Conceptualization, Methodology. Weimin Li: Supervision

Funding information

, Grant/Award Number: 2018M643504; Post-Doctor Research Project, , Grant/Award Number: 2018HXBH041.

Data availability statement

The data used to support the findings of this study are available from the corresponding author upon request.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.