Study on drug screening multicellular model for colorectal cancer constructed by three-dimensional bioprinting technology

Abstract

The existing in vitro models for antitumor drug screening have significant limitations. Many compounds that inhibit two-dimensional (2D) cultured cells do not exhibit the same pharmacological effects in vivo, thereby wasting human and material resources and time during drug development. Therefore, it is crucial to develop new models. Three-dimensional (3D) bioprinting technology has greater advantages in constructing human tissues than sandwich culture and organoid construction. We used 3D bioprinting technology to construct a 3D multicellular model of SW480 cells, tumor-associated macrophages, and endothelial cells. The biological activities of the model were evaluated by immunofluorescence, hematoxylin and eosin staining of frozen pathological sections, and transcriptome sequencing. Compared with 3D bioprinted single-cell model (3D printing-S), 3D bioprinted multicellular models (3D printing-M) showed significantly improved expression of tumor-related genes, including hub genes IL1B, FCGR2A, FCGR3A, CYBB, SPI1, CCL2, ITGAM, and ITGB2. Antitumor drug screening experiment showed that the IC₅₀ values of 5-FU, oxaliplatin, and irinotecan in 3D printing-S group/2D culture group were 31.13 μM/12.79 μM, 26.79 μM/0.80 μM, and 16.73 μM/10.45 μM, respectively. Compared with the 3D printing-S group, 3D printing-M group was significantly more resistant to chemotherapy.

1. Introduction

Drug therapy, including neoadjuvant chemotherapy, post-operative adjuvant chemotherapy, targeted drug therapy, and immunotherapy, is an important treatment for patients with advanced colorectal cancer[1,2]. Phase I–IV clinical trials are the most effective way to gradually verify the efficacy and safety of antitumor drugs[3]. Owing to drug toxicity, biosecurity, and ethical factors, clinical trials are time-consuming, labor-intensive, and not universal. Two-dimensional (2D) cultures are an important model for drug screening and have been widely used. The 2D culture model is simple, reproducible, and mature. The planar 2D structure is quite different from the three-dimensional (3D) spatial structure of the human body or animals. The influence of spatial structure on cells, the interaction between cells, and the interaction between stromal cells and tumor cells has a great influence on the morphological characteristics and biological functions of tumor cells. In practice, antitumor drugs with significant inhibitory effects in 2D culture models do not have good pharmacological effects after application in the human body[4].

A 3D culture model was developed to improve the success rate of preclinical drug screening experiments and to reduce research and development costs. The 3D culture model overcomes the limitations of 2D culture plane structures in space to some extent. At present, different types of 3D culture models have advantages and limitations[5,6]. For example, “sandwich” 3D culture still grows on a flat surface, tumor cells are stacked layer by layer, no real spatial structure has been established between cells, and cells still lack 3D interaction. Organoid culture is an important research model in the field of stem cell research. Cells can be cultured with stem cell activity through induction, reverse differentiation, and other steps to form tissues with certain organ functions. At present, the organoid model is limited by a number of factors, such as low modeling success rate, complex process, consumption of various expensive growth factors, and high cost in the culture process. Patient-derived xenograft (PDX) models have been widely used in colorectal cancer research[7]. In the PDX model, tumor tissues were inoculated subcutaneously into severely immunodeficient mice to allow for tumor growth in vivo. In this model, tumor tissues retain the characteristics of primary tumors to a large extent in terms of morphology, molecular biology, and other aspects. The PDX model showed good clinical predictability in preclinical drug screening studies. However, there are also some problems with PDXs, such as ethical disputes, more time consumption, high cost, and complicated operation[8,9]. At present, suitable in vitro tumor models for drug screening are urgently needed in clinical practice, and it is important to develop new drug screening models.

As early as 2012, a study suggested that tumor microenvironment (TME) would have an impact on tumor chemotherapeutic resistance[10]. The development of new drug screening models should fully consider the possible impact of the TME on colorectal cancer cells, which is helpful in building a real and effective drug screening model. The TME is composed of tumor cells, interstitial cells, extracellular interstitium, and cytokines secreted by each cell. TME plays an important role in tumor proliferation, invasion, and metastasis[11,12]. Tumor cells can influence proliferation through autocrine and paracrine mechanisms. Interstitial cells, such as immune and endothelial cells, can also regulate the proliferation and invasion of tumor cells through immune mediators and secreted factors[13].

3D bioprinting is a convenient, efficient, economical, and easily standardized technology[14-16]. It can use tumor cells as “cell seeds” and accurately print according to the model designed by researchers to construct a complex of cells and bio-ink. In our previous work, 3D bioprinting technology was used to construct a human liver model that had good liver function in vitro and could maintain the functional state for a long time. The 3D bioprinted tissue was used for transplantation therapy, which prolonged the survival of mice with liver failure[17]. This demonstrates the potential of using 3D bioprinted human liver tissue as a substitute for tissue from donor. For drug screening, we constructed a 3D bioprinted hepatocellular carcinoma (HCC) cell model.[18] The 3D bioprinting technology has good applicability in drug screening model construction. The results suggested that the 3D bioprinted single-cell HCC model had unique gene expression profiles and confirmed the advantages of the 3D bioprinted model over the traditional planar culture model in terms of liver function and tumor-related biological activities. We also validated primary HCC cells based on this model and predicted the treatment response of patients with HCC to specific drugs using a 3D bioprinted model to guide personalized treatment[19]. The establishment of a drug screening model for HCC based on 3D bioprinting technology brings new directions and hope for the development of antitumor drugs[20]. Based on the successful experience of 3D bioprinted single-cell models, we explored the function of stromal cells in cholangiocarcinoma by 3D bioprinting immune microenvironment model. At present, researchers have used 3D bioprinting technology to explore tumor models with various features and evaluate their application value in drug screening and cancer research[21,22]. However, current research on 3D bioprinting at home and abroad focuses on the optimization of the printing process, the selection of bio-ink, and the optimization of cell viability status[23], but there is still a lack of comprehensive and in-depth biological function evaluation and drug dose-response experiments of 3D bioprinted tumor models.

To explore the potential value of the 3D bioprinted model in colorectal cancer drug screening, we constructed a 3D bioprinted single-cell model of colorectal cancer using SW480 cells as colorectal cancer seed cells and gelatin/sodium alginate as bio-ink. The 3D bioprinted single-cell model was compared with the 2D and 3D cultures. The differences in the morphology and biological characteristics of the three culture models were evaluated. We constructed a multicellular model for colorectal cancer drug screening using 3D bioprinting technology, developed a novel tumor cell-stromal cell co-culture model, and analyzed the potential impact of TME on tumor cells in a 3D bioprinted multicellular model.

2. Materials and methods

2.1. Cell culture

The human colorectal adenocarcinoma cell line SW480, human acute monocytic leukemia cell line (THP-1), and human umbilical vein endothelial cells (HUVEC-T/T) were purchased from the Cell Resources of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences. The cells identified as having a mycoplasma infection were excluded. SW480 and HUVEC-T/T were cultured in H-DMEM medium (Gibco, Logan, USA) supplemented with 10% fetal bovine serum (FBS; Gibco), 1% penicillin G, and streptomycin (Gibco). Cells were cultured at 37°C in a 5% CO₂ incubator. After reaching approximately 80% confluence, the cells were subcultured with trypsin (0.25%; Invitrogen, Carlsbad, CA, USA), and the medium was changed every other day. Macrophage M2 was used to simulate tumor-associated macrophages in the human body. THP-1 cells were cultured at 37°C in a 5% CO₂ incubator in RPMI1640 medium supplemented with 10% FBS (Gibco) and 1% penicillin G and streptomycin (Gibco). THP-1 medium was supplemented with 12-O-tetracanoylphorbol 13-acetate, and the cells were induced to transform into macrophages M0 after 24 h. M0 was transformed to M2 48 h after induction of IL-4 and IL-13. Cells were cultured in a cell incubator (37°C, 5% CO₂). The culture method for M2 macrophages was the same as that for the SW480 cell line.

2.2. Construction of 3D bioprinted single-cell model

A single-cell model of colorectal cancer was created using the 3D bioprinter produced by SUNP. A 3D printing-S of colorectal cancer was designed in a cylindrical shape and printed with a single nozzle. SW480 cells were harvested and suspended in the medium. The cell suspension was mixed with 4% sodium alginate solution in a 2:1 volume ratio. The mixture was incubated at 37°C for 5 min and then mixed with 12% gelatin solution in the indicated volume ratio, resulting in a final cell density of 4.8 × 10⁶/mL. One milliliter of the cell/biomaterial mixture was drawn into a sterile syringe with a 23 G needle, and the SW480 single-cell model was prepared by layer-by-layer extrusion in a 3D bioprinter[24-26]. The print was immersed in a 100 mM calcium chloride solution for 3 min, and then 3 mL H-DMEM medium was added. The medium was changed every 2 days.

2.3. Construction of multicellular model for colorectal cancer using dual-nozzle 3D bioprinter

The 3D printing-M of colorectal cancer adopts concentric axis printing, namely, dual-nozzle printing, to construct a 3D concentric circle model (Figure 1). 3D cell bioprinter BIOMARKER (SUNP Biotech, Beijing, China) was used to fabricate the in vitro cell model. The ratio of sodium alginate and gelatin concentration and the printing parameters have been reported in our prior publications[17,19]. The construction of the 3D printing-M adopted in this work is realized by the 3D concentric circle model, which is also the first attempt to use such a model in our laboratory. The diameter of the inner ring is 7 mm, the diameter of the outer ring is 10 mm, and the thickness is 8 mm. The printing speed is 6 mm/s and the extrusion speed is 1.599 mm/s. The temperature of the printing platform is set to 10°C, and the temperature of the nozzle is set to 15°C. The layer height is 2 mm, the filling line distance of the nozzle is 0.2 mm, and the filling line width is 0.8 mm. The inner ring is tumor, and the outer ring is tumor stromal cells (macrophage M2 and human umbilical vein endothelial cells). 500 μL HUVEC-T/T suspension, 500 μL tumor-associated macrophage M2 cell suspension, 500 μL sodium alginate, and 1 mL gelatin were added at a ratio of 1:1:1:1. After printing, the bioprinted tissue was crosslinked and fixed in a biosafety cabinet, and 3% CaCl₂ was used as the fixative for 3 min. Fresh medium was added, and the cells were cultured in an incubator (37°C, 5% CO₂). The medium was changed every 2 days.

Figure 1.

Diagram of three-dimensional bioprinted multicellular colorectal cancer model. Tumor cells are shown in green and interstitial cells in red.

2.4. Construction of sandwich culture

The final concentration of tumor cells in the SW480 bio-ink was 4.8×10⁶/ml. The bio-ink prepared for tumor cells was tiled into a 24-well plate using a micropipette and then incubated in a cell incubator (37°C, 5% CO₂) after adding the same amount of fresh medium. The culture medium was changed every 2 days.

2.5. Cell morphology photography

The morphology of colorectal cancer cells in 3D bioprinted models from days 1 to 10 was examined using light microscopy.

2.6. Cell viability

The 3D printing-M, 3D printing-S, and sandwich culture group (3D culture) were labeled with fluorescence. Before cell viability staining, 3D bioprinted models and sandwich culture models were washed with phosphate-buffered saline and then stained with calcein-AM (1 μmol/L; Sigma) and propidium iodide (PI, 2 μmol/L; Sigma), which were kept out of the light for 15 min. After washing with PBS, cells were observed under a confocal microscope. Green indicates live cell and red indicates cell death.

2.7. Cell proliferation assay

The Cell Counting KIT-8 Kit (CCK-8) was mainly used to detect the proliferation status of 3D printing-M, 3D printing-S, and 3D culture groups. 3D bioprinted models and sandwich models were cleaned three times with PBS. Then, 1 mL of complete medium was added to the petri dish along with 0.1 mL drop of CCK-8 reagent. Cells were incubated in an incubator for 2 h. Each proliferation assay was performed in triplicates. Thereafter, 0.1 mL of culture medium was absorbed from each group and added to a 96-well plate. The 96-well plate was placed in a microplate reader for detection at 450 nm/620 nm, respectively. Cell proliferation was measured on 1^st, 3^rd, 5^th, 7^th, and 10^th days, respectively. The measurement value on the 1^st day was the baseline value, and the measurement value of the following 4 days was standardized according to the measurement value of the 1^st day.

2.8. Hematoxylin and eosin staining of frozen pathological sections

The 3D bioprinted multicellular tissue was embedded in warm 3% agarose solution and placed on ice for 3 min to solidify. Frozen tissues were sectioned with a thickness of 5 μm. The sections were fixed in xylene and absolute alcohol and cleaned with distilled water after fixation. The frozen tissue sections were stained with hematoxylin solution for 2 – 3 min, washed with running water, and then counterstained with eosin solution. The flushing time was determined based on the staining effect. After dehydration with absolute ethanol and xylene, frozen pathological sections were sealed with neutral gum. The stained tumor and stroma in 3D bioprinted multicellular tissues were observed under a light microscope.

2.9. Immunofluorescence assay

The 3D bioprinted tissue was transferred to the operating table 10 days later and fixed in 3% CaCl₂ solution for 1 min. The supernatant was discarded, and the cells were washed with PBS for five times. Afterward, the tissue was fixed in 4% paraformaldehyde for 20 min; after discarding the supernatant, it was washed with PBS for 5 times. Three groups of 0.2% Triton X-100 models were incubated for 30 min. The tissue was then blocked in 3% bovine serum albumin (BSA) at room temperature for 60 min. The primary antibody was diluted in PBS containing 1.5% BSA (Table 1), and the tissue was incubated with the antibody at 4°C overnight (16 – 20 h).

Table 1.

Information on immunofluorescent antibodies

antibody	species	antibody brand
Ki67	rabbit	Abcam; 1:200
CDX2	rabbit	Abcam; 1:200
CD68	mouse	Abcam; 1:100
CD31	mouse	Abcam; 1;100
IgG Alexa Fluor 488	Goat Anti-Rabbit IgG	Abcam; 1:500
IgG Alexa Fluor 488	Goat Anti-Mouse IgG	Abcam; 1:500
IgG Alexa Fluor 594	Goat Anti-Rabbit IgG	Abcam; 1:500
IgG Alexa Fluor 594	Goat Anti-Mouse IgG	Abcam; 1:500

Caudal-related homeobox transcription factor 2 (CDX2) plays a key role in the normal development and differentiation of intestinal epithelial cells and the development of cancer tissues, and its expression is rarely lost in colorectal cancer[27]. Ki67 is an antigen associated with cell proliferation, whose function is closely related to mitosis and is indispensable in cell proliferation[28]. Ki67 is often used to reflect the proliferation of tumors. CD68 is the most reliable marker of macrophages and is often used for immunofluorescence staining of macrophages[29]. CD31, also known as platelet-endothelial molecule-1 (PECAM-1), is often present in blood vessel endothelial cells[30].

The three groups of specimens were removed after incubation, left at room temperature for 30 min, and then cleaned 6 times with PBS for 6 times. Secondary antibodies were diluted in PBS containing 1.5% BSA (Table 1), and were used to incubate with the tissues for 1 h at room temperature in the dark. After discarding the supernatant, the tissue was washed with PBS for six times. Diluted DAPI solution (1:10000, Sigma) was added for staining for 5 min at room temperature in the dark. The supernatant was discarded, and the tissue was washed with PBS for five times. A laser confocal microscope was used for observation. The wavelength of the laser confocal excitation light was selected according to the fluorescent secondary antibody.

2.10. Transcriptome sequencing and biological information analysis

In the 2D culture group, Trizol method was used for extraction of RNA. In the 3D culture group and 3D printing-S group, the printing body was lysed to obtain lysate, and the tumor cells were obtained by centrifugation. The samples were collected after extraction using the Trizol method. The 3D printing-M group adopted the 3D concentric circle culture mode, and the outer interstitium was directly removed under the microscope. The principle of removal is to completely remove the interstitium, and part of the tumor cells can be removed, if necessary, to ensure that only tumor cells are left in the 3D bioprinted tissue, without interstitial cells. After that, the 3D bioprinted tissue was lysed, and the tumor cells were obtained by centrifugation. This experiment focused on analyzing and exploring the effects of tumor-associated macrophages M2 and endothelial cells on the gene expression of SW480 cells cultured in a 3D printing-M model. Total mRNA was isolated using the Trizol or RNeasy Mini kit (QIAGEN, Dusseldorf, Germany) and reverse-transcribed using the Ambion kit (Austin, USA). In vitro transcription was performed using 1 – 5 ng of cDNA as a template, and RNA was reverse-transcribed into the sequencing library. Sequencing libraries were prepared using the NEBNext UltraTM RNA Library Preparation Kit (Illumina). The sequencing library was then sequenced on the Illumina HiSeq platform to generate 125/150 bp peer reads. Differentially expressed genes (DEGs) were analyzed using the DeSeq2 software package. Genes with P < 0.05 adjusted for DESeq2 were assigned to DEGs. The clustering analyzer R package was used to implement the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. GO terms and KEGG pathways with P<0.05 after correction were considered to be significantly enriched by DEGs. The protein-protein interaction (PPI) network of DEGs from the STRING database was obtained, and a confidence of 0.400 was selected (version 11.0, https://string-db.org/). Cytoscape3.6.1 software was used to construct a PPI network of the top 30 differentially expressed genes. Among the five central types, more than twice as many of the top ten genes were identified as hub genes.

2.11. Pharmacodynamic evaluation of antitumor drugs

On the 7^th day, 3D bioprinted tissue was used for the antitumor drug sensitivity test. First, two groups were set: a 3D printing-S and a 2D culture group. The sensitivity of the two drug screening models to fluorouracil (5-FU), oxaliplatin, irinotecan, and other common antitumor drugs in colorectal cancer was evaluated. The concentration gradients of 5-FU, oxaliplatin and irinotecan were set to 0, 0.1, 1, 10, 50, and 100 μM in both groups. Cell proliferation was measured using the CCK8 assay, and dose-response curves were plotted using GraphPad 9. The median inhibitory concentrations of the three chemotherapeutic drugs in the two groups were calculated based on these results. 3D printing-M was treated with different concentrations of antitumor drugs (0, 0.1, 1, 10, 50, 100, 200, and 500 μM).

2.12. Statistical analysis

Data are expressed as mean ± standard deviation. The independent t-test was used for comparison of independent samples between two groups, and SPSS 26.0 software (version 26.0; IBM Corp., Armonk, NY, USA) was used for statistical analysis. Statistical significance was set at P < 0.05. Data from at least three independent samples or triplicate experiments were used for all assays.

3. Results

3.1. Construction of 3D bioprinted multicellular model of colorectal cancer

Concentric axis dual-nozzle 3D bioprinting was used to construct a 3D multicellular model of colorectal cancer in concentric circle model, with tumors in the inner ring and tumor stromal cells in the outer ring. The final concentration of interstitial cells in the bio-ink was 1.5 × 10⁶/mL. The diameter of the inner ring of the 3D printing-M was 7 mm, the diameter of the outer ring was 10 mm, and the thickness was 8 mm (Figure 2). Extrusion 3D bioprinting is used to construct stable 3D bioprinted models of colorectal cancer in high throughput.

Figure 2.

Standardized three-dimensional (3D) bioprinted colorectal cancer multicellular tissue model. The diameter of standardized 3D bioprinted tissue is 10 mm.

3.2. Morphological characteristics of 3D bioprinted colorectal cancer model

Under a low-power light microscope, the 3D bioprinted colorectal cancer model remained stable from days 1 to 10. SW480 cells were stable in the bioprinted tissue under high magnification, and no tumor cell was observed. As shown in Figure 3, from days 1 to 10, the SW480 cells in the 3D bioprinted tissue gathered into clusters, and the SW480 cell clusters became increasingly larger. On the 10^th day, dense cell aggregates were scattered in the 3D bioprinted colorectal cancer tissues.

Figure 3.

High-magnification images of three-dimensional (3D) bioprinted single-cell colorectal cancer model on day 1 and day 10 in vitro (A). High-magnification images of tumor cells in 3D bioprinted multicellular colorectal cancer model on day 1 and day 10 in vitro (B). High-magnification images of interstitial cells on day 1 and day 10 in vitro in a 3D bioprinted multicellular colorectal cancer model (C). Scale bars: 40 μm.

3.3. Cell proliferation and survival in 3D bioprinted colorectal cancer model

The 3D printing-S and 3D printing-M models were stained using calcein AM and PI to assess cell survival. The survival of SW480 cells was observed using confocal microscopy. Green indicates live cell, and red indicates cell death. As can be clearly seen from the images, a high activity of >90% was maintained by day 10 (Figure 4). The viability of macrophage and HUVEC cells in 3D hydrogel structure after 3D printing is shown in Figure S1 ^{(437.1KB, pdf)} . Ki67 immunofluorescence was used to observe the proliferation of SW480 cells in the 3D printing-S on the 7^th day, and Ki67 expression was strongly positive in the model (Figure 5). SW480 cells proliferated well in the 3D bioprinted model. This suggests that the 3D bioprinted model can provide a good environment for the growth of colorectal cancer cells. CCK-8 was used to detect the proliferation of SW480 CRC cells in the three models. Cell proliferation was measured on days 1, 3, 5, 7, and 10, respectively (Figure 6). The values measured on the 1^st day were taken as the baseline values, and the values measured on the following 4 days were standardized according to the values measured on the 1^st day. In terms of proliferation, we compared the proliferation of three groups at days 1, 3, 5, 7, and 10, with three biological replicates in each group. We found no statistical difference in proliferation between the three groups at days 1, 3, and 5. On day 7, there was a statistical difference between the 3D printing-S group and the 3D printing-M group (P = 0.022). On day 10, there was a statistical difference between the 3D printing-S group and the 3D printing-M group (P=0.034). There was no significant difference in the proliferation between the 3D culture group and the 3D printing-S group at days 1, 3, 5, 7, and 10.

Figure 4.

(A–F) Cell viability and proliferation in the 3D printing-S model of colorectal cancer. Representative live-dead staining images of 3D bioprinted SW480 structures at days 1 (A), 3 (B), 5 (C), 7 (D), and 10 (E) after printing. (G–L) Cell viability and proliferation in the 3D printing-M model of colorectal cancer. Cell viability at different times after printing (B). Representative live-dead staining images of 3D bioprinted SW480 structures at days 1 (G), 3 (H), 5 (I), 7 (J), and 10 (K) after printing. Live and dead cells were labeled with calcein AM (green) and PI (red), respectively. Scale bar: 40 μm. Histogram of cell viability at different times after printing for 3D printing S and 3D printing-M groups are shown in (F) and (L), respectively.

Figure 5.

Immunofluorescence Ki67 staining of SW480 cells in 3D bioprinted model on day 7 in vitro. Ki67 is in red and DAPI in blue. Scale bars: 40 μm.

Figure 6.

The proliferation of SW480 cells in the three models on days 1, 3, 5, 7, and 10. 3D culture, sandwich culture; 3D printing-S, 3D bioprinted single-cell model of colorectal cancer; 3D printing-M, 3D bioprinted multicellular model of colorectal cancer.

As shown in Figure 7, after 10 days of 3D bioprinting, SW480 cells, macrophage M2, and HUVEC-T/T were stably present in the 3D printing-M, which proved that the colorectal cancer multicellular model we constructed is stable.

Figure 7.

(A) Immunofluorescence staining of SW480 cells on the 10^th day in 3D printing-S. Green is CDX2 and blue is DAPI. (B) Immunofluorescence staining of SW480 cells on the 10^th day in 3D printing-M. Red is CDX2 and blue is DAPI. (C) Immunofluorescence Ki67 staining of SW480 cells in 3D printing-M on the 10^th day. Ki67 is green and blue is DAPI. (D) Immunofluorescence staining of macrophage M2 cells on the 10^th day in 3D printing-M. Red is CD68 and blue is DAPI. (E) Immunofluorescence staining of endothelial cells and SW480 cells cultured for 10 days in 3D printing-M. Red is CDX2, green is CD31, and blue is DAPI. Scale bars: 40 μm.

3.4. HE staining of frozen sections of 3D bioprinted multicellular colorectal cancer model

HE staining of frozen sections of 3D printing-M requires embedding, freeze curing, conventional paraffin embedding, sectioning, fixation, staining, and sealing with neutral gum. Tumor and interstitial cells were observed in two parts under a microscope. HE staining characterized the pathological characteristics of the tumor cell cluster in the 3D bioprinted colorectal cancer model, and multiple nuclei in the tumor cell cluster could be clearly seen, indicating that it was composed of many cells (Figure 8A). Simultaneously, the pathological characteristics of interstitial cells in the 3D printing-M were observed (Figure 8B).

Figure 8.

(A) HE staining of tumor tissues in frozen sections of 3D bioprinted multicellular tissues. Scale bar: 40 μm. (B) HE staining of interstitial tissues in frozen sections of 3D bioprinted multicellular tissues. Scale bar: 15 μm.

3.5. Transcriptional profiling of 3D bioprinted colorectal cancer model

3.5.1. 2D culture versus 3D printing-S

As can be seen from the volcano map of DEGs (Figure 9), the number of differentially expressed genes in colorectal cancer cells was 11107 between the two modes of 2D culture and 3D bioprinted culture. The number of upregulated and downregulated genes was 5498 and 5609, respectively. GO enrichment analysis and KEGG enrichment analysis were used to explore the biological processes that differential genes of colorectal cancer cells in different culture models may be involved in, for example, cell proliferation, cell metabolism, invasion, chemotherapy resistance, and immune escape. Compared to SW480 cells in 2D culture, the gene expression of SW480 cells in the 3D bioprinted model showed significant differences in biological processes, cell composition, and molecular function. KEGG enrichment analysis showed that the metabolic and oxidative phosphorylation pathways were enriched by the most significantly upregulated genes. The most significantly enriched pathways of downregulated genes were the phosphatidylinositol signaling system and the FoxO signaling pathway. The upregulated hub genes, such as CENPF, BIRC5, CDK1, and NCAPG, are listed in Table 2.

Figure 9.

(A) Differentially expressed genes (DEGs) volcano map between two dimensional (2D) culture group and three dimensional (3D) printing-S group. (B) Bar graph of GO enrichment of up regulated DEGs between 2D culture group and 3D bioprinted model group. (C) Scatter plot of KEGG pathway enrichment of up regulated DEGs between 2D culture group and 3D bioprinted model group. Q value is P-value after multiple hypothesis test; the closer the value of Q value is to zero, the more significant the gene enrichment is. (D) DEG volcano map between 3D culture group and 3D printing-S group. (E) Bar graph of GO enrichment of upregulated DEGs between 3D culture group and 3D bioprinted model group. (F) Scatter plot of KEGG pathway enrichment of upregulated DEGs between 3D culture group and 3D bioprinted model group.

Table 2.

Hub genes of differentially expressed genes in the 3D-printed SW480 model compared with the 2D model

Upregulated DEGs		Downregulated DEGs
Hub gene	Function	Hub gene	Function
CENPF	Cell apoptosis and cell proliferation	ESR1	Regulation of estrogen
BIRC5	Apoptosis inhibition	IL4	Immune regulation and cell repair
NCAPG	Cell proliferation	BDNF	Brain-Derived Neurotrophic Factor
ASPM	Cell proliferation	HIST1H4F	Histone protein coding
CDK1	Cell proliferation	HIST1H2BJ	Nucleosome assembly
TOP2A	Cell proliferation and drug resistance	MMP3	Cell migration
KIF2C	Cell proliferation	CTGF	Cell differentiation and cell adhesion
CCNA2	Cell proliferation	HDAC9	Transcriptional regulation and cell cycle regulation
RRM2	Cell cycle regulation	BTK	B-cell development
KIF20A	Cytokinesis regulation	SPI1	B-cell development

3.5.2. 3D culture versus 3D printing-S

As can be seen from the volcano map of DEGs (Figure 9), the number of DEGs in colorectal cancer cells of 3D culture and 3D bioprinted culture was 9265. There were 4622 upregulated genes and 4643 downregulated genes. Compared with SW480 cells cultured in sandwich culture, the differences in upregulated gene expression of SW480 cultured in the 3D bioprinted model mainly focused on cell composition, followed by biological process, and the differences in molecular function were small. KEGG enrichment analysis showed that the autophagy pathway and metabolic pathway were enriched by the most significantly upregulated genes, and focal adhesion and adhesion junctions were enriched by the most significantly downregulated genes. The upregulated hub genes of the 3D printing-S group compared with 2D culture include STAT1, IFIT3, OASL, and ISG15, as shown in Table 3.

Table 3.

Hub genes of differentially expressed genes in the 3D-printed SW480 model compared with the sandwich culture model.

Upregulated DEGs		Downregulated DEGs
Hub gene	Function	Hub gene	Function
STAT1	Cell apoptosis, cell proliferation	JUN	Cell apoptosis, cell proliferation and immune regulation
DDX58	Antiviral innate immune response	ATF3	Cell apoptosis and cell cycle regulation
MX1	Antiviral response	DUSP1	Cell apoptosis, cell proliferation
IFIT3	Cell apoptosis, cell proliferation	FOSB	Cell apoptosis, cell proliferation and cell differentiation
IFIT1	Viral replication inhibition and translational initiation	KLF6	Tumor suppression
OAS1	Cell growth and cell apoptosis	CXCL8	Inflammation regulation, tumor migration, invasion, angiogenesis and metastasis
RSAD2	Lipid metabolism	NR4A2	Regulation of central dopamine neurongenesis
IFIH1	Inflammatory response	SERPINE1	Fibrinolytic inhibition
OASL	Immune response regulation	CSF2	Granulocytes and macrophages differentiation
ISG15	Cell signaling	CAV1	Tumor suppression and lipid metabolism

3.5.3. 3D printing-S versus 3D printing-M

As can be seen from the volcano map of DEGs (Figure 10), the number of DEGs in colorectal cancer cells under the two 3D bioprinted models of single-cell and multicellular groups was 8532. The number of upregulated and downregulated genes was 4489 and 4043, respectively. The upregulated genes of SW480 cells between the multicellular model and single-cell model were mainly related to biological processes. There was little difference in the cell composition between the two models. KEGG enrichment analysis indicated that the upregulated genes were significantly enriched in the PD-L1/PD-1 checkpoint pathway and EGFR receptor tyrosine inhibitor resistance pathway kinase inhibitor resistance, and significant enrichment of downregulated genes was associated with metabolic pathways and ribosome pathways. Upregulated hub genes between the 3D printing-S group and 2D culture include ITGAM, IL1B, FCGR2A, and CYBB, as shown in Table 4 and Figure 11.

Figure 10.

(A) Differentially expressed genes (DEGs) volcano map between 3D printing-M group and 3D printing-S group. (B) Bar graph of GO enrichment of up regulated DEGs between 3D printing-M group and 3D printing-S group. (C) Bar graph of GO enrichment of down regulated DEGs between 3D printing-M group and 3D printing-S group. (D) Scatter plot of KEGG pathway enrichment of up regulated DEGs between 3D printing-M group and 3D printing-S group. (E) Scatter plot of KEGG pathway enrichment of down regulated DEGs between 3D printing-M group and 3D printing-S group.

Table 4.

Hub genes of differentially expressed genes in the 3D-printed multicellular model compared with the 3D-printed SW480 model.

Upregulated DEGs		Downregulated DEGs
Hub gene	Function	Hub gene	Function
ITGAM	Integrins mediate signaling and cell adhesion	DDX58	Innate immune response
IL1B	Inflammation regulation, cell apoptosis, cell proliferation and cell differentiation	OASL	Cell growth and cell apoptosis
FCGR2A	macrophages cell surface receptor and phagocytosis and clearing of immune complexes	STAT1	Cell apoptosis, cell proliferation
FCGR3A	Immune regulation and antigen-antibody complex clearance	IFIT3	Cell apoptosis, cell proliferation
SPI1	Immune cell development	IFIT1	Viral replication inhibition and translational initiation
CCL2	Immune regulation and inflammatory response	OAS1	Cell growth and cell apoptosis
CD4	Immune regulation and antigen recognition	RSAD2	Lipid metabolism
CYBB	Cytophagy and phagocytize bacteria	MX1	antiviral response
FCGR2B	Regulation of antibody production by B-cells	MX2	antiviral response
ITGB2	Immune regulation, signal transduction and cell adhesion	ISG15	Cell signaling

Figure 11.

(A) The regulatory network of top 30 up regulated differentially expressed genes (DEGs) between 2D culture group and 3D printing-S group was visualized by the Cytoscape. Red nodes represent hub differential gene. (B) The regulatory network of top 30 downregulated DEGs between 2D culture group and 3D printing-S group was visualized by the Cytoscape. (C) The regulatory network of top 30 upregulated DEGs between 3D culture group and 3D printing-S group was visualized by the Cytoscape. (D) The regulatory network of top 30 downregulated DEGs between 3D culture group and 3D printing-S group was visualized by the cytoscape. (E) The regulatory network of top 30 up regulated DEGs between 3D printing-M group and 3D printing-S group was visualized by the Cytoscape. (F) The regulatory network of top 30 downregulated DEGs between 3D printing-M group and 3D printing-S group was visualized by the cytoscape.

3.6. Effects of antitumor drugs on the 3D printed SW480 model

Seven days after 3D bioprinting the models, six concentration gradients were set to detect the sensitivity of the two drug screening models to 5-FU, oxaliplatin, and irinotecan in the 3D printing-S group and 2D culture group. The IC₅₀ value of SW480 cells for 5-FU in the 2D culture group was 12.79 μM, and that for SW480 cells in the 3D bioprinted group was 31.13 μM. The IC₅₀ of the 3D printing-S group was three times that of the 2D culture group, which was more resistant to drugs. For oxaliplatin, the IC₅₀ of 2D culture group and 3D printing-S group was 0.80 μM and 26.79 μM, respectively. For irinotecan, the IC₅₀ of 2D culture group and 3D printing-S group was 10.45 μM and 16.73 μM, respectively. The dose-response curves of the three chemotherapy drugs are shown in Figure 12A-C.

Figure 12.

Dose response curve of 5-FU (A), oxaliplatin (B), and irinotecan (C) between 2D culture group and 3D Printing-S group after 72 h of treatment. IC₅₀ represents median inhibitory concentration. Dose-effect curves of 5-FU (D), oxaliplatin (E), and irinotecan (F) in the 3D printing-M after 72 h of treatment.

Antitumor drug screening experiments for the same chemotherapy drugs were performed on the 3D printing-M group, and six concentration gradients were set as 0, 0.1, 1, 10, 50, and 100 μM in the pre-experiment. The proliferation of tumor cells was detected using CCK8 in the drug screening test. The results suggest that the 3D printing-M group was significantly resistant to chemotherapy. The CCK8 value measured at a concentration of 100 μM was similar to the CCK8 value measured at a concentration of 0.1 μm, indicating that high concentrations of chemotherapy drugs did not significantly inhibit the proliferation of colorectal cancer cells. Subsequently, we increased the drug concentration gradients to 0, 0.1, 1, 10, 50, 100, 200, and 500 μM, and the CCK8 value decreased but did not reach the median inhibitory concentration under the condition of high drug concentration, indicating that colorectal cancer cells were significantly resistant to chemotherapy in the co-culture system of tumor-associated macrophages M2, endothelial cells, and tumor cells (Figure 12D-F).

4. Discussion

The current status of drug screening models has prompted us to develop a new antitumor drug screening model, which is economical characteristics and can be used for high-throughput screening[31-33]. More importantly, drug screening should take into account the effect of the TME on tumor cells. To achieve this goal, we constructed a drug screening multicellular model for colorectal cancer using 3D bioprinting. The 3D printing-M is a co-culture system of the TME and the tumor. In this study, to construct a 3D printing-M, we first adopted a 3D concentric circle model with tumor cells and surrounding stromal cells in the center, including tumor-associated macrophages M2 and HUVEC-T cells.

Extrusion 3D bioprinting constructs items in high throughput and involves lower cost. The extrusion 3D bioprinter allows coaxial printing with a double nozzle, which meets the production requirements of the 3D concentric circle model designed for the subject[34,35]. To construct a 3D printing-M, tumor-associated M2 macrophages were selected as interstitial cells instead of monocyte THP-1 cells. The reason lies in THP-1 non-adherent cells, which are different from the SW480 cell line and endothelial cell. More importantly, we simulated tumor-related macrophages in the human body using macrophage M2 to better explore the influence of the TME on chemotherapy resistance.

After 1, 3, 5, 7, and 10 days of culture, calcein AM and PI staining was used for assessing cell survival. The results showed that the 3D bioprinted tissue remained more than 95% active on the 10^th day. On the 7^th day of culture, Ki67 proliferation immunofluorescence staining of colorectal cancer cells was strongly positive, indicating that colorectal cancer cells had good adaptability in the 3D bioprinted model. We compared the proliferation of colorectal cancer cells in 3D culture, 3D printing-S, and 3D printing-M, and the results showed that the proliferation rate of colorectal cancer cells in 3D printing-S was higher than that in the sandwich culture model. The proliferation rate of colorectal cancer in the 3D printing-M was not inferior to that in the sandwich culture model (P<0.05).

We established the multicellular colorectal cancer model using extrusion-based bioprinting. This extrusion-based 3D bioprinting model has the advantages of being simple, convenient, rapid, and economical. In addition to extrusion-based 3D bioprinting, there are other 3D bioprinting methods, mainly including droplet-based bioprinting, laser-based bioprinting, and photocuring-based bioprinting[36]. Droplet-based bioprinting can be used to build models with high resolution, but this method usually has high cell damage, and it is difficult to print biological materials with high viscosity, which affects the structural stability of 3D bioprinting. Laser-based 3D bioprinting has been reported to produce high-resolution tumor cell models[37]. The laser-based 3D bioprinting is a nozzle-free technique that avoids mechanical cell damage and can even manipulate a single cell. In addition to more available choices of printing materials, extrusion-based 3D bioprinting has advantages over droplet-based bioprinting in terms of biological ink deposition rate and scale printing. In addition, the 3D bioprinted model showed relatively high cell viability rate[18]. But extrusion-based bioprinting usually has a relatively low resolution, which depends on the diameter of the print nozzle and the characteristics of the biomaterial, and the resolution of extrusion-based bioprinting currently exceeds 100 μm[32,38]. However, compared with the extrusion-based bioprinting used in our study, laser-based bioprinting causes more damage to cells. More importantly, it is expensive and difficult to operate. Photocuring-based bioprinting has the same problems and needs further development.

Immunofluorescence staining of the 3D printing-M showed that CDX2 of SW480, CD68 of M2 macrophages, and CD31 of endothelial cells were strongly positive on the 10^th day. The results confirmed that tumor cells, macrophages M2, and endothelial cells could survive well in the 3D printing-M and that the constructed 3D printing-M of mesenchymal tumor cells was successful. In addition, through exploration and attempts, we were able to produce frozen tissue sections of 3D bioprinted models and understand the morphological characteristics of tumor cells and interstitial cells through HE staining of the frozen sections. The results show that the cell lines have a certain degree of tissue characteristics depending on the bionic characteristics of the 3D bioprinting technology.

From the perspective of transcriptome, we analyzed the differences in gene expression among the 3D printing-S group, 2D culture group, sandwich culture group, and 3D printing-M group. As can be seen from the volcano map of DEGs, the number of DEGs in SW480 cells was as high as 11107 between the two culture models of 2D culture and 3D printing-S group. GO enrichment analysis indicated that compared with colorectal cancer cells cultured in 2D culture, the gene expression of colorectal cancer cells cultured in a 3D bioprinted model was significantly different in biological processes, cell composition, and molecular function[39]. KEGG enrichment analysis showed that compared with the 2D culture group, metabolic pathway, and oxidative phosphorylation pathway were the most significantly upregulated gene enrichment pathways in the 3D printing-S group[40-42]. These results suggest that colorectal cancer cells in the 3D bioprinted model are significantly different from those in the 2D cultures. The main sources of these differences are the different cell culture models, the different structures of the cell survival space, and the different degrees of communication between cells. The functions of the hub genes between the two models were mainly focused on cell proliferation, apoptosis, cell cycle regulation, and drug resistance. Therefore, it is reasonable to use more physiological models for subsequent functional verification and drug screening.

Compared with colorectal cancer cells in sandwich culture, the differences in upregulated genes of colorectal cancer cells in the 3D printing-S were mainly concentrated in cell composition, followed by biological process, and there was little difference in molecular function.

Different types of 3D culture models have varying effects on the proliferation and signal communication of colorectal cancer cells. The 3D bioprinted model has a unique spatial structure. KEGG enrichment analysis showed that focal adhesion and adhesion junctions were the most significant pathways enriched by downregulated gene in the 3D printing-S and sandwich culture groups[43]. The results suggest that the same bio-ink material and colorectal cancer cells, the 3D bioprinted model, and the sandwich model have different cell-cell communication and adhesion mechanisms. The functions of the hub differential genes between the two models were mainly focused on signal transduction, immune response, cell proliferation, apoptosis, lipid metabolism, etc. Cell cycle regulation and signal transduction were also significantly different in 3D culture of different types.

In this study, we explored the influence of tumor-associated macrophages M2 and endothelial cells on colorectal cancer cells. The multicellular model constructed by the concentric axis dual-nozzle 3D bioprinter is conducive to the mechanistic study of the influence of the TME on tumor cells. To analyze the effect of mesenchymal cells on the tumor cell signaling pathway, we screened for important hub differential genes. Compared with colorectal cancer cells in the 3D printing-S, the upregulated DEGs in the 3D printing-M were mostly concentrated in biological processes. In terms of cell composition, the differences between them were small, considering that both were 3D bioprinted models. The difference between the two models lies in the influence of the microenvironment on interstitial cells. KEGG enrichment analysis of the enrichment pathways of DEGs in biological functions of colorectal cancer cells in 3D printing-M group and 3D printing-S group showed that in tumor cell-mesenchymal cell co-culture model, the effects of mesenchymal cells on colorectal cancer cells involve biological processes such as cell growth, cell metabolism, protein modification, signal regulation, intracellular signal transduction, and intercellular communication. Pathways with significant enrichment of upregulated DEGs include the PD-1/PD-L1 checkpoint pathway and the EGFR receptor tyrosine kinase inhibitor resistance pathway[44,45]. The PD-1/PD-L1 checkpoint pathway is an important mechanism of tumor immune escape. The PD-L1 in tumor cells and the TME binds to PD-1, restricting the host immune response. The expression of PD-L1 is dependent on tumor-associated M2 macrophages in the TME, which promotes the apoptosis of antitumor T cells and the growth of colorectal cancer cells. Tyrosine kinase inhibitors are small-molecule antitumor drugs that can cross the cell membrane and block the signaling pathway of tumor cell growth and division. Resistance to tyrosine kinase inhibitors produces a pro-cancer effect[46]. KEGG enrichment analysis suggested that the upregulation of the PD-1/PD-L1 checkpoint pathway and EGFR receptor tyrosine kinase inhibitor resistance pathway in colorectal cancer cells in the 3D printing-M was reflected in our 3D bioprinted multicellular co-culture model to a certain extent. The influence of the TME on tumor cells has been verified at the transcriptomic level. The hub differential genes between the two models (3D printing-S and 3D printing-M) include IL1B, FCGR2A, FCGR3A, CYBB, SPI1, CCL2, and other functional genes related to immune response, immune regulation, and inflammatory response[47,48]. The hub differential genes between the two models (3D printing-S and 3D printing-M) also included genes such as ITGAM, ITGB2, and other genes related to cell conduction and cell adhesion[49,50]. By analyzing hub genes, we found that in the constructed co-culture model of tumor cell-mesenchymal cells, the influence of mesenchymal cells on tumor cells was implicated in biological processes such as cell growth, cell metabolism, protein modification, signal regulation, intracellular signal transduction, and intercellular communication. The TME composed of intermediate cells in the 3D printing-M may have an impact on colorectal cancer cells, which reveals the potential value of the 3D printing-M in the functional study of the effect of TME on tumor cells.

There are also great differences between 2D culture models and 3D printing-S, and between 3D printing-S and 3D printing-M in the experimental results of antitumor drug screening. The IC₅₀ values for 5-FU in 2D culture group and 3D printing-S group were 12.79 μM and 31.13 μM, respectively. The IC₅₀ value of the 3D printing-S group was three times that of the 2D culture group, which was more resistant to drugs. For oxaliplatin, the IC₅₀ of 2D culture group and 3D printing-S group was 0.80 μM and 26.79 μM, respectively. The concentration of 0.80 μM in the 2D culture model has a significant inhibitory effect on colorectal cancer cells, which is obviously not in line with the actual situation of human body. In the drug screening experiments of three chemotherapeutic drugs, the IC₅₀ measured by the 3D printing-S group was closer to the effective blood drug concentration in the human body, indicating that the 3D bioprinted colorectal cancer model has potential value in the field of anti-colorectal cancer drug screening[51-55]. In our experiment, we found that colorectal cancer cells in the 3D printing-M were significantly resistant to chemotherapy, so we increased the concentration gradient and found that only at high concentrations did the colorectal cancer cells in the 3D printing-M group show a drug inhibition response. The addition of tumor-associated macrophages M2 and endothelial cells to the multicellular co-culture model may have a protective effect on colorectal cancer cells and make colorectal cancer cells more resistant to antitumor drugs. Further exploration in the molecular mechanism of chemotherapeutic resistance in colorectal cancer is warranted in subsequent studies.

Our study has some limitations. This is the first time that our research group developed a 3D bioprinted multicellular co-culture model using concentric axis dual-nozzle 3D biopriting, which is technically worthy of further optimization. Our 3D printing-M was incorporated with SW480 cell lines and was not constructed with primary colorectal cells. We considered using more stable cell lines in the modeling stage to ensure the stability and repeatability of the experiment to a certain extent. In the future, our research group will validate primary colorectal cancer cells based on this experimental model and plan to use this model for personalized drug prediction.

Funding

This project was supported by Beijing Natural Science Foundation (7212077), CAMS Innovation Fund for Medical Sciences (CIFMS) (No.2021-I2M-1-058), Tsinghua University-Peking Union Medical College Hospital Cooperation Project (PTQH201904552), the Non-profit Central Research Institute Fund of Chinese Academy of Medical Sciences (2021-JKCS-006), National High Level Hospital Clinical Research Funding (No. 2022-PUMCH-B-003) and CAMS Innovation Fund for Medical Sciences (2021-1-I2M-015).

Conflict of interest

The authors declare that they have no conflict of interest.

Author contributions

Conceptualization: Yilei Mao, Bin Wu, Huayu Yang

Formal analysis: Peipei Wang, Lejia Sun, Huayu Yang

Investigation: Peipei Wang, Lejia Sun, Changcan Li

Methodology: Yilei Mao, Bin Wu, Huayu Yang

Resources: Huayu Yang, Yilei Mao, Bin Wu, Bao Jin, Peipei Wang

Writing – original draft: Peipei Wang, Huayu Yang

Writing – review & editing: Huayu Yang, Pepei Wang

All authors have read and approved the final version to be published.

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors agreed to the publication of the article.

Availability of data

No additional data are available to the public.