Construction of 3D tumor in vitro models with an immune microenvironment exhibiting similar tumor properties and biomimetic physiological functionality

Abstract

Tumors pose a serious threat to people's lives and health, and the complex tumor microenvironment is the biggest obstacle to their treatment. In contrast to the basic protein matrices typically employed in 2D or 3D cell culture systems, decellularized extracellular matrix (dECM) can create complex microenvironments. In this study, a combination of physicochemical methods was established to obtain liver decellularized extracellular matrix scaffolds (dLECMs) to provide mechanical support and cell adhesion sites. By co-culturing tumor cells, tumor-associated stromal cells and immune cells, an in vitro 3D tumor model with a biomimetic immune microenvironment was constructed. By utilizing microenvironment data obtained from human liver tumor tissues and refining the double seeding modeling process, 3D in vitro liver tumor-like tissues with a tumor immune microenvironment (TIME) were obtained and designated as reconstructed human liver cancer (RHLC). These tissues demonstrated similar tumor characteristics and exhibited satisfactory physiological functionality. The results of metabolic characterisation and mouse tumorigenicity testing verified that the constructed RHLC significantly increased in vitro drug resistance while also closely mimicking in vivo tissue metabolism. This opens up new possibilities for creating effective in vitro models for screening chemotherapy drugs.

---

1. Introduction

Cancer is a leading cause of human mortality and a major impediment to increasing life expectancy.¹ Human tumors are highly complex, and their significant heterogeneity is considered the primary barrier to developing effective, patient-specific treatment approaches.^2,3 Different tumor phenotypes dynamically evolve during disease progression and clinical therapy. Therefore, deciphering tumor heterogeneity and establishing cancer research models capable of systematically defining and simulating tumor heterogeneity are of utmost importance and difficulty.

Currently, the traditional models used for cancer research in laboratories are animal models and 2D culture models—cell line models. Although cell lines are easy to manipulate, it is challenging to simulate the intricate relationships that exist between cells and the tumor microenvironment.^4–8 Animal models are an effective tool for understanding pathogenesis and therapeutic approaches.⁹ However, animal models tend to replace exogenous tumor stromal cells and immune cells with their own cells, limiting their predictability as tumor models.^10,11 Therefore, 3D bioengineered tumor models have been established in this study to better simulate the intricate internal environment of tumors, serving as a bridge between animal and human experiments.¹²

It is still a challenge to establish biomimetic 3D bioengineered tumor models. For conventional 3D hydrogel materials, the key bioactive factors needed for tissue engineering are absent, while decellularized extracellular matrix (dECM) scaffolds have been demonstrated to contain a wide variety of extracellular matrix components and a multitude of growth factors.^13,14 Therefore, extracellular matrix (ECM) has been used to create a range of 3D bioengineered tumor in vitro models. Besides, natural liver ECM is one type of ECM that can create a complex and biomimetic microenvironment.^15,16 It retains part of the morphology, ECM components, and ultrastructure of organs to support hepatocyte viability and function.^17,18 Furthermore, dLECMs are widely recognized as supporting hepatocellular carcinoma cell line culture and considered promising naturally derived biomaterials for in vitro hepatocyte culture.^19–21 Therefore, in this study, dLECMs were utilized as the foundational materials for establishing 3D tumor models. Besides, the traditional single method of dECM preparation is ineffective;²² therefore, this study employed a combination of physical and chemical methods supplemented with the use of nuclease for the effective preparation of dLECMs.

The development of tumors is influenced by the surrounding environment, in which various cell types exist in a dynamic milieu that regulates tumor growth, invasion, and spread. Currently, the disparity between 3D tumor tissues and actual tumor tissues is primarily due to the absence of their unique microenvironment. For example, the absence of tumor-related endothelial cells can lead to tissue hypoxia, resulting in suboptimal 3D tumor construction.²³Most solid tumors are composed of stromal cells, including cancer-associated fibroblasts (CAF), immune cells, tumour endothelial cells (TEC), adipocytes, ECM proteins, and soluble molecules such as cytokines, chemokines, or growth factors. These elements collectively constitute the distinct tumor microenvironment (TME).^24–26 Numerous methods exist for simulating the TME, among which multicellular co-culturing was proved to be a valuable approach.^27,28 Moreover, common 3D models usually lack an immune system, which is an incomplete model for the determination of anticancer drugs.²⁹ Immunotherapy has become a viable new treatment option for cancer in recent years. Since the therapeutic benefit is less than anticipated in a significant percentage of patients, researchers tend to examine the immune response against tumors in greater detail and develop more suitable algorithms for forecasting each patient's response to therapy. New approaches to tissue engineering can help produce humanized stroma-based models in vitro.³⁰ Therefore, in this study, activated stromal cells CAF and TEC, immune cells JURKAT, and dLECMs were used as non-tumor components to establish RHLC by multi-cell co-culturing with HepG2. Metabolic characterization and nude mouse tumorigenicity testing are utilized to value the RHLC, and the RHLC which exhibits robust metabolic functionality and resistance to chemotherapy might be a potential model for preclinical drug screening.

2. Materials and methods

2.1. Preparation of dLECMs

Fresh pig liver was acquired from a slaughterhouse, cut into 5 mm³ cubes, and placed in an ultra-low-temperature refrigerator overnight. The tissue sample was rinsed three times with deionized water for 5 minutes each time, and this process was repeated three times. The tissue samples were then placed in an ion washing buffer containing 1% (v/v) Triton X-100 (Sigma) and 0.5% (w/v) SDS (Sinopharm, China), as well as Phosphate Buffered Saline (PBS) solution with an appropriate amount of antibiotics, and stirred. After two days, the waste solution was filtered out using a funnel, and the tissue was rinsed three times with sterile PBS for 5 minutes each time. This process was repeated multiple times until the tissue became transparent. Finally, the samples were washed with nuclease (70 U ml⁻¹ nuclease, 50 mM L⁻¹ Tris-C1) (Yesen, China) for 24 hours. The rinsed transparent tissue samples were washed several times with sterile PBS to remove residual Triton X-100 and SDS. The samples were then immersed in 75% medical alcohol for two hours, exposed to UV radiation for 1 hour, and stored at −20 °C.

2.2. dLECM characterization

To validate the decellularization process of pork liver, tissue sections were subjected to component analysis using an H&E staining kit (Solarbio, China) and a Masson staining kit (Solarbio, China). Fresh liver tissue and dLECMs were paraffin-embedded for histological staining and observed by bright-field microscopy using an inverted fluorescence microscope IX73 (Research-grade) (Olympus). The morphology of the decellularized tissue and cell adhesion after cell seeding were examined by scanning electron microscopy (SEM). The dLECMs and cell-seeded scaffolds were coated with Pt and observed using SEM (EM-30/X-ACT, COXEM) at a voltage of 10 kV. The major DNA components were analyzed using DAPI (Yesen, China) and the DNeasy kit.

2.3. Cell compatibility and phenotypic assessment

The human liver cancer cells (HepG2), human umbilical vein endothelial cells (HUVEC), human embryonic lung fibroblasts (HELF), and human T lymphocyte leukemia cells (JURKAT) used in the experiments were sourced from the Chinese Academy of Sciences Cell Repository. JURKAT cells were cultured in a 1640 medium (10% FBS + 1% PS) (Gibco), while other cells were cultured in a DMEM (10% Fetal Bovine Serum (FBS) + 1% Penicillin–Streptomycin (PS) Solution) (Gibco). For the culture of 3D cells, dLECMs were first adhered to culture dishes. After the scaffolds adhered, the culture medium was added and allowed to infiltrate for 2 hours. Subsequently, the culture medium was removed, and an appropriate number of cells were seeded in the incubator for 4–6 hours. Once cell adhesion to the scaffold was observed under a microscope, a fresh culture medium was added to initiate 3D culture, and the medium was changed when it turned yellow. The viability of HepG2 cells in 3D cultures after 3 days was measured using a Calcein-AM/Propidium Iodide (PI) Live–Dead staining kit (Solarbio, China). The control group was the culture dish group (2D). To evaluate the impact of 24-hour culture under 2D and 3D conditions on the phenotype of HepG2 cells, staining was performed using an Anti-Vinculin antibody (Proteintech, China) and Anti-Ki67 antibody (Proteintech, China). For assessing the impact on the HELF cell phenotype, staining was done using an Anti-MMP2 antibody (Proteintech, China). Cells were resuspended in 500 μL of PBS (1 × 10⁵ cells per tube) and analyzed using a flow cytometer (CytoFLEX, Beckman Coulter).

2.4. Activation of stromal cells

When the HepG2 cell density reached 80%, the culture medium was replaced, and incubation was continued for another 24 hours to collect the supernatant after centrifugation, which was used as a conditioned medium (CM) (referred to as HepG2 CM). Normal stromal cells (HELF and HUVEC) were seeded in a regular culture medium, and after reaching confluency, they were switched to the HepG2 CM culture medium (100%) and incubated for an additional 48 hours. Subsequently, they were respectively activated into cancer-associated fibroblasts (CAF) and tumor-associated endothelial cells (TEC). To assess the degree of stromal cell activation, immunofluorescence was performed. Anti-α-Smooth Muscle Actin antibody (Anti-α-SMA, Proteintech, China) and DAPI (Yesen, China) were used for co-staining of HELF and CAF. An Anti-CD144 antibody (Proteintech, China), TRITC Phalloidin (Yesen, China), and DAPI (Yesen, China) were used for co-staining of HUVEC and TEC. Observations were made under an inverted fluorescence microscope IX73 (Research Grade) (Olympus). Real-time PCR was also conducted to evaluate gene expression. After 48 hours of culture in 6 cm culture dishes, RNA extraction was carried out using RNAiso Plus (Takara) and chloroform (Sinopharm, China). Reverse transcription of RNA into cDNA was performed using the inNova Uscript II All in One First-strand cDNA Synthesis SuperMix (Innovagene) (Vazyme). Taq SYBR Green qPCR Premix (with ROX) (Innovagene) was used for qPCR with β-actin as an internal reference. The expression of genes including Fibroblast activation protein-α (FAP), Fibroblast-specific protein 1 (FSP1), Collagen I, Transforming Growth Factor-β (TGF-β) for CAF, and Tumor endothelial marker 1 (TEM1), Tumor endothelial marker 8 (TEM8), Vascular Endothelial Growth Factor (VEGF), and Biglycan for TEC was analyzed (Table 1).

Table 1 Primer sequences

Genes	Forward primer sequences	Reverse primer sequences
β-Actin	TGGCACCCAGCACAATGAA	CTAAGTCATAGTCCGCCTAGAAGCA
FAP	ACGGCTTATCACCTGATCGG	AATTGGACGAGGAAGCTCATTT
TGF-β1	ACTGCAAGTGGACATCAACG	TGCGGAAGTCAATGTACAGC
Collagen I	ATCAAGGTCTACTGCAACAT	CAGGATCGGAACCTTCGCT
FSP1	GATGAGCAACTTGGACAGCAA	CTGGGCTGCTTATCTGGGAAG
Biglycan	AGGAGGCGGTCCATAAGAAT	AGGGTTGAAAGGCTGGAAAT
VEGF	TTGCCTTGCTGCTCTACCTCCA	GATGGCAGTAGCTGCGCTGATA
TEM 8	CGGATTGCGGACAGTAAGG	GCCAGAACCACCAGAGGAG
TEM 1	TCGAGTGTTATTGTAGCGAGGGACATG	AGGTGGGCTCCGGGTAGGGTAT

---

2.5. Screening of human tumor tissue information

Paraffin sections of human liver cancer tissue samples were obtained from Xiangya Hospital and subjected to immunohistochemical staining. After antigen retrieval, they were incubated with primary antibodies, including alpha-fetoprotein (Anti-AFP) and Anti-CD45 Antibodies (Proteintech, China), for 24 hours. Subsequently, they were incubated at room temperature with HRP-labeled secondary antibodies for 40 minutes, followed by incubation in a DAB working solution for 5 minutes. After staining with hematoxylin, counterstaining, dehydration, and drying, observations were made under an inverted fluorescence microscope IX73 (Research Grade) (Olympus).

2.6. Construction of tumor tissues

Four groups with consistent cell numbers were employed to establish the tumor microtissue model. Control group: a mixture of HepG2, HUVEC, and HELF cells in a 5 : 1 : 1 ratio was evenly seeded onto dLECMs. Activation group: a mixture of HepG2, TEC, and CAF cells in a 5 : 1 : 1 ratio was evenly seeded onto dLECMs. Double Seeding group: after two days of cell culture in the Activation group, a second seeding with the same ratios was performed, maintaining the same cell numbers as the first seeding. Immune Environment group: A mixture of HepG2, TEC, CAF, and JURKAT cells in a 30 : 6 : 6 : 5 ratio was evenly seeded onto dLECMs, as depicted in the schematic diagram of tumor tissue construction (Graphical abstract). Following cell culture for 3, 7, 15, and 30 days, live cell staining was conducted using the calcein-AM/propidium iodide (PI) live-dead staining reagent (Solarbio, China), and observations were made under a laser confocal microscope (Olympus).

2.7. Physiological function characterization

2.7.1. Metabolic capacity assessment. To assess the metabolic changes of in vitro tumor tissues, we collected cell culture supernatants after culturing the tumor tissues in vitro for 1, 3, 5, and 7 days. The glucose content was determined following the instructions of a glucose assay kit (Solarbio, China), and the lactate (LD) content was measured using an LD assay kit (Nanjing Jiangcheng Bioengineering Institute, China). We followed the kit instructions to set up the groups, and after the reaction with the reagents, the absorbance at the corresponding wavelengths (505 nm for glucose and 570 nm for lactate) was measured using a Nanodrop 2000 (Thermo Scientific).

2.7.2. Drug resistance capability. The dosage of two chemotherapeutic agents was set as follows: 5-Fluorouracil (5Fu) at 25 μM and Doxorubicin (Dox) at 10 μM. Prior to drug administration, it is imperative to observe the cultured cells, initiating drug treatment only when the cells have reached their optimal growth state. Following the addition of drugs to the culture medium, the culture dish was gently agitated to ensure uniform drug distribution. Subsequently, the cultivation was continued in an incubator for an additional 24 hours. To characterize the drug resistance of in vitro tumor tissues after treatment with two anti-cancer chemotherapy drugs, DOX and 5-FU, we used a CCK8 assay kit (Solarbio, China) to assess cell viability. The groups were set up according to the kit instructions, and after the reaction with the reagents, the absorbance at 450 nm was measured using a Nanodrop 2000 (Thermo SCIENTIFIC).

2.7.3. Evaluation of cell viability. To further investigate the metabolic differences between immune cells, 2D culture, and 3D culture, as well as the drug resistance of tumor tissues, flow cytometry analysis was conducted. To assess cell viability, we employed PI dye (Calcein-AM/Propidium iodide-PI Live–Dead Staining Kit, Solarbio, China). Apoptosis experiments used the Annexin V apoptosis assay kit (Thermo Fisher), and analysis was performed using a CytoFLEX flow cytometer (Beckman Coulter).

2.8. Tumorigenicity assessment of tumor microtissues

All animal procedures were performed in accordance with the Guidelines for Care and Use of Laboratory Animals of Hunan University and approved by the Animal Ethics Committee of Hunan University (Permit no. HNUBIO202101009). Five-week-old athymic nude mice were randomly divided into five groups. Group A consisted of HepG2 cells cultured in 2D. Group B included HepG2, TEC, and CAF co-cultured in 2D. Group C involved co-culturing of HepG2, TEC, CAF, and JURKAT cells in 2D. Group D included HepG2, TEC, and CAF co-cultured in a 3D environment. Group E comprised HepG2, TEC, CAF, and JURKAT cells co-cultured in a 3D environment. The number of cells was consistent among all groups. In the 3D group, cells were co-implanted with dLECMs under the skin of the nude mice. To ensure consistency, the cell-seeded scaffold was placed under the skin of the mice immediately after removing it from the culture dish. Each group of mice was housed separately, and their body weight was measured every three days. A caliper was used to measure the length and width of the tumors. The tumor volume was estimated based on the formula
where “L” represents the measured length of the tumor, and “W” represents the measured width of the tumor.

2.9. Statistical analysis

Each group included a minimum of three parallel data points, inter-group comparisons were performed using a t-test, and the comparisons between three groups or more groups were conducted by one way analysis of variance (ANOVA). Graphical analysis and plotting were conducted using Prism 8. Statistical significance was considered when p < 0.05. The significance levels between groups were represented as **** (p < 0.0001), *** (p < 0.001), ** (p < 0.01), and * (p < 0.05).

3. Results

3.1. Preparation and characterization of dLECMs

Fiber-rich dLECMs were obtained as scaffolds for RHLC. Based on previous methods,³¹ the preparation of dLECMs was optimized and the brief process is illustrated (Fig. 1A). The dLECMs displayed a porous network structure (Fig. 1B), as seen in SEM, revealing a fibrous mesh-like structure conducive to cell adhesion (Fig. 1C). The success of tissue construction is closely related to the residual components of the dLECMs. Therefore, we performed DAPI staining, HE staining, and Masson staining on the dLECMs (Fig. 1D–F). The results indicated that after the decellularization process, cellular components were removed effectively while the ECM components were retained. DAPI staining (Fig. 1D) and H&E staining (Fig. 1E) confirmed that after decellularization, nearly all intact nuclei were removed, and a significant amount of cellular debris was eliminated. Masson trichrome staining confirmed that the remaining ECM components were predominantly collagen fibers (Fig. 1F), most of the elastin fibers were removed after decellularization (Fig. 1G), and collagen fibers were abundant (Fig. 1H). Subsequent DNA content analysis showed that the ECM group had less than 50 ng mg⁻¹, meeting the standard of double-stranded DNA content below 50 ng per milligram of ECM.³² This further demonstrates that the nucleus component was mostly extracted during decellularization (Fig. 1I). These results showed that the cellular components in liver tissues were effectively removed while a large number of collagen fibers were retained, and dLECMs were successfully prepared, making them ideal scaffolds for the construction of three-dimensional RHLC.

Fig. 1 Preparation, identification, and characterization of dLECMs. (A) Schematic diagram of dLECM preparation. (B) Physical diagram of dLECMs. Dimensions of dLECMs (top) and morphology of dLECMs (bottom). (C) Scanning electron microscopy (SEM) images of dLECMs. dLECMs for low magnification (top) and dLECMs for high magnification (bottom). Scale bar = 1 μm. (D) DAPI staining results of liver tissue and dLECMs. (E) H&E staining results of liver tissue and dLECMs. (F) Masson staining results of liver tissue and dLECMs. (D–F) Nat: native liver tissue, Dec: decellularized liver tissue. Scale bar = 100 μm. (G) Quantification of elastic fibers. (H) Quantification of collagen. (I) DNA content analysis. (G–I) N liver: native liver tissue and D liver: decellularized liver tissue.

3.2. In vitro cell compatibility and cell phenotype regulation

The cytotoxicity and regulatory role of dLECMs as RHLC scaffolds were determined. Live/dead staining results demonstrate that after HepG2 cells were cultured for 3 days, there is no significant difference in the number of live and dead cells compared to the 2D culture. The cell state appears to be saturated, and no pathological changes are observed (Fig. 2A and B). This suggests that dLECMs are non-toxic to cells and exhibit good biocompatibility. Flow cytometry results showed that, compared to the 2D culture, tumor cells on dLECMs exhibit higher levels of Vinculin (Fig. 2C) and Ki67 (Fig. 2D) expression, indicating that adhesion and proliferation capabilities of tumor cells were enhanced in the dLECMs. Besides, matrix metalloproteinases (MMPs) are important mediators of fibroblasts in extracellular matrix remodeling, and dLECMs induce fibroblasts to express more MMP2 (Fig. 2E). These results suggest that dLECMs have good cytocompatibility, which is conducive to the regulation of tumor cell adhesion, proliferation and extracellular matrix remodeling, and can provide a good growth environment for cells.

Fig. 2 Toxicity assessment and functional characterization of dLECMs. (A) Live cells (green), dead cells (red), scale bar = 200 μm. (B) Quantification of live and dead cells in the image from (A) using Image J. (C) Expression of Vinculin in HepG2 cells. (D) Expression of Ki67 in HepG2 cells. (E) Expression of MMP2 in CAF (Cancer-Associated Fibroblasts). (C–E) Flow cytometry results comparing the 2D culture (blue) and the 3D culture (dLECMs) (red).

3.3. Activation of tumor-associated stromal cells

The formation of tumors is accompanied by changes in the surrounding stroma, where tumor-associated stromal cells create a suitable tumor microenvironment for tumor growth. To obtain tumor-associated stromal cells required for RHLC construction, we activated stromal cells to become tumor-associated stromal cells. After stimulating HELF with a conditioned medium from tumor cells for 48 hours, immunofluorescence results showed an upregulation of α-SMA expression in CAF, accompanied by a change in cell morphology from an elliptical shape to a spindle-shape (Fig. 3A), which resembles the appearance of CAF. Fluorescence quantitative PCR experiments were conducted for FAP, FSP, Collagen I and TGF-β. The results revealed a significant upregulation in the expression in CAF compared to that in HELF (Fig. 3C). Upon stimulation of HUVEC with a conditioned medium from tumor cells for 48 hours, immunofluorescence results demonstrated an upregulation of CD144 expression in TEC (Fig. 3B), accompanied by a change in cell morphology from rounded to dendritic (Fig. 3D), indicating a more mature vascular structure of TEC. To further validate the activation of HUVEC, we examined the expression of TEM1, TEM8, VEGF, and Biglycan. The results revealed an upregulation in expression in TEC compared to that in HUVEC (Fig. 3E), consistent with previous reports.³³ These results indicated that HELF was activated to CAF and HUVEC was activated to TEC. Successful acquisition of tumor-associated stromal cells could provide a tumor microenvironment for RHLC construction.

Fig. 3 The activation of tumor-associated stromal cells was assessed 48 hours after stimulation. (A) Immunofluorescence staining images of α-SMA (green) in normal fibroblasts (HELF) and tumor-associated fibroblasts (CAF). (B) Immunofluorescence staining images of specific protein CD144 (green) in normal endothelial cells (HUVEC) and tumor-associated endothelial cells (TEC). (C) Gene expression analysis of CAF markers FAP, FSP, collagen I, and TGF β. (D) Immunofluorescence staining images of the cytoskeleton (red) in HUVEC and TEC. Scale bar = 200 μm. (E) Gene expression analysis of TEC markers TEM1, TEM8, VEGF, and Biglycan (****, p < 0.0001; ***, p < 0.001; **, p < 0.01; n = 5).

3.4. Construction of RHLC

The cell ratio and inoculation method required to construct the RHLC were determined. In this study, the cell ratio in the tumor tissue model was determined based on human tumor tissue samples and previous laboratory modeling experience.³³ Immunohistochemical staining images and quantification results of AFP and CD45 showed that the cell ratio of hepatic cells and immune cells was around 6 : 1 in human liver tumor tissues (Fig. 4A–C). Combining previous experiences in modeling colorectal cancer,³³ the cell ratios for in vitro co-culture were established as HepG2 : CAF : TEC : JURKAT = 30 : 6 : 6 : 5. Several modeling approaches were compared and optimized for the construction of tumor microtissues (Fig. 4D), and the surviving microtissues were adhered to the dLECMs (Fig. 4E). Live cell staining results indicated that when unactivated stromal cells were co-cultured with tumor cells, microtissues survived at 3 days but died off completely at 7 days. However, when activated stromal cells were co-cultured with tumor cells, microtissues survived after 7 days. The process of aggregation of cells into tissues is time-consuming and requires a significant amount of cells to generate an “aggregation effect”. When stromal cells and tumor cells were co-cultured for 2 days and then re-seeded, small amounts of microtissues survived at 15 days. The immune microenvironment is associated with tumor progression, as immune cells secrete various cytokines that promote tumor growth. Many microtissues persisted even after 30 days when tumor cells were co-cultured with immune cells and activated stromal cells with second-seed, and distributed tumor aggregates were seen on dLECMs with the generation of tubulointerstitial structures (Fig. 4F). Besides, the H&E staining also verified the existence of cell aggregates resembling tissue structures (Fig. 4G). These results indicate that the cell ratio required for RHLC construction was HepG2 : CAF : TEC : JURKAT = 30 : 6 : 6 : 5, and the inoculation method was that HepG2, TEC, CAF, and JURKAT were co-cultured in dLECMs, and a second inoculation was done on the next day, which greatly prolonged the lifespan of the constructed RHLC, allowing it to mimic tumor tissue formation and exhibit vascular-like structures. Hence, the construction of RHLC was established following this protocol and culture condition in this article.

Fig. 4 Human liver tumor tissue cell and information screening construction of in vitro tumor tissues with an immune microenvironment. (A) Immunohistochemical staining image of the liver cancer marker AFP. (B) Immunohistochemical staining image of the immune cell marker CD45. (C) Quantification of (A) and (B), with an AFP : CD45 ratio of 6 : 1. Scale bar = 100 μm. (D) Live cell staining results of tumor tissues. Control group: co-culture of HepG2, HUVEC, and HELF on dLECMs for 3 and 7 days. Activation group: co-culture of HepG2, TEC, and CAF on dLECMs for 3, 7, and 15 days. Double Seeding group: two days of cell culture in the Activation group, followed by reseeding in the same proportions for 3, 7, 15, and 30 days. Immune Environment group: HepG2, TEC, CAF, and JURKAT were inoculated into dLECMs for co-culture, and the second inoculation was performed on the next day to continue the culture for 3, 7, 15, and 30 days. Scale bar = 200 μm. (E) Merged bright-field images of dLECMs and live cell staining. Control group: HepG2, HUVEC, and HELF were inoculated into dLECMs for 3 days (left). Activation group: co-culture of HepG2, TEC, and CAF on dLECMs for 3 days (right). Scale bar = 200 μm. (F) Live cell staining images of microtissues on dLECMs. Immune Environment group: HepG2, TEC, CAF, and JURKAT were inoculated into dLECMs for co-culture, and the second inoculation was performed on the next day to continue the culture for 30 days. Scale bar = 50 μm. (G) H&E staining images comparing human tumor tissue (hepatocellular carcinoma, HCC) (left) with in vitro constructed 3D tumor-like tissue with TIME (reconstructed human liver cancer, RHLC) (right). Scale bar = 100 μm.

3.5. Metabolic function assessment of RHLC

To explore the tumor properties of RHLC, its metabolic capacity including glucose and lactate levels and cell viability were examined. Four groups were chosen to determine glucose and lactate concentrations in the culture medium: 2D-HepG2-CAF-TEC (2D-Immunodeficiency), dLECMs (3D)-HepG2-CAF-TEC (3D-Immunodeficiency), 2D-HepG2-CAF-TEC-JURKAT (2D-Immune), and dLECMs (3D)-HepG2-CAF-TEC-JURKAT (3D-Immune, RHLC). The glucose detection results (Fig. 5A) indicate that the rate of consumption in 3D is faster than that in 2D. The glucose concentration in 3D drops below 0.1 mmol L⁻¹ after 5 days, whereas it remains around 0.1 mmol L⁻¹ in 2D even after 7 days. Moreover, 3D-Immune (RHLC) exhibits a higher glucose consumption rate compared to 3D-Immunodeficiency. Lactate detection results (Fig. 5B) indicate that more lactate was generated in 3D than that in 2D. Moreover, 3D-Immune (RHLC) demonstrates the highest efficiency in lactate production. Simultaneously, flow cytometry results demonstrate that 3D-Immune (RHLC) exhibits higher cellular viability compared to 2D-Immune (Fig. 5C). 2D and 3D-immune CM culture of immune cells (JURKAT) was used to explore the effect of different dimensions of the conditioned medium (CM) on immune cells. Immune cell (JURKAT) CM culture of 2D and 3D-Immunodeficient cells was used to explore the effect of the immune cell conditioned medium (CM) on immunodeficient cells of different dimensions. Flow cytometry revealed that the cellular activity of immune cells showed a similar tendency in 2D and 3D CM (Fig. 5D). However, the immune cell CM improved the cell activity in the 3D-Immunodeficiency group. As shown in Fig. 5E, the activity of the 3D-Immunodeficiency group was significantly enhanced when cultured with JURKAT CM. This improvement may be attributed to soluble molecules secreted by immune cells, creating an immunosuppressive microenvironment that protects tumor cells from destruction and promotes tumor cell proliferation. In summary, the introduction of the immune microenvironment improves the growth conditions for tumor cells, making tumor cells’ metabolism more vigorous. These data suggest that RHLC, consisting of 3D liver decellularized extracellular matrix scaffolds with multi-cell co-culture and immune microenvironment, exhibits similar tumor properties and is the most metabolically active compared to the other groups, since it exhibits the highest rate of glucose consumption and the highest rate of lactate production as well as the strongest cell viability.

Fig. 5 Metabolic function assessment of RHLC. (A) Glucose (505 nm) and (B) lactate (570 nm) concentrations in the culture medium of tumor microtissues cultured for 1, 3, 5, and 7 days. Flow cytometry with propidium iodide (PI) staining was employed to assess cellular activity in 2D-Immune and 3D-Immune (C). The impact of 2D-Immunodeficiency and 3D-Immunodeficiency CM on the cellular activity of immune cells (JURKAT) was examined (D). Additionally, the influence of immune cell (JURKAT) CM on the cellular activity of 2D-Immunodeficiency and 3D-Immunodeficiency was investigated (E).

3.6. In vivo tumorigenicity and in vitro drug resistance analysis

To determine the biomimetic physiological function of RHLC, its tumorigenicity and drug resistance were analyzed. For a more comprehensive analysis, five groups were set: 2D-HepG2, 2D-HepG2-CAF-TEC (2D-Immunodeficiency), 2D-HepG2-CAF-TEC-JURKAT (2D-Immune), dLECMs (3D)-HepG2-CAF-TEC (3D-Immunodeficiency), and dLECMs (3D)-HepG2-CAF-TEC-JURKAT (3D-Immune, RHLC). The results of subcutaneous tumor tissue development in nude mice after 30 days (Fig. 6B) demonstrated that 3D tumor tissues exhibited cystic morphology with a compact structure, and the darker red color indicates greater stromal vascularization, in comparison with tumor tissues without dLECMs. In all groups, 2D-HepG2 tumor volume was the smallest and 2D-Immune tumor volume was greater than that of 2D-Immunodeficiency. Tumor volume results (Fig. 6C) indicate that multicellular seeding promotes tumor tissue growth. Additionally, the incorporation of dLECM significantly enhances malignant tumor growth, leading to rapid and aggressive expansion. 3D-Immune and 3D-Immunodeficiency tumors have similar growth trends. Interestingly, in the 2D setting, the inclusion of immune cells plays a critical role in promoting tumor growth. However, in the 3D environment, the impact of immune cells on tumor growth is less pronounced. This may be attributed to altered cell–cell interactions and the barrier function of the extracellular matrix in the 3D environment. There were no significant differences in the mice's body weight between the groups (Fig. 6D). Although 3D-Immune and 3D-Immunodeficiency have similar tumor growth rates, 3D-Immune (RHLC) has a more significant cell viability compared with 3D-Immunodeficiency after treatment with chemotherapeutic medicines DOX and 5-FU (Fig. 6E). Multi-cell co-culture increases cell viability compared to single-cell culture. Experiments on apoptosis further showed that, in contrast to single-cell and 2D co-cultures, 3D tumor microtissues showed higher levels of drug resistance after chemotherapy, and that 3D-Immune (RHLC) showed lower levels of apoptosis (Fig. 6F). These results suggest that RHLC has a better bionic physiological function, since RHLC resembles the in vivo tumor microenvironment and showed excellent in vivo tumorigenicity and drug resistance.

Fig. 6 Tumorigenicity and drug resistance of microtumor tissues. (A) Subcutaneous tumor implantation model in mice. (B) Images of tumors harvested from nude mice in different groups after 30 days of subcutaneous tumorigenesis. (C) Tumor volume changes at various time points in each group. (D) Changes in body weight of the mice in different groups. (E) Changes in cell viability after treatment with chemotherapy drugs DOX and 5-FU (***, p < 0.001; n = 3). (F) Results of apoptosis flow cytometry analysis with chemotherapy drugs DOX and 5-FU.

4. Discussion

Scaffold, as one of the core elements in tissue engineering, serves the purpose of providing mechanical support for tissue growth and creating a microenvironment conducive to tissue growth.³⁴ Traditional dECM preparation may not enable complete cell elimination effectively.²² In this study, a combination of physical and chemical methods was employed for decellularization, complemented by the use of nucleases to remove surplus immunogenic materials. Results from HE staining and DAPI staining demonstrate the effectiveness of this method in thoroughly removing cellular components. Quantitative DNA analysis indicates that the enzymatic treatment method successfully eliminates immunogenic materials. dECM comprises a variety of components, and collagen and growth factors are important.³⁵ Masson staining results reveal that dLECMs contain collagen, elastin and a small quantity of elastic fibers. Besides, research studies have validated that collagen can provide mechanical support and promote cell adhesion.³⁶ Scanning electron microscopy was used to observe the dLECMs, revealing a mesh-like porous structure with a significant matrix presence, which might favor cell adhesion within dLECMs.

The dLECMs, as a natural biomaterial, could also mediate the cellular phenotype. Compared to the 2D culture, HepG2 cells in dLECMs express higher levels of Vinculin and Ki67. This suggests that dLECMs can induce cell proliferation and adhesion, thereby influencing cell phenotype changes. In the context of tumor modeling, stromal cells are indispensable, and the essential mediator of stromal remodeling by fibroblasts is matrix metalloproteinase 2 (MMP2). MMP2 holds significant physiological importance in tumor tissues.³⁷ Experimental validation indicates that dLECMs can promote the enhanced expression of MMP2 in stromal cells, which showed the positive role of dLECMs in inducing stromal remodeling phenotypes.

The tumor microenvironment (TME) plays a crucial role in altering cancer progression and treatment responses. A significant barrier to the advancement of cancer therapies is the discrepancy between the TME in tumor models and that of patients.³⁸ The tremendous heterogeneity of cell types within the TME is crucial for treatment response. The 3D organotypic culture of cancer types offers advantages in recapitulating the original tumor histology and mutational characteristics, facilitating long-term propagation for experimental research.³⁹ To model the complex tumor microenvironment, many researchers have established research platforms through multicellular co-cultures.^40–42 However, it may not fully encompass the multicellular composition of the TME.^43–45 In addition to tumor cells, research suggests that tumors are composed of multiple components such as stromal cells and immune cells that contribute to their tumoral heterogeneity.⁴⁶ Therefore, in this article, we established 3D in vitro liver tumour-like tissues with a tumour immune microenvironment (TIME). Besides, the phenotype of normal stromal cells differs from that of stromal cells within tumors. Therefore, in this study, tumor-associated stromal cells were induced by exposing stromal cells to the CM containing factors secreted by tumor cells. A comparison between activated and non-activated stromal cells revealed that the former enhanced the adhesion of co-cultured cells to the dLECMs. The development of tumors is closely linked to an immunosuppressive microenvironment. The unique immunosuppressive microenvironment of tumor tissues can protect tumor cells from destruction and create a favorable microenvironment for tumor cell growth. Since primary immune cells are difficult to obtain and expensive to culture, immune cell lines were chosen. T lymphocyte leukemia cells have been selected for their high expression of CD3, low levels of CD4, and possession of immune cell functions.^47,48

The establishment of tumor models requires consideration of various factors, and in a multicellular co-culture system, the seeding ratio of individual cells has a significant impact on tumor modeling. With the goal of creating biomimetic RHLC, cell seeding ratio data followed cellular information from hepatic tumor tissues. Based on successful laboratory experiences in the construction of 3D colorectal cancer tumors, a seeding ratio of HepG2 : CAF : TEC : JURKAT = 30 : 6 : 6 : 1³³ was adopted for the construction of the tumor tissue. In addition, the number of cells seeded also profoundly affects the efficiency of tumor tissue construction. Thus, secondary seeding was employed. Experimental results indicated that secondary seeding enabled the survival of tumor tissue for over 15 days, and the introduction of immune cells resulted in a pronounced immunosuppressive microenvironment, significantly extending the lifespan of the tumor tissue. Through in vitro metabolic function experiments and drug resistance assays, it was demonstrated that RHLC exhibits robust metabolic functionality and drug heterogeneity. Subsequent in vivo tumorigenesis experiments in nude mice further validated that the RHLC exhibits robust physiological functions, including metabolic activity and drug resistance. In general, potential new targets for immunotherapy may be discovered via the obtained RHLC, and RHLC could also be used as an in vitro experiment model for investigating drug delivery particles or nanoscale theranostics,⁴⁹ making it a valuable platform for drug screening and development.

In the current study, the obtained RHLC was proved to exhibit robust metabolic activity and high drug resistance, and a comprehensive assessment of cellular properties such as proliferation, adhesion, and metabolism has been conducted, but further research is still needed to reveal the operation mechanism of the obtained co-cultured system, such as cytoskeletal arrangement and paracrine and cell migration of the different cells which will provide crucial supplementary information for understanding the tumor microenvironment. Via analyzing the arrangement of the cytoskeleton, insights can be gained into how tumor cells adapt to and reshape their structure in a 3D microenvironment; cell migration is a critical step in tumor metastasis, and investigating the migratory and paracrine behavior of cells within 3D scaffolds can reveal the dynamic interactions between different cell types within the tumor microenvironment. Besides the hindrance might be how to distinguish or label the different cells since so many kinds of cells were in the system, and the indirect co-culture model and staining specific cell markers simultaneously might be utilized in follow-up studies.

5. Conclusion

In this study, a facile and effective approach was employed to fabricate a 3D scaffold suitable for tumour remodelling (dLECMs). The dLECMs exhibit excellent biocompatibility, promoting cell growth and effectively regulating cell phenotypes. The activation of stromal cells into tumor-associated stromal cells was successfully achieved. Through multicellular co-culturing involving secondary seeding of tumor cells, tumor-associated fibroblasts, tumor-associated endothelial cells and immune cells, a 3D in vitro liver tumor-like model with a tumor immune microenvironment (RHLC) was successfully constructed. The RHLC demonstrates robust metabolic activity and high drug resistance, making it a valuable preclinical model for drug screening. These findings offer novel insights for the further development of 3D in vitro tumor models.

Data availability

All relevant data are available and have been included in this article.

Conflicts of interest

The authors declare that they have no competing interests.

Acknowledgements

This study was financially supported by the Research and Development Plan of Key Fields in Hunan Province (No. 2022SK2027), the Natural Science Foundation of Guangdong Province (No. 2023A1515011463), the High-end Foreign Experts Recruitment Plan of China (No. G2023160019L), the Shanghai Key Laboratory of Peripheral Nerve and Microsurgery (20DZ2270200-202402), and the NHC Key Laboratory of Hand Reconstruction (Fudan University), Institute of Hand Surgery, Shanghai.

References

1. F. Bray, M. Laversanne, E. Weiderpass and I. Soerjomataram, The ever-increasing importance of cancer as a leading cause of premature death worldwide, Cancer, 2021, 127(16), 3029–3030.
2. P. L. Bedard, A. R. Hansen, M. J. Ratain and L. L. Siu, Tumour heterogeneity in the clinic, Nature, 2013, 501(7467), 355–364.
3. I. Dagogo-Jack and A. T. Shaw, Tumour heterogeneity and resistance to cancer therapies, Nat. Rev. Clin Oncol., 2018, 15(2), 81–94.
4. S. Chatterjee, G. S. Naidu, I. Hazan-Halevy, H. Grobe, A. Ezra, P. Sharma, M. Goldsmith, S. Ramishetti, D. Sprinzak, R. Zaidel-Bar and D. Peer, Therapeutic gene silencing of CKAP5 leads to lethality in genetically unstable cancer cells, Sci. Adv., 2023, 9(14), eade4800.
5. D. Hanahan and R. A. Weinberg, Hallmarks of cancer: the next generation, Cell, 2011, 144(5), 646–674.
6. K. S. Smalley, M. Lioni and M. Herlyn, Life isn't flat: taking cancer biology to the next dimension, In Vitro Cell. Dev. Biol.: Anim., 2006, 42(8–9), 242–247.
7. J. P. Gillet, S. Varma and M. M. Gottesman, The clinical relevance of cancer cell lines, J. Natl. Cancer Inst., 2013, 105(7), 452–458.
8. M. E. Katt, A. L. Placone, A. D. Wong, Z. S. Xu and P. C. Searson, in vitro Tumor Models: Advantages, Disadvantages, Variables, and Selecting the Right Platform., Front. Bioeng. Biotechnol., 2016, 4, 12.
9. K. Yan, H. Gao, X. Liu, Z. Zhao, B. Gao and L. Zhang, Establishment and identification of an animal model of long-term exercise-induced fatigue, Front. Endocrinol., 2022, 13, 915937.
10. P. Malaney, S. V. Nicosia and V. Davé, One mouse, one patient paradigm: New avatars of personalized cancer therapy, Cancer Lett., 2014, 344(1), 1–12.
11. S. C. Frederico, X. Zhang, B. Hu and G. Kohanbash, Pre-clinical models for evaluating glioma targeted immunotherapies., Front. Immunol., 2023, 13, 1092399.
12. A. Sharma, S. Sances, M. J. Workman and C. N. Svendsen, Multi-lineage Human iPSC-Derived Platforms for Disease Modeling and Drug Discovery, Cell Stem Cell, 2020, 26(3), 309–329.
13. R. Wang, X. He, Z. Chen, S. Su, J. Bai, H. Liu and F. Zhou, A nanoparticle reinforced microporous methacrylated silk fibroin hydrogel to promote bone regeneration, Biomater. Sci., 2024, 12(8), 2121–2135.
14. A. Almalla, L. Elomaa, L. Bechtella, A. Daneshgar, P. Yavvari, Z. Mahfouz, P. Tang, B. Koksch, I. Sauer, K. Pagel, K. H. Hillebrandt and M. Weinhart, Papain-Based Solubilization of Decellularized Extracellular Matrix for the Preparation of Bioactive, Thermosensitive Pregels, Biomacromolecules, 2023, 24(12), 5620–5637.
15. L. P. Ferreira, V. M. Gaspar and J. F. Mano, Decellularized Extracellular Matrix for Bioengineering Physiomimetic 3D in vitro Tumor Models, Trends Biotechnol., 2020, 38(12), 1397–1414.
16. X. Ma, C. Yu, P. Wang, W. Xu, X. Wan, C. S. E. Lai, J. Liu, A. Koroleva-Maharajh and S. Chen, Rapid 3D bioprinting of decellularized extracellular matrix with regionally varied mechanical properties and biomimetic microarchitecture, Biomaterials, 2018, 185, 310–321.
17. P. M. Baptista, M. M. Siddiqui, G. Lozier, S. R. Rodriguez, A. Atala and S. Soker, The use of whole organ decellularization for the generation of a vascularized liver organoid, Hepatology, 2011, 53(2), 604–617.
18. H. C. Ott, T. S. Matthiesen, S. K. Goh, L. D. Black, S. M. Kren, T. I. Netoff and D. A. Taylor, Perfusion-decellularized matrix: using nature's platform to engineer a bioartificial heart, Nat. Med., 2008, 14(2), 213–221.
19. Y. Wang, C.-B. Cui, M. Yamauchi, P. Miguez, M. Roach, R. Malavarca, M. J. Costello, V. Cardinale, E. Wauthier, C. Barbier, D. A. Gerber, D. Alvaro and L. M. Reid, Lineage restriction of human hepatic stem cells to mature fates is made efficient by tissue-specific biomatrix scaffolds, Hepatology, 2011, 53, 293–305.
20. K.-M. Park, K. H. Hussein, S.-H. Hong, C. Ahn, S.-R. Yang, S.-M. Park, O.-K. Kweon, B.-M. Kim and H.-M. Woo, Decellularized liver extracellular matrix as promising tools for transplantable bioengineered liver promotes hepatic lineage commitments of induced pluripotent stem cells, Tissue Eng. Part A, 2016, 22, 449–460.
21. Y. Cheng, Y. Wang, Y. Z. Kang, P. Y. Hu, Y. Gao and M. X. Pan, in vitro culture of tumour-derived hepatocytes in decellularised whole-liver biological scaffolds, Digestion, 2013, 87, 189–195.
22. X. Zhang, X. Chen, H. Hong, R. Hu, J. Liu and C. Liu, Decellularized extracellular matrix scaffolds: Recent trends and emerging strategies in tissue engineering, Bioact. Mater., 2021, 10, 15–31.
23. N. W. Choi, M. Cabodi, B. Held, J. P. Gleghorn, L. J. Bonassar and A. D. Stroock, Microfluidic scaffolds for tissue engineering, Nat. Mater., 2007, 6(11), 908–915.
24. D. Hanahan and R. A. Weinberg, Hallmarks of cancer: the next generation, Cell, 2011, 144(5), 646–674.
25. A. Labani-Motlagh, M. Ashja-Mahdavi and A. Loskog, The Tumor Microenvironment: A Milieu Hindering and Obstructing Antitumor Immune Responses, Front. Immunol., 2020, 11, 940.
26. T. R. Cox, The matrix in cancer, Nat. Rev. Cancer, 2021, 21(4), 217–238.
27. D. Öhlund, A. Handly-Santana, G. Biffi, E. Elyada, A. S. Almeida, M. Ponz-Sarvise, V. Corbo, T. E. Oni, S. A. Hearn, E. J. Lee, I. I. Chio, C. I. Hwang, H. Tiriac, L. A. Baker, D. D. Engle, C. Feig, A. Kultti, M. Egeblad, D. T. Fearon, J. M. Crawford, H. Clevers, Y. Park and D. A. Tuveson, Distinct populations of inflammatory fibroblasts and myofibroblasts in pancreatic cancer, J. Exp. Med., 2017, 214(3), 579–596.
28. G. Biffi, T. E. Oni, B. Spielman, Y. Hao, E. Elyada, Y. Park, J. Preall and D. A. Tuveson, IL1-Induced JAK/STAT Signaling Is Antagonized by TGFβ to Shape CAF Heterogeneity in Pancreatic Ductal Adenocarcinoma, Cancer Discovery, 2019, 9(2), 282–301.
29. P. Shukla, S. Yeleswarapu, M. A. Heinrich, J. Prakash and F. Pati, Mimicking tumor microenvironment by 3D bioprinting: 3D cancer modeling, Biofabrication, 2022, 14(3), 032002.
30. C. Vitale, M. Marzagalli, S. Scaglione, A. Dondero, C. Bottino and R. Castriconi, Tumor Microenvironment and Hydrogel-Based 3D Cancer Models for In Vitro Testing Immunotherapies, Cancers, 2022, 14(4), 1013.
31. S. Sabetkish, A. M. Kajbafzadeh, N. Sabetkish, R. Khorramirouz, A. Akbarzadeh, S. L. Seyedian, P. Pasalar, S. Orangian, R. S. Beigi, Z. Aryan, H. Akbari and S. M. Tavangar, Whole-organ tissue engineering: decellularization and recellularization of three-dimensional matrix liver scaffolds, J. Biomed. Mater. Res., Part A, 2015, 103(4), 1498–1508.
32. S. Qiao, T. Peijie and J. Nan, Crosslinking strategies of decellularized extracellular matrix in tissue regeneration, J. Biomed. Mater. Res., Part A, 2024, 112(5), 640–671.
33. H. Chen, Y. Cheng, X. Wang, J. Wang, X. Shi, X. Li, W. Tan and Z. Tan, 3D printed in vitro tumor tissue model of colorectal cancer, Theranostics, 2020, 10(26), 12127–12143.
34. O. Habanjar, M. Diab-Assaf, F. Caldefie-Chezet and L. Delort, 3D Cell Culture Systems: Tumor Application, Advantages, and Disadvantages, Int. J. Mol. Sci., 2021, 22(22), 12200.
35. J. Y. Won, M. H. Lee, M. J. Kim, K. H. Min, G. Ahn, J. S. Han, S. Jin, W. S. Yun and J. H. Shim, A potential dermal substitute using decellularized dermis extracellular matrix derived bio-ink, Artif. Cells, Nanomed., Biotechnol., 2019, 47(1), 644–649.
36. I. N. Amirrah, Y. Lokanathan, I. Zulkiflee, M. F. M. R. Wee, A. Motta and M. B. Fauzi, A Comprehensive Review on Collagen Type I Development of Biomaterials for Tissue Engineering: From Biosynthesis to Bioscaffold, Biomedicines, 2022, 10(9), 2307.
37. N. Théret, O. Musso, A. L'Helgoualc'h, J. P. Campion and B. Clément, Differential expression and origin of membrane-type 1 and 2 matrix metalloproteinases (MT-MMPs) in association with MMP2 activation in injured human livers, Am. J. Pathol., 1998, 153(3), 945–954.
38. J. Lv, X. Du, M. Wang, J. Su, Y. Wei and C. Xu, Construction of tumor organoids and their application to cancer research and therapy, Theranostics, 2024, 14(3), 1101–1125.
39. M. M. Haykal, C. Nahmias, C. Varon and O. C. B. Martin, Organotypic Modeling of the Tumor Landscape, Front. Cell Dev. Biol, 2020, 8, 606039.
40. B. Malacrida, S. Nichols, E. Maniati, R. Jones, R. Delanie-Smith, R. Roozitalab, E. J. Tyler, M. Thomas, G. Boot, J. Mackerodt, M. Lockley, M. M. Knight, F. R. Balkwill and O. M. T. Pearce, A human multi-cellular model shows how platelets drive production of diseased extracellular matrix and tissue invasion, iScience, 2021, 24(6), 102676.
41. A. V. Taubenberger, L. J. Bray, B. Haller, A. Shaposhnykov, M. Binner, U. Freudenberg, J. Guck and C. Werner, 3D extracellular matrix interactions modulate tumour cell growth, invasion and angiogenesis in engineered tumour microenvironments, Acta Biomater., 2016, 36, 73–85.
42. M. Sameni, A. Anbalagan, M. B. Olive, K. Moin, R. R. Mattingly and B. F. Sloane, MAME models for 4D live-cell imaging of tumor: microenvironment interactions that impact malignant progression, J. Vis. Exp., 2012, 3661.
43. F. Weeber, M. van de Wetering, M. Hoogstraat, K. K. Dijkstra, O. Krijgsman, T. Kuilman, C. G. M. Gadellaa-van Hooijdonk, D. L. van der Velden, D. S. Peeper, E. P. Cuppen, R. G. Vries, H. Clevers and E. E. Voest, Preserved genetic diversity in organoids cultured from biopsies of human colorectal cancer metastases, Proc. Natl. Acad. Sci. U. S. A., 2015, 112(43), 13308–13311.
44. M. van de Wetering, H. E. Francies, J. M. Francis, G. Bounova, F. Iorio, A. Pronk, W. van Houdt, J. van Gorp, A. Taylor-Weiner, L. Kester, A. McLaren-Douglas, J. Blokker, S. Jaksani, S. Bartfeld, R. Volckman, P. van Sluis, V. S. Li, S. Seepo, C. S. Pedamallu, K. Cibulskis, S. L. Carter, A. McKenna, M. S. Lawrence, L. Lichtenstein, C. Stewart, J. Koster, R. Versteeg, A. van Oudenaarden, J. Saez-Rodriguez, R. G. Vries, G. Getz, L. Wessels, M. R. Stratton, U. McDermott, M. Meyerson, M. J. Garnett and H. Clevers, Prospective derivation of a living organoid biobank of colorectal cancer patients, Cell, 2015, 161(4), 933–945.
45. H. Clevers, Modeling Development and Disease with Organoids, Cell, 2016, 165(7), 1586–1597.
46. E. Lavallee, J. P. Sfakianos and D. J. Mulholland, Tumor Heterogeneity and Consequences for Bladder Cancer Treatment, Cancers, 2021, 13(21), 5297.
47. J. P. Roose, M. Diehn, M. G. Tomlinson, J. Lin, A. A. Alizadeh, D. Botstein, P. O. Brown and A. Weiss, T cell receptor-independent basal signaling via Erk and Abl kinases suppresses RAG gene expression, PLos Biol., 2003, 1(2), E53.
48. J. Borst, T. Ahrends, N. Bąbała, C. J. M. Melief and W. Kastenmüller, CD4+ T cell help in cancer immunology and immunotherapy, Nat. Rev. Immunol., 2018, 18(10), 635–647.
49. Z. Wu, H. Zhang, J. Yan, Y. Wei and J. Su, Engineered biomembrane-derived nanoparticles for nanoscale theranostics, Theranostics, 2023, 13(1), 20–39.

---

Footnote

† These authors have contributed equally to the work.

---