Establishment of patient‐derived organotypic tumor spheroid models for tumor microenvironment modeling

Abstract Patient‐derived cancer models that reconstitute the characteristics of the tumor microenvironment may facilitate efforts in precision immune‐oncology and the discovery of effective anticancer therapies. Organoids that have recently emerged as robust preclinical models typically contain tumor epithelial cells and lack the native tumor immune microenvironment. A patient‐derived organotypic tumor spheroid (PDOTS) is a novel and innovative ex vivo system that retains key features of the native tumor immune microenvironment. Here, we established and characterized a series of colorectal cancer PDOTS models for use as a preclinical platform for testing effective immunotherapy and its combinations with other drugs. Partially dissociated (> 100 μm in diameter) tumor tissues were embedded in Matrigel‐containing organoid media and subsequently formed into organoid structures within 3 to 7 days of culture. The success rate of growing PDOTS from fresh tissues was ~86%. Morphological analysis showed that the PDOTSs varied in size and structure. Immunofluorescence and flow cytometry analysis revealed that the PDOTSs retained autologous tumor‐infiltrating lymphoid cells and tumor‐infiltrating lymphoid cells were continually decreased through serial passages. Notably, PDOTSs from tumors from a high‐level microsatellite instability‐harboring patient were sensitive to anti‐PD‐1 or anti‐PD‐L1 antibodies. Our results demonstrate that the PDOTS model in which the tumor immune microenvironment is preserved may represent an advantageous ex vivo system to develop effective immune therapeutics.

success rate of growing PDOTS from fresh tissues was ~86%. Morphological analysis showed that the PDOTSs varied in size and structure. Immunofluorescence and flow cytometry analysis revealed that the PDOTSs retained autologous tumor-infiltrating lymphoid cells and tumor-infiltrating lymphoid cells were continually decreased through serial passages. Notably, PDOTSs from tumors from a high-level microsatellite instability-harboring patient were sensitive to anti-PD-1 or anti-PD-L1 antibodies.
Our results demonstrate that the PDOTS model in which the tumor immune microenvironment is preserved may represent an advantageous ex vivo system to develop effective immune therapeutics.
Hye Kyung Hong and Nak Hyeon Yun are co-first authors.

| INTRODUCTION
Conventional human cancer cell lines have been widely used for preclinical studies as in vitro models to evaluate anticancer drugs. However, these cell line-based cancer models may not reflect the characteristics of the original tissue, and most drugs that are effective in cell line-based cancer models often fail in clinical trials. 1 Human cancer cell lines are frequently passaged for long periods of time, which may lead to genetic alterations in these cells. In addition, almost all cancer cell lines are cultured in monolayer conditions, which are not physically representative of cancer tissues. 2,3 Patient-derived cancer models, including primary patient-derived cancer cells, organoids, multicellular tumor spheroids, and patient-derived xenografts (PDXs), which recapitulate the molecular and histological features of each patient's tumor, have been developed for various types of cancer. [4][5][6][7] These patient-derived cancer models are considered the most appropriate for evaluating the efficacy of anticancer drugs and facilitating a better understanding of cancer biology. However, PDXs are expensive and labor-intensive to generate. Organoid or multicellular tumor spheroids cultured three-dimensionally are less time-consuming and cheaper than PDXs, but they do not contain the native tumor microenvironment (TME). [8][9][10][11] There is growing interest in the study of the TME with regard to where tumor cells interact with various stromal cells, including immune cells, fibroblasts, epithelial cells, and the extracellular matrix, which play crucial roles in tumor growth and metastasis. 12 The dynamic interactions between tumor and stromal cells in the TME impact therapy responses and are considered novel and effective therapeutic targets. [13][14][15] Given the emerging importance of the TME, more appropriate cancer models that include features of the TME must be developed. In particular, immune therapeutics that accelerate anticancer immune responses through immune checkpoint inhibition have necessitated the establishment of a patientderived model that recapitulates the TME. [16][17][18] Recently, patient-or murine-derived organotypic tumor spheroids (PDOTSs/MDOTSs) that retain the native TME, including tumor-infiltrating lymphoid and myeloid cells, have been reported as novel and innovative ex vivo systems. 19,20 Using PDOTSs/MDOTSs cultured in 3D microfluidic devices, the responses to an immune checkpoint blockade were evaluated. 19,20 In this study, we established and characterized a series of colorectal cancer PDOTS models that preserved the TME for use as a preclinical platform for testing effective immunotherapies and their combinations with other drugs.

| Establishment of PDOTSs
Tumor samples were collected immediately after surgery, washed with 70% ethanol and ice-cold PBS with 3% penicillin-streptomycin (Thermo Fisher Scientific Inc., Waltham, MA, USA), and minced in a 10 cm culture dish on ice using forceps and a scalpel. The minced tumor samples were resuspended in digestion buffer [DMEM (Gibco, Thermo Fisher Scientific, Inc.) containing 1% penicillinstreptomycin (Thermo Fisher Scientific, Inc.), 2.5% FBS (Biowest, Nuaille, France), 75 U/ml collagenase type IV (Gibco, Thermo Fisher Scientific, Inc.), and 125 μg/ml Dispase II (Life Technologies, Carlsbad, CA, USA)] and incubated for 20-30 min at 37°C on a tube rotator. Following digestion, samples were pelleted and resuspended in fresh DMEM, passed through a cell strainer with a pore size of 100 μm (Falcon; BD Biosciences, NJ, USA), and centrifuged at 1000 rpm for 3 min. Then, the pellet was resuspended in DMEM and centrifuged again at 1000 rpm for 3 min to remove the debris and collagenase.

| Hematoxylin and eosin (H&E) staining
PDOTS-Matrigel domes were fixed with 4% paraformaldehyde for 15 min at room temperature, and the fixed domes were carefully retrieved with a spatula, placed in a mold containing optimal cutting temperature compound, and stored at −80°C. For preparing the paraffin blocks, the fixed domes were broken up by pipetting, and the PDOTSs were collected and embedded in 1% agarose. The PDOTS-containing agarose samples were embedded in paraffin. Frozen PDOTS sections (thickness, 10 μm) were subjected to immunofluorescence staining and the paraffin-embedded sections (thickness, 6 μm) were subjected to routine H&E staining for the examination of cell morphology.

| Establishment of PDOTSs from colorectal cancer tissue
To establish PDOTSs, we cultured dissociated tumor tissues from 15 patients with colorectal cancer (Table 1). Following the digestion of fresh tumor samples with collagenase, organotypic tumor spheroids with autologous TMEs were isolated and cultured in Matrigel with human organoid media (Figure 1). Growth of PDOTSs was obtained with traditional cell culture media supplemented with FBS, 19,20 and we used organoid media containing growth factors to improve PDOTS growth for long-term culture. [8][9][10][11] The success rate of growing PDOTS from fresh tissues was ~86%. The organotypic tumor spheroids obtained from two patients showed poor cell growth, followed by cell death over time. The PDOTSs usually formed cystic structures within 3 to 7 days of culture. When the PDOTSs grew to a size of 400 μm or became densely packed, they were passaged by either trituration with a pipette or dissociation with trypsin to establish each patient-derived organoid ( Figure 1). The established organoids were expanded for several months and cryopreserved at a low passage number (2-5 passages).

| Characterization of the PDOTSs
Morphological analysis showed that the PDOTSs varied in size and structure between patient samples, ranging from thin-walled cystic structures to compact organoids devoid of a lumen (Figure 2A). The structure and morphology of the PDOTSs were preserved through several passages. Some PDOTS continued to grow beyond 7 days and formed large thin-walled cystic structures (> 400 μm in size) with lumens. The initial size of the PDOTSs was approximately 100 μm following 2 days of culture; they increased in size by twofold  Figure 2B). H&E staining confirmed that most PDOTS had a well-differentiated epithelial layer with a luminal area ( Figure 2C).

| The PDOTSs preserved the immune elements of the TME
The initial culture of the PDOTSs included a mixed population of epithelial tumor cells and various stromal cells. We investigated the presence of immune cells from the PDOTSs using confocal microscopy. The PDOTSs contained tumor-infiltrated CD3+ T cells that were embedded in the luminal area or in close proximity to tumor epithelial cells (Figures 3). These CD3+ T cells existed in the PDOTSs but their numbers decreased after serial passages ( Figure 4A). The PDOTSs also contained various levels of CD45+, CD8+, CD68+ macrophages, and CD31+ endothelial cells ( Figure 4B). To compare the immune cell populations between tumor tissues and the PDOTSs, T cells, CD4+, CD8+, myeloid cells, especially those of the monocyte lineage (CD14+), and NK cells (CD56+) were analyzed by flow cytometry ( Figure 5A). The PDOTSs showed a smaller myeloid cell population, including monocyte lineage CD14+ cells, than the tumor tissues due to the death of myeloid cells in the PDOTSs. The relative T-cell ratio was higher in the PDOTSs than in the tumor tissues, and this was caused by the decreased numbers of live myeloid cells in the PDOTSs. In the T-cell subpopulation, a higher number of CD4+ T cells than CD8+ cells were maintained in both the PDOTSs and tumor tissues. NK cells existed in tumor and OTS but they were relatively minor immune cell population than T cells ( Figure 5B).

| Sensitivity of the PDOTSs toward immune checkpoint blockades
To demonstrate the response of PDOTSs to immune checkpoint blockades, PDOTSs from a tumor from a high-level microsatellite instability (MSI-H)-harboring patient (CRC13) were treated with anti-PD-1 or anti-PD-L1 antibodies for 7 days. Acridine orange (live cells) and propidium iodide (dead cells) labeling were used for live/dead cell staining of the PDOTSs after drug treatment. The PDOTSs with MSI-H were sensitive to anti-PD-1 and anti-PD-L1 antibodies ( Figure 6). Flow cytometric analysis revealed larger CD4+ and CD8+ T-cell populations in the PDOTSs (CRC13); these cells are necessary effectors following treatment with anti-PD-L1 antibodies ( Figure 5A).

| DISCUSSION
The development of a human cancer model for predicting the anticancer drug sensitivity of individual patients is needed F I G U R E 3 Immunofluorescent Characterization of PDOTSs. PDOTS contained CD3+ T cells integrally embedded in close proximity to tumor cells. Tumor cells were readily detected using EpCAM antibody (green). CD3+ T cells were readily detected using CD3 antibody (red). DAPI was used for nuclear staining (Blue). First panel figures (CRC5) were adapted from reference 6. Scale bar, 100 μm to improve the clinical outcomes. Patient-derived cancer models, which include organoids, multicellular tumor spheroids, and PDXs, are considered to reflect the original tumors more closely than conventional cancer models and are used to screen for novel anticancer drugs. 5,[21][22][23][24] However, the establishment of PDXs is time-consuming, expensive, and associated with a low-throughput. Organoids or other tumor spheroids do not contain the TME, which crucially affects drug responses. [8][9][10][11] Immunotherapy is a major paradigm shift in current cancer drug development and cancer biology. In particular, immune checkpoint blockades targeting PD-1, PD-L1, and CTLA4 have demonstrated promising clinical activity against several tumor types. 25 Although current immunotherapies show impressive clinical responses in melanomas and other cancers, the majority of patients with cancer do not respond to these therapeutics. 26 Currently, there is no appropriate model available that represents the interactions of tumor and immune cells in the TME, and it is currently unclear which patients will respond to immunotherapies. 27 Recently, Jenkins et al 19 demonstrated that PDOTSs and MDOTSs contain relatively similar subsets of cells from lymphoid and myeloid populations that grow readily in collagen hydrogels in a 3D microfluidic culture system. 19,20 Using these models, the authors evaluated the response to the PD-1/PD-L1 blockade and recapitulated the features of ex vivo sensitivity and resistance to the PD-1 blockade.
In this study, we established a series of PDOTSs from partially dissociated (> 100 μm in diameter) tumors of patients with colorectal cancer. Characterization of the PDOTSs showed that autologous tumor-infiltrating T lymphocytes (CD3+, CD4+, and CD8+) were retained during short-term in vitro culture. PDOTSs do not require days or weeks of culture. Despite concerns that organoid media would be unsuitable for immune cell viability, our study shows that relevant tumor-infiltrating CD3+, CD4+, and CD8+ cells can be successfully retained in organoid media. Organoids have been rapidly adapted to patient-derived cancer modeling. 28 Early patient-derived organoid models of colorectal cancer comprised only tumor epithelial cells without stromal components. 5,11,29,30 Recently, various organoid culture methods in which cancer cells are grown with TME-associated immune components have been reported. 28 Epithelial tumoronly organoids can be reconstituted with exogenously added immune components from autologous peripheral blood or tumor tissues. 31 In addition, the air-liquid interface method can recapitulate the native TME by preserving diverse endogenous immune cells without reconstitution. 32 These models have been applied to make them more suitable for evaluating the response to immunotherapies and for assessing novel therapeutics.
Our results demonstrate the potential of PDOTSs as a patient-derived cancer model for the predictive assessment of individualized patient responses to current immunotherapies. Additional studies are required to extend the long-term preservation of other cell types, such as fibroblasts, endothelial cells, and other immune cells that are located in the TME. In addition, future studies will be conducted to establish definitive correlations between PDOTSs and immunotherapy responses by analogy to 3D microfluidic approaches. Threedimensional microfluidic culture of PDOTSs may provide a more realistic environment in which tumor-TME interactions could be evaluated; they could be used as an ex vivo system to drive translational efforts to develop personalized immunotherapies.

ETHICS APPROVAL STATEMENT
Patient samples were obtained from the Department of Surgery at the Samsung Medical Center in accordance with protocols approved by the Institutional Review Board (IRB approval no. SMC2019-11-129-001). Surgical tumor specimens and clinical records were obtained from patients with informed written consent.