The submerged Matrigel technique is widely used to culture dissociated cancer cells from tumor biopsies underneath tissue culture medium in a mixture with a dome or flat gel of 3D Matrigel. In this culture method, various growth factors and/or pathway inhibitors are added to the culture medium depending on the type of tumor tissue (25, 50–52). Customized culture conditions have been adapted for many different tissues (25, 32, 53–64), but often include ligands and additives, such as Wnt3a and/or R-spondin, bone morphogenetic (BMP) inhibitor Noggin, and epidermal growth factor (EGF) (25), which allow the stem cells to undergo long-term self-renewal and differentiation into various cell lineages (25). These niche factor requirements are mainly determined by genetic variations that increase tumorigenicity (51). These supplementations are also been utilized in other organoid culture strategies such as ALI (42). Dissociation of tissue during organoid preparation leads to activation of Rho kinase (ROCK)-dependent programmed cell death, therefore the addition of ROCK inhibitors to the medium can efficiently increase the success rates of organoid generation (25, 65, 66). It should be noted that conventional submerged Matrigel methods exclusively enrich epithelial tumor cells, but fail to preserve stromal components (25). Hence, TIME modeling in these techniques needs a co-culture of PDOs with exogenous (immune) cells. Seino et al. (50), by using this technique showed that co-culturing of human pancreatic ductal adenocarcinoma (PDAC) organoids with cancer-associated fibroblasts (CAFs), which shows immunosuppressive activity in TME (67), trigger organoid growth of WNT-nonproducing PDAC subtypes by CAF produced WNT (50). On the other hand, TGFβ and IL-1 ligands secreted by PDAC organoids can increase CAF heterogeneity and induce distinct myofibroblast and inflammatory CAF subtypes, respectively (68). Therefore, selective targeting of tumor-promoting CAFs can be improved by understanding the CAF heterogeneity mechanisms (68). Chakrabarti et al, designed a complex submerged Matrigel culture system by co-culturing of mouse tumor organoids with cytotoxic T lymphocytes and bone marrow-derived dendritic cells pulsed by conditioned media (tumor antigen) collected from tumor organoids. The promotion of apoptosis by activated CTLs was observed in cancer cell in the presence of PD-L1 neutralizing antibody (69). This method, co-culture of immune cells and human tumor organoids, is also used for the generation of tumor-reactive T cells from autologous peripheral blood lymphocytes (33) and survey on the Helicobacter pylori infection process (34, 70). Forsythe et al. (71), utilized a collagen-based extracellular matrix (ECM, hydrogel) to fabricate PDOs. They generate immunocompetent organoids with coculturing of tumor cells with patient-matched immune cells derived from peripheral blood mononuclear cells (PBMC), spleen, and lymph nodes. They showed that immunocompetent organoids can be useful in the preclinical study of personalized immunotherapy efficacy.

In microfluidic 3D devices, murine- or patient-derived organotypic tumor spheroids (MDOTS/PDOTS) are cultured in a mixture of collagen gel (72). For MDOTS/PDOTS culture (44), tumor tissues specimen is obtained from the patient and dissociated mechanically and enzymatically. This procedure ultimately gives a heterogeneous mixture of single cells, spheroids, and macroscopic tumor fragments. Then, this mixture is filtered through 100 μm and 40 μm filters to obtain spheroids with 40–100 μm in diameter, afterward, this fraction is pelleted in ultra-low-attachment plates and mixed with collagen gel, and inoculated into the center region of the microfluidic device. To feed the spheroids, the culture medium is added into the media channels located on both sides of the central channel. In this approach, spheroids maintain the native cancer tissue complexity and cellular diversity such as autologous myeloid populations (tumor-associated macrophages [TAM], monocyte, DC, and MDSC), lymphocytes (B cell and T cell), and cancer cells without reconstitution (45). T cell infiltration into cancer spheroids and tumor-immune cell interactions and cross-talk can be studied in these devices by adding exogenous T cells such as Jurkat cells into the media channels (46, 73). The composition of the devices and the size of media and central channels are variables that can reduce the validity and reproducibility of studies (44). The composition of the devices such as PDMS (polydimethylsiloxane) is one of the interfering factors in testing immune checkpoint blockade (ICB) in a mixture with small molecules (generally prepared in dimethyl sulfoxide, DMSO), because PDMS can adsorb the small hydrophobic molecules (74) that prevent the drug delivery to tumor spheroids.

In this approach, two inner and outer dishes are used. The inner dish consists of two layers, bottom and top, for preparation of the bottom layer the collagen gel matrix was added to the inner dish (43). For the preparation of primary tissues, after obtaining the tissue specimen, it immediately are placed in an ice-cold medium (43). After rinsing the tissue, it mechanically dissociated into small fragments and then are cultured in a mixture with collagen gel as a top layer of the inner transwell dish by pouring the mixture onto the inner dish with the bottom layer gel matrix (43). Then prepared inner dish is placed into an outer dish, transferred to a 37°C incubator, and allowed the gel of the inner dish to solidify (43). Afterward, culture media is added to the outer dish that can diffuse into the inner dish through a permeable membrane. In addition, the top layer of culture is in direct air exposure that supply tissue organoid oxygen efficiently (26, 43, 75). ALI allows the growth of large multicellular tissue fragments that retain native tissue architecture, such as cancer cells en bloc with endogenous stromal and immune cells without reconstitution, which is different from submerged Matrigel culture (76). In addition, ALI PDOs preserve not only the architecture and complex cellular composition of the TME, but also the genetic alterations of the native tumor (42). These features of ALI, in contrast to submerged Matrigel Culture, make it suitable for TME modeling (32). It has been observed that in ALI, probably essential endogenous niche factors produced by stromal cells are sufficient to support tissue fragment growth without growth factor supplementation (26).

Organoid culture without physical restrictions resulted in organoid fragments with differences in shape, size, geometry, and cell number. As well, the mechanical factors such as fluid shear stress, tension, and compression are not recapitulated the native condition of tumor in 2D or 3D culture models. Therefore, these limitations lead to the non-reproducibility of results in organoid cultures (3). Organoid-on-a-chip models can overcome this challenge by increasing of uniformity of organoids and mimicking the physical conditions of the body such as providing perfusion of culture media. In addition, microfluidic devices can monitor and control the culture condition and assay variables by integrated sensors and actuators (77). The multi-organoid-on-a-chip have also been built (77, 78), nevertheless, optimization of culture media and mechanical condition for different tissue remain as a major challenge.

In some cases, immune cells are not added, but the immune cells within the organoid are maintained and expanded. Zumwalde et al. (79), showed the presence of immune cell populations and characterized the intraepithelial lymphocyte compartment within the organoid culture. They demonstrated that leukocyte populations of the breast organoid differed from those in peripheral blood, and conserved T lymphocytes can be expanded in response to the bisphosphonate. As noted above, cultured tumor spheroids in microfluidics devices are able to preserve autologous myeloid and lymphoid cell populations (45). The air-liquid interphase (ALI) culture strategy is another approach that retains native immune and stromal components. In ALI, the cancer epithelium and its complex microenvironment including fibroblasts, immune cells such as cytotoxic T lymphocytes (CTLs), helper T cells, TAMs, natural killer (NK) cells, B cells, and NK T cells are maintained for over a 1- to 2-month period (42). In addition, the T cell receptor (TCR) spectrum of the native tumor is accurately preserved in organoid culture (42).

In some organoid culture strategies, such as the submerged Matrigel technique, stromal and immune cells are not preserved. Therefore, to perform immunotherapy studies, it is necessary to add immune cells to the organoid. Tsai et al. (80), generated pancreatic cancer organoids using primary organoid co-culture with patient-matched peripheral blood lymphocytes and CAFs that were relevant to the immunotherapy and tumor-immune cell interaction studies. Activation of myofibroblast-like CAFs and tumor-dependent lymphocyte infiltration were detected in these complex organotypic models. In another study (33), co-cultures of autologous tumor organoids and peripheral blood lymphocytes were established for the enrichment of tumor-reactive T cells from peripheral blood of colorectal cancer (CRC) and non-small-cell lung carcinoma (NSCLC) patients. Thus, this approach can be used to evaluate the efficiency of the killing of tumor organoids by autologous tumor-reactive T cells (33). Many factors can affect the outcome and our success in the generation of complex organoids. For instance, in some cancer immune cells located in the surrounding stroma of the tumor or completely lack in TME. Therefore, this will impact the immune cell population of primary organoid culture (3). Furthermore, the addition of exogenous immune components to organoids will not be able to recapitulate the complexity of the patient-specific immune cell composition within TME. If allogeneic immune cells are used as a source of exogenous immune components, because of interpersonal differences in HLA, it will lead to high background killing compared with patient-specific PBMCs (3). Ultimately, choosing the proper strategy for co-culture depend on the downstream assays.

Lack of perfusion flow and vascular networks in organoids remains a major challenge in the preparation of complex organoids as an ideal cancer model. The absence of vasculature limits the size of the organoid, and organoids larger than 100-200 μm in diameter suffer from the diffusion of nutrients, oxygen, and metabolites to the central region of the organoid. Furthermore, organoid fragments larger than 500 μm in diameter show necrosis (81, 82). Many studies have been performed to overcome this problem and establish perfusion flow in the organoid (23, 27–29). Wörsdörfer et al. (29), co-cultured the mesodermal progenitor cells (MPCs) with tissue-specific (progenitor) cell types for the establishment of complex neural as well as human tumor organoids. They showed the directed incorporation of MPCs into organoids and the formation of vascular networks in these organoids. They observed the expansion of vascular networks during organoid growth. In addition, this endothelial network was responsive to pro-angiogenic conditions and anti-angiogenic agents (29). Organoid vascularization has been previously well-reviewed by Grebenyuk and Ranga (28).

In ACT immunotherapy, TILs or circulating lymphocytes are collected, high-affinity TCRs recognizing tumor antigens are selected or genetically engineered, subsequently, these cells expanded and activated ex vivo followed by reinfusion into patients (8, 83). CAR T cells are genetically modified T cells that produce an artificial T-cell receptor, targeting a specific antigen on the surface of tumor cells and circumventing MHC restriction (84). Studies have shown that PDOs are efficient platforms to evaluate the tumor-specific cytotoxicity of T cells (e.g., CAR T cell, TCR T cell). In the study of Michie et al. (85), PDOs were exploited to evaluate the effect of combination therapy of CAR T cells and birinapant, the inhibitor of apoptosis antagonist. They observed that the combination of CAR T cells with birinapant significantly reduce PDOs growth in a tumor necrosis factor (TNF)-dependent manner, while CAR T cells alone were relatively ineffective. Schnalzger et al. (35), developed a sensitive preclinical model, 3D PDOs, that allows assaying the CAR-mediated cytotoxicity in the native TIME mimicking model. Furthermore, they established a confocal live-cell imaging protocol for dynamic monitoring of cytotoxic activity toward organoids at a single organoid level. They demonstrated a stable effector-target cell interaction in the co-culture of NK cells with CRC or normal organoids on an ECM layer. In addition, CRC organoids were utilized to monitor the tumor antigen-specific cytotoxicity of EGFRvIII or FRIZZLED receptors-targeting CAR-engineered NK-92 cells. In sum, they established a sensitive platform to assay CAR efficacy and tumor specificity in a personalized manner (35). In addition, Epithelial-only PDOs, while lacking stromal and immune components, can be used for the selection of tumor-reactive T cells (33). This co-culture strategy can be served to enrichment, stimulation, and efficacy evaluation of tumor-reactive lymphocytes. Dijkstra et al. (33), cocultured CRC or NSCLC organoids with autologous circulating T cells (PBMCs), in medium supplemented with anti-PD1, anti-CD28, and IL-2 for generation of tumor-reactive CD8⁺ populations. After two weeks of co-culture, T cell-mediated killing, MHC-dependent cytotoxicity, against autologous tumors organoids, and upregulation of CD107a and IFNγ secretion in CD8⁺ T cells were observed. But, these tumor-reactive CD8⁺ cells did not affect the survival of matched healthy organoids. Therefore, they established a platform for the expansion of tumor-reactive T cells and evaluation of the specificity and sensitivity of cancer cell killing by autologous T cells in a personalized manner. Additionally, tumor-reactive T cells can be isolated from TILs that are more specific than the non-infiltrating lymphocytes, subsequently, these cells can be infused back into the patient (86).

Clinical benefits of ICIs that target CTLA-4, PD1/PD-L1 have been observed in various advanced tumors, such as renal cell carcinoma (RCC) (87), cutaneous squamous cell cancer (88), melanoma (11), head and neck cancer (89), and NSCLC (10). Epithelial-only organoid biobanks have been established from diverse malignancies and are widely available through entities such as the Human Cancer Models Initiative (HCMI). But lack of immune compartments in these types of PDOs hinders their immunotherapy application. Some studies have been performed to overcome this issue. For instance, Kong and colleagues (36), cocultured epithelial-only submerged Matrigel organoids with autologous TILs and demonstrated the TILs migration toward organoids and cytotoxic activity of T cells. They also showed the rescue of TIL function after ICB. But since, co-culturing the epithelial only PDOs with exogenous immune cells, as a reconstitution approach, cannot fully recapitulate the complex interaction and crosstalk between diverse cell populations in the TME, especially when using immunomodulatory drugs. To tackle this problem, a holistic culture approach, such as 3D microfluidic and ALI culture strategies, can be utilized for TIME modeling. The dynamic response and resistance to ICB (such as PD-1 blockade) can be recapitulated using organotypic spheroids in a short-term 3D microfluidic culture that maintains autologous myeloid and lymphoid cell populations similar to the original donor human or mouse tumors (45). As well, the cytokine secretion profiles of PDOTS/MDOTS matched the profile of the donor tumors (45). Therefore, PDOTS/MDOTS profiling facilitates the evaluation of ICB using clinically relevant models. Neal et al. (42), utilized WENR (WNT3A, EGF, NOGGIN, and RSPO1) base medium to expand and serially passage physically processed cancer fragments as ALI organoids. They demonstrated that ALI PDOs, like the microfluidic approach, preserve the stromal and immune cell populations, and effectively recapitulate the expansion, activation, and tumor cytotoxicity of tumor antigen-specific TILs in response to PD-1/PD-L1 ICB. They observed the CD8⁺ TIL expansion, activation, and tumor cell killing after 1 week of anti-PD-1 treatment of ALI PDOs from various human tumor biopsies, including RCC, NSCLC, and melanoma. It should also be noted that the material and composition of devices used in organoid culture can affect the results of immunotherapy such as ICB studies (74).

PDOs can be used as predictive tools to study the specific infectivity, and cytotoxicity of the oncolytic virus alone or in combination with chemotherapy (90). In the study of Raimondi et al, Oncolytic adenovirus (OA) displayed a good selectivity of replication in PDAC organoids, but not in healthy pancreas tissue organoids. Patient-specific responses were also observed, indicating that PDOs are reliable in vitro tumor models to assay preclinical responses to oncolytic viruses.

Tumor organoids can also be utilized to study the specificity and efficacy of antibody-based immunotherapy. As mentioned in the previous section, many studies used the organoid models in antibody-based ICB therapies (42, 45, 69, 91). Courau et al. (92), showed that both NKG2D-MICA/B and NKG2A-HLA-E pathways are involved in the infiltration process of activated/memory T and NK cells into organoids, subsequently these activated cells can kill cancer cells and disrupt the 3D structure. They demonstrated that anti-MICA/B and a combination of both anti-MICA/B and anti-NKG2A antibodies were able to induce immune-mediated destruction of colorectal tumor organoids during cocultures with autologous TILs. Gonzalez-Exposito et al. (93), established seven PDOs from treatment-refractory metastatic CRC and one from a treatment naïve primary CRC, to investigate on resistance and sensitivity mechanisms of cibisatamab, a bispecific monoclonal antibody that binds carcino-embryonic antigen (CEA) on tumor cells and CD3 on T cells. They designed a co-culture of organoids and allogeneic CD8⁺ T cells to evaluate cibisatamab efficacy. Using this platform, they demonstrated that CEA_low PDOs were resistant whereas CEA_high PDOs were sensitive to cibisatamab. Indeed, CEA_low cells maintain tumor cell growth, and an increased WNT/β-catenin pathway activity was detected in CEA_low cells by RNA-sequencing. They suggested the use of WNT/β-catenin pathway inhibitors in combination with cibisatamab as a potential strategy to increase the treatment success.