Dendritic cell phenotype and function in a 3D co-culture model of patient-derived metastatic colorectal cancer organoids

Abstract

Colorectal cancer (CRC) remains one of the most aggressive and lethal cancers, with metastasis accounting for most deaths. As such, there is an unmet need for improved therapies for metastatic CRC (mCRC). Currently, the research focus is shifting towards the reciprocal interactions within the tumor microenvironment (TME), which prevent tumor clearance by the immune system. Dendritic cells (DCs) play a key role in the initiation and amplification of anti-tumor immune responses and in driving the clinical success of immunotherapies. Dissecting the interactions between DCs and CRC cells may open doors to identifying key mediators in tumor progression, and possible therapeutic targets. This requires representative, robust and versatile models and tools. Currently, there is a shortage of such in vitro systems to model the CRC TME and its tumor-immune cell interactions. Here we develop and establish a dynamic organotypic 3D co-culture system to recapitulate and untangle the interactions between DCs and patient-derived mCRC tumor organoids. To our knowledge, this is the first study investigating human DCs in co-culture with tumor organoids in a 3D, organotypic setting. This system reveals how mCRC organoids modulate and shape monocyte-derived DCs (MoDCs) behavior, phenotype, and function, within a collagen matrix, using techniques such as brightfield and fluorescence microscopy, flow cytometry, and fluorescence-activated cell sorting. Our 3D co-culture model shows high viability and extensive interaction between DCs and tumor organoids, and its structure resembles patient tissue sections. Furthermore, it is possible to retrieve DCs from the co-cultures and characterize their phenotypic and functional profile. In our study, the expression of activation markers in both mature and immature DCs and their ability to activate T cells were impacted by co-culture with tumor organoids. In the future, this direct co-culture platform can be adapted and exploited to study the CRC-DC interplay in more detail, enabling novel and broader insights into CRC-driven DC (dys)function.

Keywords: 3D co-culture; dendritic cell dysfunction; human dendritic cells; immunosuppression; metastatic colorectal cancer; patient-derived tumor organoids; tumor microenvironment.

Figure 1

3D co-culture of PDTOs and DCs: Setup, morphology, and viability. (A) Schematic representation of the co-culture system and the three different conditions used in this study: DCs cultured alone, DCs and PDTOs co-cultured, and PDTOs cultured alone. (B) Closeup and morphology of iDCs and mDCs. The PDTOs are derived from CRC liver metastasis of two different patients, PDTO cystic as the name suggests presents a cystic morphology, whereas PDTO dense has a compact morphology. (C) Quantification of mDCs and iDCs viability when cultured alone in 3D in the collagen matrix after 48h, based on NucBlue™ Live and NucGreen™ Dead stainings, in two different experiments/donors. (D) The viability of mDCs, iDCs, and PDTOs was evaluated, after 48h of co-culture, during live imaging with NucBlue™ Live reagent (staining the nuclei of all cells) and NucGreen™ Dead reagent (staining only dead cells). The large majority of DCs seem to be viable alone, and in co-culture. iDCs - immature MoDCs, mDCs - mature MoDCs, PDTOs - patient-derived tumor organoids.

Figure 2

Visualization of DCs - PDTOs interactions within the co-culture. (A) Time series frames: iDCs establish direct contacts with co-cultured PDTOs by migrating towards and agglomerating near the tumor organoid borders (examples pinpointed by the arrows). (B) 3D immunofluorescence stainings with DAPI, CD11c, and PanCK, to distinguish DCs and tumor cells. (C) Examples of iDCs in close proximity to and surrounding/engulfing tumor-derived fragments. iDCs - immature MoDCs, mDCs - mature MoDCs, PDTOs - patient-derived tumor organoids.

Figure 3

DCs distribution in relation to the tumor lesions within the co-culture in comparison to patient tumor samples by immunofluorescence. (A) Parallel between H&E and immunofluorescence stainings of DCs and organoids/tumors in fixed sections of the 3D co-culture and of CRC liver metastasis in patients. On the lF panel, representative examples of immunofluorescence stainings of CD11c and PanCK in the co-culture and of CD1c and PanCK in liver metastasis tumor sections. DCs are present agglomerating around, surrounding and infiltrating tumor organoids and tumor lesions in patients. Additional and larger images are included in Supplementary Figures 3 and 4. (B) Analysis of DCs distribution within the co-culture by image processing including segmentation and distance maps (normal and inverted). Each DC was assigned a positive, 0 or negative value depending on whether they were found outside, at the border, or inside the tumor, respectively. (C) The scatter dot plot shows differences in DC distribution around and inside the tumor organoids. Each dot represents one DC, line at the median. The p values were determined using an unpaired t-test. Statistical significance was annotated as follows: ****p < 0.0001 based on two sections from two independent experiments. (D) The scatter dot plot shows differences in DC distribution around and inside the tumor lesions, based on sections from 3 different patients. Each dot represents one DC, line at the median. The p values were determined using an unpaired t-test. Statistical significance was annotated as follows: ***p < 0.001, ****p < 0.0001 based on sections from 3 different patients. iDCs - immature MoDCs, mDCs - mature MoDCs, PDTOs - patient-derived tumor organoids.

Figure 4

Recovery of DCs after co-culture with tumor PDTOs - viability, gating strategy, and phenotypic characterization to assess tumor-induced phenotypical changes. (A) DCs viability was assessed by trypan blue staining after collagenase treatment to disassemble the collagen scaffold (before centrifugation). (B) For flow cytometry analysis, cells were gated based on size, single cells, and live cells. Depicted is the HLA-DR-based gating strategy to distinguish PDTOs and DCs, using three conditions: DCs only, DCs and PDTOs co-culture, and PDTOs only. (C) Representative histogram plots to exemplify basal expression of CD86, HLA-DR, and PD-L1 markers in iDCs and mDCs. (D) Representative histogram plot of CD86, HLA-DR and PDL-1 to highlight the phenotypic shift of mDCs cultured in the presence of PDTOs. (E) Scattered dot plots showing normalized MFI values to iDCs and mDCs, respectively. Each dot/triangle represents a different donor, 4 donors were used in total. Data plotted as normalized values of raw MFI, mean with SD. The statistical significance between different conditions (mDCs/iDCs with and without PDTOs) was analyzed by a mixed-effects model followed by a Dunnett’s post-hoc multiple comparisons test on the log2 transformed ratio values. The statistical significance was annotated as follows: *p < 0.05, **p < 0.01. (Raw data can be found in Supplementary Figure 5 ).

Figure 5

Sorting and functional characterization of DCs after co-culture with PDTOs – Allogeneic T cell assay. (A) Isolation of DCs, using EpCAM to sort out PDTOs. DCs gate defined based on EpCAM expression. (B) Representative CFSE histogram plots are shown. Numbers indicate the percentage of gated proliferating T cells. (C) Proliferation of allogeneic T cells after 6 days of co-culture with sorted DCs. Scattered dot plots show the percentage of proliferating T cells in each condition (average of technical replicates), normalized to proliferating T cells in the conditions with only iDCs and mDCs, respectively, mean with SD. The statistical significance between different conditions (mDCs/iDCs with and without PDTOs) was analyzed by a mixed-effects model followed by a Dunnett’s post-hoc multiple comparisons test on the log2 transformed ratio values. The statistical significance was annotated as follows: *p < 0.05. (Raw data can be found in Supplementary Figure 6 ).