Prostate Stroma Increases the Viability and Maintains the Branching Phenotype of Human Prostate Organoids

Abstract

The fibromuscular stroma of the prostate regulates normal epithelial differentiation and contributes to carcinogenesis in vivo. We developed and characterized a human 3D prostate organoid co-culture model that incorporates prostate stroma. Primary prostate stromal cells increased organoid formation and directed organoid morphology into a branched acini structure similar to what is observed in vivo. Organoid branching occurred distal to physical contact with stromal cells, demonstrating non-random branching. Stroma-induced phenotypes were similar in all patients examined, yet they maintained inter-patient heterogeneity in the degree of response. Stromal cells expressed growth factors involved in epithelial differentiation, which was not observed in non-prostatic fibroblasts. Organoids derived from areas of prostate cancer maintained differential expression of alpha-methylacyl-CoA racemase and showed increased viability and passaging when co-cultured with stroma. The addition of stroma to epithelial cells in vitro improves the ability of organoids to recapitulate features of the tissue and enhances the viability of organoids.

Keywords: Bioengineering; Biological Sciences; Cell Biology; Tissue Engineering.

Graphical abstract

Figure 1

3D Co-culture with Prostate Stroma Supports Organoid Growth and Increases Organoid Branching (A) Gene expression of PrS mix 1 from microarray of stroma-specific genes TIMP3, von Willebrand factor (VWF), α-SMA (ACT2 shown as SMA), and EpCAM. Quadruplicate measures reported; error bars represent one standard deviation of the mean. Stromal cells are a heterogeneous mixture of fibroblasts and myofibroblasts by immunofluorescent staining of α-SMA and vimentin in PrS mix 3. Scale bar, 100 μm. (B) Schematic of mono-culture and co-culture workflow and conditions. PrE and PrS cells seeded into 33% Matrigel on top of 50% Matrigel base layer. (C) Fluorescent images of PrS mix 1 incorporation of an EdU analog cultured in 3D conditions. Scale bar, 200 μm. (D) Bright-field images of organoids in mono- and co-culture derived from dissociated epithelial cells isolated from benign prostate tissue specimens of two patients (PrE-1, top; PrE-2, bottom). Hematoxylin and eosin staining showing histopathology of patient tissue. Scale bar, 200 μm. (E) Percentage of branched organoids in mono- and co-cultures on days 7 and 14 according to a branching classification scheme used to categorize each organoid by extent of branching (right). Quantification of organoid branching in mono- and co-cultures according to the branching classification scheme in organoids derived from two patients (PrE-1, top; PrE-2, bottom). Triplicate wells per condition, n reported as sum of branched organoids from triplicate wells per condition. (F) Bright-field images of time course showing PrE-1 organoid branching from days 6–10 in co-culture with PrS mix 4. Scale bar, 200 μm.

Figure 2

Organoids in Mono-culture and Co-culture Form Mixed Epithelial Differentiation and Contain Lumens (A) H&E and immunofluorescent images of differentiation markers p63 and Ck8 of PrE-1 organoids in mono-culture and co-culture with PrS mix 1. Inset shows Ck8+ cells co-expressing p63 (green arrow) or p63− (white arrow). Scale bars, 50 μm. (B) Immunohistochemical stain for AR in PrE-1 organoids in mono-culture and co-culture with PrS mix 1. Insets show high magnification of adjacent AR+ (green arrow) and AR− (black arrow) cells. Scale bars, 50 μm. (C) Whole-mount immunofluorescent staining of Ck5 (green), Ck8 (blue), and phalloidin (red) in day 14 PrE-2 organoids co-cultured with PrS mix 1. Scale bars, 500 μm (merge) and 100 μm (zoom). (D) Whole-mount immunofluorescent staining of organoid with PrS-GFP cells (green), Ck8 (blue), and phalloidin (red) in day 12 PrE-1 organoids cultured with GFP-labeled PrS mix 3. Scale bars, 500 μm (merge) and 100 μm (zoom).

Figure 3

Differential Effects of the Epithelial-to-Stromal Cell Ratio and Epithelial Seeding Density on Organoid Formation Efficiency, Size, and Branching Morphologies (A) Pipeline for whole-well imaging and analysis. Image acquisition at 25 z-planes per well, generation of a single enhanced depth-of-field image, conversion to grayscale, and identification of each organoid. The area, circularity, and maximum/minimum radius ratio were calculated for each organoid in every well. Calculation details can be found in Methods. (B) Organoid formation efficiency percentage (number of organoids per number of PrE seeded) for mono and co-cultures. Triplicate wells per condition; errors bars represent one standard deviation of the mean. (C) Area quantification of organoids in mono and co-cultures. Triplicate wells per conditions; error bars show mean with 95% confidence interval. (D) Representative bright-field images depicting branching morphologies of organoids in co-culture derived from two patients, PrE-1 and PrE-2, at various PrE seeding densities. Scale bars, 500 μm. (E) Organoid formation efficiency (number organoids per number of PrE cells seeded) in mono-culture (white bar) and co-culture (black bars) seeded at 100 PrE cells/well with increasing numbers of PrS cells. Triplicate wells per condition; error bars represent one standard deviation of the mean. (F) Area quantification of organoids in mono-culture (gray circles) and co-culture (white squares) with increasing numbers of PrS cells. Each dot represents a single organoid; data from triplicate wells combined. Error bars show mean with 95% confidence interval. (G) Quantification of organoid area (left), circularity (middle), and maximum/minimum radius ratio (right) at days 7 and 14. Enhanced depth-of-field images (bottom) of a single organoid in mono-culture and co-culture and their corresponding area and circularity metrics on the graphs above (Mono, blue circle; co, red square). Each data point represents a single organoid from triplicate wells. Error bars show mean with 95% confidence interval for area and circularity and standard error of the mean for maximum/minimum radius ratio. Scale bar, 100 μm. All data from PrE-1 and PrS mix 2 except where noted. p value <0.1 was considered statistically significant. *p < 0.1, **p < 0.05, ***p < 0.01, ****p < 0.001; ns, not significant.

Figure 4

Stromal Cells Influence Organoid Branching Morphogenesis via Cell-Cell Contact and Express Unique Factors Compared with Non-prostate Fibroblasts (A) Bright-field image on sequential days of stromal cell (boxed and zoom) taxis toward an organoid followed by distal bud formation (black arrow). Large scale bar, 200 μm; small scale bar, 50 μm. PrE-1 in co-culture with PrS-2. (B) PrE-1 organoids grown in mono-culture (mono), Transwell co-culture (Trans), or direct co-culture (Co) with PrS mix 4 stromal cells. Circularity and maximum/minimum radius ratio measurements quantified from triplicate wells per condition; error bars show mean with 95% confidence interval for circularity and standard error of the mean for maximum/minimum radius ratio. (C) Bright-field and fluorescent images of GFP-labeled PrS mix 3 cells in co-culture with PrE-1 organoids at days 3 and 10. Fluorescent images of GFP-labeled PrS mix 3 cells in co-culture with PrE-1 organoids at day 12 with phalloidin (red) in bottom panel. Scale bars, 200 μm. (D) Fluorescent image of GFP-labeled PrS mix 3 cells in co-culture with PrE-1 epithelial organoid incorporation of an EdU analog (pink). Scale bar, 200 μm. (E) Bright-field and fluorescent images of GFP-labeled NIH3T3 cells co-cultured with PrE-1 organoids. Circularity and maximum/minimum radius ratio measurements of organoids in mono-culture and co-culture with NIH3T3 cells. Triplicate wells per condition; error bars show mean with 95% confidence interval for circularity and standard error of the mean for maximum/minimum radius ratio. Scale bar, 200 μm. (F) Comparative gene expression profiling of the human growth factors and TGF-β pathway by qPCR arrays in PrS mix 1 and NIH3T3 cells grown in 3D conditions. Array normalization details in Methods section. (G) Circularity and maximum/minimum radius ratio measurements of organoids in mono-culture and co-culture treated with Noggin (NOG) or FGFR inhibitor (BGJ398). PrS mix 4 and PrE-1 organoids used in triplicate wells per condition; error bars show mean with 95% confidence interval for circularity and standard error of the mean for maximum/minimum radius ratio. p value < 0.1 was considered statistically significant. *p < 0.1, **p < 0.05, ***p < 0.01, ****p < 0.001; ns, not significant.

Figure 5

Co-culture Organoids Grown from Benign and Cancer Regions of the Same Patient Maintain Histologic Phenotype and Improve Passaging (A) Isolation and co-culture of matched primary prostate organoids derived from benign and cancer regions of the same patient, with PrS cells. (B) Matched benign (PrE-5) and cancer (PrECa-5) tissue from the patient from which cells were isolated for organoid growth. Histology shown by hematoxylin and eosin staining (top panel), and presence of AMACR shown by immunohistochemistry (bottom panel). 10X image shown with 40X inset for AMACR. (C) Matched organoids derived from benign regions (PrE-5) and cancer regions (PrECa-5) from patient tissue shown in (B) grown in mono-culture and co-culture in 3D with PrS mix 5. Histology shown by hematoxylin and eosin staining (top panel), and presence of AMACR shown by immunohistochemistry (bottom panel). 40X images shown with 100-μM scale bar. (D) Whole-well images of passage 0 and passage 2 mono- and co-culture PrECa-7 organoids derived from cancer regions. High magnification of selected organoid shown in inset. Inset shown with 100-μm scale bar. (E) Expression of AMACR by qRT-PCR in total RNA isolated from matched organoids derived from benign and cancer tissue shown in (C) and passage 2 organoids shown in (D). Note that the benign organoids did not survive passage. p value < 0.1 was considered statistically significant. *p < 0.1, **p < 0.05, ***p < 0.01, ****p < 0.001; ns, not significant.