Adipose stem cell niche reprograms the colorectal cancer stem cell metastatic machinery

Abstract

Obesity is a strong risk factor for cancer progression, posing obesity-related cancer as one of the leading causes of death. Nevertheless, the molecular mechanisms that endow cancer cells with metastatic properties in patients affected by obesity remain unexplored.Here, we show that IL-6 and HGF, secreted by tumor neighboring visceral adipose stromal cells (V-ASCs), expand the metastatic colorectal (CR) cancer cell compartment (CD44v6 + ), which in turn secretes neurotrophins such as NGF and NT-3, and recruits adipose stem cells within tumor mass. Visceral adipose-derived factors promote vasculogenesis and the onset of metastatic dissemination by activation of STAT3, which inhibits miR-200a and enhances ZEB2 expression, effectively reprogramming CRC cells into a highly metastatic phenotype. Notably, obesity-associated tumor microenvironment provokes a transition in the transcriptomic expression profile of cells derived from the epithelial consensus molecular subtype (CMS2) CRC patients towards a mesenchymal subtype (CMS4). STAT3 pathway inhibition reduces ZEB2 expression and abrogates the metastatic growth sustained by adipose-released proteins. Together, our data suggest that targeting adipose factors in colorectal cancer patients with obesity may represent a therapeutic strategy for preventing metastatic disease.

Fig. 1. Tumor-infiltrating VAT boosts the metastatic potential of CR-CSphCs.

a Forest plot of survival changes in high (>30, obesity) versus low (≥18.5 and <25, lean) BMI CRC patients. Data represent the risk ratio ± 95% CI. Statistical significance was calculated by a Random-effect meta-analysis model. b Kaplan–Meier of progression-free survival (PFS) curve in a cohort of 511 CRC patients, based on BMI status. Healthy weight indicates 18,5<BMI < 30, and obesity BMI > 30. Statistical significance was calculated using the log-rank (Mantel–Cox) test. c H&E analysis and CDX2 expression on primary and liver metastasis in CRC patients with healthy weight or affected by obesity. Black arrow heads indicate tumor-infiltrating adipose cells. Li: liver; T = tumor. d Immunohistochemical analysis of CD34 (brown color), CD31 (green color), and CD45 (red color) in tissues as in c. For c, d one representative of 9 independent experiments is shown. e Phase-contrast analysis of CMS2 cells (CSphC #9) treated with medium or V-ASC CM. For (c–e) scale bars, 100 µm. One representative of three independent experiments is shown. f ELDA software analysis of the clonogenic activity in CMS2 CR-CSphCs following treatment with medium or V-ASC CM. g Clonogenic assay of CMS2 CR-CSphC lines TOP–GFP^high and TOP–GFP^low (15% highest/lowest TOP–GFP levels) treated with medium or V-ASC CM. For (f–g) statistical significance was calculated using the two-tailed t test and data are mean ± standard error of three independent experiments performed with CR-CSphCs isolated from three different CRC patients (CSphC #8, #9). h Percentage of TOP–GFP positive cells, in CMS2 cells treated with medium or V-ASC CM (left panel). Box plots show min-to-max values, with line indicating the mean value. Flow cytometry analysis of TOP–GFP (black color indicates Wnt^- cells; green color scale indicates low, intermediate, and high Wnt⁺ cells) (right panel). Statistical significance was calculated using the paired two-tailed t test. Data are mean ± standard error of independent experiments performed with different CR-CSphCs (#1, #4, #5, #8, #9, #11, #21). i Number of mouse tumor xenografts generated by subrenal capsule injection of 10, 100, 1000, or 10,000 CR-CSphCs, alone or in combination with 50,000 V-ASCs (upper panel). Percentage of cancer-initiating cell (CIC) and its fold increase of cells (lower panels). Data are mean ± standard error (95% confidence interval) of 12 independent experiments performed with CR-CSphCs injected as described above. Statistical significance was calculated by ELDA software (http://bioinf.wehi.edu.au/software/elda/). j In vivo imaging and CK20 immunohistochemistry analysis of xenograft tumor formation obtained by subrenal capsule injection of 100 CR-CSphCs alone or together with V-ASCs at the indicted time points. Photon signal of all metastatic sites (kidney, liver, and lungs) at 12 weeks. A yellow dotted line indicates a tumor xenograt lesion. Tumor (T), kidney (K), liver (Li), and lung (Lu) are indicated. One representative of 12 independent experiments is shown. Scale bars, 100 µm.

Fig. 2. Adipose-derived factors expand CD44v6 + cell fraction that secretes NGF and potentiates the migration capacity of ASCs.

a Cytokines secreted by CR-CSphCs (n = 4: #1, #8, #9, #21), S-ASCs (n = 6: #3, #5, #6, #8, #14, #20), V-ASCs (n = 6: #3, #5, #6, #8, #14, #18), or primary adipose tissue (AT) (n = 4). Data are the mean of 3 independent experiments. b Cell growth of CR-CSphCs treated for 5 days with IL-6 and HGF alone or in combination. The dotted red line shows the cell number at day 0. c Colony size of CR-CSphCs treated as indicated. n represents the number of colonies. Statistical significance was calculated using the two-tailed t test. d Invasion assay of CR-CSphCs pretreated with the indicated cytokines for 48 h. For b–d data show mean ± S.D. of three independent experiments using fourdifferent CR-CSphCs (CSphC #1, #8, #9, #21). e mRNA expression levels of CSC-related genes in CMS2 CR-CSphCs (CSphC #8, 9) exposed to vehicle (Medium) or IL-6 in combination with HGF for 48 h. f Immunoblot analysis of ZEB2 in CMS2 CR-CSphCs (CSphC #8) treated as indicated. Data are mean ± S.D. of three independent experiments using two different CSphCs (CSphC #8, #9). Samples were run on the same gel and images were cropped only for the purpose of this figure. Source data are provided as a Source Data file. g Kinetic growth of CR-CSphCs treated as indicated. h Number of invading CR-CSphCs at 48 h, pretreated with V-ASC CM and the indicated neutralizing antibodies for 48 h. i Flow cytometry analysis of CD44v6 positivity in CR-CSphCs treated as indicated, for 48 h. For g–i data are mean ± S.D. of six independent experiments performed with 2 different CR-CSphC lines (#8 and #9). For (b–d and g-i) statistical significance was calculated using the unpaired two-tailed t test. j, CD44v6 expression in CD44v6^- and GFP-transduced CD44v6 + cells after 3 days of exposure to V-ASC CM. One representative of 6 independent experiments is shown. k NGF, BDNF, NTF3, and NTF4 mRNA expression levels on CD44v6^- and CD44v6 + cells. Results show mean ± S.D. of three independent experiments performed with enriched cells from two different CR-CSphC lines (CSphC #8, #9). l Lollipop plot showing NGF, BDNF, NT-3, and NT-4 production by the indicated cells treated as indicated. m, Invasion assay of RFP transduced ASCs, using the indicated cells/media as chemoattractant agents. Scale bars, 100 µm. n Number of invading ASCs in presence of the indicated cells/media as chemoattractant agents. For l–n data are mean ± SD of three independent experiments using CR-CSphCs from different patients (CSphC #1, #8, #9, #21).

Fig. 3. VEGF induces endothelial differentiation of ASCs, which activate the EMT program of CRC sphere cells.

a Clustergram of tumor microenvironment-related genes in CR-CSphCs (CSphC #1, #8, #9, #21) and CD44v6 ⁻or CD44v6 + enriched cells. Data are presented as normalized expression values. b VEGF production in cells as indicated. Data are mean ± SD of 4 independent experiments. Box and whiskers show min-to-max values, with line indicating the mean value. c, Gating strategy of CD271/VEGFR expression on ASCs (upper panels). Dot-plots of CD271/VEGFR staining with or without the indicated antibody (FMO-APC control, minus CD271-PE-Cy7) (middle panel). Flow cytometry analysis of CD271 and VEGFR in ASCs. Data are representative of 3 independent experiments performed with 10 different ASC lines (lower panel). d, Percentage of CD31 positivity, by flow cytometry analysis, on CD34 + /CD31^-/CD45^- enriched ASCs exposed to vehicle (Medium), CD44v6 + CR-CSCs CM (CSC #1, #8, #9, #21), in presence or absence of VEGF neutralizing antibody, or VEGF for 14 days. Data are mean ± SD of three independent experiments using 3 different ASC cultures. e Phase-contrast micrographs of capillary-like tubular structures of Huvec cells treated as indicated for 16 h. Scale bars, 500 µm. One representative of 3 independent experiments is shown. f Immunohistochemical analysis of CD31 (red) and CD44v6 (green) on tumor xenografts generated by subcutaneous injection of CR-CSphCs alone or in combination with S-ASCs or V-ASCs (upper panel). Percentage of vascular surface area, based on CD31 positivity, in tumor xenografts (lower panel). Scale bars, 200 µm. Data are representative of 3 independent experiments. For b, d, and f statistical significance was calculated using the unpaired two-tailed t test. g Transcriptomic profile correlation between CMS2 CR-CSphCs (CSphC #8, #9) treated with S-ASCs or V-ASC conditioned medium (CM) and CMS4-associated gene signature. h GSEA of CMS4-associated gene signature in CMS2 CR-CSphCs (CSphC #8, #9) treated with V-ASC CM (upper panel). Top ten significantly up- and downregulated CMS4 signature genes in treated cells (lower panel). Statistical significance between two groups was determined by unpaired Student’s t test (2-tailed). i, Kinetics and whole-body in vivo imaging analysis of mice (n = 6) intrasplenically injected with LUC-GFP CMS4, or CMS2 CR-CSphCs alone or co-injected with V-ASCs and treated as indicated. Data are mean ± S.D. of independent experiments performed with CR-CSphCs isolated from two different CMS2 (CSphC #8, #9) and CMS4 (#1, #21) CRC patients.

Fig. 4. V-ASCs enhance the expression of ZEB2 sustaining the metastatic activity of CR-CSphCs.

a Up- (red) and down- (blue) regulated genes and their relative top twenty significantly enriched gene sets (FDR q value ≤ 0.05), common in CMS4 CR-CSphCs (CSphC #1, #21) and CMS2 cells (CSphC #8, #9) treated with V-ASC CM, selected from all gene sets within MSigDB (H, CP Biocarta, CP Kegg, MIR, CGN, CM, BP, CC, MF, C6, C7). b Heatmap of EMT-related genes in CMS2 (CSphC #8, #9) and CMS4 (CSphC #1, #21) CR-CSphCs treated for 48 h as indicated. c, Venn diagrams of up- and downregulated genes in untreated CMS4 (CSphC #1, #21) and V-ASC CM-treated CMS2 (CSphC #8, #9) CR-CSphCs, compared to untreated CMS2 cells. d ZEB1 and ZEB2 mRNA expression levels in CMS4 and CMS2 CR-CSphCs treated as indicated. e Immunofluorescence analysis of CR-CSphCs expressing nuclear ZEB1 and ZEB2 (CMS2 #8, CMS4 #21) treated as indicated. Nuclei were counterstained with Toto-3. Scale bars, 20 µm. For d and e data represent mean ± S.D. of three independent experiments using CMS2 (#8, #9) and CMS4 (#1, #21) CR-CSphC lines. f Global gene expression profile of miRNAs in CMS4 (CSphC #1, #21) and CMS2 (#8, #9) CR-CSphCs treated as indicated. g Network of most differentially expressed miRNAs and their targets inferred from miRTarBase in CMS2 CR-CSphCs (CSphC #8, #9) treated with V-ASCs CM for 48 h. Bold colors represent miRNAs with a fold-change >8. Orange area within dashed line highlights ZEB1 and ZEB2 as direct targets of miR-200a. h Immunoblot analysis of ZEB2 in CMS4 (CSphC #1, #21) and CMS2 (CSphC #8, #9) CR-CSphCs treated as indicated. Data are mean ± S.D. of three independent experiments performed with CR-CSphCs isolated from 2 different CMS2 (CSphC #8, #9) and CMS4 (CSphC #1, #21) CRC patients. Samples were run on the same gel and images were cropped only for the purpose of this figure. Source data are provided as a Source Data file. For (d and h) statistical significance was calculated using the two-tailed t test. i Flow cytometry analysis of CD44v6 in CMS2 CR-CSphCs (CSphC #8, #9) transduced with empty vector (EV) or ZEB2 synthetic gene. Bars represent means ± S.D. of three independent experiments using two CR-CSphCs. j In vivo whole-body imaging analysis of mice (n = 6) following intrasplenic injection of CR-CSphCs transduced as in i at 30 min and 8 weeks after splenectomy (left panel). Luciferase signal measured as ph/s/cm²/sr (right panel). Data are mean ± S.D. of independent experiments performed with two CMS2 CR-CSphCs (#8, #9).

Fig. 5. IL-6 and HGF induce a mesenchymal phenotype by activation of STAT3.

a GSEA of STAT3 targets gene signature in CMS2 CR-CSphCs (CSphC #8, #9) exposed to V-ASC CM. Statistical significance was calculated as described in Subramanian et al. (doi: 10.1073/pnas.0506580102). b Immunoblot analysis of phosphorylated STAT3 (p-STAT3^{SER727/TYR705}) and STAT3 in CMS2 CR-CSphCs (CSphC #9) treated as indicated. β-actin was used as a loading control. Samples were run on the same gel and images were cropped only for the purpose of this figure. Source data are provided as a Source Data file. One representative of three independent experiments is shown. c Immunohistochemical analysis of p-STAT3 on paraffin-embedded sections of CRC patients with healthy weight or affected by obesity. One representative experiment of nine is shown. Scale bars, 100 µm. Box and whiskers show min-to-max values, with a line indicating the mean value. d Immunoblot analysis of CMS2 CR-CSphCs (CSphC #9) treated as indicated. β-actin was used as a loading control. One representative of four independent experiments is shown. Samples were run on the same gel and images were cropped only for the purpose of this figure. Source data are provided as a Source Data file. e Proliferation rate of CMS2 CR-CSphCs exposed to the indicated treatment. Data are mean ± S.D. of three independent experiments using two different CR-CSphC lines (CSphC #8, #9). f Colony-forming analysis of CR-CSphCs treated as indicated, at 21 days. Boxes and whiskers represent mean ± S.D. of colony size performed in four independent experiments using four different CR-CSphC lines (CSphC #1, #8, #9, #21). n represents the number of colonies. g miR-200a expression levels in CMS2 cells treated as indicated using two different CR-CSphC lines (CSphC #8, #9). U6 was used as housekeeping control gene. Histograms represent mean ± S.D. of three independent experiments. For (c and e-g) statistical significance was calculated using the two-tailed t test. h Clustergram of stemness-related genes in CMS2 CR-CSphCs (CSphC #8, #9) treated for 48 h as indicated. GAPDH and HPRT1 were used as housekeeping control genes. i Phase-contrast analysis of CMS2 CR-CSphCs grown in matrigel and treated as indicated at 10 days. One representative of four independent experiments carried out with two different CR-CSphC lines (CSphC #8, #9) is shown. Scale bars, 20 µm. For (d-i) STAT3 inhibitor C188-9 was used at 10 µM concentration.

Fig. 6. IL-6 and HGF targeting hampers the VAT-induced metastatic capacity of CR-CSphCs.

a Growth kinetics of CMS2 CR-CSphCs treated with V-ASC CM, alone or in combination with tocilizumab (Toc) and crizotinib (Criz). b Colony-forming assay of CR-CSphCs following the indicated treatment, at 21 days. c, miR-200a expression in CMS2 CR-CSphCs treated as indicated for 48 h. U6 was used as housekeeping control gene. d ZEB1 and ZEB2 expression levels in cells treated for 72 h as indicated. GAPDH was used as housekeeping control gene. e Number of invading CMS2 CR-CSphCs pretreated as indicated for 48 h. Statistical significance between two groups was determined by unpaired Student’s t test (2-tailed). For (a-e) data are mean ± S.D. of three independent experiments using two different CR-CSphC lines (CSphC #8, #9). Statistical significance was calculated using the two-tailed t test. f Schematic model of intrasplenic injection of CR-CSphCs showing time points of treatments and in vivo bioluminescence detection. g Kinetics and whole-body imaging analysis of mice (n = 6) following intrasplenic injection of LUC-GFP-transduced CMS2 CR-CSphCs alone or co-injected with V-ASCs untreated or treated with the indicated pharmaceutical compounds. Insets represent spleen collected 30 min after cell injection, and liver, lung, and bowel collected at the time of sacrifice. Data are mean ± S.D. of independent experiments using two different CR-CSphC lines (CSphC #8, #9), and 2 S- (#3, #6) or V-ASC (#5, #14) lines. h RFS rate of CMS2/MSS/Stage1-2 CRC patients according to ZEB2 expression levels. Statistical significance was calculated using the log-rank (Mantel–Cox) test. i Univariate and Multivariate analysis of relapse-free survival (RFS) according to regression Cox model in CRC patients as in h. Statistical significance was calculated using the Wald test. j Schematic representation of bidirectional crosstalk between ASCs and CRC cells. Visceral adipose factors enhance the expression of CD44v6; CD44v6 + -released NGF/NT-3 drives the intra-tumor recruitment of ASCs; adipose-released proteins induce EMT of CRC cells through the activation of STAT3; CD44v6 + -released VEGF promotes the endothelial transdifferentiation of ASCs. Clinically available drugs targeting HGF and IL-6 are highlighted in red.