Metastatic tumor growth in steatotic liver is promoted by HAS2-mediated fibrotic tumor microenvironment

Abstract

Steatotic liver enhances liver metastasis of colorectal cancer (CRC), but this process is not fully understood. Steatotic liver induced by a high-fat diet increases cancer-associated fibroblast (CAF) infiltration and collagen and hyaluronic acid (HA) production. We investigated the role of HA synthase 2 (HAS2) in the fibrotic tumor microenvironment in steatotic liver using Has2ΔHSC mice, in which Has2 is deleted from hepatic stellate cells. Has2ΔHSC mice had reduced steatotic liver-associated metastatic tumor growth of MC38 CRC cells, collagen and HA deposition, and CAF and M2 macrophage infiltration. We found that low-molecular weight HA activates Yes-associated protein (YAP) in cancer cells, which then releases connective tissue growth factor to further activate CAFs for HAS2 expression. Single-cell analyses revealed a link between CAF-derived HAS2 and M2 macrophages and CRC cells through CD44; these cells were associated with exhausted CD8+ T cells via programmed death-ligand 1 and programmed cell death protein 1 (PD-1). HA synthesis inhibitors reduced steatotic liver-associated metastasis of CRC, YAP expression, and CAF and M2 macrophage infiltration, and improved response to anti-PD-1 antibody. In conclusion, steatotic liver modulates a fibrotic tumor microenvironment to enhance metastatic cancer activity through a bidirectional regulation between CAFs and metastatic tumors, enhancing the metastatic potential of CRC in the liver.

Keywords: Extracellular matrix; Fibrosis; Hepatology; Liver cancer; Oncology.

Figure 1. HFD-induced steatotic liver increases HA accumulation and HAS2 expression in tumors.

(A) Representation of the mouse model illustrating the induction of steatotic liver and subsequent splenic injection of MC38 cells to form liver metastases. (B) Macroscopic appearance of the liver with arrows indicating tumor sites. Scale bar: 1 cm. (C) Representative images of hematoxylin and eosin–stained (H&E-stained) tumor and quantitative assessment of tumor area based on H&E staining. (n = 4–5 per group.) Scale bar: 500 μm. (D) RNA-Seq analysis. Top: Gene set enrichment analysis (GSEA) of gene expression for ECM and liver fibrosis in tumors from mice fed an LFD or an HFD. Bottom: Enrichment plot for ECM organization. FDR, false discovery rate; NES, normalized enrichment score; NOM, nominal. (E) A heatmap of the HA-related genes. (n = 5.) (F) Representative microscopic images depicting liver sections stained for α-smooth muscle actin (α-SMA), with Sirius red, and for HA-binding protein (HABP). Scale bars: 200 μm. Quantification of α-SMA–positive area, Sirius red–positive area, and HABP-positive area. (n = 4–5 per group.) (G) Comparison of mRNA expression levels of Has2 in nontumor (NT) and tumor tissues from mice fed an LFD or an HFD. (n = 8 per group.) (H) Representative images of RNAscope in situ hybridization for Has2. Scale bar: 50 μm. Data are presented as mean ± SEM. Statistical significance was calculated with a 2-tailed Student’s t test (C and F) and 1-way ANOVA followed by Tukey’s post hoc test (G). *P < 0.05, **P < 0.01.

Figure 2. Suppression of metastatic liver tumor growth enhanced by metabolic dysfunction–associated steatotic liver disease through HSC-specific Has2 deficiency.

(A) Macroscopic appearance of the liver. WT and Has2^ΔHSC mice were intrasplenically injected with MC38 cells after 6 weeks of either LFD or HFD feeding. Mice were maintained on their respective diets for an additional 2 weeks. Arrows indicate the tumors. Scale bar: 1 cm. (B) Measurement of liver weight, maximal tumor diameter, and number of nodules. (n = 10–12 per group.) KO, knockout. (C) Representative images of H&E staining of liver tissue sections. Scale bar: 200 μm. (D) Representative images and quantification of HABP staining. (n = 6.) Scale bar: 200 μm. (E) Representative images and quantification of α-SMA and Sirius red staining of liver tissue sections. (n = 6.) Scale bar: 200 μm. Data are presented as mean ± SEM. Statistical significance was calculated with 1-way ANOVA followed by Tukey’s post hoc test. *P < 0.05, **P < 0.01.

Figure 3. HAS2 and LMW-HA play an important role in cancer cell aggressiveness and YAP activation.

(A) Serum HA levels (left) and tumor HA levels (right) in tumor-bearing mice on an LFD and those on an HFD. (Serum, n = 5–6 per group; tumor, n = 8 per group.) (B) Fractionated analysis of HA content, distinguishing between HMW-HA (>300 kDa), medium–molecular weight (MMW) HA (100–300 kDa), and LMW-HA (<100 kDa). Serum and tissue homogenate samples from HFD-fed, tumor-bearing mice were fractionated using columns. (Serum, n = 4; tumor, n = 8.) (C and D) The influence of LMW-HA and HMW-HA on the colony formation (C) and invasion ability (D) of MC38 cells. (n = 3.) (E) The effect of LMW-HA and HMW-HA on Yap1 and Ccn2 mRNA expression in MC38 cells. (n = 3.) Con, control. (F) The correlation between LMW-HA levels and Ccn2 mRNA expression in tumors from mice on an LFD and on an HFD. The Pearson’s correlation coefficient (r) was calculated. (n = 5.) (G) Effect of HSC-specific Has2 deletion on YAP expression in tumors from WT mice or knockout mice. (n = 8.) Representative images of YAP staining are shown. NT, nontumor; T, tumor. Scale bar: 100 μm. (H) The effect of Yap1 knockdown in MC38 cells on LMW-HA–induced cancer cell invasion. The number of invaded cells per field is shown. (n = 3.) sh, short hairpin. Data are presented as mean ± SEM. Statistical significance was calculated with Student’s t test (A and H) and 1-way ANOVA followed by Tukey’s post hoc test (B–E). *P < 0.05, **P < 0.01.

Figure 4. YAP knockdown attenuates CRC aggressiveness and CAF activation.

(A) After 6 weeks of LFD or HFD feeding, MC38 cells with either shCon or shYap1 were intrasplenically injected into mice. Left: Liver weight. Middle: Maximum tumor size. Right: Number of nodules. (n = 7–8 per group.) (B) Gene set enrichment analysis. NES, normalized enrichment score; NOM, nominal. (C) Representative Sirius red (top), α-SMA (middle), and HABP (bottom) in metastatic liver tumors. (D) mRNA expression of Col1a1 and Acta2 in tumor tissues. (n = 8.) (E) Quantification of HABP-positive area. (n = 5–6 per group.) (F) Coculture experiments. Has2 and Col1a1 mRNA levels in mouse primary HSCs are shown. ShCon-MC38 or shYap1-MC38 cells were placed in the upper chamber, and primary HSCs were seeded in the lower chamber. (n = 3.) (G) Ccn2 mRNA levels in tumor tissues. (n = 8.) (H) CTGF treatment in primary HSCs. (n = 3.) (I) Has2 mRNA levels in mouse primary HSCs. MC38 cells were transiently transfected with small interfering RNA for control (siCon) or Ccn2 (siCcn2) and treated with vehicle or LMW-HA. MC38 cells were loaded in the upper chamber. HSCs were seeded in the lower chamber 1 day before coculture. Coculture lasted 48 hours. (n = 4.) (J) Illustration showing the bidirectional regulation between HSCs and CRC. Data are presented as mean ± SEM. Statistical significance was calculated with Student’s t test (D and F–H) and 1-way ANOVA followed by Tukey’s post hoc test (A, E, and I). *P < 0.05, **P < 0.01.

Figure 5. CAF-derived HAS2 and cancer-derived YAP contribute to a prometastatic immune TME in steatotic liver.

(A and B) Representative immunohistochemistry images for F4/80 (A) and CD206 (B) from tumors in Figure 2. Scale bars: 200 μm. (C) Quantification of F4/80-positive (top) and CD206-positive (bottom) areas. (n = 5–7 per group.) Data are presented as mean ± SEM. Statistical significance was calculated with 1-way ANOVA followed by Tukey’s post hoc test. *P < 0.05, **P < 0.01. (D) Tumor-infiltrating CAF and immune cell populations. Uniform manifold approximation and projection (UMAP) of single-cell RNA-Seq from 46,577 cells showing 25 clusters determined by integrated analysis, colored by cluster. Cells were from metastatic liver tumors of LFD-fed and HFD-fed mice. (n = 3 per group.) (E, J, and M) The proportion of CAF (E), M1 and M2 (J), and T cell (M) clusters in metastatic liver tumors of LFD-fed and HFD-fed mice. (F and G) Expression of Has1, Has2, Has3, and Cd44 genes (columns) by specific CAF subpopulations (rows). Dot size represents the cell fraction within the CAF subpopulations. Fill color indicates average expression (Ave. exp.). (H and I) CellChat (53) receptor-ligand analysis of the predicted intercellular communication networks for cells from metastatic liver tumors of LFD-fed and HFD-fed mice. Arrows are proportional to the interaction strength between CAF2 and other cell clusters; node size indicates the number of cells within that population. (K and L) Expression of Cd44 and Cd274 genes (columns) by specific M2 subpopulations (rows). Dot size represents the cell fraction within the M2 subpopulations. (N and O) Expression of key immunomodulatory genes (columns) by specific T cell subpopulations (rows). (P) Proposed model representing cancer YAP regulation of HSC-derived HAS2 for the immunosuppressive TME in steatotic liver.

Figure 6. Increased CAF infiltration and immunosuppressive TAM and T cell phenotypes in patients with CRC liver metastasis with steatotic liver.

(A) Representative images for HABP staining (left) and CTGF staining (right) using tissue microarray sections of metastatic CRC patients. Scale bars: 200 μm. (B) Quantification of HABP-positive area. (Normal, n = 16; MASLD, n = 13.) NT, nontumor; T, tumor. (C) Correlation between HABP-positive area and CTGF-positive area. Pearson’s correlation coefficient (r) was calculated. (Normal, NT, n = 15; MASLD, NT, n = 12; normal, tumor, n = 14; MASLD, tumor, n = 11.) (D) Top: Representative IMC images for metastatic liver tumors for FAP, α-SMA, CD8, CD44, and CDX2 expression. Scale bars: 100 μm. Bottom: Per-patient proportions of CAFs and α-SMA–positive CAFs. (Normal, n = 17; MASLD, n = 13.) (E and F) Spatial analysis of IMC data to evaluate density of CAFs surrounding CD44-positive or -negative (E) or YAP-positive or -negative (F) cancer cells. (G and H) Dot plot for expression of CD44, YAP, Ki67, and immunomodulatory molecules (columns) on CD44-positive and -negative (G) or YAP-positive and -negative (H) cancer cells (rows). Dot size represents the cell fraction within each cell population. Fill color indicates average expression (Ave. exp.). (I) Dot plot for expression of immunomodulatory molecules (columns) by macrophage subpopulations from patients with or without MASLD (rows). (J and K) Spatial analysis of IMC data to evaluate the relationship between macrophage PD-L1 expression and macrophages’ distance from CAFs (J) or density of CAFs (K). (L) Illustration of the proposed model. Data are shown as mean ± SEM (B and D) or mean ± SD (E, F, J, and K). Statistical significance was calculated with 1-way ANOVA followed by Tukey’s post hoc test (B) or with 2-tailed Student’s t test (D) or generalized linear models using the sample as a clustering variable to obtain robust standard error (E, F, J, and K). *P < 0.05, **P < 0.01.

Figure 7. Inhibition of HA synthesis alleviates metastatic liver tumor growth and CAF activation in the steatotic liver disease condition.

(A) In vivo preventive experimental protocol. 4-Methylumbelliferone (4-MU) was administered orally (PO) at 450 mg/kg, 5 times a week for 4 weeks, while 4-methylumbelliferyl glucuronide (4-MUG) was provided in drinking water at 2 mg/mL for 4 weeks. (B) Macroscopic appearance of the liver. Arrows indicate tumor sites. Veh, vehicle. Scale bar: 1 cm. (C) Analysis of maximal tumor diameter and number of nodules. (n = 6–9 per group.) (D) Quantitative assessment of HABP-, Sirius red–, and α-SMA–positive areas. (n = 6–9 per group.) (E) Measurement of mRNA expression levels for profibrogenic genes in HFD-fed mice treated with the respective drugs. (n = 6–8 per group.) (F) Western blot analysis of mature collagen I and α-SMA. (G) Evaluation of mRNA expression levels for Has1, Has2, and Has3. (n = 6–8 per group.) (H) In vivo treatment experimental protocol. (I) Macroscopic appearance of the liver from tumor-bearing mice. Scale bar: 1 cm. (J) Quantification of the maximal tumor diameter and the number of tumor nodules. (n = 7 per group.) (K) Kaplan-Meier survival curves. Statistical significance was determined using the log-rank test. (n = 8 per group.) Data are presented as mean ± SEM. Statistical significance was calculated with 1-way ANOVA followed by Tukey’s post hoc test (C–E, G, and J). *P < 0.05, **P < 0.01.

Figure 8. Improved effects of anti–PD-1 antibody treatment in combination with HA synthesis inhibition on metastatic liver tumor growth in the steatotic liver disease condition.

(A) Experimental protocol for the in vivo combination treatment of 4-MU and anti–PD-1 antibody. 4-MU was administered orally (PO) at 450 mg/kg daily for 2 weeks, while anti–PD-1 antibody (200 μg) was administered intraperitoneally (i.p.) every 3 days for a total of 4 injections. (B) Analysis of maximal tumor diameter and number of nodules. (n = 8–9 per group.) Data are presented as mean ± SEM. Statistical significance was calculated with 1-way ANOVA followed by Tukey’s post hoc test. **P < 0.01. (C) Macroscopic appearance of the liver. Arrows indicate tumor sites. Scale bar: 1 cm. IgG, control IgG.