IL-6/STAT3 axis dictates the PNPLA3-mediated susceptibility to non-alcoholic fatty liver disease

Abstract

Background & aims: A number of genetic polymorphisms have been associated with susceptibility to or protection against non-alcoholic fatty liver disease (NAFLD), but the underlying mechanisms remain unknown. Here, we focused on the rs738409 C>G single nucleotide polymorphism (SNP), which produces the I148M variant of patatin-like phospholipase domain-containing protein 3 (PNPLA3) and is strongly associated with NAFLD.

Methods: To enable mechanistic dissection, we developed a human pluripotent stem cell (hPSC)-derived multicellular liver culture by incorporating hPSC-derived hepatocytes, hepatic stellate cells, and macrophages. We first applied this liver culture to model NAFLD by utilising a lipotoxic milieu reflecting the circulating levels of disease risk factors in affected individuals. We then created an isogenic pair of liver cultures differing only at rs738049 and compared NAFLD phenotype development.

Results: Our hPSC-derived liver culture recapitulated many key characteristics of NAFLD development and progression including lipid accumulation and oxidative stress, inflammatory response, and stellate cell activation. Under the lipotoxic conditions, the I148M variant caused the enhanced development of NAFLD phenotypes. These differences were associated with elevated IL-6/signal transducer and activator of transcription 3 (STAT3) activity in liver cultures, consistent with transcriptomic data of liver biopsies from individuals carrying the rs738409 SNP. Dampening IL-6/STAT3 activity alleviated the I148M-mediated susceptibility to NAFLD, whereas boosting it in wild-type liver cultures enhanced NAFLD development. Finally, we attributed this elevated IL-6/STAT3 activity in liver cultures carrying the rs738409 SNP to increased NF-κB activity.

Conclusions: Our study thus reveals a potential causal link between elevated IL-6/STAT3 activity and 148M-mediated susceptibility to NAFLD.

Impact and implications: An increasing number of genetic variants manifest in non-alcoholic fatty liver disease (NAFLD) development and progression; however, the underlying mechanisms remain elusive. To study these variants in human-relevant systems, we developed an induced pluripotent stem cell-derived multicellular liver culture and focused on a common genetic variant (i.e. rs738409 in PNPLA3). Our findings not only provide mechanistic insight, but also a potential therapeutic strategy for NAFLD driven by this genetic variant in PNPLA3. Our liver culture is therefore a useful platform for exploring genetic variants in NAFLD development.

Keywords: Disease modelling; Genetic variant; IL-6/STAT3 signalling; Liver culture; Multicellular culture; NAFLD; PNPLA3; Stem cells.

Fig. 1.. Generation of a hPSC-derived multicellular liver culture.

(A) Top: scheme of HSCs differentiation, followed by replating on plastic surface. Middle: representative images at specified stages. Bottom: staining of lipid droplets. (B) Auto-fluorescence (blue) of retinoic acid under UV light. Negative control: Mesoderm (MES) cells; Nuclei: Histone H3(red). (C, D) Analysis of HSC markers by western blot (C) and RT-qPCR (D, n=4 for hPSC-derived HSCs and 2 technical replicates for pHSC each donor) in hPSC-derived HSCs from day 13 (qHSC) and 18 (aHSC) and primary HSCs (pHSCs) from two donors (pri-1 and pri-2) at early (p2) and late passage (p5). (E) Collagen secretions in supernatants from re-plated hPSC-HSCs. (F) Phagocytosis analysis of hPSC- and primary-macrophages using fluorescently pre-labeled E. coli (Green). Negative control: Cytochalasin D. (G) Cytochrome P450 activity in hPSC-derived hepatocytes (HEP, n=6) and primary human hepatocytes (PHH, n=4 technical replicates) from two donors. (H-I) Left: Scheme of multicellular liver culture (H). Right: staining of cell-type markers. Staining of HEPs and HSCs in bottom compartment (I). (J) Cell numbers at the indicated time points post-coculture, n=3. Data are mean ± SD and compared with Student’s t-test (two-tailed) with Welch’s corrections, p < 0.05 were considered statistically significant.

Fig. 2.. Characterization of the multicellular liver culture.

(A) Albumin secretion from three-cell (HEP/HSC/Mac), two-cell (HEP/HSC), and HEP monoculture, n=3. (B) Analysis of CYP3A4 and CYP2B6 transcripts in hepatocytes, n=4. (C) Analysis of LRAT transcript in HSCs from three-cell (HSC/HEP/Mac), two-cell (HSC/HEP, HSC/Mac) and HSC monoculture, n=4. (D) Staining of lipid droplets, hepatocytes (ALB), and HSCs (PCDH7). (E) Analysis of IL10 transcript in individual cell-type at three-weeks post-coculture, n=4. (F) Analysis of IL6 and TNF transcripts in macrophages, n=4. Positive control: IFNγ-treated macrophages. (G) Cytokines analysis in liver cultures treated with LPS or control, with or without macrophages, n=4. Data are mean ± S.D; Student’s t-test (two-tailed) with Welch’s corrections (B); One-way ANOVA with Tukey’s multiple comparisons test (A, C, E-G). p < 0.05 were considered statistically significant.

Fig. 3.. Modeling NAFLD in multicellular liver cultures.

(A) Scheme of NAFLD liver cultures, with healthy (Heal) or lipotoxic milieu (Lipo). (B-E) Analysis of oil-red intensity (B, n=10), TAG (C, n=4), baseline glucose secretion (D, n=4) and insulin-induced glucose secretion (E, n=5) in hepatocytes. (F). Analysis of AKT, FOXO1, phosphorylated AKT and FOXO1 in insulin-treated hepatocytes. (G-H). Analysis of markers in HSCs (G) or cytokines in supernatants (H, n=4) from three-cell (HEP/HSC/Mac), two-cell (HSC/Mac, HSC/HEP) or HSC monocultures. (I). Analysis of TNF, TGFB, and IL6 transcripts in macrophages, n=4. (J). Analysis of CK18 fragments in supernatants with either glucose or fructose, n=3. (K). Scheme of liver cultures with hPSC-derived or primary cells. (L-O). Analysis of CHI3L1 and IL6 transcripts (macrophages, L, n=3), CXCL10 transcript and ATP levels (hepatocytes, M-N, n=3), and HSC markers (HSCs, O) in hPSC- or primary cells (two donors: Pri-1 and Pri-2)-derived liver cultures. Data are mean ± S.D; Student’s t-test (two-tailed) with Welch’s corrections (B-D, I); One-way ANOVA with Tukey’s multiple comparisons test (E, H, J, L-N). p < 0.05 were considered statistically significant.

Fig. 4.. Generation of an isogenic pair of co-cultures harboring either PNPLA3^WT or PNPLA3^I148M.

(A). Introduction of rs738409 C>G into hPSC^WT to generate hPSC^I148M line. (B). Proliferation analysis of hPSC^WT and hPSC^I148M pluripotent cells, n=3. (C-F). Analysis of albumin secretion (hepatocytes, C, n=4); IL10 transcript (macrophages, n=3) and CXCL10 and IL6 transcripts (macrophages treated with LPS, E, n=3) and HSC markers (with or without LPS treatment, control: pHSCs from two donors without LPS treatment, F) in differentiated individual cell-types. (G). Expression of PNPLA3, PDGFRβ, ALB at three-weeks under healthy condition. (H-I). Analysis of IL10 transcript (macrophages, H, n=3) and LRAT (HSCs, I, n=3) in liver cultures under healthy condition. Week 0: macrophages and HSCs before co-culture. Data are mean ± S.D; Student’s t-test (two-tailed) with Welch’s corrections. p < 0.05 were considered statistically significant.

Fig. 5.. Enhanced susceptibility to NAFLD phenotypes in co-cultures harboring PNPLA3^I148M.

(A). Analysis of oil-red intensity (left, n=10) and TAG (right, n=4) in purified hepatocytes. (B-D). Baseline glucose secretion (B, n=4), insulin-induced glucose secretion (C, n=4), and ATP level (D, n=3) in hepatocytes under lipotoxic condition. (E). Fibronectin (FN) secretion in culture supernatants, n=3. (F). Analysis of CHI3L1 transcript in macrophages, n=3. (G). Cytokines analysis in both liver cultures, n=3. Data are mean ± S.D; One-way ANOVA with Tukey’s multiple comparisons test. p < 0.05 were considered statistically significant.

Fig 6.. Elevated IL6/STAT3 activity in PNPLA3^I148M liver cultures and NAFLD patients.

(A). Analysis of cytokine secretions at day 6 post-lipotoxic exposure, n=3. (B). Analysis of STAT3 and phosphorylated STAT3 in purified individual cells. (C). Heatmap: Top 30 DEGs across PNPLA3 wild-type, I148M heterozygous and homozygous groups. (D).GO analysis of DEGs between PNPLA3 I148M homozygous and wild-type patients (p<0.01). (E-F). Selected transcripts in pooled three-cell (E, n=4) or IL6-treated liver cultures (F, n=4). Data are mean ± SD; Student’s t-test (two-tailed) with Welch’s corrections. p < 0.05 were considered statistically significant.

Fig 7.. IL6/STAT3 hyperactivation mediates the enhanced susceptibility to NAFLD phenotypes in PNPLA3^I148M liver cultures.

(A-D). Analysis of TAG (A, n=4), ACC1 and FASN transcripts (B, n=3), baseline glucose secretion (C, n=4) in hepatocytes, and activation markers in HSCs (D) purified from liver cultures transduced with shRNA-control (sh-Ctrl) or shRNA-IL6 (sh-IL6). (E). Analysis of soluble IL6R (sIL6R) and gp130 (sgp130) in supernatants, n=3. (F-G). Analysis of ACC1 and FASN transcripts (F, n=3) and HSC markers (G) in liver culture treated with sgp130-Fc protein. (H). HSC activation markers in PNPLA3^WT liver cultures supplemented with IL6. Data are mean ± S.D; One-way ANOVA with Tukey’s multiple comparisons test (A-C, E-F). p < 0.05 were considered statistically significant.

Fig 8.. IL6/STAT3 hyperactivation induced by NF-κB activation in PNPLA3^I148M liver cultures.

(A). Analysis of IL6 transcript in purified individual cells, n=3. (B). Analysis of transcription factors in wild-type macrophages treated with both supernatants collected in Fig.6A. (C). Analysis of IL6 transcript in macrophages collected from (B), n=4. (D). NF-κB luciferase activity in transduced macrophages, n=4. Data are mean ± S.D; Student’s t-test (two-tailed) with Welch’s corrections (C-D); One-way ANOVA with Tukey’s multiple comparisons test (A). p < 0.05 were considered statistically significant.