The Hippo pathway links adipocyte plasticity to adipose tissue fibrosis

Abstract

Fibrosis disrupts adipose tissue (AT) homeostasis and exacerbates metabolic dysfunction upon chronic caloric excess. The molecular mechanisms linking adipocyte plasticity to AT fibrosis are largely unknown. Here we show that the Hippo pathway is coupled with TGFβ signaling to orchestrate a cellular and/or functional shift of adipocytes from energy storage to extracellular matrix (ECM) remodeling in AT fibrosis. We found that Lats1/2-knockout adipocytes could dedifferentiate into DPP4⁺ progenitor cells and convert to DPP4^- myofibroblasts upon TGFβ stimulation. On the other hand, Hippo pathway inhibition during obesity impaired adipocyte identity while promoted ECM remodeling activity of adipocytes. Macrophages recruited by CCL2 produced TGFβ to accelerate AT fibrosis. YAP and TAZ, the Hippo downstream effectors, enhanced SMAD2 stability to promote fibrotic responses. Importantly, inhibition of YAP/TAZ activity in obese mice markedly relieved AT fibrosis and improved metabolic homeostasis. Together, our findings identify the Hippo pathway as a molecular switch in the initiation and development of AT fibrosis, implying it as a therapeutic target.

Fig. 1. Lats1/2 deficiency leads to fat remodeling.

a–d Male WT or ob/ob mice were analyzed at 12 weeks old. a, Real-time quantitative PCR (RT-qPCR) analysis of genes involved in Hippo pathway and fibrotic response in scWAT (n = 5 mice). b, Immunoblot analysis of LATS1/2, YAP/TAZ, YAP phosphorylated at Ser112 (p-YAP) and TAZ phosphorylated at Ser89 (p-TAZ) in scWAT. c Quantification of protein expression from scWAT shown in b (n = 4 mice). d Quantification of phosphorylation levels of YAP/TAZ (n = 4 mice). e–h Male WT mice were fed a HFD or ND for 18 weeks. e RT-qPCR analysis of genes involved in Hippo pathway and fibrotic response in scWAT (n = 5 mice). f Immunoblot analysis of LATS1/2, YAP/TAZ, p-YAP and p-TAZ in scWAT. g Quantification of protein expression from scWAT shown in f (n = 3 mice). h Quantification of phosphorylation levels of YAP/TAZ (n = 3 mice). i–m Male Lats1^f/fLats2^f/f (L1L2-FF) and Lats1^f/fLats2^f/f Adipoq^Cre (L1L2-AKO) mice were analyzed at 5 weeks old. i Quantification of body fat mass by DXA (n = 6 mice). j Quantification of scWAT, vWAT and BAT weight (n = 10 mice). k Representative sections of scWAT with H&E staining or Masson’s trichrome staining at the indicated age. l Immunoblot analysis of adipocyte identity and fibrotic protein expression. m RT-qPCR for adipocyte and fibrotic marker gene expression in scWAT (n = 6 mice). Data are means ± SEM. Two-tailed unpaired Student’s t test; *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values are provided in a Source data file.

Fig. 2. TGFβ stimulation is coupled with Hippo pathway inactivation to promote AT fibrosis.

a Quantification of phosphorylation levels of SMAD2 in scWAT during growth (n = 4 mice). b Eight-week-old male L1L2-FF or Lats1^f/fLats2^f/fAdipoq^CreERT2 (L1L2-iAKO) mice were locally injected with AAV-CAG-GFP and AAV-CAG-TGFβR1 (T204D) respectively in scWAT. After 3 days, all mice were intraperitoneally (i.p.) administered 3 doses of tamoxifen every other day and then analyzed 4 weeks later. c Immunostaining of αSMA⁺ cells in the indicated mice. Independent experiments were performed twice with similar results. d mRNA expression of fibrotic and adipocyte markers in L1L2-FF or L1L2-iAKO scWAT transduced with AAV-CAG-GFP (n = 5 mice) or AAV-CAG-TGFβR1 (T204D) (n = 6 mice) followed by tamoxifen administration. e, f L1L2-FF or L1L2-iAKO mice were locally injected with AAV-CAG-TGFβ1 (2CS) in scWAT. e mRNA expression of fibrotic and adipocyte markers (n = 5 mice). f Immunostaining of αSMA⁺ cells. g AAV vectors for inducible expression of TGFβR1 (T204D) by Double-Floxed Inverted Open reading frame system. h, i, mRNA expression of fibrotic and adipocyte markers (h) (n = 4 mice) or immunostaining of αSMA⁺ cells (i) in L1L2-FF or L1L2-iAKO scWAT transduced with AAV-CAG-GFP or AAV-EF1A-DIO-TGFβR1. j Eight-week-old male L1L2-FF and L1L2-iAKO mice were subcutaneously co-injected with AAV-ADP-FLPo and AAV-EF1A-FIO-TGFβR1 (T204D) for 3 weeks, followed by i.p. injection of tamoxifen. k, l mRNA expression of fibrosis markers (k) (n = 6 mice) or immunostaining of αSMA⁺ cells (l) of mice in j. m mRNA expression of Acta2 of Cas9^Tg/+ SVF transduced with vector (n = 4 biologically independent cell cultures) or YT-gRNA (n = 3 biologically independent cell cultures) in the presence or absence of TGFβ1 for 36 h. Data are means ± SEM. Two-way analysis of variance (ANOVA) with Bonferroni’s multiple-comparisons test in d, e, h, k, m; *P < 0.05, **P < 0.01, ***P < 0.001; NS, not significant. Exact P values are provided in a Source data file.

Fig. 3. Lats1/2 deficiency induces AT fibrosis in obese mice.

a mRNA expression of Tgfb1/2/3 in scWAT of male mice fed a ND or HFD for 18 weeks from 6 weeks old (n = 5 mice). b Immunoblot analysis of phosphorylation of SMAD2/3 (p-SMAD2/3) of mice fed a ND or HFD for 18 weeks. Right, quantification of protein and relative phosphorylation levels of SMAD2/3 (n = 3 mice). c mRNA expression of Tgfb1/2/3 in scWAT of 12-week-old male WT or ob/ob mice (n = 5 mice). d Immunoblot analysis of protein expression of p-SMAD2/3 of mice in 12-week-old male WT or ob/ob mice. Right, quantification of protein and relative phosphorylation levels of SMAD2/3 (n = 4 mice). e, f Eight-week-old male Lats1^f/fLats2^f/f mice were injected with AAV-ADP-GFP or AAV-ADP-Cre in scWAT for 4 weeks. e Immunoblot analysis of LATS1 or LATS2 knockout efficiency in scWAT. f mRNA expression of fibrotic and adipocyte markers in scWAT (n = 6 mice). g Experimental outline: male L1L2-FF mice were fed a HFD for 16 weeks from 6 weeks old, followed by a local injection with AAV-ADP-GFP or AAV-ADP-Cre in scWAT, and all were analyzed after 4 weeks. h RT-qPCR analysis of fibrosis markers in scWAT of mice in g (n = 5 mice). i RT-qPCR analysis of fibrosis markers in scWAT of 12-week-old male Lats1^f/fLats2^f/fob/ob mice injected with AAV-ADP-GFP or AAV-ADP-Cre (n = 5 mice). j, k Representative scWAT sections with Masson’s trichrome staining (j) or Picrosirius red staining (k) of mice in i. Data are means ± SEM. Two-tailed unpaired Student’s t test; *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values are provided in a Source data file.

Fig. 4. Lats1/2 deficiency in adipocytes elicits a CCL2/CCL7-macrophage feedforward loop that increases TGFβ expression in macrophages.

a Expression of F4/80, Col1a1, and Acta2 of L1L2-FF (n = 6) and L1L2-AKO (n = 5 at P1, n = 6 at P7, P14 and P21) mice during growth. b Representative scWAT sections stained for F4/80 of 3-week-old male L1L2-FF or L1L2-AKO mice. Independent experiments were performed three times with similar results. c Quantification of F4/80⁺ cell percentages in scWAT SVF of 3-week-old male L1L2-FF or L1L2-AKO mice (n = 6). d, e Quantification of CD206⁺CD11c⁻ (d) and CD206⁻CD11c⁺ (e) cell percentages in scWAT SVF of 3-week-old male L1L2-FF or L1L2-AKO mice (n = 6). f mRNA expression of inflammatory markers in scWAT of 3-week-old male L1L2-FF or L1L2-AKO mice (n = 6). g mRNA expression of inflammatory genes in male L1L2-FF and L1L2-iAKO mice that were locally injected with AAV-CAG-GFP (n = 5) or AAV-CAG-TGFβR1 (T204D) (n = 6) in scWAT. h mRNA expression of inflammatory genes in male L1L2-FF or L1L2-iAKO mice that were locally injected with AAV-CAG-TGFβ1 (2CS) in scWAT (n = 5). i mRNA expression of inflammatory genes in male L1L2-FF or L1L2-iAKO mice that were locally injected with AAV-EF1A-DIO-TGFβR1 (T204D) in scWAT (n = 4). j Eight-week-old male L1L2-iAKO mice were subcutaneously injected with AAV-CAG-GFP (n = 7) or AAV-CAG-CCL2/CCL7 (n = 5) and analyzed for mRNA expression of inflammatory genes and fibrosis markers 4 weeks later. k F4/80⁻ and F4/80⁺ cells were magnetically sorted from pooled WT scWAT SVF and analyzed for mRNA expression of Tgfb1/2/3 (n = 3). scWAT from three male mice were pooled as one sample. l BMDMs were treated with Veh, and IL4 plus IL13 for 24 h and analyzed for Mrc1, Ccl2 and Ccl7 mRNA expression (n = 4 biologically independent cell cultures). Data are means ± SEM. Two-tailed unpaired Student’s t test in a, c–f, h–l; two-way ANOVA with Bonferroni’s multiple-comparisons test in g; *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values are provided in a Source data file.

Fig. 5. Lats1/2 deficiency promotes cell fate conversion from adipocytes to myofibroblasts.

a Immunoblot analysis of protein expression of GFP in SVF and adipocyte from 8-week-old Adipoq^CreRosa26^mTmG scWAT. Independent experiments were performed twice with similar results. b Flow cytometry analysis of percentages of live CD45⁻CD31⁻GFP⁺DPP4⁺ cells in scWAT SVF of P21 L1L2-FF^{LSL-CAS9-EGFP} or L1L2-AKO^{LSL-CAS9-EGFP} reporter mice (n = 5). c Representative sections of scWAT of P14 or P28 L1L2-AKO^{LSL-CAS9-EGFP} mice stained for αSMA (red), DPP4 (purple), nucleus (DAPI, blue) and GFP (green). Arrowheads indicate examples of GFP⁺DPP4⁺ cells. Arrows point to GFP⁺αSMA⁺ cells. Independent experiments were performed three times with similar results. d Sorted cell subsets from P21 L1L2-AKO^mTmG scWAT SVF were plated and stained for αSMA (AF647, red) and nucleus (DAPI, blue). e Quantification of percentages of GFP+DPP4+ (n = 6), GFP+DPP4- (n = 7), GFP-DPP4+ (n = 7), GFP-DPP4- (n = 7) cells. Dots represent cell percentages of random visual fields from P21 L1L2-AKOmTmG mice. Independent experiments were performed twice with similar results. f Sorted GFP⁺DPP4⁺ cells from P21 L1L2-AKO^{LSL-CAS9-EGFP} scWAT SVF were treated with TGF-β1 (10 ng/ml) for 0 h or 24 h. Cells were stained with αSMA (red), DPP4 (purple), nucleus (DAPI, blue) and GFP (green). An arrowhead points to the GFP⁺αSMA⁻DPP4⁺ cell. Independent experiments were performed twice with similar results. g Sorted GFP⁺DPP4⁺ cells from L1L2-AKO^{LSL-CAS9-EGFP} reporter scWAT SVF were analyzed for DPP4 expression in GFP⁺ cells before or after transplantation into scWAT of 8-day-old L1L2-AKO recipient mice. One representative transplant from n = 3 biological replicates. Data are means ± SEM. One-way ANOVA with Bonferroni’s multiple-comparisons test in e; two-tailed unpaired Student’s t test in b. *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values are provided in a Source data file.

Fig. 6. The YAP/TAZ–TEADs axis cooperates with SMAD2 to regulate fibrotic responses.

a Representative scWAT sections with Masson’s trichrome staining of indicated mouse strains at 5 weeks old. Independent experiments were performed three times with similar results. b mRNA expression of fibrosis and adipocyte markers (n = 6 mice). Data are shown as fold change in L1L2-AKO, L1L2YT-AKO, L1L2Y-AKO, and L1L2T-AKO compared to corresponding WT littermates. The column of L1L2-AKO had also been shown in Fig. 1m. c–n Differentiated adipocytes from Cas9^Tg/+ scWAT was transduced with the indicated gRNAs on day 2 post differentiation, and then analyzed 5 days later. c mRNA expression of adipocyte markers in differentiated adipocytes transduced with Vec (n = 4 biologically independent cell cultures), Lats1/2 gRNA (L1L2-gRNA) (n = 3 biologically independent cell cultures) and LATS1/2-YAP/TAZ gRNA (L1L2YT-gRNA) (n = 4 biologically independent cell cultures). d Immunoblot analysis of protein expression in cells in c. e Schematic diagram of the Lats1/2-Tead1/2/3/4 Tandem gRNA construct. f mRNA expression of adipocyte marker in differentiated adipocytes transduced with Vec, L1L2-gRNA, or Lats1/2-Tead1/2/3/4 gRNA (L1L2-Tds-gRNA) (n = 3 biologically independent cell cultures). g YAP and TAZ protein level in the nuclear/cytoplasmic fractionation of differentiated adipocytes transduced with Vec, L1L2-gRNA and L1L2-Tds-gRNA. h Immunoblot analysis of protein expression of SMAD2, YAP and TAZ in differentiated adipocytes transduced with Vec and L1L2-gRNA. i Quantification of relative density of SMAD2 normalized to HSP90 (n = 3 biologically independent cell cultures). j mRNA expression of Smad2 of cells in i (n = 3 biologically independent cell cultures). k Immunoblot analysis of SMAD2 protein expression in differentiated adipocytes (from SVF of Cas9^Tg/+ scWAT) transduced with the indicated gRNAs. l Immunoblot analysis of SMAD2 protein expression in a cycloheximide chase experiment. m, Regression analysis of SMAD2 protein stability in l. n Differentiated adipocytes were transduced with MYC-Ub, L1L2-gRNA-SMAD2-HA and YT-gRNA. Immunoprecipitation (IP) assay showing an increased ubiquitination of SMAD2 in Yap/Taz-deficient cells. Data are means ± SEM. One-way ANOVA with Bonferroni’s multiple-comparisons test in c, f; two-tailed unpaired Student’s t test in b, i, j; *P < 0.05, **P < 0.01, ***P < 0.001; NS, not significant. Exact P values are provided in a Source data file.

Fig. 7. Targeting YAP/TAZ to relieve AT fibrosis in obese mice.

a–e Yap^f/fTaz^f/fAdipoq^CreERT2 (YT-iAKO) or YT-FF mice were fed a HFD for 11 weeks followed by i.p. injection of 5 doses of tamoxifen (50 mg/kg) every day (n = 6). a Experimental scheme. b mRNA expression of fibrosis markers in scWAT from YT-FF or YT-iAKO mice (n = 6). c Representative scWAT sections with Picrosirius red staining of YT-FF or YT-iAKO mice. d, e Glucose tolerance test (GTT) (d) or insulin tolerance test (ITT) (e) of YT-FF (n = 6) or YT-iAKO (n = 7) mice. f–l Eight-week-old ob/ob mice were i.p. injected with Veh or verteporfin (VP) (25 mg/kg) for 5 doses every other day. f Experimental scheme. g mRNA expression of fibrosis markers and YAP target genes in scWAT of ob/ob mice (n = 6). h Immunoblot analysis (left) and quantification (right) of Col1a1 (n = 3 mice). i Representative sections of scWAT stained for Col1a1 (red) and nucleus (DAPI, blue). j Representative scWAT sections with Picrosirius red staining of ob/ob mice. k GTT (n = 6 mice). l ITT (n = 6 mice). Data are means ± SEM. Two-tailed unpaired Student’s t test in b, g, h; two-way ANOVA with Bonferroni’s multiple-comparisons test in d, e, k, l; *P < 0.05, **P < 0.01, ***P < 0.001. Exact P values are provided in a Source data file.

Fig. 8. Hippo pathway and TGFβ signaling synergistically orchestrate adipocyte plasticity and AT fibrosis.

In genetic model, Lats1/2-knockout adipocytes can dedifferentiate into DPP4⁺ progenitors and convert to DPP4⁻ myofibroblasts. Loss of Lats1/2 activates YAP/TAZ that in turn enhance SMAD2 stability through inhibiting its ubiquitination. In the presence of CCL2 and CCL7, monocyte-derived macrophages are recruited to scWAT to produce TGFβ, which activates SMAD2 in dedifferentiated adipocytes to form a complex with YAP/TAZ. The complex translocates into the nucleus to induce transcription of inflammatory and fibrotic genes. In obesity model, the Hippo pathway is inhibited while TGFβ signaling is activated, which impairs adipocyte identity and promotes AT fibrosis. Moreover, CCL2/7 and macrophage form a feedforward loop to amplify inflammatory responses and accelerate AT fibrosis. Together, Hippo pathway inactivation is coupled with TGFβ stimulation to promote AT fibrosis.