Hepatic Snai1 and Snai2 promote liver regeneration and suppress liver fibrosis in mice

SUMMARY

Liver injury stimulates hepatocyte replication and hepatic stellate cell (HSC) activation, thereby driving liver regeneration. Aberrant HSC activation induces liver fibrosis. However, mechanisms underlying liver regeneration and fibrosis remain poorly understood. Here, we identify hepatic Snai1 and Snai2 as important transcriptional regulators for liver regeneration and fibrosis. Partial hepatectomy or CCl4 treatment increases occupancies of Snai1 and Snai2 on cyclin A2 and D1 promoters in the liver. Snai1 and Snai2 in turn increase promoter H3K27 acetylation and cyclin A2/D1 expressions. Hepatocyte-specific deletion of both Snai1 and Snai2, but not one alone, suppresses liver cyclin A2/D1 expression and regenerative hepatocyte proliferation after hepatectomy or CCl4 treatments but augments CCl4-stimulated HSC activation and liver fibrosis. Conversely, Snai2 overexpression in the liver enhances hepatocyte replication and suppresses liver fibrosis after CCl4 treatment. These results suggest that hepatic Snai1 and Snai2 directly promote, via histone modifications, reparative hepatocyte replication and indirectly inhibit liver fibrosis.

In brief

Wang et al. show that hepatic Snai1 and Snai2 proteins support liver regeneration and defend against liver fibrosis. Snai1 and Snai2 stimulate regenerative hepatocyte replication by regulating transcriptions of cell-cycle genes. The Snai1/Snai2 pathway provides a genomic paradigm for studying liver injury, regeneration, and fibrosis.

Graphical Abstract

INTRODUCTION

The liver is an essential metabolic organ, and liver disease is detrimental. In response to liver injury, hepatocytes, unlike many other types of differentiated cells, are rapidly activated to proliferate, thereby generating new hepatocytes to replace lost ones.^1,2 In rodents with 70% of partial hepatectomy (PH), the remaining liver grows to a normal size within 10 days. Hepatic stellate cells (HSCs) are also activated during liver injury and then proliferate and differentiate into myofibroblasts.³ Myofibroblasts secrete extracellular matrix (ECM) that helps liver repair, but excessive HSC activation and ECM production cause liver fibrosis.^4,5 Of note, hepatocyte replication and liver regeneration are impaired in chronic liver disease, which accelerates liver disease progression.^6–9 However, mechanisms underlying liver regeneration impairment and liver fibrosis remain poorly understood.

Snai1 and Snai2 (also called Slug) are related transcription factors that promote epithelial-to-mesenchymal transition (EMT).^10,11 Snai1 and Snai2 bind, via the C-terminal zinc-finger domain, to DNA at a consensus E2 box (CAGGTG or CACCTG).^10,11 The N-terminal Snai1 and Gfi-1 domain of Snai1 or Snai2 binds to various epigenetic modifiers that elicit histone modifications on target promoters, including histone 3 lysine-9 (H3K9) acetylation (H3K9ac), H3K27ac, and H3K27 methylation (H3K27me).^10,12–14 Snai1 and Snai2 play pleiotropic functions in addition to stimulating EMT. Snai2 in hypothalamic neurons induces leptin resistance and obesity, presumably through repressing leptin receptor LepRb expression.¹⁴ Adipose Snai1 inhibits lipolysis by suppressing adipose triglyceride lipase expression.¹⁵ Hepatic Snai2 potentially promotes de novo lipogenesis and metabolic dysfunction-associated steatohepatitis (MASLD) by stimulating expression of fatty acid synthase.¹² Surprisingly, hepatic Snai1 exerts the opposite effects on MASLD progression,¹³ presumably by recruiting different epigenetic modifiers to target promoters. In contrast, Snai1 and Snai2 act redundantly to support stem cell function and tissue repair.^16–19 Snai1 and Snai2 act in concert to promote mesenchymal stem cell self-renewal and osteoblastogenesis.^20,21 They also support skin stem cell survival and skin wound healing.^22–24 However, the roles of hepatic Snai1 and Snai2 in liver regeneration and liver fibrosis are poorly understood. In this study, we generated and characterized hepatocyte-specific Snai1 (Snai1^Δhep), Snai2 (Snai2^Δhep), and Snai1/Snai2 double-knockout (Snai1/2^Δhep) mice. We examined reparative hepatocyte proliferation and liver regeneration using both 70% PH and carbon tetrachloride (CCl4)-induced liver injury paradigms. We explored HSC activation and liver fibrosis using a chronic CCl4 treatment model. We found that hepatic Snai1 and Snai2 act coordinately to promote hepatocyte replication and liver regeneration by inducing, at least in part, histone modifications and transcriptional reprogramming. Moreover, Snai1/Snai2-stimulated hepatocyte replication restrains improper activation of HSCs, thereby protecting against liver fibrosis.

RESULTS

Hepatocyte-specific deletion of Snai1 and Snai2 impairs liver regeneration after PH

Considering that Snai1 and Snai2 act redundantly to promote muscle stem cell proliferation,²⁵ we reasoned that deletion of both hepatic Snai1 and Snai2, but not one alone, might impair liver regeneration. Snai1^f/f mice were crossed with Snai2^f/f mice to produce Snai1^f/f;Snai2^f/f (Snai1/2^f/f) mice. Snai1/2^Δhep mice (Snai1^f/f;Snai2^f/f;albumin-Cre^+/−) were generated by crossing Snai1/2^f/f mice with albumin-Cre drivers. Snai1/2^Δhep mice were relatively normal. To stimulate liver regeneration, 70% of the liver was removed at 8 weeks of age following a PH procedure.⁹ BrdU was injected 10 h before euthanization to assess DNA synthesis. Plasma alanine aminotransferase (ALT) levels, a liver injury marker, were similar between Snai1/2^Δhep and Snai1/2^f/f mice during the entire experimental period (Figure S1A). Liver sections were stained with antibodies to BrdU and Ki67 (a proliferation marker). Baseline liver cell proliferation was barely detectable. PH increased hepatocyte replication in both Snai1/2^Δhep and Snai1/2^f/f mice (Figure S1B). Importantly, Ki67⁺ and BrdU⁺ cell numbers were substantially lower in Snai1/2^Δhep than in Snai1/2^f/f littermates (Figure 1A). Ki67⁺ and BrdU⁺ cell contents (normalized to DAPI⁺ cells) were significantly lower in Snai1/2^Δhep than in Snai1/2^f/f mice 2 days after PH (Figure 1B). To further examine hepatocyte replication, we costained liver sections with antibodies to HNF4α (a hepatocyte marker) and BrdU. Proliferating hepatocytes were identified as BrdU⁺HNF4α⁺ cells. BrdU⁺HNF4α⁺ hepatocyte percentages (percentage of total HNF4α⁺ hepatocytes) were significantly lower in Snai1/2^Δhep than in Snai1/2^f/f littermates 2 days post-PH (Figure 1B). We also detected binucleate hepatocytes in liver sections, some of which are dividing cells (Figure 1A, H&E panels). The number of binucleate hepatocytes was significantly lower in Snai1/2^Δhep than in Snai1/2^f/f mice 2 days post-PH (Figure 1B). Liver cell death, as assessed by TUNEL assays, was comparable between Snai1/2^Δhep and Snai1/2^f/f mice (Figures S1C and S1D). Importantly, liver growth (liver weight/body weight ratio) was significantly lower in Snai1/2^Δhep mice relative to Snai1/2^f/f littermates after PH (Figure 1C). Liver weight was significantly lower in Snai1/2^Δhep (0.84 ± 0.07 g, n = 8) than in Snai1/2^f/f mice (1.11 ± 0.03 g, n = 8, p = 0.003) 7 days after PH. To examine liver architectures, liver sections were stained with antibody to glutamine synthetase, a zone 3 marker (central vein hepatocytes). Liver zonation structures were normal in Snai1/2^Δhep mice (Figure S1E). Snai1/2^Δhep mice gained significantly less body weight compared to Snai1/2^f/f mice after PH (Figure 1D), likely due to impaired liver regeneration. Likewise, liver Ki67⁺ cell number was also significantly lower in Snai1/2^Δhep than in Snai1/2^f/f female mice 2 days after PH (Figures 1E and S1H). Liver weight/body weight ratio was significantly lower in Snai1/2^Δhep females 7 days post-PH (Figure 1F). These results unveil hepatic Snai1 and Snai2 as important molecular promoters for liver regeneration.

Figure 1. Hepatocyte-specific ablation of both Snai1 and Snai2 impairs PH-stimulated liver regeneration.

PH was performed on Snai1/2^f/f and Snai1/2^Δhep male mice at 8 weeks of age. BrdU was injected 10 h before euthanization.

(A and B) Liver sections were stained with antibodies to Ki67, BrdU, or HNF4α (2 days after PH). Positive cells were normalized to total (DAPI⁺) or hepatocytes (HNF4α⁺). Arrows: binucleate hepatocytes. Scale bar: 200 mm. Day 2: Snai1/2^f/f: n = 8, Snai1/2^Δhep: n = 6; day 7: n = 8 per group.

(C) Liver weight post-PH (normalized to body weight). Day 0: Snai1/2^f/f:n = 6, Snai1/2^Δhep: n = 7; day 2: Snai1/2^f/f:n = 8, Snai1/2^Δhep: n = 6; days 7 and 14: n = 8 per group.

(D) Body weight changes after PH. Snai1/2^f/f: n = 8, Snai1/2^Δhep: n = 8.

(E and F) PH was performed on Snai1/2^f/f and Snai1/2^Δhep female mice at 8 weeks of age.

(E) Liver sections were stained with anti-Ki67 antibody (2 days after PH). Ki67⁺ cell number was normalized to total cells (DAPI⁺). Snai1/2^f/f:n = 5, Snai1/2^Δhep: n = 5.

(F) Liver weight after PH. Day 0: Snai1/2^f/f: n = 6, Snai1/2^Δhep: n = 5; day 2: n = 5 per group; day 7: Snai1/2^f/f: n = 8, Snai1/2^Δhep: n = 10.

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, two-tailed unpaired Student’s t test (B and E) or two-way ANOVA (C, D, and F).

Hepatocyte-specific deletion of Snai1 and Snai2 inhibits hepatocyte proliferation in CCl4-induced acute liver injury

To test if hepatic Snai1 and Snai2 are involved in liver regeneration in drug (or hepatotoxin)-induced liver injury, Snai1/2^Δhep and Snai1/2^f/f male littermates were treated with a single dose of CCl4 at 8 weeks of age. BrdU was injected 10 h before liver isolation. CCl4 rapidly increased plasma ALT levels in both Snai1/2^Δhep and Snai1/2^f/f mice. ALT levels were modestly higher in Snai1/2^Δhep mice 48 h post-CCl4 injection (Figure S2A). Liver weight (normalized to body weight) was comparable between Snai1/2^Δhep and Snai1/2^f/f mice within 2 days after CCl4 injection (Figure S2B). CCl4 induced liver cell death to a similar degree between Snai1/2^Δhep and Snai1/2^f/f mice (Figures S2C and S2D). To examine hepatocyte proliferation, we stained liver sections with antibodies to Ki67, BrdU, and phospho-histone H3 (pH3, pSer10). CCl4 markedly increased liver Ki67⁺, BrdU⁺, and pH3⁺ cells in both Snai1/2^Δhep and Snai1/2^f/f mice; however, Ki67⁺, BrdU⁺, pH3⁺, and binuclear cell numbers were significantly lower in Snai1/2^Δhep mice 48 and 72 h after CCl4 injection (Figures 2A and 2B). These results suggest that hepatic Snai1 and Snai2 support hepatocyte regeneration during drug-induced liver injury.

Figure 2. Ablation of hepatic Snai1 and Snai2 attenuates hepatocyte proliferation in mice with CCl4-induced acute liver injury.

Snai1/2^f/f and Snai1/2^Δhep male mice (8 weeks) were injected with CCl4 (1 μL/g body weight). BrdU was administrated 10 h before euthanization. Liver sections were stained with the indicated reagents.

(A) Representative liver sections 72 h after CCl4 treatment. Scale bar: 200 μm.

(B) Ki67⁺, BrdU⁺, and pH3⁺ cells were normalized to total liver cells (DAPI⁺). Proliferating hepatocytes (BrdU⁺ HNF4α⁺) were normalized to total hepatocytes (HNF4α⁺). Binucleate hepatocyte number was normalized to total liver cells (H&E staining). CCl4 for 24 h: n = 4 per group; 48 h: Snai1/2^f/f: n = 6, Snai1/2^Δhep: n = 8; 72 h: n = 8 per group; 96 h: n = 6 per group.

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, two-tailed unpaired Student’s t test.

Deletion of hepatic Snai1 and Snai2 impairs liver regeneration in chronic liver injury

To model chronic liver injury, Snai1/2^Δhep and Snai1/2^f/f male littermates were treated with CCl4 for 24 days (0.6 μL/g body weight, twice a week). BrdU was injected 10 h before euthanization. Chronic CCl4 treatment seemed to decrease liver Snai2 levels and increase Snai1 levels in Snai1/2^f/f mice (Figures 3A and 3B). Likewise, PH (for 2 days) also decreased Snai2 levels and increased Snai1 levels in the liver (Figure S1G). Hepatocyte proliferation was assessed by immunostaining liver sections. Ki67⁺, BrdU⁺, pH3⁺, and binucleate cell numbers were significantly lower in Snai1/2^Δhep mice relative to Snai1/2^f/f littermates (Figures 3C and 3D). We specifically assessed hepatocyte proliferation by costaining liver sections with antibodies to HNF4α and BrdU. The amount of BrdU⁺HNF4α⁺ proliferating hepatocytes (percentage of total HNF4α⁺ hepatocytes) was significantly lower in Snai1/2^Δhep than in Snai1/2^f/f mice (Figure 3D). Liver size was smaller, and liver weight was significantly lower in Snai1/2^Δhep than in Snai1/2^f/f mice after CCl4 treatment for 24 days (Figure 3E). We reasoned that impaired hepatocyte replication might stimulate proliferation and differentiation of hepatic progenitor cells as a compensatory liver regeneration mechanism in Snai1/2^Δhep mice. Considering that hepatic progenitor cells express cholangiocyte markers such as keratin 19 (K19), we costained liver sections with antibodies to K19 and Ki67. K19⁺Ki57⁺ (proliferating) cell percentage (percentage of total K19⁺ cholangiocytes/hepatic progenitors) was significantly higher in Snai1/2^Δhep than in Snai1/2^f/f mice after CCl4 treatment for either 72 h or 24 days (Figures S2E and S2F). Hence, impaired hepatocyte replication indeed increases compensatory hepatic progenitor proliferation and differentiation in Snai1/2^Δhep mice. Body weight was lower in Snai1/2 ^Δhep than in Snai1/2^f/f mice after chronic CCl4 treatment (Figure 3F), likely due to impaired liver regeneration.

Figure 3. Hepatocyte-specific ablation of Snai1 and Snai2 suppresses liver regeneration in mice with CCl4-induced chronic liver injury.

(A and B) C57BL/6J male mice (8 weeks) were treated with CCl4 (0.6 μL/g body weight, twice a week) or vehicle for 24 days. Liver nuclear extracts were immunoblotted with the indicated antibodies. Snai1 and Snai2 levels were normalized to Lamin A/C levels. a.u., arbitrary units. Basal: n = 6; CCl4: n = 7.

(C–F) Snai1/2^f/f and Snai1/2^Δhep male mice (8 weeks) were treated with CCl4 for 24 days.

(C and D) Liver sections were stained with the indicated antibodies (CCl4 for 24 days). Positive cells were counted. Binucleate hepatocyte number was normalized to total liver cells (H&E staining). Scale bar: 200 mμm. Snai1/2^f/f: n = 9, Snai1/2^Δhep: n = 7.

(E) Representative livers and liver weight (CCl4 for 24 days). Snai1/2^f/f: n = 9, Snai1/2^Δhep: n = 7.

(F) Body weight after CCl4 treatment. Snai1/2^f/f: n = 9, Snai1/2^Δhep: n = 7.

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, two-tailed unpaired Student’s t test (B, D, and E) or two-way ANOVA (F).

To test if Snai1 and Snai2 support liver repair in alcohol-associated liver disease, we placed Snai1/2^Δhep and Snai1/2^f/f male littermates on a Lieber-DeCarli diet supplemented with 5% ethanol for 8 weeks. The Lieber-DeCarli diet is widely used to model alcohol-associated liver disease in mice.^26,27 Like chronic CCl4 treatment, ethanol feeding also decreased liver Snai2 expression (Figures S3A). Body weight and plasma ALT levels were comparable between Snai1/2^Δhep and Snai1/2^f/f mice after alcohol consumption (Figures S3B and S3C). Unexpectedly, liver Ki67⁺ and BrdU⁺ cells (normalized to total liver cells) were similar between Snai1/2^Δhep and Snai1/2^f/f mice (Figures S3D and S3E). We then analyzed hepatocyte populations. Proliferating hepatocytes (percentage of total HNF4α⁺ hepatocytes), which were identified as Ki67⁺HNF4α⁺ cells and BrdU⁺HNF4α⁺ cells, were significantly lower in Snai1/2^Δhep than in Snai1/2^f/f mice (Figure S3E). Collectively, these results uncover a Snai1/Snai2/hepatocyte proliferation axis that drives liver regeneration in response to liver injury.

Deletion of hepatic Snai1 or Snai2 alone is insufficient to impair liver regeneration

We asked whether deletion of Snai1 or Snai2 alone is sufficient to impair liver regeneration. Snai1^Δhep and Snai2^Δhep mice were generated by crossing Snai1^f/f mice and Snai2^f/f mice with albumin-Cre mice, respectively. We previously confirmed hepatocyte-restricted deletion of Snai1 (Snai1^Δhep) and Snai2 (Snai2^Δhep).^12,13 Snai1^Δhep, Snai2^Δhep, and wild-type male mice (8 weeks old) were treated with PH for 7 days or with CCl4 for 72 h and 24 days. Liver weight was comparable between Snai1^Δhep and Snai1^f/f mice and between Snai2^Δhep and Snai2^f/f mice after PH (Figure S4A). Liver cell proliferation was comparable between Snai2^Δhep and Snai2^f/f mice after CCl4 treatment for 72 h (Figure S4B). Liver weight was also indistinguishable between Snai1^Δhep and Snai1^f/f mice and between Snai2^Δhep and Snai2^f/f mice after CCl4 treatment for 24 days (Figure S4C). Liver Ki67⁺ and BrdU⁺ cells were similar between Snai1^f/f and Snai1^Δhep mice and between Snai2^f/f and Snai2^Δhep mice after CCl4 treatment for 24 days (Figures S4D and S4E). Liver expression of cyclins A2, D1, D2, and E1 and CDK4 was comparable between Snai1^f/f and Snai1^Δhep mice and between Snai2^f/f and Snai2^Δhep mice (Figure S4F). These results support the notion that hepatic Snai1 and Snai2 act redundantly or synergically to promote liver regeneration.

Overexpression of Snai2 augments liver regeneration

Considering that hepatic Snai2 is downregulated by CCl4 treatment and PH, we reasoned that restoration of hepatic Snai2 might rescue impairment in liver regeneration. C57BL/6J male mice (8 weeks old) were transduced with adeno-associated virus (AAV)-CAG-Snai2 or AAV-CAG-GFP vectors (control). Two weeks later, the mice were treated with CCl4 for 72 h or 24 days. AAV-CAG-Snai2 transduction increased Snai2 levels in the liver but not the muscle, lung, heart, white adipose tissues, or kidney (Figures 4A and S5A). Hepatocyte proliferation was examined by immunostaining liver sections. Liver Ki67⁺, BrdU⁺, and pH3⁺ cells (normalized to total liver cells) were significantly higher in AAV-CAG-Snai2-than in AAV-CAG-GFP-transduced mice after CCl4 treatment for 72 h (Figures 4B and 4C). Liver weight (normalized to body weight) and BrdU⁺HNF4α⁺ (proliferating) hepatocytes (percentage of total HNF4α⁺ hepatocytes) were significantly higher in the AAV-CAG-Snai2 group relative to the AAV-CAG-GFP group (Figures 4C and 4D). Likewise, liver weight and hepatocyte proliferation rates were also significantly higher in the AAV-CAG-Snai2 group than in the AAV-CAG-GFP group after CCl4 treatment for 24 days (Figure 4E). Snai2-overexpressing mice gained more body weight (Figure S5C), supporting the notion that Snai1/Snai2-enhanced liver regeneration confers health benefits.

Figure 4. Snai2 overexpression increases liver growth.

C57BL/6J male mice (8 weeks old) were transduced with AAV8-CAG-Snai2 or AAV8-CAG-GFP vectors for 2 weeks and then treated with a single dose of CCl4 for 72 h (1 μL/g body weight) (A–D) or with CCl4 for 24 days (0.6 μL/g body weight, twice a week) (E). BrdU was injected 10 h before euthanization (A–D).

(A) Liver extracts were immunoblotted with antibodies to Snai2 and GAPDH.

(B and C) Liver sections were stained with the indicated antibodies. Ki67⁺, BrdU⁺, and pH3⁺ cells were normalized to total liver cells (DAPI⁺). Proliferating hepatocytes (BrdU⁺ HNF4α⁺) were normalized to total hepatocytes (HNF4α⁺). AAV-CAG-GFP: n = 6, AAV-CAG-Snai2: n = 7.

(D) Liver weight 72 h post-CCl4 treatment. AAV-CAG-GFP: n = 6, AAV-CAG-Snai2: n = 7.

(E) Liver cell proliferation and liver weight after CCl4 treatment for 24 days. AAV-CAG-GFP: n = 6, AAV-CAG-Snai2: n = 5.

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, two-tailed unpaired Student’s t test.

Snai1 and Snai2 stimulate expression of cyclins A2 and D1 in hepatocytes

We set out to identify transcriptional targets of Snai1 and Snai2. We previously performed an Affymetrix GeneChip analysis on livers from Snai2^f/f and Snai2^Δhep mice on a high-fat diet.¹² We reanalyzed the liver transcriptomes and detected several upregulated and downregulated pathways (Figure S6A). Interestingly, gene set enrichment analysis revealed that the cell-cycle pathway and the DNA replication pathway were suppressed in Snai2^Δhep mice compared to Snai2^f/f mice (Figure S6B). These results prompted us to further examine the cell-cycle machinery. Liver levels of cyclins B1/D2/E1, CDK2/4/6, p21^Cip1, p27^Kip1, Stat3, and p65 were comparable between Snai1/2^Δhep and Snai1/2^f/f mice after either PH (for 2 days) or CCl4 treatment (for 24 days) (Figures S1F and S7A). By contrast, liver cyclins A2 (encoded by Ccna2) and D1 (encoded by Ccnd1) levels—both mRNA and protein—were markedly lower in Snai1/2^Δhep males relative to Snai1/2^f/f males after PH for 2 days (Figures 5A, 5B, and S1F). Likewise, mRNA and protein levels of cyclins A2 and D1 were also significantly lower in Snai1/2^Δhep than in Snai1/2^f/f females 2 days after PH (Figures 5C and S1I). Baseline cyclin D1 and E1 levels were comparable between Snai1/2^Δhep and Snai1/2^f/f mice (Figure S7B). After CCl4 treatment for 48 h (Figures 5D, S2G, and S2H) or 24 days (Figures 5E and S7A), liver cyclin A2 and D1 levels (mRNA and protein) were also significantly lower in Snai1/2^Δhep than in Snai1/2^f/f mice. To test if overexpression of Snai1 or Snai2 directly stimulates expression of cyclin A2 and D1, we transduced primary hepatocytes with Snai1 or Snai2 adenoviral vectors. Overexpression of either Snai1 or Snai2 markedly increased mRNA and protein levels of cyclin A2 and D1 (Figures 5F and 5G). To examine Snai2 in vivo, C57BL/6J mice were transduced with AAV-CAG-Snai2 or AAV-CAG-GFP vectors and then treated with CCl4 for 72 h. Liver-restricted overexpression of Snai2 substantially increased mRNA and protein levels of cyclin A2 and D1 in the liver (Figures 5H and S5B). Thus, Snai1 and Snai2 activate Ccna2 and Ccnd1 transcriptions, thereby driving hepatic cell-cycle progression.

Figure 5. Snai1 and Snai2 directly stimulate expression of cyclins A2 and D1 in hepatocytes.

(A–C) PH was performed on littermates, and the liver was harvested 2–14 days later. Liver cyclins A2 and D1 mRNA levels were measured by qPCR (normalized to 36B4 expression). Liver extracts were prepared 2 days after PH and immunoblotted with antibodies to cyclin A2, cyclin D1, cyclin E1, and GAPDH. Cyclin levels were normalized to GAPDH levels.

(A) Day 0: Snai1/2^f/f: n = 6, Snai1/2^Δhep: n = 7; day 2: Snai1/2^f/f: n = 8, Snai1/2^Δhep: n = 6; days 7 and 14: n = 8 per group.

(B) n = 4 mice per group.

(C) mRNA levels: n = 5 mice per group; protein levels: n = 4 mice per group.

(D) Snai1/2^f/f and Snai1/2^Δhep males (8 weeks) were treated with CCl4 for 48 h (1 μL/g body weight). Liver extracts were immunoblotted with antibodies to cyclin A2, cyclin D1, cyclin E1, and GAPDH. Cyclin levels were normalized to GAPDH levels (n = 4 mice per group).

(E) Snai1/2^f/f and Snai1/2^Δhep males (8 weeks) were treated with CCl4 for 24 days (0.6 μL/g body weight, twice a week). Liver cyclin A2, cyclin D1, and cyclin E1 mRNA (normalized to 36B4, Snai1/2^f/f n = 9, Snai1/2^Δhep: n = 7) and protein (normalized to GAPDH, Snai1/2^f/f: n = 7, Snai1/2^Δhep: n = 5) levels were measured by qPCR and immunoblotting, respectively.

(F and G) Primary hepatocytes were transduced with the indicated adenoviral vectors for 48 h.

(F) Cyclin A2, cyclin D1, and cyclin E1 mRNA levels (normalized to 36B4 levels, n = 3 repeats per group).

(G) Cell extracts were immunoblotted with the indicated antibodies.

(H) C57BL/6J males (8 weeks) were transduced with AAV-CAG-Snai2 or AAV-CAG-GFP vectors for 2 weeks and then treated with CCl4 for 72 h (1 μL/g body weight). Liver cyclin A2, cyclin D1, and cyclin E1 mRNA (normalized to 36B4, n = 6 mice per group) and protein (normalized to GAPDH, n = 3 mice per group) levels were measured by qPCR and immunoblotting, respectively.

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, two-tailed unpaired Student’s t test.

Snai1 and Snai2 activate CCNA2 and CCND1 promoters through histone modifications

Mouse and human CCNA2 and CCND1 promoters contain putative Snai1/Snai2 binding motifs (CAGGTG or CACCTG) (Figure S8A). To test if Snai1 and Snai2 bind to Ccna2 and Ccnd1 promoters, primary hepatocytes were transduced with Snai1 or Snai2 adenoviral vectors and then subjected to chromatin immunoprecipitation (ChIP)-qPCR. Snai1 and Snai2 bound to both Ccna2 and Ccnd1 promoters (Figures 6A and 6B). To test if liver injury (inducing liver regeneration) recruits Snai1 and Snai2 to the promoters, mice were treated with PH for 2 days or CCl4 for 48 h. PH and CCl4 significantly increased the occupancies of Snai1 and Snai2 on Ccna2 and Ccnd1 promoters in the liver (Figures 6C and 6D). We previously reported that Snai1 and Snai2 induce histone modifications via recruiting epigenetic modifiers,^12–14 prompting us to examine Snai1/Snai2-induced histone modifications in Ccna2 and Ccnd1 promoters in the liver. Overexpression of either Snai1 or Snai2 markedly increased H3K27ac levels (an active epigenetic marker) in Ccna2 and Ccnd1 promoters in primary hepatocyte cultures (Figure 6E). In HEK293T cells, Snai2 also bound to human CCNA2 and CCND1 promoters and increased H3K27ac2 levels (Figures S8B and S8C). In contrast, Snai2 overexpression did not change H3K27 di-methylation and tri-methylation (H3K27me3), H3K9ac, or H3K9me3 in CCNA2 and CCND1 promoters (Figure S8C). To examine genomic action of Snai1 and Snai2 in vivo, C57BL/6J male mice (8 weeks old) were transduced with AAV-CAG-Snai2 or AAV-CAG-GFP vectors. Two weeks later, the mice were treated with CCl4 for 72 h. Snai2 bound to Ccna2 and Ccnd1 promoters and increased H3K27ac levels in the liver (Figure 6F). Conversely, hepatocyte-specific ablation of Snai1 and Snai2 markedly decreased liver H3K27ac levels in Ccna2 and Ccnd1 promoters in Snai1/2^Δhep mice compared to Snai1/2^f/f mice (Figure 6G). H3K27me3 levels were comparable between Snai1/2^Δhep and Snai1/2^f/f mice (Figure S8D).

Figure 6. Snai1 and Snai2 activate Ccna2 and Ccnd1 promoters via histone modifications.

(A and B) Primary hepatocytes were transduced with Snai1 (n = 4), Snai2 (n = 4), or β-galactosidase (β-gal) (n = 3) adenoviral vectors for 2 days.

(A) ChIP-qPCR images.

(B) Occupancies of Snai1 and Snai2 on the Ccna2 and Ccnd1 promoters.

(C and D) Wild-type male mice were treated with PH or CCl4 for 48 h. Occupancies of Snai1 and Snai2 on the Ccna2 and Ccnd1 promoters in the liver were measured using ChIP-qPCR. PH: n = 6 per group, control (Con): n = 6, CCl4: n = 5.

(E) H3K27ac levels in Ccna2 and Ccnd1 promoters (n = 4 per group).

(F) C57BL/6J males (8 weeks) were transduced with AAV8-CAG-Snai2 (n = 6) or AAV8-CAG-GFP (n = 6) vectors for 3 weeks. Snai2 occupancies and H3K27ac levels on Ccna2 and Ccnd1 promoters were measured in the liver by ChIP-qPCR.

(G) Livers were harvested from Snai1/2^f/f (n = 4) and Snai1/2^Δhep (n = 4) males (8 weeks old) and subjected to ChIP to measure H3K27ac levels in Ccna2 and Ccnd1 promoters.

(H) Snai2 was coexpressed with p300 in HEK293T cells, immunoprecipitated with anti-Snai2 (or anti-300) antibody, and immunoblotted with anti-p300 (or anti-Snai2) antibody.

(I) CBP, p300, or β-gal plasmids were cotransfected with or without Snai2 plasmids in HEK293T cells (with pGL3-Ccna2 cotransfection). CCNA2 promoter H3K27ac levels were measured by ChIP-qPCR (n = 4 per group).

(J) Ccna2 luciferase reporter plasmids (pGL3-Ccna2) or pGL3-basic plasmids (Con) were cotransfected with GFP, Snai1, Snai2, or ΔN30 into Huh 7 cells. Luciferase activity was measured 48 h later (n = 3 repeats per group).

(K) Ccna2 luciferase reporter plasmids were cotransfected into Huh 7 cells with GFP, CBP, or p300 plasmids in the presence of GFP or Snai2 plasmids. Luciferase activity was measured 48 h later (n = 3 per group).

(L) Snai1 and Snai2 recruit CBP and p300 that catalyze H3K27ac (active epigenetic mark) in cyclin A2 and D1 promoters, thereby enhancing hepatocyte proliferation and liver regeneration. Snai1/2-induced hepatocyte renewal suppresses liver fibrosis.

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, two-tailed unpaired Student’s t test (B–G) and one-way ANOVA (I–K).

Considering that p300 and related CBP bind to Snai1,²⁸ we tested if the two histone acetyltransferases mediate the genomic action of hepatic Snai1 and Snai2. We verified that Snai2 also coimmunoprecipitated with p300 (Figure 6H). Overexpression of either p300 or CBP modestly increased H3K27ac levels in the Ccna2 promoter; importantly, coexpression of Snai2 markedly enhanced the ability of p300 and CBP to increase H3K27ac levels (Figure 6I). We generated Ccna2 luciferase reporter plasmids as well as an epigenetic-defective Snai2 mutant called ΔN30, which is unable to bind to epigenetic modifiers.¹² We cotransfected Ccna2 luciferase reporter plasmids with Snai1, Snai2, or ΔN30 plasmids into HEK293T cells and confirmed expressions of Snai2 and ΔN30 (Figure S8E). Snai1 and Snai2, but not ΔN30, robustly stimulated Ccna2 promoter activity (Figure S8F). Snai1 and Snai2, but not ΔN30, also increased Ccna2 luciferase reporter activity in Huh 7 cells, a hepatocyte line (Figure 6J). Overexpression of p300 or CBP increased Ccna2 luciferase reporter activity in Huh 7 and HEK293T cells, and coexpression of Snai2 markedly increased the ability of p300 and CBP to activate the Ccna2 promoter (Figures 6K and S8G). Taken together, these results suggest that Snai1 and Snai2 recruit p300/CBP to CCNA2 and CCND1 promoters, where p300/CBP catalyze H3K27ac to increase expression of cyclins A2 and D1 (Figure 6L).

Deletion of hepatic Snai1 and Snai2 augments liver fibrosis in chronic liver injury

Liver injury also activates HSCs in addition to stimulating hepatocyte proliferation, prompting us to examine the role of Snai1 and Snai2 in liver fibrosis. Snai1/2^f/f and Snai1/2^Δhep male littermates were treated at 8 weeks of age with CCl4 for 24 days. CCl4 increased plasma ALT levels to a higher degree in Snai1/2^Δhep than in Snai1/2^f/f mice (Figure 7A). Liver cell death, as assessed by TUNEL and cleaved caspase-3, was comparable between Snai1/2^Δhep and Snai1/2^f/f mice (Figures 7B and 7C). Liver F4/80⁺ Kupffer cells/macrophages and cytokine expression were also comparable between Snai1/2^Δhep and Snai1/2^f/f mice (Figures 7B, 7C, and S7C). Similarly, liver cell death was also comparable between Snai1/2^Δhep and Snai1/2^f/f mice after acute CCl4 treatment or PH (Figures S1C and S2C). To assess liver fibrosis, liver sections were stained with Sirius red or antibody to α-smooth muscle actin (αSMA). Both Sirius red and αSMA (an HSC activation marker) areas were significantly higher in Snai1/2^Δhep than in Snai1/2^f/f mice (Figures 7B and 7C). Liver aSMA protein and mRNA levels were higher in Snai1/2^Δhep mice (Figures 7D and 7E). Liver expression of Colα1 and vimentin (fibrosis genes) was significantly higher in Snai1/2^Δhep mice (Figure 7E). In agreement, liver hydroxyproline levels (a fibrosis marker) were significantly higher in Snai1/2^Δhep than in Snai1/2^f/f mice (Figure 7F). In contrast, liver fibrosis was comparable between Snai1^Δhep and Snai1^f/f mice and between Snai2^Δhep and Snai2^f/f mice after CCl4 treatment for 24 days (Figures S4D and S4E). Thus, hepatic Snai1 and Snai2 act redundantly or synergistically to defend against liver fibrosis.

Figure 7. Hepatocyte-specific ablation of Snai1 and Snai2 enhances CCl4-induced liver fibrosis.

(A–F) Snai1/2^f/f and Snai1/2^Δhep male mice (8 weeks) were treated with CCl4 (0.6 μL/g body weight, twice a week) for 24 days.

(A) Plasma ALT levels. Snai1/2^f/f: n = 9, Snai1/2^Δhep: n = 7.

(B and C) Liver sections were stained with the indicated reagents, and positive cells were counted (CCl4 for 24 days). Scale bar: 200 μm. Snai1/2^f/f:n = 9, Snai1/2^Δhep: n = 7.

(D) Liver extracts were immunoblotted with the indicated antibodies.

(E) Liver mRNA abundance (normalized to 36B4 expression (CCl4 for 24 days). Snai1/2^f/f: n = 9, Snai1/2^Δhep: n = 7.

(F) Liver hydroxyproline levels (normalized to liver weight). Snai1/2^f/f: n = 9, Snai1/2^Δhep: n = 7.

(G and H) C57BL/6J male mice were transduced with AAV8-CAG-Snai2 or AAV8-CAG-GFP vectors for 2 weeks and then treated with CCl4 for 21 days (0.6 μL/g body weight, twice a week). Liver sections were stained with anti-αSMA antibody or Sirius red. Signal intensity was normalized to view areas. Scale bar: 500 μm. AAV-CAG-Snai2: n = 6, AAV-CAG-GFP: n = 8.

Data are presented as mean ± SEM. *p < 0.05, **p < 0.01, and ***p < 0.001, two-tailed unpaired Student’s t test.

Given that hepatic Snai2 is downregulated in chronic liver injury, we reasoned that restoration of liver Snai2 might rescue liver fibrosis. C57BL/6J males were transduced with AAV-CAG-Snai2 or AAV-CAG-GFP vectors and then treated with CCl4 for 21 days. Snai2 overexpression considerably suppressed HSC activation and liver fibrosis (Figures 7G and 7H). To gain insight into the underlying mechanism, we measured expression of cytokines and growth factors known to activate HSCs. Liver expression of osteopontin (Opn), Pdgfβ, Bmp8b, and Il-11 was lower in AAV-CAG-Snai2 than in AAV-CAG-GFP mice after CCl4 treatment for 21 days (Figure S9A). Conversely, liver expression of these ligands was substantially higher in Snai1/2^Δhep than in Snai1/2^f/f mice after CCl4 treatment for 24 days (Figure S9B). Notably, liver Opn, Pdgfβ, Bmp8b, and Il-11 expressions were comparable between Snai1/2^Δhep and Snai1/2^f/f mice after a short CCl4 treatment for 72 h (Figure S9C), and HSC activation and liver fibrosis were also indistinguishable between these two groups (Figure S9D). These results suggest that hepatic Snai1 and Snai2 act redundantly or synergistically to suppress HSC activation and liver fibrosis, likely by a paracrine mechanism.

DISCUSSION

We here identify hepatic Snai1 and Snai2 as important transcriptional regulators for liver regeneration and liver fibrosis. Deletion of hepatic both Snai1 and Snai2, but not one alone, suppressed liver regeneration while augmenting liver fibrosis. Using multiple proliferation markers (Ki67, BrdU, pH3), we provided strong evidence that reparative hepatocyte proliferation and liver regeneration were markedly impaired in Snai1/2^Δhep mice but not in Snai1^Δhep and Snai2^Δhep mice. These findings suggest that hepatic Snai1 and Snai2 act redundantly or synergistically to support liver regeneration. Of note, small interfering RNA (siRNA)-based knockout of Snai1 was reported to enhance liver regeneration.²⁹ The reason for the discrepancy between the two studies is currently unclear. We noticed that in the other report, a large volume of an siRNA solution was rapidly injected into the liver, known as a hydrodynamic delivery.²⁹ Hydrodynamic injection is expected to induce liver injury. Liver phenotypes may be influenced by different knockdown approaches and associated conditions. To increase scientific rigor, we examined multiple distinct liver injury models, including PH, acute and chronic CCl4 treatments, and alcohol consumption. In all models examined, reparative hepatocyte proliferation was impaired in Snai1/2^Δhep mice compared to Snai1/2^f/f littermates. Liver weight and body weight were also lower in Snai1/2^Δhep than in Snai1/2^f/f mice after PH or CCl4 treatment. Conversely, liver-specific overexpression of Snai2 increased hepatocyte replication, liver weight, and body weight after CCl4 treatment. These observations suggest that Snai1/Snai2-promoted liver regeneration critically supports liver homeostasis and liver function.

Cyclins A2 and D1 critically regulate liver regeneration.^30,31 Human and mouse CCNA2 and CCND1 promoters contain multiple putative Snai1/2 binding motifs. Using ChIP-qPCR, we demonstrated that Snai1 and Snai2 directly bound to Ccna2 and Ccnd1 promoters in hepatocyte culture as well as in the liver. PH or CCl4 treatment markedly increased the occupancies of Snai1 and Snai2 on Ccna2 and Ccnd1 promoters in the liver. Snai1 and Snai2 directly stimulated Ccna2 promoter luciferase activity. Overexpression of Snai2 also increased expression of endogenous cyclins A2 and D1 both in hepatocyte culture and in the liver. Conversely, ablation of hepatic Snai1 and Snai2 decreased liver expression of cyclins A2 and D1. These data suggest that hepatic Snai1 and Snai2 support liver regeneration at least in part by activating expression of cyclins A2 and D1. An epigenetic-defective Snai2 mutant, which was unable to bind to epigenetic modifiers, was also unable to stimulate Ccna2 promoter luciferase activity. These results suggest that Snai2 increases cyclin A2/D1 expression and liver regeneration by an epigenetic mechanism (e.g., histone-modification-based chromatin editing). In line with this notion, overexpression of Snai2 substantially increased H3K27ac levels in Ccna2 and Ccnd1 promoters in hepatocyte culture as well as in the liver. Conversely, ablation of hepatic Snai1 and Snai2 decreased H3K27ac levels in Ccna2 and Ccnd1 promoters in the liver. Snai1 and Snai2 bound to p300, a histone acetyltransferase. Overexpression of p300 or related CBP substantially increased the ability of both Snai1 and Snai2 to increase H3K27ac levels in the Ccna2 promoter and to stimulate Ccna2 promoter luciferase activity. Hence, we propose a Snai1/Snai2-based epigenetic liver regeneration model (Figure 6L). Snai1 and Snai2 recruit p300/CBP to Ccna2 and Ccnd1 promoters, where p300/CBP catalyze H3K27ac to activate expression of cyclins A2 and D1. The cyclins in turn drive the hepatic cell cycle, thereby promoting hepatocyte replication and liver regeneration.

Snai1/2^Δhep mice developed more severe liver fibrosis compared to Snai1/2^f/f littermates after chronic CCl4 treatment. Consistently, HSCs were activated to a higher level in Snai1/2^Δhep than in Snai1/2^f/f mice. In contrast, deletion of either hepatic Snai1 or Snai2 alone was unable to increase liver fibrosis. These observations suggest that hepatic Snai1 and Snai2 act redundantly or synergistically to defend against HSC activation and liver fibrosis. In line with this notion, liver-restricted overexpression of Snai2 inhibited HSC activation and liver fibrosis after chronic CCl4 treatment. Ablation of hepatic Snai1 and Snai2 increased liver expression of Opn, Pdgfβ, Bmp8b, Il-11, and Ctgf in mice with chronic CCl4 treatment. Conversely, overexpression of Snai2 in the liver had the opposite effects. These profibrotic ligands may link hepatic Snai1 and Snai2 to HSC activation and liver fibrosis.

We found that K19⁺ cholangiocyte/progenitor proliferation was higher in Snai1/2^Δhep than in Snai1/2^f/f mice. These results suggest that hepatocyte replication dysfunctions, owing to deficiency of hepatic Snai1 and Snai2, stimulate proliferation and differentiation of K19⁺ cholangiocyte/progenitor cells as a compensatory liver regeneration mechanism. Of note, excessive K19⁺ cholangiocyte/progenitor proliferation and expansion, known as ductular reaction, also induce HSC activation and liver fibrosis.³² We observed that in the short-term CCl4 treatment (72 h) paradigm, hepatocyte replication was severely blunted in Snai1/2^Δhep mice, while expression of profibrotic ligands, HSC activation, and liver fibrosis were indistinguishable between Snai1/2^Δhep and Snai1/2^f/f mice. These results suggest that prolonged hepatocyte replication blockage is a driving force for K19⁺ cholangiocyte/progenitor proliferation, ductular reaction, HSC activation, and liver fibrosis. Collectively, our data suggest that hepatic Snai1/Snai2-promoted reparative hepatocyte replication restrains pathogenic ductular reaction and HSC activation, thus protecting against liver fibrosis.

Limitations of the study

In albumin-Cre mice, Cre is believed to be expressed in some hepatoblasts. Thus, Snai1 and Snai2 may be deleted in a subset of cholangiocytes in Snai1/2^Δhep mice. However, K19⁺ cholangiocyte proliferation was higher in Snai1/2^Δhep mice, arguing against the role of biliary Snai1 and Snai2 in the observed phenotypes. It is unclear how hepatic Snai1 and Snai2 regulate ductular reaction. Beyond cyclins A2 and D1, Snai1 and Snai2 may regulate expression of additional targets that mediate hepatocyte replication. Investigations are warranted to study these putative targets in the future.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Dr. Liangyou Rui (ruily@umich.edu).

Materials availability

The study did not generate new reagents.

Data and code availability

• Original western blot images and microscopy data reported in this paper will be shared by the lead contact upon request. The authors declare that the data supporting the findings of this study are available within the paper and supplementary information files.

• This paper dose not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon reasonable request.

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Animals

Snai1^f/f, Snai2^f/f, and albumin-Cre mice were on the C57BL/6J background and used to generate hepatocyte-specific Snai1^Δhep and Snai2^Δhep knockout mice.^12,15 Snai1/2^f/f mice were obtained by crossing Snai1^f/f mice with Snai2^f/f mice. Snai1/2^Δhep mice (Snai1^f/f; Snai2^f/f;albumin-Cre^+/−) were generated by crossing Snai1/2^f/f mice with albumin-Cre drivers. Snai2^f/f mice were crossed with Snai1/2^Δhep mice to produce Snai2^f/f and Snai1/2^Δhep littermates for experiments. Mice were housed on a 12 h light-dark cycle at ambient temperature of ~22–24°C. They were fed ad libitum a normal chow diet (9% fat; TestDiet, St. Louis, MO).

Ethics statements

Animal research was complied with relevant ethic regulations and conducted following the protocols approved by the University of Michigan Institutional Animal Care and Use Committee (IACUC).

METHOD DETAILS

PH

Ventral midline incision was performed on mice (8–10 weeks) under anesthetization with isoflurane. The median and left lateral lobes were resected by pedicle ligations (70% PH).⁹ Mice were intraperitoneally injected with BrdU (100 μg/g body weight, ip) 10 h prior to euthanization, and livers were harvested for various assays.

CCl4 treatments

To induce acute liver injury, mice (8 weeks old) were intraperitoneally injected with a single dose of CCl4 (1 μL/g body weight), and liver were harvested 24–96 h after CCl4 injection. To induce chronic liver injury, Snai1/2^f/f and Snai1/2^Δhep mice (8 weeks old) were injected with CCl4 (0.6 μL/g body weight, twice a week) for 24 days. Blood samples were collected via tail veins. Plasma ALT was measure using an ALT reagent set (Pointe Scientific Inc., Canton, MI). Mice were intraperitoneally injected with BrdU (100 μg/g body weight, ip) 10 h prior to euthanization.

AAV-mediated overexpression of Snai2 in mouse livers

C57BL/6J male mice (8 weeks old) were injected with adenoviral vectors via tail veins (10¹¹ viral particles per mouse). Two weeks after transduction, the mice were intraperitoneally injected with CCl4 (1 μL/g body weight). Livers were harvested 72 h later. Mice were intraperitoneally injected with BrdU (100 μg/g body weight, ip) 10 h prior to euthanization. Two separated cohorts were treated with CCl4 for 21 or 24 days (twice a week) and used to assess liver regeneration and liver fibrosis, respectively.

Ethanol feeding

Mice (8 weeks old) were fed a Lieber-DeCarli diet (Bio-Serv, F1259SP, Frenchtown, NJ) supplemented with 5% (v/v) ethanol (200% proof) for 8 weeks.^26,27 In the first 5 days, the mice were acclimated to the Lieber-DeCarli liquid diet by gradually increasing alcohol concentrations by 1% daily to 5% concentrations.

Immunoblotting

Liver tissue or hepatocyte cultures were homogenized in ice-cold lysis buffer (50 mM Tris HCl, pH 7.5, 0.5% Nonidet P-40, 150 mM NaCl, 2 mM EGTA, 1 mM Na₃VO₄, 100 mM NaF, 10 mM Na₄P₂O₇, 1 mM phenylmethylsulfonyl fluoride, 10 μg/mL aprotinin, 10 μg/mL leupeptin). Tissue or cell extracts were resolved by SDS-PAGE and immunoblotted with the indicated antibodies (Table S1).

Luciferase assays

The Ccna2 promoter was isolated using PCR and mouse genomic DNAs and inserted into pGL3-basic plasmids (Ccna2 promoter luciferase reporters). PCR primer sequences were: Ccna1-forward: ATAGGTACCACTATTCGGCAAGCACAGGG and Ccna1-reverse: TTACTCGAGCCAGCCTGGACTGCAGAGTTA. Ccna2 luciferase reporter plasmids were cotransfected with Snai1 or Snai2 plasmids and β-galactosidase (β-gal) plasmids into Huh 7 (a human hepatocyte line) or HEK293T cells using polyethylenimine (Sigma-Aldrich, St. Louis, MO). Huh 7 cells were grown in DMEM supplemented with 5% fetal bovine serum (FBS), 100 units ml⁻¹ penicillin, and 100 μg mL⁻¹ streptomycin. Luciferase activity was measured 48 h later, using a kit (Progema, Madison, WI), and normalized to β-gal expression (internal control).

Chromatin immunoprecipitation (ChIP)

Liver samples or hepatocyte cultures were treated with 1% formaldehyde for 10 min. Genomic DNA was extracted and sheared to 200–500 bp fragments using a sonicator (Model Q800R, QSONICA).¹² DNA-protein complexes were immunoprecipitated with the appropriate antibodies (Table S1). Crosslink was reversed by heating at 65°C for 4 h or overnight. DNA was recovered using commercial kits or chemical purifications and used for qPCR analysis using Ccna1 and Ccnd1 primers (Table S2).

Primary hepatocyte cultures and adenoviral transduction

Primary hepatocytes were isolated from mice using type II collagenase (Worthington Biochem, Lakewood, NJ) and grown on William’s medium E (Sigma) supplemented with 2% FBS, 100 units ml⁻¹ penicillin, and 100 μg mL⁻¹ streptomycin.³³ At 2 h after plating in culture dishes, hepatocytes were incubated with adenoviral vectors in culture medium for 4 h and then, the adenoviral medium were replaced with fresh growth medium.

Immunostaining and TUNEL assays

Liver frozen sections were prepared using a Leica cryostat (Leica Biosystems Nussloch GmbH, Nussloch, Germany), fixed in 4% paraformaldehyde for 30 min, blocked for 3 h with 5% normal goat serum (Life Technologies) supplemented with 1% BSA, and incubated overnight at 4°C with the indicated antibodies (Table S1). Liver sections were stained with TUNEL reagents using an In Situ Cell Death Detection Kit (Roche Diagnostics, Indianapolis, IN, 11684817910).

Sirius red staining and hydroxyproline assays

Liver paraffin sections were stained with 0.1% Sirius-red (Sigma, 365548) and 0.1% Fast green (Sigma, F7252) (in saturated picric acid). Liver samples were homogenized in 6 N HCl, hydrolyzed at 100°C for 18 h, and centrifuged at 10000 rpm for 5 min. Supernatant was dried in speed-vacuum, dissolved in H₂O, and neutralized with 10 N NaOH. Samples were incubated in a chloramine-T solution (60 mM chloramines-T (Sigma, 857319), 20 mM citrate, 50 mM acetate, pH 6.5) for 25 min at room temperatures, and then in Ehrlich’s solution (Sigma, 038910) at 65°C for additional 20 min. Hydroxyproline content was measured using a Beckman Coulter AD 340 Plate Reader (570 nm) and normalized to liver weight.

Quantitative real time RT-PCR (qPCR)

Total RNA was extracted using TRIzol reagent (Invitrogen life technologies, Carlsbad, CA). The first-strand cDNAs were synthesized using random primers and M-MLV reverse transcriptase (Promega, Madison, WI). qPCR was performed using Radiant SYBR Green 2X Lo-ROX qPCR Kits (Alkali Scientific, Pompano Beach, FL), a StepOnePlus RT PCR Systems (Life Technologies Corporation, NY, USA), and respective primers (Table S2).

QUANTIFICATION AND STATISTICAL ANALYSIS

Data were presented as means ± SEM. Difference was analyzed by 2-tailed Student’s t test (2 groups), two-way ANOVA/Bonferroni posttest (>2 groups), and one-way ANOVA/Tukey posttest using GraphPad Prism 8. A p value less than 0.05 was considered statistically.

KEY RESOURCES TABLE.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-αSMA	Sigma	Cat#A5228; RRID: AB_262054
Mouse monoclonal anti-BrdU	Cell Signaling Technology	Cat#5292; RRID:AB_10548898
Rabbit polyclonal anti-CDK2	Proteintech Group	Cat#10122–1-AP; RRID:AB_2078556
Rabbit polyclonal anti-CDK4	Proteintech Group	Cat#11026–1-AP; RRID:AB_2078702
Mouse monoclonal anti-CDK6	Cell Signaling Technology	Cat#3136; RRID:AB_2229289
Rabbit monoclonal anti-Cleaved caspase-3	Cell Signaling Technology	Cat#9664; RRID:AB_2070042
Rabbit polyclonal anti-Cyclin A2	Proteintech Group	Cat#18202–1-AP; RRID:AB_10597084
Rabbit polyclonal anti-Cyclin B1	Santa Cruz Biotechnology	Cat#sc-752; RRID:AB_2072134
Rabbit monoclonal anti-Cyclin D1	Cell Signaling Technology	Cat#2978; RRID:AB_2259616
Rabbit monoclonal anti-Cyclin D2	Cell Signaling Technology	Cat#3741; RRID:AB_2070685
Mouse monoclonal anti-Cyclin E1	Cell Signaling Technology	Cat#4129; RRID:AB_2071200
Mouse monoclonal anti-F4/80	eBioscience	Cat#14–4801-82; RRID:AB_467558
Rabbit monoclonal anti-GAPDH	Cell Signaling Technology	Cat#5174; RRID:AB_10622025
Rabbit monoclonal anti-Glutamine synthetase	ABclonal	Cat#A19641; RRID:AB_2862711
Rabbit monoclonal anti-H3K27ac	Cell Signaling Technology	Cat#8173; RRID:AB_10949503
Rabbit monoclonal anti-H3K27me2	Cell Signaling Technology	Cat#9728; RRID:AB_1281338
Rabbit monoclonal anti-H3K27me3	Cell Signaling Technology	Cat#9733; RRID:AB_2616029
Rabbit monoclonal anti-H3K9ac	Cell Signaling Technology	Cat#9649; RRID:AB_823528
Rabbit monoclonal anti-H3K9me3	Cell Signaling Technology	Cat#13969; RRID:AB_2798355
Rabbit polyclonal anti-HNF4α	Santa Cruz Biotechnology	Cat#sc-8987; RRID:AB_2116913
Rabbit monoclonal anti-Hsp90	Cell Signaling Technology	Cat#4877; RRID:AB_2233307
Rabbit anti-Ki67	Vector lab	Cat#VP-RM04; RRID:AB_2336545
Mouse monoclonal anti-Lamin A/C (4C11)	Cell Signaling Technology	Cat#4777; RRID:AB_10545756
Mouse monoclonal anti-p21Cip1	Sigma	Cat#MAB88058; RRID:AB_2291542
Rabbit monoclonal anti-p27Kip1	Cell Signaling Technology	Cat#3688; RRID:AB_2077836
Mouse monoclonal anti-p300	Upstate Biotechnology	Cat#05–267; RRID: AB_11213111
Rabbit monoclonal anti-NF-κB p65	Cell Signaling Technology	Cat#8242; RRID:AB_10859369
Mouse monoclonal anti-PCNA	Sigma	Cat#P8825; RRID:AB_477413
Rabbit polyclonal anti-Phospho-Histone H3	Santa Cruz Biotechnology	Cat#sc-8656-R; RRID:AB_653256
Rabbit monoclonal anti-Phospho-NF-κB p65	Cell Signaling Technology	Cat#3033; RRID:AB_331284
Rabbit monoclonal anti-Phospho-Stat3	Cell Signaling Technology	Cat#9145; RRID:AB_2491009
Mouse monoclonal anti-Snai1	Cell Signaling Technology	Cat#3895S; RRID: AB_2191759
Rabbit monoclonal anti-Snai2	Cell Signaling Technology	Cat#9585; RRID:AB_2239535
Mouse monoclonal anti-Stat3	Cell Signaling Technology	Cat#9139; RRID:AB_331757
Bacterial and virus strains
Biological samples		N/A
Critical commercial assays
ALT reagent set	Pointe Scientific Inc.	N/A
Lieber-DeCarli diet	Bio-Serv	F1259SP
type II collagenase	Worthington Biochem	N/A
William’s medium E	Sigma	N/A
Ehrlich’s solution	Sigma	038910
chloramines-T	Sigma	857319
Fast green	Sigma	F7252
Sirius-red	Sigma	365548
In Situ Cell Death Detection Kit	Roche Diagnostics	11684817910
Deposited data
Experimental models: Cell lines
Huh7	University of Michigan	N/A
HEK293T	ATCC	CRL-1573
Experimental models: Organisms/strains
Mice(C57/BL6)	Jackson Lab	N/A
Oligonucleotides
Primers used for Real-time PCR reactions		See Table S2. Primer lists
Recombinant DNA
Ccna2 luciferase reporter	Ccna2-forward: ATAGGTACCACTATTCGGCAAGCACAGGG; Ccna2-reverse: TTACTCGAGCCAGCCTGGACTGCAGAGTTA	N/A
Software and algorithms
GraphPad Prism 9.2.0 software	GraphPad Software	https://www.graphpad.com/scientific-software/prism/
Photoshop software(22.3.3)	Adobe	https://www.adobe.com/products/photoshop.html
Illustrator software (26.5)	Adobe	https://www.adobe.com/products/illustrator.html

Highlights.

• Snai1 and Snai2 activate cyclin A2 and D1 genes via histone modifications

• Deletion of both hepatic Snai1 and Snai2, but not one alone, impairs liver regeneration

• Deletion of both hepatic Snai1 and Snai2, but not one alone, augments liver fibrosis

• Snai2 overexpression accelerates liver regeneration and suppresses liver fibrosis