Deletion of Smad2 in mouse liver reveals novel functions in hepatocyte growth and differentiation

Abstract

Smad family proteins Smad2 and Smad3 are activated by transforming growth factor beta (TGF-beta)/activin/nodal receptors and mediate transcriptional regulation. Although differential functional roles of Smad2 and Smad3 are apparent in mammalian development, the relative functional roles of Smad2 and Smad3 in postnatal systems remain unclear. We used Cre/loxP-mediated gene targeting for hepatocyte-specific deletion of Smad2 (S2HeKO) in adult mice and generated hepatocyte-selective Smad2/Smad3 double knockouts by intercrossing AlbCre/Smad2(f/f) (S2HeKO) and Smad3-deficient Smad3ex8/ex8 (S3KO) mice. All strains were viable and had normal adult liver. However, necrogenic CCL4-induced hepatocyte proliferation was significantly increased in S2HeKO compared to Ctrl and S3KO livers, and transplanted S2HeKO hepatocytes repopulated recipient liver at dramatically increased rates compared to Ctrl hepatocytes in vivo. Using primary hepatocytes, we found that TGF-beta-induced G1 arrest, apoptosis, and epithelial-to-mesenchymal transition in Ctrl and S2HeKO but not in S3KO hepatocytes. Interestingly, S2HeKO cells spontaneously acquired mesenchymal features characteristic of epithelial-to-mesenchymal transition (EMT). Collectively, these results demonstrate that Smad2 suppresses hepatocyte growth and dedifferentiation independent of TGF-beta signaling. Smad2 is not required for TGF-beta-stimulated apoptosis, EMT, and growth inhibition in hepatocytes.

FIG. 1.

Targeted disruption of Smad2 gene. (A) Targeting and deletion strategy of the Smad2 locus. A part of the genomic Smad2 locus containing exon2, which harbors the ATG codon, is shown. A loxP-flanked neomycin resistance cassette was introduced into the BglII site. The third loxP site was inserted into PstI site. By crossing mice containing one targeted allele (Smad2^3loxP) with CMV-cre transgenic mice we generated Cre-mediated recombinations as listed. (B) Southern blot analysis of genomic DNA isolated from total liver of Smad2^f/f (Ctrl) mice without (lane 1) or with Albumin-cre transgene (lane 2). The BamHI restriction fragment size change from 11.7 to 9.1 kb is seen in the Albcre/Smad2^f/f (Cre) mice using a 1-kb BamHI-BglII flanking probe (see panel A). (C) Western blot analysis of total protein lysates extracted from the livers of Albcre/Smad2^f/f (Cre) (lanes 1 and 2) and Smad2^f/f (Ctrl) (lanes 3 and 4) mice using anti-Smad2 antibody. β-Tubulin was used as a loading control. (D) Western blot analysis of TGF-β1-treated Smad2^f/f (Ctrl) and S2HeKO hepatocytes for 0, 1, and 4 h using anti-Smad2 antibody. (E) C-terminal phospho-Smad2 was demonstrated in TGF-β1 treated Ctrl but not in S2HeKO hepatocytes (upper panel), while a truncated phosphor-Smad2 protein was detected in S2HeKO cells only with overnight film exposure time (lower panel).

FIG. 2.

Molecular analysis of Smad2^dex2 allele. (A) Aberrant splicing in S2HeKO cells leads to a 227-amino-acid mutant protein, which consists of part of the linker region adjacent to the MH2 domain. The domain structure of Smad2 protein, exons, and sequence of the mutant transcript and putative amino acids are indicated. Spotted area indicate deleted region. (B) qRT-PCR analysis of Smad2 and Smad3 expression in Ctrl, S2HeKO, S3KO, and DKO hepatocytes. The S2 2-3 primer pair detects S2-FL, whereas the S2 10-11 primer pair can detect S2Δ1-240 transcripts in S2HeKO cells. (C) Western blot analysis of ectopic expression of Myc-Smad2FL and Myc-Smad2Δ1-240 in fibroblasts using anti-Myc, anti-Smad2, and anti-phospho-Smad2 antibodies as indicated. (D) Bar graphs demonstrate the average ± the standard deviation of luciferase activities of Smad2-dependent ARE-Lux reporter construct, cotransfected with pMyc-Smad2FL or pMyc-Smad2Δ1-240 in WT and Smad2KO fibroblasts left untreated or treated with TGF-β1 (5 ng/ml). (E) Bar graphs depict average ± the standard deviation of luciferase activities of Smad3-dependent SBE4-Luc reporter construct in untreated or TGF-β-treated fibroblasts cotransfected with pMyc-Smad2FL or pMyc-Smad2Δ1-240. (D and E) Representative triplicate results from three independent experiments are shown.

FIG. 2.

Molecular analysis of Smad2^dex2 allele. (A) Aberrant splicing in S2HeKO cells leads to a 227-amino-acid mutant protein, which consists of part of the linker region adjacent to the MH2 domain. The domain structure of Smad2 protein, exons, and sequence of the mutant transcript and putative amino acids are indicated. Spotted area indicate deleted region. (B) qRT-PCR analysis of Smad2 and Smad3 expression in Ctrl, S2HeKO, S3KO, and DKO hepatocytes. The S2 2-3 primer pair detects S2-FL, whereas the S2 10-11 primer pair can detect S2Δ1-240 transcripts in S2HeKO cells. (C) Western blot analysis of ectopic expression of Myc-Smad2FL and Myc-Smad2Δ1-240 in fibroblasts using anti-Myc, anti-Smad2, and anti-phospho-Smad2 antibodies as indicated. (D) Bar graphs demonstrate the average ± the standard deviation of luciferase activities of Smad2-dependent ARE-Lux reporter construct, cotransfected with pMyc-Smad2FL or pMyc-Smad2Δ1-240 in WT and Smad2KO fibroblasts left untreated or treated with TGF-β1 (5 ng/ml). (E) Bar graphs depict average ± the standard deviation of luciferase activities of Smad3-dependent SBE4-Luc reporter construct in untreated or TGF-β-treated fibroblasts cotransfected with pMyc-Smad2FL or pMyc-Smad2Δ1-240. (D and E) Representative triplicate results from three independent experiments are shown.

FIG. 3.

Characterization of primary hepatocytes isolated from livers of WT, S2HeKO, S3KO, and DKO mice. (A) Histogram depicts mean ± the standard deviation of cell viability of primary hepatocytes as determined by trypan blue exclusion in four independent experiments. (B) Phase-contrast microscopy indicates representative morphology of primary hepatocytes isolated from livers of four different genotypes as indicated.

FIG. 4.

Effect of Smad2 and Smad3 deficiency on EMT and wound healing. (A and B) Indirect immunofluorescence labeling of Ctrl and S2HeKO hepatocytes treated without or with TGF-β1 for 24 h using anti-E-cadherin (A) and fluorescein isothiocyanate-phalloidin (B). Solid arrow shows E-cadherin-positive adherens junctions; nonconsecutive arrows denote actin stress fibers. (C) Merged images. Nuclei were labeled by DAPI (blue) staining. (D) The motility and/or migratory behavior of primary Ctrl, S2HeKO, and S3KO hepatocytes was analyzed in an in vitro wound model. Representative phase-contrast photographs were taken 24 h after wounding. (E) Western blot analysis of epithelial (E-cadherin) and mesenchymal (Vimentin) markers in total protein lysates of TGF-β1-treated Ctrl, S2HeKO, and S3KO primary hepatocytes.

FIG. 4.

Effect of Smad2 and Smad3 deficiency on EMT and wound healing. (A and B) Indirect immunofluorescence labeling of Ctrl and S2HeKO hepatocytes treated without or with TGF-β1 for 24 h using anti-E-cadherin (A) and fluorescein isothiocyanate-phalloidin (B). Solid arrow shows E-cadherin-positive adherens junctions; nonconsecutive arrows denote actin stress fibers. (C) Merged images. Nuclei were labeled by DAPI (blue) staining. (D) The motility and/or migratory behavior of primary Ctrl, S2HeKO, and S3KO hepatocytes was analyzed in an in vitro wound model. Representative phase-contrast photographs were taken 24 h after wounding. (E) Western blot analysis of epithelial (E-cadherin) and mesenchymal (Vimentin) markers in total protein lysates of TGF-β1-treated Ctrl, S2HeKO, and S3KO primary hepatocytes.

FIG. 5.

TGF-β1-induced apoptosis was abolished by Smad3 deficiency but not by Smad2 deficiency. (A) DNA fragmentation (laddering) shown in WT, S2HeKO, and S3KO hepatocytes maintained without or with TGF-β1 for 48 and 72 h. (B) Histogram shows the mean ± the standard deviation obtained in three independent experiments of caspase 3 activity in WT, S2HeKO, and S3KO hepatocytes maintained without or with TGF-β1 (5 ng/ml).

FIG. 6.

Effect of Smad2 and Smad3 deficiency on DNA synthesis and cell cycle regulator expression in hepatocytes. (A) Histogram shows the mean ± the standard deviation of the effect of TGF-β1 concentrations, as indicated, on [³H]thymidine incorporation in Ctrl, S2HeKO, and S3KO hepatocytes after 24, 48, and 72 h. The y axis shows the relative change of [³H]thymidine incorporation in TGF-β1-treated normalized to untreated cells, respectively. (B) A line graph depicts the means ± the standard deviations of three independent experiments of baseline [³H]thymidine incorporation in untreated primary hepatocytes cultured for 24, 48, and 72 h. ✽, P < 0.05. (C) Representative Western blot analysis demonstrating cyclin D1, p21^CIP1, p15^INK4b, p27^KIP1, and c-Myc protein levels in Ctrl, S2HeKO, and S3KO hepatocytes treated with TGF-β1 (5 ng/ml) for 0, 8, and 24 h. β-Tubulin was used as a loading control.

FIG. 7.

Proliferation of transplanted Ctrl and S2HeKO hepatocytes in DPPIV^−/−/Rag2^−/− mice after liver injury by CCL4. (A to F) Histochemistry staining of DPPIV⁺ cells shows proliferation of transplanted Ctrl (A to C) and S2HeKO (D to F) donor hepatocytes in liver segments of recipient mice at 1, 2, and 3 months after transplantation. (G) Line graphs show the mean ± the standard deviation of the percentage of recipient liver repopulated by WT (C57BL/6J), Ctrl (Smad2^f/f; MLH1^f/f), or S2HeKO donor hepatocytes as indicated. Repopulation was significantly different in S2HeKO compared to WT or Ctrl at all three time points (1 month after transplantation, ✽, P < 0.01; 2 and 3 months after transplantation, ✽✽ and ✽✽✽, respectively, P < 0.001). (H) Number of DPPIV⁺ donor cells per repopulation cluster. The means ± the standard deviations of DPPIV⁺ cell number per donor cell cluster were significantly higher in livers transplanted with S2HeKO donor cells compared to WT and Ctrl cells at 2 months (✽✽, P < 0.001) and 3 months (✽✽✽, P < 0.001) after transplantation.

FIG. 8.

CCL4-induced liver damage in Ctrl, S2HeKO, S3KO and DKO mice. (A to D) Representative hematoxylin-eosin staining of liver sections of Ctrl (A), S2HeKO (B), S3KO (C), and DKO (D) mice 2 days after CCL4 injection. Arrows indicate centrilobar necrosis in Ctrl, S2HeKO, and S3KO livers in panels A, B, and C, respectively. The arrowhead in panel D indicates bridging necrosis in DKO livers.

FIG. 9.

Loss of Smad2 increased Ki67-positive hepatocytes in CCL4-induced liver damage in vivo. Representative Ki67 immunohistochemistry staining of liver sections of Ctrl, S2HeKO, and S3KO mice 2 days after oil (A to C) or CCL4 (D to F) treatment. Arrows denote Ki67 positive nuclei. (G) Histogram shows the mean ± the standard deviation of percentage of Ki67-positive hepatocytes in liver sections on days 1, 2, and 3 after CCL4 injection. Thirty fields were chosen randomly in liver sections per each animal in at least four Ctrl, S2HeKO, or S3KO animals, respectively (✽, P < 0.05 [Student t test]).

FIG. 10.

Effect of TGF-β on relative transcript expression of TGF-β target genes in primary Ctrl, S2HeKO, S3KO, and DKO hepatocytes. (A) Relative expression profiles of genes induced by TGF-β in Ctrl hepatocytes; (B) relative expression profiles of genes repressed by TGF-β in Ctrl cells. Values were normalized to untreated Ctrl hepatocytes and plotted as the log of fold change.