Mice with Platelet/Megakaryocyte-Specific Deletion of TGF-β1 Are Partially Protected from Liver Fibrosis in a CCl₄-Induced Liver Injury Model

Platelets contain a high concentration of transforming growth factor β1 (TGFβ1). Platelet TGFβ1 has been shown to contribute to plasma TGFβ1 levels and resultant cardiac fibrosis and systolic dysfunction in a murine pressure overload model. TGFβ1 has also been reported to be involved in the development of liver fibrosis, but the cellular source of TGFβ1 was not known. In this study, we tested whether platelet TGFβ1 contributes to liver fibrosis by comparing three groups of mice: 1) mice with platelet-specific deletion of TGFβ1 (PF4Cre/Tgfb1^flox/flox), 2) littermate control Tgfb1^flox/flox, and 3) WT C57Bl/6 mice. Mice were injected with CCl₄ (1ml/kgBW) with peanut oil (1:100) i.p. 2 times a week for 6 weeks. CCl₄-induced liver injury was evaluated by in situ hybridization of two liver-specific genes, albumin and Cyp2e, which are expressed in periportal and pericentral areas of the liver, respectively. Liver fibrosis was quantified by slide stitching picrosirius red stained whole liver sections under polarized light. Total collagen content in liver homogenate was evaluated by hydroxyproline assay. Blood was collected before and after CCl₄ injections by retrobulbar puncture method into tubes containing 0.1 vol of 3.8% Na-citrate containing10µM PGE₁to minimize platelet activation in vitro. TGFβ1 levels in plasma were measured by ELISA.

PF4Cre/Tgfb1^flox/flox mice had > 90% lower levels of total TGFβ1 in their platelets and ~50% lower plasma TGFβ1 levels than WT or littermate control mice, (TGFβ1 levels in 10⁹ platelets/ml were 73 ± 10 ng in WT and 4.0 ± 1.0 ng in PF4CreTgfb1^flox/flox mice, p<0.0001; and in plasma 2.5 ± 0.7 ng/ml in WT and 1.3 ± 0.2 ng/ml in PF4CreTgfb1^flox/flox; p<0.001). Plasma TGFβ1 levels increased 24 hours after CCl₄ injection in WT (3 ng/ml before and 6 ng/ml after CCl₄ injection), but not in PF4Cre/Tgfb1^flox/flox mice (1.2 ng/ml before and 1.5 ng/ml after CCl₄ injection). In situ hybridization showed high expression of Cyp2e gene primarily in the pericentral vein area where damage occurred within 3 days after the first CCl₄ injection. However, albumin gene expression was normal in all non-damaged periportal areas. Liver fibrosis was ~ 20% less in PF4Cre/Tgfb1^flox/flox mice (n=11) compared to WT (n=14), or littermate control mice as quantified by both picrosirius red staining and hydroxyproline assays (picrosirius staining area fraction was 2.6 ± 0.4 % in PF4Cre/Tgfb1^flox/flox, 3.2 ± 0.6 % in WT, and 3 ± 0.4 % in littermate controls; p=0.004 between WT and PF4Cre and P=0.07 for littermate control vs. PF4Cre). There was a lower trend in total collagen content in PF4CreTgfb1^flox/flox mice compared to combined controls mice (collegen in PF4Cre/Tgfb1^flox/flox was 0.8 ± 0.1 ng/ml and 1.6 ± 1 ng/ml in controls (p=0.07). Staining platelets in frozen sections of liver after 1 day of CCl₄ injection with anti-CD41 (eBioscience) showed platelet accumulation in the damaged area of the liver, so we created a transient thrombocytopenia in WT mice by injecting a single dose of anti-GP1bα (2.5 mg/kg) antibody (R300, Emfret), which caused an 80% decrease in platelet count within 24 hours, and we found a 20% reduction in fibrosis in the CCl₄ treated thrombocytopenic mice compared to isotype-matched antibody-injected mice (p=0.02). Because myofibroblast accumulation is a hallmark of tissue fibrosis, we quantified this in the damaged area by counting aSMA and vimentin double-stained cells and found reduced myofibroblast accumulation in PF4Cre/Tgfb1^flox/flox mice vs. WT or littermate control mice. Evidence of active TGFβ1 signaling was detected as transient loss of phosphoSmad2 signaling in damaged tissue after 1 day of CCl₄ challenge, but intensity increased as the liver tissue began to repair after 3 days. PhosphoSmad2 signaling intensity was 20% less in PF4Cre/Tgfb1^flox/flox mice compared to WT mice after 3 days. Liver function was evaluated by measuring plasma levels of albumin and globin and we found that albumin to globin ratio was decreased in WT mice, but not in PF4Cre/Tgfb1^flox/flox mice after 3 days of CCl₄ injection.

We conclude that platelet-derived TGFβ1 contributes at least partially to the development of liver fibrosis and the deterioration of liver function in the CCl₄-induced liver injury model. These data have important implications for understanding TGFβ1 biology in various tissue injuries and in assessing the role of TGFβ1 in murine models of human diseases.