Improved Recovery from Liver Fibrosis by Crenolanib

Chronic liver diseases are associated with excessive deposition of extracellular matrix proteins. This so-called fibrosis can progress to cirrhosis and impair vital functions of the liver. We examined whether the receptor tyrosine kinase (RTK) class III inhibitor Crenolanib affects the behavior of hepatic stellate cells (HSC) involved in fibrogenesis. Rats were treated with thioacetamide (TAA) for 18 weeks to trigger fibrosis. After TAA treatment, the animals received Crenolanib for two weeks, which significantly improved recovery from liver fibrosis. Because Crenolanib predominantly inhibits the RTK platelet-derived growth factor receptor-β, impaired HSC proliferation might be responsible for this beneficial effect. Interestingly, blocking of RTK signaling by Crenolanib not only hindered HSC proliferation but also triggered their specification into hepatic endoderm. Endodermal specification was mediated by p38 mitogen-activated kinase (p38 MAPK) and c-Jun-activated kinase (JNK) signaling; however, this process remained incomplete, and the HSC accumulated lipids. JNK activation was induced by stress response-associated inositol-requiring enzyme-1α (IRE1α) in response to Crenolanib treatment, whereas β-catenin-dependent WNT signaling was able to counteract this process. In conclusion, the Crenolanib-mediated inhibition of RTK impeded HSC proliferation and triggered stress responses, initiating developmental processes in HSC that might have contributed to improved recovery from liver fibrosis in TAA-treated rats.

Supplemental Figure S4: GATA4 expression by HSC. (A) Immunofluorescence of GATA4 in tissue sections of normal rat liver. Co-staining of desmin (green) and GATA4 (red) indicated that GATA4 was already detectable in HSC in situ (arrows). In addition to HSC, GATA4 was also expressed to various degrees by hepatocytes. (B) Strong nuclear localization of GATA4 was maintained in freshly isolated HSC cultured for 1 day (scale bars in A and B: 50 μm).

Supplemental Figure S5: Evaluation of p38 MAPK/MAPKAP2 and JNK inhibitors in Crenolanib-treated HSC.
(A) The suitability of 1 μM SB203580 to suppress p38 MAPK signaling was tested by Western blot using antibodies against p38 MAPK and its downstream signaling element MAPKAPK2 in its phosphorylated form. Densitometry analysis of protein bands revealed that Crenolanib-mediated p38 MAPK phosphorylation was not significantly altered by SB203580, but MAPKAPK2 phosphorylation was significantly prevented by this inhibitor, indicating successful inhibition of p38 MAPK signaling by SB203580 in HSC (n = 3; *p < 0.05). (B) The suitability of the JNK inhibitor SP600125 (5 μM) to diminish Crenolanib-induced JNK phosphorylation in HSC was also tested by Western blot and was found to be able to reduce phosphorylated JNK significantly (n = 3; *p < 0.05). The protein γtubulin as well as total p38 MAPK and total JNK served as a loading controls and were used for normalization. Figure S6: Expression of dual-specificity phosphatases (Dusp/mitogen-activated protein kinase phosphatase/Mkp) in HSC after short-and long-term treatment with Crenolanib. (A-J) HSC were treated with 1 μM Crenolanib and analyzed with respect to Dusp expression by qPCR at indicated time points (n = 4; *p < 0.05). Short-term stimulation of HSC with Crenolanib for up to 120 min indicated that some Dusp might be involved in elevated p38 MAPK and JNK signaling as indicated by downregulation of mRNA amounts. (K) However, analysis of long-term stimulated HSC over 7 days by qPCR revealed either no alteration or an increase in Dusp expression, which could not explain the sustained kinase activation in response to Crenolanib (n = 3; *p < 0.05).

Supplemental
Supplemental Figure S7: DUSP1 protein levels in HSC after Crenolanib treatment. Since Dusp1 expression was found to be highly upregulated in HSC by Crenolanib, this phosphatase was also analyzed at protein level by Western blot. (A) DUSP1 protein levels were analyzed in HSC cultured in medium without ITS at indicated time points (n = 3). (B) The presence of ITS in the culture medium had no obvious effect on the DUSP1 levels after treatment of HSC with 0.1 μM Crenolanib for 7 days (n = 3; p < 0.05).
Supplemental Figure S8: Inhibition of FGF signaling in Crenolanib-treated HSC. To identify the mechanism responsible for Crenolanib-mediated endodermal specification of HSC, (A, C) growth factors and (B, D) receptors known to control developmental fate decisions in stem/progenitor cells were analyzed by pPCR in HSC treated with 0.1 or 1 μM Crenolanib for 7 days (n = 4-10; *p < 0.05). Hepatocyte growth factor (Hgf), Fgf7 and Fgf10 were found to be upregulated in HSC in response to Crenolanib treatment, while only the mRNA levels of Fgfr3 increased significantly at 0.1 μM Crenolanib. Elevating the Crenolanib amount to 1 μM prevented increased Fgfr3 expression.
(E) To evaluate the hypothesis that autocrine or paracrine stimulation of HSC by growth factors is involved in Crenolanib-mediated endodermal specification of HSC, the FGFR inhibitor BGJ398 was used. (F) HSC were pretreated with 1 μM BGJ398 before 1 μM Crenolanib was applied. Western blot analysis indicated no inhibition of p38 MAPK and JNK phosphorylation by BGJ398 after short-term stimulation with Crenolanib (15-60 min; n = 3; p < 0.05; significant differences are indicated by different letters). (G) This suggested that FGFR signaling had little or no effect on Crenolanib-mediated p38 MAPK and JNK activation. Other signaling pathways seem to be responsible for this process.
Supplemental Figure S9: IL1β and LPS-mediated p38 MAPK and JNK activation in HSC. IL1β and LPS are known to trigger p38 MAPK and JNK activation via IL1R and TLR4 signaling. Therefore, activated HSC were treated with 1 ng/ml IL1β or 100 ng/ml LPS, to evaluate the concept that Crenolanib might exert its effects on p38 MAPK and JNK signaling through these receptors. (A, B) IL1β and LPS were able to activate p38 MAPK and JNK in HSC but increased also ERK1/2 phosphorylation. In contrast, AKT phosphorylation remained unchanged in HSC after stimulation with IL1β and LPS (n = 3; *p < 0.05). (C) IRAK4 is a downstream factor of IL1R and TLR4 pathways. Therefore, the IRAK4 gene was deleted by CRISPR/Cas9-mediated knockout in HSC. IRAK4 protein levels were significantly reduced by this method, but HSC treated with 1 μM Crenolanib still showed p38 MAPK and JNK phosphorylation compared to the mock control transfected with Cas9 only (n = 3; p < 0.05; significant differences are indicated by different letters).