SCIENCE@home

Reply to: "The many faces of Hedgehog signalling in the liver: recent progress reveals striking cellular diversity and the importance of microenvironments".

C. A. Grzelak,N. D. Sigglekow,D. Watkins,G. McCaughan

Published 2014 in Journal of Hepatology

ABSTRACT

To the Editor: We thank Matz-Soja and colleagues for their interest and critique of our article [1]. We agree with Matz-Soja et al. that generalisations regarding the contribution of Smoothened (SMO)-dependent and SMO-independent GLI-mediated signals during liver injury processes should be made with caution. We concur that more work is required to delineate how, within various liver cell niches, Hedgehog (Hh) pathway signalling components elicit canonical or ‘non-canonical’ Hh signalling responses during liver injury. The focus of our article was to highlight the assumption that GLI2 expression alone as an accurate indicator of canonical Hh pathway activity in the liver, may be an oversimplification. Data from our in vivo model suggest that canonical SMO-dependent GLI-mediated signals can occur within primary cilia (Pc) positive/GLI2 positive liver progenitor cell (LPC) populations, while in Pc negative/GLI2 positive populations GLI-mediated signals may be driven in a SMO-independent manner. Hh signalling pathway components coordinate diverse responses (both GLI-dependent and GLI-independent), categorised into five specific pathways (Table 1), recently reviewed [2,3]. Table 1 illustrates the complexity of Hh pathway signalling; a comprehensive understanding of Hh signalling processes within the liver would require that all five aspects be addressed. In our paper, we specifically studied canonical Hh (SMO-dependent GLI-mediated) signalling in vivo. Our results also point towards a role for SMO-independent GLI-mediated signalling involvement in chronic liver injury. We acknowledge that we have not incorporated type I (PTCH1) or type II (SMO-dependent GLI-independent) non-canonical Hh signals within our working model (see Table 1), which does not suggest the irrelevance of these in chronic liver injury. Indeed, Matz-Soja recently showed the importance of SMO-dependent GLI-independent signals within hepatocytes [4]. However, within our current understanding GLI is either activated via Pc/SMO, or through multiple SMOindependent mechanisms such as PI3K/AKT, S6K, RAS, TGFb/ SMAD, MEK/ERK, RTK, and EWS/FLI1 (reviewed in [3]). Currently, there is no evidence of Hh/PTCH1/SMO/GLI ‘canonical’ signalling occurring independently of the Pc. We are perplexed by a few conclusions drawn by Matz-Soja et al. from the current literature. Matz-Soja et al. wrote, ‘besides canonical (ciliaand SMO-dependent) signalling in endothelial cells, cholangiocytes, activated hepatic stellate cells (HSC) and progenitor cells, there is obviously cilia-independent but SMOdependent signalling in mature healthy hepatocytes, Kupffer cells and, most probably, quiescent HSCs.’ The conclusion that canonical Hh signalling occurs in liver endothelial cells (ECs) and activated HSCs has been drawn prematurely, as the presence of Pc (required for canonical Hh signalling) was not evaluated in the study referenced [5] or in previous studies. The presence of Pc on HSCs is debatable. Confusingly, Matz-Soja et al. inferred that activated but not quiescent HSCs express Pc. Our data from TAA-induced chronic liver injury demonstrate that vimentin positive HSCs are Pc negative, suggesting that activated HSCs do not express a Pc, at least in mice. Further, a recent publication suggested myofibroblasts disassemble Pc during activation in vitro [6]. In human liver however, a minority of HSCs can assemble Pc under certain circumstances (http://www.bowserlab.org/ primarycilia/cilialist.html) and we have observed Pc on a minority of vimentin positive HSCs in human alcoholic liver disease (unpublished data). Other electron microscopy (EM) studies have also reported <5% HSCs express a Pc in human cirrhotic injury [7]. The concern over the use of pan Hh antibodies is unwarranted, as the pan Hh antibody used detects both Sonic Hh (SHH) and Indian Hh (IHH). Ligand detection was corroborated separately with in situ Shh mRNA and IHH ligand protein detection. IHH positive ballooned hepatocytes were denoted as ‘damaged’ on pathologist recommendations, due to altered cell morphology and significantly increased AST/ALT serum levels in 20 week TAA mice, in keeping with previous studies that identify damaged hepatocytes as a source of ligand [8]. We note that Fig. 1 of the ‘Letter to the Editor’ inaccurately references our data with regard to LPC’s ‘known’ involvement

PUBLICATION RECORD

CITATION MAP

EXTRACTION MAP

CLAIMS

  • No claims are published for this paper.

CONCEPTS

  • No concepts are published for this paper.

REFERENCES

Showing 1-8 of 8 references · Page 1 of 1

CITED BY