Title: GLI2 as a marker of hedgehog‐responsive cells
Abstract: Potential conflict of interest: Nothing to report. To the Editor: The article by Guy et al.1 nicely illustrated improved nonalcoholic steatohepatitis (NASH) patient outcomes correlated with a decrease in Sonic Hedgehog (SHH+) cell numbers. Further, the number of GLI2+ SOX9+ liver progenitor cells also tended to decrease.1 We are excited by these data, which suggest that an association between a decrease in exposure to SHH ligand reduces the intensity of the ductular reaction. However, we would like to address the limitations of using only GLI2 as a marker of Hedgehog (Hh)‐responsive cells. The canonical Hh pathway signal is transduced following Hh ligand binding to its receptor, Patched1 (PTCH1), leading to the derepression and activation of Smoothened (SMO). The result is stabilization of GLI proteins (GLI1, GLI2, GLI3) into transcriptional activators, and Hh target gene expression. In vertebrates, transduction of Hh ligand signal into a GLI‐mediated transcriptional response occurs following SMO translocation into the primary cilium (Pc).2 This is pertinent to consider, as GLI‐mediated transcriptional responses can be driven independently of the Hh/PTCH1/Pc/SMO‐cascade.3 Therefore, although GLI expression is an indicator that cells may be canonically responding to Hh ligand, to definitively identify an Hh‐responsive cell in a panel of markers is required. This is what we have used within our own study.4 Using a combination of Pc, SMO, and GLI2 expression, we were able to confirm the importance of canonical Hh signals within the liver progenitor population (EpCAM+ Pc+ SMO+ GLI2+) in the thioacetamide mouse liver injury model. However, we also showed that there were a number of other GLI2+ cell populations including myofibroblasts, leukocytes, and hepatocytes in chronic liver injury that were Pc–.4 Furthermore, in Fig. 3E of Guy et al.,1 there seems to be widespread hepatocyte GLI2 staining posttreatment. This is highly suggestive that not all GLI‐mediated transcriptional responses are driven canonically in liver disease. Indeed, it is well recognized that SMO‐independent GLI transcription can occur by way of growth factors such as transforming growth factor beta (TGF&bgr);, EGF, HGF, and FGF in other systems,3 and the widespread hepatocyte staining in improved liver injury may represent the hepatocyte response to “repair/regenerative”‐orientated growth factors. Thus, we would encourage the use of a panel of phenotypic markers including Pc, GLI, and SMO before claiming to identify “canonical Hh‐responsive” cells in vivo.