The role of mesenchymal cell populations in the development of liver fibrosis has been extensively studied in the last decades15,16,17,18,19, but oftentimes interpretation of the results is limited by the models used.

Both the Lrat- and the PDGFR promoters were shown to label HSCs in the context of liver fibrosis of different etiologies3,7. However, both the previously described LratCre and PDGFR-Cre reporter mice have a constitutively active Cre recombinase, which limits their use. This can result in unspecific activity if the promoter is only temporarily active in the course of cell differentiation, but not specific for the differentiated cell20. Furthermore, the time point of Cre activity can be a concern once a specific deletion results in cell lethality during early development.

Those concerns could be addressed by using a transgenic mouse model with inducible Cre expression. However, no such model that targets mesenchymal cell populations has been published so far. We therefore investigated whether the inducible PDGFR-P2A-CreERT2 mouse11 can be used as a reliable tool to express transgens in mesenchymal cell populations in the liver.

To address this question, we generated a triple transgenic mouse model containing the PDGFR-P2A-CreERT211, a red fluorescent tdTomato Cre reporter21, and a Col1a1-driven GFP22. Our data revealed that tamoxifen-induced activation of PDGFR-P2A-CreERT2 efficiently induced reporter gene expression in pericytes of the liver, which lasted up to one year after activation. Experiments with vehicle-treated mice showed practically no reporter expression, demonstrating no significant leakiness of the PDGFR-P2A-CreERT2 construct which has been observed for other CreER transgenic mice, such as the RipCreER, a beta-cell specific mouse line that can be used to manipulate gene expression in insulin-producing cells of the endocrine pancreas23. PDGFR-P2A-CreERT2 induced reporter expression was tested via immunostaining for markers of different liver resident cell types, in which fluorescent reporter expression only overlapped with desmin, a marker commonly used to stain pericytes in different organs, including HSCs3,24,25,26. Staining for the portal fibroblast marker Thy1.2 revealed overlap with PDGFR-P2A-CreERT2 induced reporter expression, which is in line with previous data showing that fibroblasts and VSMC express Thy1 to a certain extent5. It has been suggested that HSCs and Thy1.2 positive cells are two distinct cell populations13,27, however it cannot be excluded that also some HSCs express Thy15.

Furthermore, PDGFR-P2A-CreERT2 induced reporter expression remained specific for mesenchymal cells even under fibrogenic conditions in the CCl4 toxic liver fibrosis model. Overlap of alpha smooth muscle actin, a common myofibroblast marker, and overlap with endogenous collagen 1a1 driven GFP further confirmed that fibrogenic cells in the liver are PDGFR-P2A-CreERT2 derived. As we achieve a high recombination of PDGFR-P2A-CreERT2 in retinoid positive HSCs of over 90% and a similarly high percentage of pericyte derived myofibroblasts in three different liver fibrosis models, the inducible PDGFR-P2A-CreERT2 model can be used once an inducible Cre mouse model for liver mesenchymal cell populations is required with a similar efficiency as the constitutive PDGFRCre7 or the well accepted LratCre transgenic mouse model3.

However, PDGFRCre as a marker for fibrogenic cells in the liver has some limitations. In recent years, single cell RNA sequencing studies using a Pdgfrb-GFP transgenic mouse have revealed a spatial zonation of HSCs with central-vein-associated HSCs and portal vein-associated HSCs, whereby central-vein-associated HSCs were the dominant collagen-producing cells in CCl4 induced toxic liver injury5. Without other markers, PDGFRCre cannot distinguish between these different HSC populations with distinct functions and the different PDGFR-positive cell populations including fibroblasts, HSCs and VSMC5. Furthermore, another study identified several clusters of fibroblasts in the liver, with some of them (Fib-3 and Fib-4 clusters) expressing low levels of Pdgfrb and thus being underestimated in studies using Pdgfrb as a promoter for Cre recombinases or GFP14. We also performed immunohistochemistry for Slit2, a marker for portal fibroblasts with mesenchymal stem cell features (PMSCs)14. In contrast to this study, we observed Slit2 expression not only restricted to the portal area, but also in the liver parenchmya. Of note, PDGFR-P2A-CreERT2 driven tdTomato expression overlapped with Slit2, both in normal and fibrotic liver indicating that Slit2 might not only be a precursor for PMSCs but also for HSCs. Further studies need to address this finding.

Nevertheless, our data demonstrates that the tamoxifen inducible PDGFR-P2A-CreERT2 mouse model is similarly efficient to the established constitutive LratCre and PDGFR-Cre mouse models and can be applied once an inducible Cre recombinase is required to study liver fibrogenesis.

See more here:
An inducible model for genetic manipulation and fate-tracing of ... - Nature.com