A major goal of the current study was to assess wound healing and fibrogenic responses to acute and chronic DSS induced mucosal injury and inflammation based on changes in extent and site of ECM deposition. We compared responses to DSS in two inbred mouse strains commonly used to derive genetically manipulated mice and assessed the utility of a procollagen α1(I)-GFP reporter mouse as a readout for activation of a gene encoding a major ECM protein during DSS or TNBS induced disease. Our results provide important information relevant to use of the DSS model and a new reporter mouse to test impact of interventions or genetic manipulations on normal mucosal healing and fibrosis.
Our findings that males on both FVB-N and C57BL6 backgrounds exhibit more weight loss, more persistent hemoccult positivity and higher mortality/need for euthanasia demonstrate that males develop more serious disease or disease complications during the recovery period after DSS, especially after one cycle of DSS. Practically, such findings suggest that data in males and females need to be analyzed separately when evaluating impact of therapeutic/prophylactic interventions or genetic manipulations on disease severity, wound healing or fibrosis. However, lower amounts of DSS (2 or 2.5% of DSS) than the 3% DSS used here may improve survival in males and may be a useful strategy for investigators including males in their study. Our findings of sex-dependent differences in susceptibility to DSS are consistent with prior reports that male C3H/HeJ mice developed more severe DSS-induced colonic lesions 
and that across a number of mouse strains, females were more resistant to DSS-induced disease 
. Our findings are also relevant to emerging evidence that IBD susceptible loci including Toll-like receptor 8 reside on the X chromosome 
In FVB-N mice, the most severe colitis and fibrosis was observed after 3 days of recovery following one cycle of DSS. By 7 days of recovery, DSS treated FVB-N mice showed reduced colitis and fibrosis scores which did not differ significantly from water controls. This suggests that FVB-N mice exhibit only transient increases in ECM deposition and mucosal healing between 3 and 7 days after one DSS cycle. This pattern of disease provides evidence that evaluation of the recovery period after one DSS cycle in FVB-N mice could be particularly useful to test whether particular interventions or genetic manipulations accelerate or improve mucosal healing, which is currently considered an important therapeutic and prognostic indicator in CD 
. FVB-N mice subjected to one DSS cycle and recovery could also prove useful to explore differences in molecular or functional phenotypes of mesenchymal cells at times after DSS associated peak mucosal damage and ECM deposition versus times associated with mucosal healing and restoration of normal mucosal architecture. In FVB-N mice, two DSS cycles and a 7 day recovery period resulted in transmural increases in ECM in a subset of animals but overall did not consistently lead to more severe or sustained fibrosis. This suggests that FVB-N mice may need more DSS cycles to induce severe transmural fibrosis or may be resistant to development of fibrosis. Alternative mouse models of chronic intestinal inflammation and fibrosis exist including one or multiple cycles of intrarectal TNBS treatment 
or salmonella infection 
. Each of these models and the DSS model have strengths and limitations as models of IBD. However, the technical ease of the DSS model, and the transient fibrogenic changes and subsequent mucosal healing in FVB-N mice during recovery from one DSS cycle, provides a potentially useful and readily accessible model for initial testing of interventions or therapies that may impact these processes.
Our studies in procollagen α1(I)-GFP reporter mice aimed to test whether the reporter gene is activated during acute DSS-induced injury or more chronic inflammation and provide proof of principle as to whether levels or sites of reporter gene activation in whole colon or histologically could provide a useful readout of mesenchymal cell activation. Procollagen α1(I)-GFP mice have proved useful to investigate the cellular mechanisms of liver fibrosis 
. Increased expression and deposition of type I collagen is a major component of fibrosis in CD and intestinal stricture formation 
. Histology on procollagen α1(I)-GFP mice demonstrated that in contrast to FVB-N mice, mice on C57BL6 background did not show mucosal healing through 7 days of recovery after a single DSS cycle and disease progressively worsened to severe and transmural colitis after 2 DSS cycles. These findings are consistent with prior reports that C57BL6 mice do not show mucosal healing but develop chronic colitis after a single DSS cycle 
. Importantly, transmural colitis was associated with transmural activation of the procollagen α1(I)-GFP reporter gene, suggesting activation of mesenchymal cells to increase expression of a major component of fibrotic ECM. Strong and significant correlations between whole colon GFP intensity and histologic fibrosis scores, and between GFP mRNA and endogenous collagen α1(I) mRNA provide important evidence that procollagen α1(I)-GFP reporter activation provides a rapid and easy measure of activation of fibrogenic processes and induction of a major ECM gene, and could be potentially useful to study responses to interventions aimed at limiting fibrosis. We recognize that more studies will be required to definitively establish if procollagen α1(I)-GFP activation faithfully reflects development of fibrosis or increased expression or deposition of ECM proteins. We back-crossed the procollagen α1(I)-GFP mice onto the C57BL6 background to facilitate cross-breeding of this model with genetically manipulated mice on this same background in order to test genetic manipulations that may impact activation of this key fibrogenic gene. Confirmation that the procollagen α1(I)-GFP reporter is activated in the alternate TNBS and T cell driven model of colitis further supports the utility of the procollagen α1(I)-GFP reporter model. Histological findings that procollagen α1(I)-GFP expressing cells are located mainly in mucosa and submucosa after one DSS cycle and recovery, followed by transmural activation in cells throughout the bowel wall, including muscularis and serosa after two DSS cycles, demonstrate that the reporter will provide a useful tool to visualize and study the molecular or functional phenotypes of mucosal, submucosal and muscularis- or serosa-based mesenchymal cells that are activated to express this key fibrogenic gene.
Current views implicate myofibroblasts or modified smooth muscle cells as major mediators of transmural fibrosis in CD 
. However, it is important to note that one detailed study in tissues from patients with CD documented increases in vimentin+
fibroblasts as well as vimentin+
cells in regions of fibrosis in CD 
. Our comparison of the cellular sites of procollagen α1(I)-GFP expression and α-SMA or vimentin immunoreactive cells indicates that in response to DSS or TNBS colitis, vimentin positive cells are by far the most abundant cell types exhibiting activation of procollagen α1(I)-GFP in mucosa, submucosa, serosa and, muscularis layers. This suggests that cells with fibroblast phenotype are the primary cell type activated to express procollagen α1(I) in the DSS and TNBS models. Similar conclusions were reached by Suzuki et al 
who colocalized type I collagen with vimentin or α-SMA. After two DSS cycles, however, in both DSS and TNBS models, we did observe α-SMA+
cells with activated procollagen α1(I)-GFP reporter in the lamina propria but not submucosa or muscularis. This may suggest that mucosal rather than submucosal fibroblasts are activated to myofibroblast phenotype in both models because of injury initiated in epithelium and mucosa. Given that transmural activation of cells to myofibroblast phenotype is involved in fibrosis associated with IBD, this should be considered a limitation of the DSS and acute TNBS models in terms of relevance to CD fibrosis. However, it is also important to consider the possibility that immunofluorescence approaches may underestimate the number of myofibroblasts co-expressing reporter and α-SMA if levels of SMA are, for example, lower than in smooth muscle cells. Preliminary studies in a new dual reporter model expressing an α-SMA-RPF reporter as well as the procollagen α1(I)-GFP reporter suggest that this may be the case. In this model, dual labeled cells were observed in both submucosa and disorganized muscularis as well as mucosa. This emphasizes that characterization of mesenchymal cell types involved in injury or inflammation-induced fibrosis may require multiple approaches to definitively assess cell phenotype. Clearly more will need to be done to definitively identify the cell types expressing procollagen α1(I)-GFP reporter and their relevance to wound healing after acute injury or to transmural fibrosis associated with chronic inflammation. However, the procollagen α1(I)-GFP reporter provides a potentially useful and valuable system to apply flow cytometry based methods to compare, quantify and isolate cells showing activation of a key fibrogenic gene at different times during wound healing or fibrosis in DSS or other IBD and intestinal injury models. Fluorescence activated cell sorting of cells expressing the procollagen α1(I)-GFP reporter, combined with antibody based sorting for α-SMA or vimentin may provide a more complete or sensitive evaluation of the numbers of cells co-expressing GFP reporter and myofibroblast or fibroblast biomarkers at different stages of DSS or TNBS induced disease and fibrogenic changes. Such studies will also provide unique opportunities to define molecular and functional changes in different populations of fibrogenic cells during injury, fibrosis or therapeutic interventions. We should also emphasize that the visualization of GFP by fluorescence microscopy may complement and have benefits relative to localization of cell specific antigens on Sirius red or Masson’s Trichrome stained sections, because GFP is intracellular, while secretion and deposition of ECM in the extracellular compartment can make it difficult to definitively establish the precise cell types that are sources of ECM/collagen. The dual reporter model offers additional opportunities to compare, quantify and isolate α-SMA positive and α-SMA negative procollagen α1(I)-GFP expressing cells in DSS, TNBS or other intestinal injury and inflammation models. Our current preliminary studies provide an important first step towards more extensive characterization of this model and its value and validity for studies of the cellular and molecular basis of intestinal fibrosis.
In conclusion, the current study characterizes DSS-induced mucosal injury, and mucosal healing in FVB-N mice and DSS or TNBS-induced colitis and fibrosis in C57BL6 procollagen α1(I)-GFP reporter mice as an important step to use of these models to better understand mechanisms of mucosal healing and fibrosis. Our study demonstrates activation of the procollagen α1(I)-GFP reporter during acute mucosal injury or chronic, transmural inflammation. This provides an essential first step to further use of this model to better define cellular mediators and molecular phenotypes of mesenchymal cells, or potentially other cell types, that are activated during acute mucosal injury, or during chronic inflammation and progressive fibrosis, and to test interventions that may impact these processes.