Soluble Guanylate Cyclase Reduced the Gastrointestinal Fibrosis in Bleomycin-induced Mouse Model of Systemic Sclerosis.

Background: Systemic sclerosis (SSc) is a chronic autoimmune-mediated connective tissue disorder. Although the etiology of the disease remains undermined, SSc is characterized by brosis and proliferative vascular lesions of the skin and internal organs. SSc involves the gastrointestinal tract in more than 90% of patients. Soluble guanylate cyclase (sGC) is used to treat pulmonary artery hypertension (PAH), and has been shown to inhibit experimental skin brosis. Methods: Female C57BL/6J mice were treated with BLM or normal saline by subcutaneous implantation of osmotic minipump. These mice were sacriced on day 28 or day 42. Gastrointestinal pathologies were examined by Masson Trichrome staining. The expression of brosis-related genes in gastrointestinal tract were analyzed by real-time PCR, and the levels of collagen in the tissue was measured by Sircol collagen assay. To evaluate peristaltic movement, the small intestinal transport (ITR%) was calculated as [Dyeing distance×(Duodenum- Appendix)] -1 ×100 (%). We treated BLM-treated mice with soluble guanylate cyclase (sGC) or DMSO orally and analyzed them on day 42. Results: Histological examination revealed that brosis from lamina propria to muscularis mucosa in the esophagus was signicantly increased in BLM-treated mice, suggesting that BLM induces esophageal brosis in C57BL/6J mice. In addition, the levels of Col3a1 and CTGF were signicantly increased in BLM-treated mice. More severe brosis was observed in the mice sacriced on day 42 than the mice sacriced on day 28. The ITR% was found to be signicantly lower in BLM-treated mice, suggesting that gastrointestinal peristaltic movement was reduced in BLM-treated mice. Furthermore, we demonstrated that sGC treatment signicantly reduced brosis of esophagus and intestine in BLM-treated mice, by histological examination and Sircol collagen assay. Conclusions: These ndings We examined gene expression in the esophagus and intestine by qPCR to indicate brotic gene expression. Total RNA was isolated from the esophagus and intestine using an RNeasy Mini kit purchased from Qiagen, and complementary DNA was reverse-transcribed using a QuantiTect Reverse Transcription kit (Qiagen). PCR reaction mixtures were prepared using the QuantiTect SYBR Green PCR kit (Qiagen). The results were indicated on a PikoReal system. The following primer pairs were used: glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 5’-AACTTTGGCATTGTGGAAG-3’ (forward) and 5’-ACACATTGGGGGTAGGAACA-3’ (reverse); collagen 3a1 (COL3A1), 5’-CAAGGTCTTCCTGGTCAGCCT-3’ (forward) and 5’-TGCCACCAGGAGGAGATCCATCTC-3’ (reverse); collagen tissue growth factor (CTGF), 5’-CACTCCGGGAAATGCTCCATGTTG-3’ (forward) and 5’-GTTGGGTCTGGGCCAAATGT-3’ (reverse); interleukin-6 (IL-6), 5’TTCCATCCAGTTGCCTTCTTG-3’ (forward) and 5’-TCATTTCCACGATTTCCCAGAG-3’ (reverse); smooth muscle actin (αSMA), 5’-AGAGACTCTCTTCCAGCCATC-3’ (forward) and 5’-ACACATTGGGGGTAGGAACA-3’ (reverse); collagen 1a1 (COL1A1), 5’-TGACTGGAAGAGCGGAGAGTACT-3’ (forward) and 5’-GGTCTGACCTGTCTCCATGTTG-3’ (reverse). GAPDH was


Background
Systemic sclerosis (SSc) is a connective-tissue disease of unknown etiology.SSc is characterized by autoimmunity, microvascular impairment, chronic in ammation, and brotic changes in various organs [1].There are several mouse models of SSc which develop dermal thickening and brosis, the most obvious feature of human SSc.However, they exhibit only some aspects of the disease or develop additional abnormalities not associated with SSc in humans [2].Bleomycin (BLM) is a chemotherapeutic agent that is used in the management of some human malignancies such as lymphomas and squamous cell carcinomas.The major limitation of BLM therapy is pulmonary toxicity and skin brosis [3].On the other hand, BLM-treated mice are widely accepted as an experimental model of SSc and are mainly used for estimating skin brosis.Lee et al. recently reported that systemic delivery of BLM using osmotic minipumps caused lung brosis from a peripheral lung lesion, which is similar to the lung brosis observed in human SSc patients [4].
On the other hand, more than 90% of SSc patients develop gastrointestinal tract brosis.The brosis extends from the mouth to the anus, and the esophagus and anorectum are most frequently affected [5] [6].The esophageal brosis causes a reduced quality of life in SSc patients due to gastroesophageal re ux disease and decreased peristaltic movement.Nevertheless, only symptomatic treatment is currently available for gastrointestinal manifestations in SSc [6].In addition, there have been few reports of mouse models that reproduce the gastrointestinal lesions of SSc.[5] [7].An epithelial Fli1-de cient mouse has been reported to develop skin, lung, and esophagus brosis [8].The transgenic mouse strain TβRIIδk-b is characterized by ligand-dependent up-regulation of TGFβ signaling and has been shown to develop skin and lung brosis.This TG mouse model was previously shown to develop colonic brosis [9].However, there have been no reports about gastrointestinal brosis in BLM-treated mice.
A soluble guanylate cyclase (sGC) stimulator is known as a drug for treatment of pulmonary arterial hypertension in SSc patients.This treatment stimulates soluble guanylate cyclase, increasing cyclic guanosine monophosphate (cGMP) levels and activating protein kinase G (PKG) in the cytosol, resulting in subsequent relaxation of vascular smooth muscle cells.In addition, recent studies have shown that sGC dose-dependently inhibits the brosis of kidney, skin, liver, and intestine in several mouse models, thus playing a critical role in brotic disease [10].For example, sGC reduces skin and intestinal brosis in experimental sclerodermatous chronic graft-versus-host-disease (Scl-GvHD) [11].
In this study, we provide the rst demonstration that continuous subcutaneous administration of BLM induced gastrointestinal brosis in mice, which histologically resembled human SSc.In addition, peristaltic movement was signi cantly impaired in the mice.Furthermore, we revealed that treatment with an sGC agonist ameliorated gastrointestinal brosis in the esophagus and intestine of BLM-treated mice.

Animals.
Female C57BL/6 mice aged 8 to 9 weeks old were obtained from CLEA Japan, Inc. (Osaka, Japan).We used the mice from 9 to 10 weeks of age.These mice were housed in the animal facility of Kobe University, with a 12-h dark/light cycle at a constant temperature and were provided with food and water ad libitum.All procedures were carried out in accordance with the recommendations of the Institutional Animal Care Committee of Kobe University.

BLM administration.
BLM was dissolved in normal saline (NS).BLM or NS were administered with osmotic minipumps, as described in previous reports, with minor modi cations [4] [12] [13].The osmotic minipumps containing 200 µL of BLM (125 mg/kg) or NS were implanted subcutaneously under the loose skin on the backs of C57BL/6 mice on day 0. The pumps delivered 1.0 mg/h for 7 days.The mice were euthanized on day 28 (4 w) or on day 42 (6 w).
For the sGC treatment experiment, we removed these pumps on day 7, then administered 200 µL daily DMSO or sGC orally to mice from day 14 to day 42.
The mice were sacri ced on day 42 (6 w) and gastrointestinal lesions were harvested.

Histology and immunohistochemistry.
The esophageal and intestinal samples were xed in 4% paraformaldehyde, embedded in para n, sectioned, and stained with Masson's trichrome (MT).To evaluate the esophageal and intestinal brosis caused by BLM treatment, we measured the thickness between the top of the brosis layer and the muscularis mucosa at 40 × magni cation under a BZ-X700 uorescence microscope (Keyence, Osaka, Japan).We analyzed the full length of the esophagus and the intestine up to 5 cm from the pylorus.
Changes in body weight.
We evaluated the change in body weight of the mice from day 0 to day 28 or day 42.The change was calculated using the following formula: Body weight change (%) = [(body weight on day 28 or on day 42) -(body weight on day 0)] × 1/ (body weight on day 0) × 100 (%).

Gastrointestinal transit.
We examined the small intestinal transport rate (ITR%) as an indicator of gastrointestinal movement [14].First, the mice were fasted overnight but given free access to water.Two hundred microliters of Evans blue solution (5% w/v in NS) was then orally administered to each mouse and dye distance was evaluated.All animals were sacri ced 30 min after Evans blue solution administration, and the rate of gastrointestinal transit was calculated by dividing the distance of the Evans blue migration by the total length of the small intestine.Speci cally, the full length of intestine from the pylorus to the ileocecum and the length between the pylorus and the forefront of the transported dye (distance of dye movement) were measured.The ITR% was calculated using the following formula.

Measurement of soluble collagen content.
Sircol collagen assay (Biocolor Ltd, Belfast, UK) was used to quantify soluble collagen content of the esophagus and intestine.Brie y, we measured the weight of the esophagus and intestine, then homogenized each tissue.We mixed the homogenate with 100 mL of acid-neutralizing reagent, and 200 mL of cold isolation & concentration reagent, then added 1 mL of Sircol dye reagent, mixed and allowed to stand for 30 minutes.After centrifugation, the pellets were dissolved in 750 mL of ice-cold acid salt wash regent and 250 mL of Sircol alkali reagent and vortexed.Relative absorbance was measured at 540 nm.

Statistical analysis.
Data are presented as mean ± standard error of the mean (SEM).Differences between groups were analyzed by unpaired t test with Welch's correction, one way ANOVA and Tukey's multiple comparison test using GraphPad Prism 5 software (GraphPad Software Inc., La Jolla, CA, USA).

BLM caused esophageal and intestinal brosis.
We rst evaluated whether BLM induced gastrointestinal brosis in female C57BL/6J mice.We administered BLM (125 mg/kg) or NS by subcutaneous implantation of an osmotic mini-pump on day 0, and sacri ced the mice on day 28 (4 w) or day 42 (6 w) (Fig. 1A).BLM-treated mice exhibited decreased body weight (Fig. 1B), but no mice died as a result of the implantation itself or the drug administration.We found that BLM treatment increased the distance from the top of the lamina propria to the muscularis mucosa in the esophagus and small intestine (Fig. 1C-F).These results con rmed that BLM-treated mice had signi cantly increased esophageal and intestinal brosis compared with NS-treated mice.Additionally, we showed that more severe esophageal brosis was observed in the mice sacri ced at 6 w compared to those sacri ced at 4 w.Also, we con rmed that BLM treatment caused lung and skin brosis in mice as previously reported (data not shown).
BLM induced brotic gene expression in murine esophagus.
Next, we evaluated brotic gene expression in murine esophagus and intestine.The gene expression levels of COL3A1, CTGF and IL-6 in esophagus were signi cantly increased in BLM-treated mice compared with NS-treated mice (Fig. 2).These results indicated that BLM treatment caused esophageal brosis, as re ected in gene expression levels.
We also evaluated gene expression levels of the intestine.There was no signi cant difference in the expression levels of brotic genes between BLM-treated mice and NS-treated mice, although their expression levels tended to be higher in BLM-treated mice than in NS-treated mice (Additional le 1; Additional Fig. 1).
BLM treatment reduced the peristaltic distance of the upper gastrointestinal tract.
Most SSc patients have decreased peristaltic movement of the intestine, leading to chronic intestinal pseudo obstruction (CIPO).Consequently, we next investigated the effect of BLM on peristaltic movement in mice.We analyzed the ITR% to evaluate peristalsis in the upper gastrointestinal tract (Fig. 3A).The ITR% of mice treated with BLM for 6 w was signi cantly reduced compared to that of NS-treated mice (**P < 0.01) (Fig. 3B).This result suggested that BLM-induced gastrointestinal brosis impaired peristaltic movement of gastrointestinal involvement.Our results showed that BLM caused gastrointestinal brosis both histologically and functionally, resembling the gastrointestinal lesions of SSc patients.
sGC treatment improved esophageal and intestinal brosis.
We next examined the effect of methyl (4, 6-diamino-2-(1-(2-uorobenzyl)-1H-pyrazolo[3, 4-b] pyridine-3yl) pyrimidin-5-yl) (methyl) carbamate; Riociguat, BAY 63-2521, a stimulator of sGC, which is a drug used to treat pulmonary arterial hypertension in SSc patients.We administered DMSO or BAY 63-2521 orally to BLM-treated mice.The study design of the sGC-stimulation experiment is described in Fig. 4A.No mice died as a result of the experimental procedures.BLM treatment caused a decrease in body weight, while BAY 63-2521 had no effect (Fig. 4B).We found that histological brosis was signi cantly reduced in the esophagus of BAY 63-2521-treated mice compared to DMSO-treated mice (Fig. 4C, 4E).We next measured soluble collagen content of the esophagus and intestine by Sircol collagen assay.BLM signi cantly increased the soluble collagen content of the esophagus, and BAY 63-2521 signi cantly decreased the soluble collagen content induced by BLM (Fig. 4G).In addition, we con rmed that BAY 63-2521 treatment decreased the thickness from the top of the lamina propria to the muscularis mucosa and the soluble collagen content in the intestine of BLM-treated mice (Fig. 4D, 4F, 4H).These results demonstrated that oral sGC-stimulation therapy ameliorated the gastrointestinal brosis induced by BLM.

Discussion
In this study, we provided the rst evidence that BLM caused gastrointestinal brosis, demonstrated both histologically and functionally.The gastrointestinal lesions of SSc patients included esophageal dysmotility, lower esophageal sphincter insu ciency, gastroesophageal re ux, esophageal stricture, a reduction in motility in the intestine, wide-mouthed diverticula in the large intestine and rectal atonia in advanced cases [15].BLM-induced gastrointestinal lesions resembled several patterns of gastrointestinal lesions observed in SSc patients.In particular, we found that esophageal brosis induced by BLM in mice were similar to the esophageal lesions of SSc patients in terms of increasing thickness of the muscularis mucosa observed by MT staining.
BLM causes in ammation by activating the TGFβ and p53 pathways.These activations result in proliferation of broblasts and induce apoptosis of epithelial cells, which in turn cause brosis.Previously, BLM has been reported to cause intestinal in ammation by increasing tumor necrosis factor α (TNFα), lipopolysaccharide (LPS), and IL-1β [16].In our experiments, the gene expression level of IL-6 in the esophagus of BLM-treated mice was signi cantly increased compared with those of NS-treated mice.This in ammation is a pre-requisite for the initiation of brotic lesions [17].
In this mouse model, BLM administration caused an increase in the thickness of gastrointestinal brosis and reduced peristaltic distance.The gene expression levels of COL3A1 and CTGF in the esophagus of BLM-treated mice were signi cantly increased compared with those of NS-treated mice.Mishra et al. previously reported that esophageal brosis affected the area from the lamina propria to the muscularis mucosa of trichrome-stained para n-embedded esophagus in mice and human [18].Therefore, we measured the brotic thickening between the top of the lamina propria and the muscularis mucosa after histopathological staining.Additionally, colon brosis is an important part of gastrointestinal brosis.A dreaded complication in individuals with SSc is CIPO, which is characterized by bowel dilatation and abnormal motility, and colon brosis is an important cause of chronic intestinal failure in patients with SSc [19].Although brosis in the colon is an important feature of gastrointestinal lesions in humans with SSc, we found no evidence of colon brosis in this mouse model (data not shown).
The mice treated with BLM for 4 weeks exhibited signi cantly different patterns of esophageal brosis compared to those treated for 6 weeks.However, the severity of in ammation during the onset of brogenesis did not correlate with collagen deposition in the mouse model of intestinal brosis [17].Lee et al. previously reported that the lung brosis induced by BLM spontaneously decreased after more than 6 weeks [4].We considered that the period after in ammation might be particularly important.For these reasons, we decided to sacri ce the mice at 6 weeks after BLM administration; however, this might not have been enough time to observe colon brosis.
In previous reports, animal models of intestinal brosis were classi ed into seven categories: spontaneous, gene-targeted, chemical-, immune-, bacteria-, and radiation-induced as well as postoperative brosis [20].However, we were unable to nd a suitable mouse model for human SSc which caused brosis in the esophagus and intestine simultaneously after chronic in ammation.For example, the dextran sodium sulfate (DSS) -induced intestinal brosis model mouse is the easiest and the most reproducible protocol to induce colonic in ammation with associated brosis.However, in this mouse model, brosis is induced after acute chemical injury and no esophageal lesion has been documented [20].Another model, the TGFβ1-overexpression mouse, develops colonic brosis with obstruction.However, the intestinal brosis of this mouse is focal [20].There have not been any reports of a mouse model in which esophageal and intestinal brosis are caused simultaneously.In our mouse model, it is signi cant that the esophagus and the intestine both exhibit brotic changes at the same time.sGC induces production of cGMP and PKG, which cause vasorelaxation in vascular smooth muscle cells.The de novo synthesis of collagen type I is reduced by sGC due to the inhibition of TGFβ-induced ERK1/2 signaling in human lung broblasts [21].In previous reports, treatment with the sGC agonist Riociguat, improved the histological brosis and hydroxyproline content in intestine compared with control in Scl-GVHD mice.TGFβ plays a central role in brosis in Scl-GVHD mice [11] [22].We demonstrated that BLM induced gastrointestinal brosis in mice.TGFβ1 is known to be involved in BLM-induced organ brosis, and activation of TGFβ1 causes broblast proliferation [23].There are several common mechanisms by which cGMP elevation can elicit anti-brotic effects.First, cGMP elevation inhibits TGFβ-induced ECM production.Second, cGMP elevation inhibits TGFβ-induced broblast to myo broblast differentiation.Third, cGMP elevation inhibits TGFβ-induced cell proliferation.[10] [24].As sGC stimulation inhibited gastrointestinal brosis in our BLM-treated mice, TGFβ1 may also be involved in BLM-induced gastrointestinal lesions.In our experiments, the gene expression levels of TGFβ1 in the esophagus of BLM-treated mice were signi cantly increased compared with those of NS-treated mice.Also, the gene expression levels of TGFβ in the esophagus of sGC-stimulation treated mice tended to be lower than those of BLM-treated mice (data not shown).
On the other hand, Hemnes et al. recently reported that PKG activity was decreased by BLM exposure in the lung [25].sGC stimulation might increase PKG activation in lesions exposed to BLM, which may have been responsible for the therapeutic effect.However, we do not consider that sGC stimulation alone provided the perfect treatment.In fact, sGC stimulation did not signi cantly improve ITR% (data not shown).These may have been affected by the duration of sGC stimulant treatment.
Riociguat (BAY 63-2521) is known to improve pulmonary arterial hypertension associated with connective tissue diseases (PATENT-1 and PATENT-2) [26].However, it remains unknown whether Riociguat improves organ brosis in SSc.In the RISE-SSc trial and a pilot study, Riociguat did not show a signi cant effect on the cutaneous lesions of SSc [27] [28].There has been no report of the effect of Riociguat on gastrointestinal involvement of SSc patients partly because intestinal lesions are di cult to evaluate.Our mouse model may be a simple and usable model to study the gastrointestinal lesions of SSc.

Conclusion
This study demonstrated that BLM induced gastrointestinal brosis which was ameliorated by sGC stimulation.Our model may be a novel mouse model of gastrointestinal brosis in C57BL/6J mice corresponding to human systemic sclerosis.

Figures
Figures

Figure 4 Treatment
Figure 4