Vitamin D modulates biliary fibrosis in ABCB4-deficient mice

Purpose Impaired vitamin D receptor signaling represents an aggravating factor during liver injury, and recent studies suggest that vitamin D might exert a protective role in chronic hepatobiliary diseases. We hypothesized that vitamin D supplementation would ameliorate liver fibrosis in ATP-binding cassette transporter B4 knockout (Abcb4 −/−) mice as a preclinical model of sclerosing cholangitis. Methods Abcb4 −/− and wild-type mice were fed a regular chow diet (600 IU vitamin D/kg food) or diets with lower (100 IU/kg) and higher (2,400 IU/kg) vitamin D concentrations for 12 weeks. Serum 25-hydroxyvitamin D concentrations were measured by chemiluminescence immunoassays. Liver injury and biliary fibrosis were assessed by liver enzyme activities, histopathology and hepatic collagen contents. Hepatic mRNA expression of markers for fibrosis, vitamin D and bile acid metabolism were analyzed by quantitative PCR. Results Different vitamin D concentrations were observed depending on genotype and diet group, with Abcb4 −/− mice on the control diet showing lower vitamin D concentrations compared to wild-type mice. Abcb4 −/− animals on the low vitamin D diet demonstrated the most advanced liver fibrosis and highest hepatic collagen contents. Feeding Abcb4 −/− mice a high vitamin D diet enriched serum vitamin D levels, lowered liver enzyme activities, altered expression levels of profibrogenic genes and ameliorated, in part, liver injury. Conclusions This is the first report to demonstrate that fibrogenesis in the established Abcb4 −/− model is influenced by vitamin D supplementation. Since vitamin D modulates sclerosing cholangitis in vivo, we speculate that sufficient vitamin D intake might improve liver damage and induce antifibrotic effects in chronic cholestasis in humans. Electronic supplementary material The online version of this article (doi:10.1007/s12072-014-9548-2) contains supplementary material, which is available to authorized users.


Introduction
Vitamin D deficiency occurs in up to 90 % of patients with chronic liver diseases, irrespective of the underlying etiology of liver injury [1,2]. Moreover, in these patients, low vitamin D levels are associated with increased grades of necroinflammation and stages of fibrosis and a heightened mortality risk [3][4][5][6]. Despite these observations, and in particular the inverse correlation between vitamin D levels and the severity of liver diseases, to date few studies have investigated the beneficial and adverse effects of vitamin D supplementation in chronic liver diseases [7][8][9]. Therefore, data regarding the efficacy of vitamin D supplementation are still lacking, and preclinical models for assessing the effects of vitamin D are needed.
Vitamin D is a steroid hormone with pleiotropic effects that undergoes hydroxylation in liver and kidney [6]. The major circulating form, 25-hydroxyvitamin D, is bound to vitamin D-binding protein, whereas effects of 1a,25-dihydroxyvitamin D on target genes in many organs are mediated by a ligand-activated nuclear receptor, the vitamin D receptor (VDR) [10], which forms a heterodimer with the retinoid X receptor (RXR) to modulate processes and networks ranging from immune responses to mineral homoeostasis ( Supplementary Fig. 1). In the liver, VDR expression is restricted to non-parenchymal cells and biliary epithelial cells [11]. Besides activation of VDR by vitamin D, lithocholic acid and its derivatives have been demonstrated to function as VDR ligands and activators [12]. Furthermore, regulation of bile acid synthesis and enzymes responsible for bile acid detoxification are influenced by vitamin D-VDR-signaling [13][14][15]. These findings collectively point to a potential modulatory role of vitamin D-VDR signaling in biliary-type liver injury.

Materials and methods
Animals FVB/N-Abcb4 tm1bor mice and FVB/NJ control mice were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). The mice were housed and bred in individually ventilated cages with a 12-h light-dark cycle under incandescent lighting free from UVB radiation. Temperature and humidity were regulated at 22 ± 1°C and 55 ± 5 %, respectively. Water and food were provided ad libitum. Mice were genotyped by polymerase chain reaction (PCR) of tail DNA using neo (5 0 -CTT GGG TGG AGA GGC TAT TC-3 0 ; 5 0 -AGG TGA GAT GAC AGG AGA TC-3 0 ) and Abcb4 (5 0 -CAC TTG GAC CTG AGG CTG TG; TCA GGA CTC CGC TAT AAC GG-3) specific primer pairs. The PCR reaction included PCR buffer (Applied Biosystems, Darmstadt, Germany), 2 mM MgCl 2 , 10 lM dNTPs, 10 lM primer, 1.25 U Taq DNA polymerase (Invitrogen, Darmstadt, Germany) and 20-100 ng DNA in 25-ll reactions. PCR cycling conditions were 94°C/30 s, 55°C/60 s and 72°C/30 s for 35 cycles and a final extension step of 10 min at 72°C.
The animal experiments were performed with permission from the federal state of Saarland according to §8 of the German Law for the Protection of Animals and the Directive 2010/63/EU of the European Parliament. All institutional and national guidelines for the care and use of laboratory animals were followed.

Diets
The vitamin D diets were commenced after weaning in the 4th week of age, coinciding with the initiation of liver pathology including sclerosing cholangitis, which at this point is not yet fully developed [20]. Wild-type controls and Abcb4 -/mice of both sexes were divided into three groups, with each receiving a different diet: The control group was fed a regular chow diet containing 600 IU vitamin D (cholecalciferol)/kg food, which is based on the established nutrient requirements for laboratory mice [21]. The low vitamin D group obtained a diet with 100 IU vitamin D/kg, and the high vitamin D diet group was fed a diet enriched with 2,400 IU vitamin D/kg. All diets were given for 12 weeks. With the exception of vitamin D content, all three diets were otherwise equal in nutrient composition and total energy, and were obtained from Altromin (Lage, Germany; see Supplementary Tables 1, 2 and 3 for detailed dietary composition). In total, 113 animals were analyzed, with a minimum of 15 mice per genotype and diet group. Table 1 provides the number of animals analyzed per genotype, sex and diet group. Food intake was controlled by monitoring weekly food consumption per cage and did not differ (3-4 g/day) among the groups. Survival rates were 100 %. Control vitamin D diet (600 IU/ kg) 10 7 10 9 High vitamin D diet (2,400 IU/kg) 12 11 9 11 Phenotypic characterization of biliary fibrosis For histopathological evaluation, liver samples were preserved in 4 % neutral-buffered formaldehyde solution at 4°C and embedded in paraffin. Paraffin sections (2 lm) were stained with haematoxylin-eosin (H&E) or Sirius red for the detection of collagen. The stages of liver fibrosis and relative collagen areas were assessed by a pathologist blinded to the study protocol using a semiautomatic system for image analysis (Stingray F146C IRF Medical camera, 00 type progressive scan CCD, Germany) and HistoQuant image morphometry software (3DHistech, Budapest, Hungary). Hepatic fibrosis was staged (F-score) using a scale adapted from the Batts and Ludwig as well as Ishak scoring systems [22,23]. The F-scores are subdivided into five classes: 0 = no fibrosis; 1 = scatter periportal and perineoductular fibrosis; 2 = periportal, perineoductular fibrosis; 3 = periportal, perineoductular fibrosis with portal-portal septa; 4 = complete cirrhosis.
Relative collagen content was determined as the mean percentage of the collagen-stained area to the whole area (field of view). Therefore, we screened representative microscopic fields (magnification, 1009), which were randomly chosen from each liver section (avoiding arteries of [100 lm diameter) after setting a threshold capturing Sirius red-stained areas of collagen. In addition, collagen in liver was quantified by colorimetric measurement of the collagen-specific amino acid hydroxyproline, as described previously [24,25].

Serum biochemical assays
Blood was collected from the inferior vena cava after harvesting. Samples were left for 10 min at room temperature in darkness and centrifuged for 20 min at 2,0009g. Serum was stored at -80°C until analysis. Serum calcium concentrations, alanine aminotransferase (ALT), aspartate aminotransferase (AST) and alkaline phosphatase (AP) activities were measured in the Cobas Ò 8000 analyzer (Roche Diagnostics, Mannheim, Germany) by standardized methods following the recommendations of the International Federation of Clinical Chemistry. Serum 25-hydroxyvitamin D concentrations were determined using the chemiluminescence immunoassay LIAISON Ò 25 OH Vitamin D TOTAL Assay (DiaSorin, Stillwater, MN, USA).

Statistics
The results were analyzed using SPSS 20.0 (IBM, Ehningen, Germany). Quantitative data are presented as means ± standard errors (SE) or median and ranges, as appropriate. All data were analyzed for the effects of diet and genotype. Means were compared with Student's t tests or one-way analyses of variance (ANOVA), followed by post hoc Bonferroni correction. The medians of non-parametric data were compared with Mann-Whitney U or Kruskal-Wallis tests, respectively. Two-way ANOVA was applied to assess the interaction of diet and sex after exclusion of extreme values. Correlation coefficients were calculated according to Pearson or Spearman as appropriate. p values \ 0.05 were considered significant.

Results
Vitamin D supplementation exerts no toxic effects in Abcb4 -/mice After weaning, all mice continued to develop normally and displayed no obvious signs of developmental or behavioral abnormalities. Body weight and liver-to-body weight ratios were not affected by dietary composition, with the exception of female Abcb4 -/mice on the high vitamin D diet displaying increased body weight compared to knockout mice on the control diet (24.6 ± 0.9 g vs. 20.2 ± 0.9 g, p \ 0.01). Serum calcium levels were slightly higher in Abcb4 -/mice as compared to wild-type controls in all dietary groups, and no hypercalcemia was induced (Supplementary Fig. 2).
Vitamin D serum concentrations depend on genotype, sex and diet As shown in Fig. 1, mice displayed different serum vitamin D concentrations depending on genotype and diet group. On the low vitamin D diet, both strains displayed serum vitamin D levels \10 ng/ml. Abcb4 -/mice on the control diet showed significantly (p \ 0.01) lower vitamin D concentrations compared to FVB/NJ wild-type mice on this diet (38.4 ± 1.9 vs. 46.8 ± 1.7 ng/ml). In contrast, Abcb4 -/mice on the high vitamin D diet demonstrated significantly (p \ 0.001) increased vitamin D levels in comparison to wild-type counterparts (67.2 ± 2.4 vs. 42.0 ± 2.3 ng/ml). The vitamin D concentrations in female mice were significantly (p \ 0.001) higher than in male mice in the high vitamin D group, and a significant (p \ 0.01, two-way ANOVA) interaction between sex and diet was observed for vitamin D levels but not for other phenotypes.
Vitamin D supplementation influences hepatic fibrosis in Abcb4 -/mice Figure 2 illustrates that the quantification of hepatic collagen contents revealed a significant (p \ 0.01) increase in Abcb4 -/mice in all three diet groups as compared to controls. Hepatic collagen contents were highest in Abcb4 -/mice receiving the vitamin D-deficient diet and displayed a dose-related trend towards lower levels in knockout mice on the regular and the high vitamin D diets. In contrast, no impact of diet on hepatic collagen levels was detected in wild-type mice (Fig. 2).
Panel A of Fig. 3 shows that these histopathological observations were reflected by the relative collagen areas, as determined after Sirius red staining. Representative liver sections stained with Sirius red are shown in Fig. 3b. As expected, we observed genotype-dependent differences, with increased collagen areas in Abcb4 -/mice as compared to wild-type mice. Collagen areas were significantly (p \ 0.01) larger in Abcb4 -/mice receiving the low vitamin D diet in comparison to those on the control diet   Serum parameters for liver damage and cholestasis (ALT, AST, AP) were significantly elevated in Abcb4 -/compared to corresponding wild-type mice in all diet groups (Fig. 4). Notably, enzymatic activities of both ALT and AST in Abcb4 -/mice significantly decreased on the high vitamin D diet as compared to knockout mice fed the control diet. Wild-type mice receiving the high vitamin D diet displayed elevated ALT activities in comparison to those on the control diet (68 U/l, range 37-314 vs. 48 U/l, range 11-77; p \ 0.01). AP activities did not differ between diet groups in either knockout or wild-type mice (Fig. 4c).
Dietary vitamin D affects hepatic expression of profibrogenic genes in Abcb4 -/mice Panel A of Fig. 5 illustrates that hepatic steady-state mRNA expression levels of collagen type I (Col1a2) were significantly (p \ 0.001) higher in ABCB4-deficient mice as compared to wild-type controls, irrespective of the diet. Timp1 expression increased significantly (p \ 0.001) in Abcb4 -/mice fed the control or the high vitamin D diets, but not in animals fed the low vitamin D diet as compared to corresponding wild-type mice (Fig. 5b). Panel C of Fig. 5 shows that a similar trend was observed for transforming growth factor (Tgf)-b1 expression, with significantly (p \ 0.05) reduced levels in Abcb4 -/mice on the low vitamin D diet as compared to the control diet. Tgf-b1 expression correlated significantly (q = 0.92, p \ 0.01) with expression of Vdr in Abcb4 -/mice.

Diet and genotype modify the hepatic expression of Vdr and vitamin D hydroxylases
The hepatic gene expression of Vdr was increased in ABCB4deficient mice compared with the corresponding wild-type controls in all diet groups. Similar to Tgf-b1 and Timp1 expression, vitamin D-deficient Abcb4 -/mice displayed lower expression levels of Vdr as compared to Abcb4 -/mice on the control diet (Fig. 6a). Panel B of Fig. 6 illustrates that hepatic gene expression of the antimicrobial peptide cathelicidin (Cramp) was increased in Abcb4 -/animals on control and low vitamin D diets, whereas knockout mice fed the high vitamin D diet did not differ from wild-type animals. The highest Cramp expression levels were observed in vitamin D-deficient Abcb4 -/mice.  As circulating vitamin D levels depend on the Abcb4 genotype and diet (Fig. 1), we assessed expression differences of the major vitamin D hydroxylase Cyp2r1 and Cyp27a1 in liver. The mRNA levels of both enzymes were markedly increased in Abcb4 -/mice challenged with low and high vitamin D diets (data not shown). The expression of the genes Cyp7a1 and Cyp8b1, encoding the rate-limiting enzymes for bile acid synthesis and conversion to cholic acid, respectively, tended to decrease in Abcb4 -/mice on diets with higher vitamin D contents, but no significant differences were observed (data not shown).

Vitamin D supplementation does not appear to influence hepatic adherens junction integrity
As determined by double immunofluorescence microscopy (Fig. 7), genotype-specific changes of bile duct integrity could be observed. Abcb4 -/animals from all vitamin D diet groups displayed an increase in ductular proliferation, as shown by the higher number of K19-positive cholangiocytes and overexpression of E-cadherin. At the hepatocyte level, a diffuse E-cadherin expression within the liver lobe was detected in Abcb4 -/mice (Fig. 7d-f), which contrasted with the findings in wild-type controls. However, an influence of dietary vitamin D intake on changes in the E-cadherin distribution pattern on the hepatocyte or even bile duct level was not apparent in either strain.

Discussion
In this study, we examined the modulatory effects of dietary vitamin D supplementation on biliary fibrosis in vivo in Abcb4 -/mice. The following major results were demonstrated: (1) vitamin D deficiency is associated with increased collagen accumulation and fibrogenesis in this preclinical model; (2) a vitamin D-sufficient diet ameliorates chronic liver injury; and (3) high dose vitamin D supplementation exerts a further benefit only by reduction of aminotransferase activities. Taken together, these findings suggest that low vitamin D levels aggravate liver fibrosis, but supranormal vitamin D levels have no additive antiinflammatory or -fibrotic effects in Abcb4 -/mice. This may indicate that vitamin D deficiency should be avoided to decrease the risk of progressive fibrosis in chronic cholangitis and that an adequate vitamin D intake has, at least to some extent, antiinflammatory and antifibrotic effects. These observations are relevant for future translational studies, because chronic cholestatic liver injury in humans is frequently accompanied by vitamin D deficiency [7,28]. More specifically, Abcb4 -/mice receiving the low vitamin D diet developed more advanced stages of fibrosis as compared to mice consuming diets with higher vitamin D contents. Of note, maximum vitamin D supplementation did not further improve the fibrosis scores in our model, but in comparison to the control group it ameliorated hepatocellular damage, as illustrated by lower serum aminotransferase activities. These observations are also reflected by enhanced hepatic collagen accumulation in Abcb4 -/mice with vitamin D deficiency, whereas the lowest collagen contents were observed in mice receiving the high vitamin D diet. The mRNA expression analyses did not identify differences in collagen expression, pointing to vitamin D-dependent post-transcriptional regulatory mechanisms. On the other hand, in vitro studies and a rat model of toxin-induced hepatic fibrosis have demonstrated possible antifibrotic effects of vitamin D supplementation through diminished hepatic stellate cell (myofibroblast) activation and collagen expression [29][30][31]. These different observations might be due to the diverse experimental setups, in particular homogeneous cell cultures versus heterogeneous cell populations from total liver, given that in cholestatic disorders portal fibroblasts are predominantly responsible for hepatic collagen deposition [32]. Despite identical dietary intake, Abcb4 -/mice demonstrated lower vitamin D serum concentrations on control chow diet as compared to wild-type mice. Serum vitamin D levels increased markedly in Abcb4 -/mice receiving the high vitamin D diet, whereas it remained constant in wildtype mice. These findings allude to a potential dysregulation of intestinal vitamin D uptake. Additionally, we observed higher expression of the vitamin D activating enzymes Cyp2r1 and Cyp27a1 in livers of ABCB4-deficient mice on the high vitamin D diet. Vitamin D levels are regulated through degradation via 24-hydroxylase, and vitamin D induces the expression of this enzyme, thus controlling its own catabolism [33,34]. Of note, bile acids such as lithocholic acid have been shown to decrease this stimulatory effect of vitamin D and influence a broad spectrum of signaling pathways via the nuclear receptors VDR, FXR and PXR [35][36][37][38]. In fact, minor changes in bile salt composition have been shown to affect hepatic fibrogenesis in ABCB4-deficient mice [39], although the expression of the genes encoding the rate-limiting enzymes for bile salt synthesis (Cyp7a1) and conversion to the hydrophobic cholic acid (Cyp8b1) showed no major differences across the groups. Further studies, however, would be required to elucidate the molecular mechanisms underlying the genotype-diet interactions.
The expression levels of the fibrosis markers Col1a2, Timp1 and Tgf-b1 did not strictly correlate with the severity of liver injury and fibrosis. Some of these discrepancies could be related to the fact that we measured mRNA levels in total liver, although vitamin D might exert specific effects on different hepatic cell populations. Altered vitamin D-VDR signaling, which has been reported in cholestatic conditions, was corroborated in this study, as illustrated by increased Vdr expression in ABCB4-deficient as compared to wild-type mice. Furthermore, Vdr expression was reduced in Abcb4 -/mice on the low vitamin D diet, and these animals displayed the highest fibrosis scores. Recently, elegant mechanistic studies have demonstrated that the activation of VDR antagonizes TGF-binduced recruitment of SMAD3 via co-occupation of regulatory sites in key profibrogenic genes (including COL1A2, TIMP1 and TGFB1) in the presence of TGF-b [31]. In this context it is interesting to note that in comparison to Abcb4 -/mice on control diet, our data show lower Tgf-b1 expression levels both in vitamin D deficiency and upon excess dietary supplementation, illustrating the interaction of vitamin D and TGF-b signaling in the setting of biliary inflammation and fibrosis.
Given the potential antiinflammatory effects of vitamin D, we also assessed the hepatic expression of the antimicrobial peptide cathelicidin (Cramp) [40]. Since the vitamin D response element is absent in the murine Cramp promoter (in contrast to humans) [41], we conclude that the decrease of Cramp expression upon vitamin D supplementation to levels observed in healthy controls is more likely to reflect the amelioration of hepatic inflammation than direct effects on Cramp. Further investigations into the regulation of antimicrobial peptides in sclerosing cholangitis would be interesting but were beyond the scope of this study.
To protect parenchymal liver cells from toxic bile compounds, epithelial cells form a physical barrier by apical junctional complexes (tight junction and adherens junction). Distinct functional and morphological alterations of these junctional complexes have been described in cholestasis in both mice and humans [19,42]. Since vitamin D has been shown to be involved in the regulation of apical junctional complexes in other organs [43][44][45], Firrincieli et al. [46] recently investigated the role of VDR in the maintenance of bile duct integrity in mice with biliary-type injury. Their study demonstrated altered E-cadherin staining and loss of cell adhesion in biliary epithelial cells in Vdr -/mice subjected to bile duct ligation [46]. Therefore, we also assessed hepatic adherens integrity in our model, which revealed genotype-specific differences in biliary epithelial cell proliferation and E-cadherin staining. However, vitamin D-dependent effects could not be substantiated. Nevertheless, this does not exclude the possibility that truncated E-cadherin, which was shown to be present in higher amounts in Vdr -/mice [46], might be relevant.
In conclusion, our findings indicate that vitamin D modulates biliary injury and fibrogenesis in vivo and that the severity of biliary fibrosis in Abcb4 -/mice is influenced by vitamin D status. In this model, vitamin D deficiency appears to aggravate liver fibrosis; however, the beneficial effects of vitamin D supplementation are limited. We speculate that excess vitamin D supplementation does not fully protect against liver fibrosis, but an adequate vitamin D intake abates hepatic injury and confers antifibrotic effects in patients with cholestatic liver diseases.