Overexpression of TNFα in TNF^∆ARE+/− mice increases hepatic periportal inflammation and alters bile acid signaling in mice

Abstract

Background:

Intestinal inflammation is a common factor in ~70% of patients diagnosed with primary sclerosing cholangitis. The TNF^∆ARE+/− mouse overexpresses TNFα and spontaneously develops ileitis after weaning. The aim of this study was to examine the influence of ileitis and TNFα overexpression on hepatic injury, fibrosis, inflammation, and bile acid homeostasis.

Methods:

Using serum, hepatic, and ileal tissue isolated from 24- to 26-week-old C57BL/6 and TNF^∆ARE+/− mice, hepatic injury and fibrosis, inflammation, ductal proliferation, and regulation of bile acid synthesis were assessed by immunohistochemical and quantitative PCR methods.

Results:

Compared to age-matched C57BL/6 mice, TNF^∆ARE+/− mice exhibited increased serum AST, ALT, and serum bile acids, which corresponded to increased hepatic picrosirius red staining, and an increase in hepatic mRNA expression of Tgfb, Timp1, Col1a1, and MMP9 supporting induction of fibrosis. Examining inflammation, immunohistochemical staining revealed a significant periportal increase in MPO+ neutrophils, CD3+ lymphocytes, and a panlobular increase in F4/80+ macrophages. Importantly, periportal inflammation corresponded to significantly increased proinflammatory chemokines as well as hepatic cytokeratin 7 staining supporting increased ductular proliferation. In the liver, increased mRNA expression of bile acid transporters was associated with suppression of classical but not alternative bile acid synthesis. In the ileum, increased inflammation correlated with suppression of Nr1h4 and increased Fgf15 and Nr0b2 mRNA expression.

Conclusions:

Increased TNFα expression is sufficient to promote both intestinal and hepatobiliary inflammation and fibrotic injury and contributes to hepatic dysregulation of FXR signaling and bile acid homeostasis. Overall, these results suggest that the TNF^∆ARE+/− mouse may be a useful model for studying chronic hepatic inflammation.

INTRODUCTION

The cholestatic liver disease accounts for ~9% of adult and 43% of pediatric liver transplants in the United States.¹ Cholangiopathies include primary sclerosing cholangitis (PSC), primary biliary cholangitis (PBC), autoimmune cholangitis, and pediatric biliary atresia and are characterized by extensive biliary inflammation,² a strong periportal ductular reaction leading to the development of fibrosis and obstruction of the biliary system. To date, a paucity of therapies have been approved for cholangiopathy treatment; only ursodeoxycholic acid (effective in ~40% of patients with PBC) and obeticholic acid (Ocaliva-approved but with a black box warning for PBC) are therapeutic options. Thus, liver transplantation is required for progressive, nonresponsive cholangiopathies.^3,4 Even following transplantation, patients with PSC and PBC frequently recur (20%–40% PSC and 29% PBC), supporting an urgent unmet need for effective medical therapies to reverse or cure inflammatory cholangiopathies.⁵

PSC is frequently associated with autoimmune disorders, such as inflammatory bowel disease (70%), and overlaps with autoimmune hepatitis.⁶ Unfortunately, the initial insult and/or causal factors underlying most cases of cholangiopathies have not been elucidated and are likely multifactorial, with genetic predisposition, environmental exposures, and immune dysregulation as key contributors. Combined with the resiliency of the liver to insult, this often results in diagnosis at later stages of disease progression when cirrhosis is already present, making the determination of factors that will mitigate disease progression even more important. With limited therapeutic options, an improved mechanistic understanding of these diseases is urgently needed to develop new strategies that ameliorate liver injury after diagnosis.

By its ability to trigger the activation of proinflammatory pathways, the proinflammatory cytokine TNFα plays a central role in the pathogenesis of chronic liver diseases. Cholangiopathies are no exception; in human PSC, TNFα expression is increased in whole liver tissue as well as primary biliary epithelial cells.⁷ Furthermore, increased TNFα is also evident in the serum isolated from patients with AIH, biliary atresia, and PBC.^8–10

Although TNFα signaling has been demonstrated in chronic cholangiopathies, the effects of chronic upregulation of TNFα on the liver and bile ducts, and more specifically, bile acid homeostasis, which is especially relevant in cholangiopathies, are not well understood. In murine models, recent reports have shown that the TNFα transgenic mouse developed hepatic fibrosis and increased hepatic mRNA expression of inflammatory cytokines at 5.5 months.¹¹ The TNF^∆ARE mouse model of TNFα overexpression is a well-characterized model of ileitis. TNF^∆ARE mice possess a deletion of AU-rich elements (ARE) in the 3′ untranslated region of the TNFα mRNA transcript that results in greater stability of TNFα mRNA and, therefore, increased production of TNFα. TNF^∆ARE+/− mice develop ileitis, which is similar to Crohn’s in patients with PSC by 12 weeks of age, and ultimately develop intestinal fibrosis by 26 weeks of age.¹² Although this is an established model for inflammatory bowel disease, the effects of TNF^∆ARE+/− on hepatic function and pathology have not been examined.

In the present study, we hypothesized that there would be increased liver injury and dysregulation of bile acid homeostasis in the TNF^∆ARE+/− mouse. We find that chronic hepatic TNFα production results in increased hepatocellular injury and periportal fibrosis that corresponded to increased periportal inflammation and dysregulation of bile acid homeostatic regulatory genes, which was associated with increased ileal FXR activation.

METHODS

Murine sample procurement

All animal care and procedures were approved by the University of Colorado Anschutz Medical Campus Institutional Animal Care and Use Committee (protocol number 00000879). All studies involving animal experiments conformed with the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines.¹³ Mice were housed under pathogen-free conditions with a 12/12-hour light/dark cycle and ad libitum access to standard chow. For short-term TNFα exposure studies, male C57BL/6 wild-type (WT) mice (n = 6) were administered recombinant TNFα (i.p. 8 mg/kg in PBS, ∼200 ng/mouse [#410-MT/CF; R&D Systems]) and euthanized 4 hours later.¹⁴ For chronic TNFα expression studies, serum, hepatic, and ileal tissue were removed at the time of sacrifice male 24- to 26-week-old C57BL/6 and TNF^{∆ARE+/− 12} (N = 4–9 per condition).

Immunohistochemical evaluation

Formalin-fixed slides were analyzed for hepatic macrophages using antibodies directed against F4/80 (1:250, catalog number T-2028, Rat anti-mouse monoclonal, BMA Biomedicals), myeloperoxidase (MPO, 1:100, catalog number AF3667, Millipore), and CD3 (1:100, catalog number A0452, Dako). Heat-induced antigen retrieval was performed in Dako target Antigen Retrieval Solution (Dako). Following incubation with primary antibodies overnight, slides were washed 3 × 5 minutes in tris-buffered saline 1% tween and incubated in Horseradish peroxidase conjugated goat anti-rat or goat anti-rabbit (MP-7444, MP7401, Vector Laboratories) secondary antibodies for 30 minutes. The peroxidase substrate used was IMMPACT-DAB (SK-4105, Vector Laboratories). Histologic images were captured on an Olympus BX51 microscope equipped with a 4-megapixel Macrofire digital camera (Optronics) using the Picture Frame Application 2.3 (Optronics). All images were cropped and assembled using Adobe Photoshop Elements (Adobe Systems, Inc.). For quantification of immunohistochemical staining, 10 images per animal were obtained at ×100 magnification (tiling). All images were imported into SlideBook 6.0 (Intelligent Imaging Innovations; Colin Monks), and pixels per image were quantified. Quantification of CD3, MPO, and Ki67 staining was completed by counting the number of positive cells/×200 field with a minimum of 4 different fields counted per animal and 3 animals per group/condition.

Biochemical assessment of liver injury

The serum was separated by centrifugation at 5000 rpm for 5 minutes at 4°C and was assayed for ALT, AST, and ALP activity by the University of Colorado Anschutz Hospital clinical laboratory as described.¹⁵ Total serum bile acids were determined using a Mouse Total Bile Acids Assay Kit (Catalog# 80471, Crystal Chem).

Quantification of serum IL-1β and IL-6 concentrations

Serum IL-1β and IL-6 concentrations were determined using a mouse IL-1β/IL-1F2 Quantikine ELISA Kit and mouse IL-6 Quantikine ELISA kit (R&D Systems) according to the manufacturer’s instructions.

Mass spectrometric quantification of hepatic bile acids

Quantification of hepatic bile acids was performed by the University of Colorado School of Pharmacy Mass Spectrometry Core using fresh frozen hepatic tissue and an Agilent 6520 quadrupole time-of-flight mass spectrometer in negative ionization mode as described.¹⁶ The drying gas was 300 °C at a flow rate of 12 mL/min. The nebulizer pressure was 30 psi. The capillary voltage was 4000 V. Fragmentor voltage was 200 V. Spectra were acquired in the mass range of 50–1700 m/z with a scan rate of 2 spectra/sec. Quantification was determined using known standards and expressed as pmol/mg hepatic tissue.

Quantitative PCR

qRT-PCR to examine mRNA expression was performed using TaqMan probes from Applied Biosystems as described (Supplemental Table S1, http://links.lww.com/HC9/B94).¹⁴

Western blotting

Western blots were performed using separate gels for each protein analyzed and liver homogenate using 30 μg of protein per lane as described.^15,17 The antibodies used in this study are glyceraldehyde 3-phosphate dehydrogenase (GAPDH (1:20,000, MAB374, Millipore)), SHP (1:1000, catalog number PA5-76632, Invitrogen/Thermofisher), BSEP (1:1000, catalog number NBP1-89319, Novus Biologicals), FXR (1:1000, catalog number sc-25309, Santa Cruz Biotechnologies), and NTCP (1:1000, catalog number PA5-80001, Invitrogen/Thermofisher). The secondary antibodies were obtained from Cell Signaling and used at a dilution of 1:10,000. Densitometric analysis was performed using ImageJ with GAPDH expression used as a loading control. Relative changes in expression are presented as compared to WT mice. To validate the use of GAPDH as a loading control, blots were subsequently stained using Ponceau S (Supplemental Data, http://links.lww.com/HC9/B94). No significant differences were evident between the 2 methods of normalization.

Statistical analysis

The data are presented as means ± SEM. Comparisons between genotypes or treatments were accomplished by a Student t test. Statistical significance was set at p < 0.05. Prism 5 for Windows (GraphPad Software) was used to perform all statistical tests.

RESULTS

We have previously reported that short-term TNFα exposure suppressed mRNA expression of hepatic bile acid and sterol transporters.¹⁴ To determine if FXR activation is suppressed by short-term TNFα, WT mice were injected with TNFα (~200 ng/mouse i.p.) and sacrificed after 4 hours. Liver and ileal tissue were isolated, and mRNA expression of the FXR target gene Nr0b2 and expression of classical and alternative bile acid synthesis pathways were examined. Interestingly, short-term TNFα exposure increased Cyp7b1 but did not result in significant changes in Nr0b2, Cyp7a1, Cyp8b1, Cyp27a1, or Cyp3a11 mRNA expression (Figure 1A). To further explore the impact of short-term TNFα on the regulation of bile acid signaling, mRNA expression of FXR signaling intermediates was examined in ileal tissue. Notably, relative ileal expression of Nr1h4 (FXR) was suppressed, but there was a significant increase in both Nr0b2 and Fgf15 expression after 4 hours of TNFα exposure; however, Slc10a2 was not significantly different (Figure 1B).

FIGURE 1.

Short-term TNFα exposure impacts hepatic bile acid signaling. (A) Relative mRNA expression of hepatic bile acid synthesis pathways. (B) Relative mRNA expression of ileal FXR signaling pathway. (C) Relative mRNA expression of hepatic Nlrp3 complex genes. (D) Relative mRNA expression of ileal Nlrp3 complex genes. Values are mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

In macrophages, exposure to TNFα results in increased IL-1b secretion.¹⁸ We have previously shown that i.p. injection of TNFα upregulated IL-1b mRNA after only 4 hours.¹⁴ Following its production, IL-1β is cleaved and released by the Nlrp3 inflammasome complex.¹⁹ To determine if TNFα impacted Nlrp3 inflammasome expression, mRNA quantification was performed on Nlrp3, Gasdermin D (GsdmD), and Asc-pycard in the liver and ileal tissue isolated from control and mice injected i.p. with TNFα. In the liver, mRNA expression of inflammasome components Nlrp3, GsdmD, and Asc-pycard was significantly increased (Figure 1C). Interestingly, in the ileum, mRNA expression of the inflammasome complex was not upregulated (Figure 1D).

Data herein demonstrate clear differences in bile acid signaling molecule expression in both the ileum and liver following short-term TNFα i.p. injection; however, most hepatic diseases are chronic. The effects of long-term TNFα exposure on hepatic injury and bile acid signaling have not been extensively examined. The TNF^∆ARE+/− mouse is a model of chronic TNFα production that results in ileitis.¹² To determine the effects of chronic TNFα overproduction on hepatic injury and bile acid homeostasis, serum, liver, and ileal tissue were isolated from 24- to 26-week-old male TNF^∆ARE+/− mice and controls. To validate hepatic TNFα overexpression, mRNA of TNFα was increased 6-fold in TNF^∆ARE+/− liver tissue compared to age-matched WT controls (Figure 2A). Examining liver injury, TNF^∆ARE+/− mice had increased serum AST, ALP, and total serum bile acids, but no significant differences were evident in ALT activity (Figure 2B).

FIGURE 2.

Chronic production of TNFα increases hepatic biochemical injury. WT or TNF^∆ARE+/− mice were sacrificed at 24 weeks. (A) Relative TNFα mRNA expression. (B) Serum measures of hepatic injury and cholestasis were obtained and compared (ALT, AST, and Alk Phos). Values are mean ± SEM. *p< 0.05, ****p< 0.0001. Abbreviations: Alk Phos, alkaline phosphatase; WT, wild type.

Previous reports have shown that the TNF^∆ARE+/− mouse develops significant histologic evidence of ileitis.¹² To determine the effects of chronic TNFα exposure on liver histology in this model, tissue sections were stained with hematoxylin and eosin. TNF^∆ARE+/− mice developed a dramatic increase in inflammatory cell infiltration around portal triads, but interestingly, inflammation did not extend into the centrilobular region (Figure 3A). To further characterize the inflammatory response, formalin-fixed paraffin-embedded tissue sections were immunohistochemically stained for monocytes (F4/80), neutrophils (myeloperoxidase, MPO), and lymphocytes (CD3). In agreement with Hematoxylin and Eosin staining, increased periportal staining was evident for all 3 inflammatory markers in TNF^∆ARE+/− liver tissue (Figure 3A). In contrast to both CD3 and MPO, which exhibited a mild increase in staining outside of the periportal region, F4/80+ cells were significantly increased across the lobule with notable periportal F4/80-positive cells surrounding and among cholangiocytes (Figure 3B, yellow arrows). To support cell-specific immunohistochemical data, serum concentrations of IL-1β and IL-6 were determined using ELISA. Compared to WT mice, both IL-1β and IL-6 were significantly increased in TNF^∆ARE+/− mice (Figure 3C). Given that there is increased serum proinflammatory cytokines in the TNF^∆ARE+/− mice, we next sought to determine the mRNA expression of key inflammatory mediators by qPCR (Figure 3D). In TNF^∆ARE+/− mice, mRNA expression of Ly6c, Ccr2, Ccl2, Itgam, and Emr1, indicators of macrophage infiltration, were all significantly increased (Figure 3D). Given the aforementioned increase in inflammatory cell infiltration, we hypothesized that there would be increased expression of the Nlrp3 inflammasome complex and its downstream targets IL-1β and IL-18. Quantification of mRNA expression revealed a significant increase in Il1b, Il18, Il6, Ifng, Nlrp3, Asc-pycard, and GsdmD mRNA supporting inflammasome activation (Figure 3D).

FIGURE 3.

Chronic production of TNFα results in increased periportal inflammation. WT or TNF^∆ARE+/− mice were sacrificed at 24 weeks. (A) Formalin-fixed paraffin-embedded liver sections were stained using H&E or immunohistochemically examined for CD3+ lymphocytes, MPO+ neutrophils, or F4/80+ monocytes. Yellow arrows indicate infiltration of inflammatory cells. (B) Quantification of CD3+, MPO+, and F4/80+ staining in WT and TNF^∆ARE+/− livers. (C) Serum concentrations of IL-1β and IL-6. (D) Relative mRNA expression of inflammatory markers. Values are mean ± SEM, n = 3–6/group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Abbreviations: H&E, Hematoxylin and Eosin; WT, wild type.

In the ileum, at 24 weeks, TNF^∆ARE+/− mice developed intestinal fibrosis.¹² To determine if fibrosis was similarly induced in the liver, tissue sections were stained with picrosirius red; the same field was then examined with brightfield or exposures taken under polarized light, and fibrosis was quantified.²⁰ A significant increase in picrosirius red staining was evident in portal tracts and in periportal septae, supporting increased hepatic fibrosis (Figures 4A, B). In addition, a mild increase in portal-portal bridging was visible (arrows). To further support picrosirius red staining data, the relative expression of selected fibrosis genes was examined. Compared to WT, mRNA expression of Tgfb, Timp1, Col1a1, and Mmp9 was increased, but the expression of Mmp13 significantly decreased (Figure 4C).

FIGURE 4.

Chronic production of TNFα increases hepatic fibrosis. (A) Picrosirius red staining of formalin fixed paraffin embedded liver sections prepared from WT and TNF^∆ARE+/− mice (yellow arrows indicate increased PSR staining). (B) Quantification of PSR using polarized light exposures. (C) Relative mRNA expression of markers of fibrosis in liver tissue. Values are mean ± SEM, N = 4–9/group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Abbreviations: FFPE, formalin fixed paraffin embeded; PSR, picrosirius red; WT, wild type.

From the Hematoxylin and Eosin stained tissue sections, we hypothesized that chronic TNFα production and exposure may also contribute to increased periportal proliferation and a ductular reaction. Tissue sections from WT and TNF^∆ARE+/− mice were immunohistochemically stained for cytokeratin 7 (a pan cholangiocyte marker) and Ki67 to assess cellular proliferation (Figure 5A). There was a notable increase in cytokeratin 7 staining, and there were significantly more Ki67-positive nuclei (predominantly periportal) present in TNF^∆ARE+/− liver sections (Figure 5B). These data were supported by significantly upregulated mRNA expression of proliferating cell nuclear antigen (PCNA, Figure 5C).

FIGURE 5.

Chronic production of TNFα induces a ductular reaction and increased periportal proliferation. (A) Immunohistochemical analysis of CK7 and Ki67 expression in formalin fixed paraffin embedded liver sections prepared from WT and TNF^∆ARE+/− mice. Yellow arrows indicate increased CK7 or Ki67 staining. (B) Quantification of CK7 and Ki67-positive cells in WT and TNF^∆ARE+/−livers. (C) Relative mRNA expression of PCNA in liver tissue. Values are presented as mean ± SEM, n = 4–5/group. *p < 0.05, **p < 0.01, ****p<0.001. Abbreviations: CK7, cytokeratin 7; FFPE, formalin fixed paraffin embeded; WT, wild type.

To determine the effect of chronic TNFα exposure in TNF∆ARE^+/−mice on hepatic bile acid homeostasis and transport, relative mRNA expression of FXR-dependent signaling and bile acid homeostasis genes were examined. Chronic TNFα exposure resulted in a significant increase in Nr1h4 (FXR), Nr1h3 (LXR), Nr0b2 (Shp), Nr5a2 (Lrh1), Nr2a1 (HNF4α), Slc10a1 (Ntcp), Abcc2 (Mrp2), Abcc3 (Mrp3), Abcb4 (Mdr2), and Fgfr4 hepatic mRNA expression; however, the expression of Abcb11 (Bsep) was not significantly changed (Figure 6A). Examining bile acid synthesis genes, the expression of Cyp7a1, Cyp7b1, and Cyp8b1 was suppressed, whereas paradoxically, there was a significant increase in Cyp27a1 and Cyp3a11 expression (Figure 6A). We next examined the protein expression of BSEP, FXR, NTCP, and SHP by western blotting and found that both BSEP and SHP expression was suppressed, whereas FXR and NTCP were significantly increased in TNF^∆ARE+/− mice (Figure 6B). To further support changes in bile acid synthesis and bile acid transport, hepatic bile acid concentrations were determined. Interestingly, from the panel of bile acids that were quantified using liquid chromatography mass spectrometry, cholic acid was significantly suppressed in TNF^∆ARE^+/− livers, but all other bile acids that were measured were not significantly different (Figure 6C).

FIGURE 6.

Chronic production of TNFα dysregulates hepatic FXR and bile acid synthesis. (A) Relative mRNA expression of hepatic FXR and bile acid synthesis pathways. (B) Western analysis and quantification of BSEP, FXR, NTCP, and SHP protein expression with representative GAPDH loading control. All proteins were normalized to GAPDH expression. (C) LC-MS of hepatic bile acid concentrations. Data are presented as mean ± SEM and were statistically analyzed using the Student t test, *p < 0.05, **p < 0.01, ***p < 0.001, N = 3–9/group. Abbreviations: FXR, farnesoid X receptor; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; LC-MS, liquid chromatography mass spectrometry.

Previous reports have shown increased ileal inflammation in TNF^∆ARE+/− mice.¹² To confirm intestinal inflammation, tissue sections were stained for F4/80+ monocytes and CD3+ lymphocytes. Compared to WT ileum, TNF^∆ARE+/− mice exhibited increases in both F4/80+ and CD3+ staining (Figure 7A). In the liver, the expression of Cyp7a1 is regulated by the hormone FGF15 which originates from the ileum of the mouse.²¹ We hypothesized that the suppression of Cyp7a1 may be related to increased production of FGF15 in the ileum. Relative mRNA expression of Nr1h4, Fgf15, Nr0b2, Slc10a2 (ASBT), and Slc51a (Ostα) was examined in ileal tissue (Figure 7B). In TNF^∆ARE+/− mice, there was transcriptional suppression of Nr1h4, but a significant increase in Fgf15 and Nr0b2 mRNA expression. No significant differences were evident with respect to Slc10a2 or Slc51a. It has been previously reported that at 12 weeks of age, TNF^∆ARE has no effect on intestinal permeability but suppresses Paneth cell responses.²² Given the pathology of ileal tissue in 26-week-old TNF^∆ARE+/− mice,¹² we hypothesized that intestinal permeability would be impacted. The relative mRNA expression of genes regulating intestinal permeability and innate immune responses was examined in WT and TNF^∆ARE+/− ileal tissue. Interestingly, at 26 weeks, there was a significant decrease in Ocln and Cldn2, whereas Hp (zonulin) and the ratio of Cldn1:Ocln were significantly increased, supporting increased permeability (Figure 7C). Examining Paneth cell innate immune responses, marked suppression of Reg3g, Defa1, Defa5, and Mmp7 was evident in TNF^∆ARE+/− mouse ileum (Figure 7D).

FIGURE 7.

Chronic production of TNFα decreases ileal innate immune response, increases gut permeability, and activates ileal FXR. (A) Hematoxylin and eosin staining and immunohistochemical analysis of F4/80+ monocytes and CD3+ lymphocytes in formalin fixed paraffin embedded ileal sections prepared from WT and TNF^∆ARE+/− mice. (B) Relative mRNA expression of ileal FXR signaling. (C) Relative mRNA expression of genes regulating gut permeability. (D) Relative mRNA expression of ileal Paneth cell innate immune responses. Values are mean ± SEM, n = 4/group. *p < 0.05, **p < 0.01. Abbreviations: FFPE, formalin fixed paraffin embeded; FXR, farnesoid X receptor; WT, wild type.

DISCUSSION

The proinflammatory cytokine TNFα has been shown to be increased in the serum in patients with AIH, biliary atresia, and PBC.^8–10 Furthermore, in human PSC, TNFα expression is increased in whole liver tissue as well as primary biliary epithelial cells.⁷ The long-term effects of increased TNFα exposure on the liver are not well described. In one study, using a TNFα transgenic mouse, overexpression of TNFα resulted in fibrosis and increased inflammatory cytokines at 5.5 months.¹¹ In the current study, we sought to determine the effects of chronic TNFα production on hepatocellular injury, inflammation, and bile acid signaling using the TNF^∆ARE+/− model of Tnfa overexpression.

A hallmark of chronic cholangiopathies is an increase in periportal inflammation that contributes to biliary injury and cholestasis in humans which is supported by murine data.^23,24 In the Mdr2^KO model, periportal infiltration of lymphocytes, neutrophils, and macrophages has been shown to contribute to hepatocellular injury and fibrosis.^20,25 We observed that overexpression of TNFα similarly results in increased periportal infiltration and cholangiocyte proliferation that correlates with an increase in periductal fibrosis and associated biliary injury. Notably, lymphocytes and monocytes can be seen surrounding and infiltrating cholangiocytes in the TNF^∆ARE+/− group.

In humans, cholestatic liver injury decreases ileal FXR activation contributing to decreased Fgf19 (also known as Fgf15 in mice) production.²⁶ This is also evident in murine models following bile duct ligation as well as in Mdr2^KO mice. The lack of Fgf15/19 production contributes to a failure to suppress hepatic bile acid synthesis through Fgfr4/β-klotho–mediated suppression of Cyp7a1 expression. We found that although there is extensive periportal inflammation, ileal FXR activation is still present in the ileum of TNF^∆ARE+/− mice, as evidenced by increased mRNA expression of Fgf15 and Nr0b2 (Shp). Yet, with the exception of cholic acid, which is suppressed, hepatic bile acids are not significantly different, whereas serum bile acids increased, suggesting that while classical bile acid synthesis is suppressed, the alternative pathway is upregulated. These data, combined with the increase in ileal Fgf15 following 4-hour i.p. injections of TNFα, support a TNFα-dependent mechanism of FXR regulation in the gut and the liver. Research has shown that activation of FXR has been shown to reduce intestinal inflammation,²⁶ and we hypothesize that ileal FXR activation may be a protective (but insufficient) response due to increased TNFα production in the gut.

We have recently reported that increased TNFα signaling can suppress the expression of hepatic bile acid transporters.¹⁴ Yet, in the chronic model, protein expression of BSEP was decreased in contrast to NTCP, which was increased and corresponded to increased mRNA expression of all other bile acid transporters (Abcc2, Abcc3, and Abcb4) as well as an increase in serum bile acids which support cholestasis. This suggests that during chronic TNFα exposure, hepatocyte bile acid transport is upregulated. Overall, these data demonstrate contrasting effects with respect to short-term and chronic exposure to TNFα on hepatic bile acid signaling. In the liver, infiltrating lymphocytes have been shown to suppress Cyp7a1 and Cyp8b1 through increased production of both TNFα and IFNγ.²⁷ Data herein further support the contribution of TNFα or IFNγ in the suppression of hepatic bile acid transport as well as primary bile acid synthesis despite increased mRNA expression of Nr5a2 (LRH1), which has been shown to increase expression of Cyp7a1 and Cyp8b1.^28,29 We hypothesize that this is partially due to increased Fgf15 production which would downregulate Cyp7a1 through increased Fgfr4 signaling in hepatocytes.³⁰

In conclusion, in this study, the hepatocellular, biochemical, and pathological changes in the liver were examined in the TNF^∆ARE+/− mouse. We found that there was an increase in hepatic periportal inflammation, bile ductular reaction, and fibrosis that correlated with suppression of classical bile acid synthesis and, with the exception of BSEP, increased expression of hepatic bile acid transporters (Figure 8). Mechanistically, suppression of classical bile acid synthesis (Cyp7a1) was supported by the suppression of hepatic cholic acid concentrations and increased ileal FXR activation of expression of Fgf15 mRNA. In summary, this study supports the pathological role of TNFα in chronic hepatic inflammatory diseases and provides an important model for their study.

FIGURE 8.

Chronic TNFα expression directly impacts the gut-liver axis. In the ileum, increased TNFα contributes to decreased intestinal permeability and Paneth cell defensin production, but also activates FXR increasing FGF15 mRNA expression. In the liver, chronic overproduction of TNFα contributes to an increase in hepatic injury, inflammation, and fibrosis, but bile acid homeostasis is dysregulated. Abbreviation: FXR, farnesoid X receptor.

Supplementary Material

hc9-8-e0589-s001.pdf ^{(687.6KB, pdf)}