Mutant p53 in concert with an IL27 Receptor α deficiency causes spontaneous liver inflammation, fibrosis, and steatosis

Abstract

Background

The cellular and molecular etiology of unresolved chronic liver inflammation remains obscure. While mutant p53 has gain-of-function properties in tumors, the role of this protein in liver inflammation is unknown.

Main

Herein, mutant p53^R172H is mechanistically linked to spontaneous and sustained liver inflammation and steatosis when combined with the absence of IL27 signaling (IL27RA), resembling the phenotype seen in non-alcoholic fatty liver (NAFLD) and non-alcoholic steatohepatitis (NASH) patients. Indeed, these mice develop with age hepatocyte necrosis, immune cell infiltration, fibrosis, micro- and macro-steatosis; however, these phenotypes are absent in mutant p53^R172H or IL27RA^−/− mice. Mechanistically, ETAR-positive macrophages are highly accumulated in the inflamed liver, and chemical inhibition of ETAR signaling reverses the observed phenotype and negatively regulates mutant p53 levels in macrophages.

Conclusion

The combination of mutant p53 and IL27RA^−/− causes spontaneous liver inflammation, steatosis, and fibrosis in vivo, while either gene alone in vivo has no effects on the liver.

The most commonly mutated gene in cancer cells is the oncogene p53. Many investigators have shown in vivo that mice carrying mutant p53 in a Li-Fraumeni mouse model spontaneously develop a tumor profile with metastatic disease that is different from mice carrying null p53 (1, 2). Aside from the most-commonly known function of oncogenes in propelling tumor progression, a new notion of oncogene-induced inflammation or oncogene-sustained inflammation in modulating human disease is emerging (3). Indeed, one of the mutant p53 gain-of-function (GOF) properties was attributed to chronic NF-κB activation upon the inflammatory stimuli of TNF in cultured cancer cells and in vivo (4–8). While these mutant p53 mice do not develop spontaneous inflammatory disease, they are highly susceptible to DSS-associated colitis and inflammation-associated colorectal cancer (4).

Chronic liver inflammation can result in the deadly diseases such as cirrhosis and hepatocellular carcinoma (HCC). Spontaneous hepatocyte apoptosis, infiltration of inflammatory cells, and compensatory proliferation are the initial steps in sustaining chronic liver inflammation which when left unresolved results in fibrosis. Although the knowledge of the key players that spontaneously drive this process is limited, the evidence suggests that whichever role p53 plays, it is an important one. Many investigators have shown that p53 protein accumulates in patients with various inflammatory liver diseases (9–12). Mechanistically, Kodama and colleagues increased p53 expression in hepatocytes by deleting MDM2 (the enzyme that binds and degrades p53) which promoted fibrosis via inducing connective tissue growth factor (CTGF) expression (13). Meanwhile, another group has shown that p53 expression in activated stellate cells will limit liver fibrosis by inducing cell death of these cells (14). Because the role of mutant p53 in inflammation is so poorly understood, one imminent question is whether endogenous mutant p53 has any effect in liver inflammation and fibrosis. If so, under what circumstance mutant p53 can affect liver problems such as spontaneous inflammation, fibrosis and/or fatty liver since mice carrying mutant p53 gene do not develop any liver abnormalities (2)?

The role of Interleukin 27 (IL27) in modulating the shape and the strength of the immune response has been well-appreciated. This cytokine possesses either pro- or anti-inflammatory functions in a context-dependent matter (15). IL27 is tightly involved in liver disease; mice absent in IL27RA signaling are highly sensitive to liver inflammation, ConcovalinA-induced liver disease or LPS stimulation, albeit these mice are inflammation-free in absence of stimuli (16, 17). We hypothesized that p53^515A “knock-in” mice carrying one germline mutant p53 allele encoding p53R172H (heterozygote, referred as p53^H/+ from herein), the mouse equivalent of the human hot spot mutant p53R175H, plus mice IL27RA knockout signaling might spontaneously develop liver disease. Indeed, these mice (IL27RA^−/−p53^H/+) mice spontaneously develop hepatocyte necrosis, immune cells inflammatory, fibrosis, and fatty liver with age, while neither IL27RA^−/− nor p53^H/+ mice develop any of these phenotypes. Of interest is one particular population of immune cell, Endothelin A receptor (ETAR)-positive macrophages, which increasingly accumulate in the inflamed liver. Chemical inhibition of ETAR signaling reverses the inflammatory phenotype seen on these mice, suggesting that mutant p53 and absence of IL27 signaling results in liver inflammation and steatosis via ETAR-positive macrophages.

MATERIALS AND METHODS

Transgenic Mice

All mice strains were maintained in C57Bl/6 cohort. IL27RA^−/− mice and P53^H/+ mice were bred to create cohorts. IL27RA^−/− mice were previously donated by Dr. Frederic de Sauvage (Genentech). Genotyping was performed as described below for both p53 and IL27RA. For the cohort studies, mice were euthanized once mice reached euthanasia criteria mainly due to tumor burden. For the time course study, mice from each genotype were euthanized at the indicated times. Liver tissue was fixed in neutralized paraformaldehyde prior to sectioning. To inhibit ETAR signaling, Zibotentan (ZD4054, Selleck Biologicals), 5mg/kg/mouse or vehicle control was given IP to mice three times a week starting at age 3 month, diluted at 100 ul of saline per mouse. The drug was given to the mice for 8 consecutive weeks. Consequently, mice were euthanized via carbon dioxide asphyxiation, and liver tissue was used for further studies. All the procedures involving animals were approved by the Institutional Animal Care and Use Committee.

Western Blot

Briefly, liver tissue was snap frozen in liquid nitrogen. Tissue was homogenized in lysis buffer, and supernatants were quantified for protein amount (18). 160 ug of cell lysate was loaded in 10% SDS-PAGE, transferred to a nitrocellulose membrane and incubated with primary antibody overnight at 4° C (1000x dilution for anti-p53 (Novus), 2500x anti-GAPDH, 5000x anti-mouse IgG-HRP (Santa Cruz), 500x anti-pERK1/2 (Santa Cruz), 500x anti-MDM2 (Santa Cruz).

Picrius Red Staining

Picrius Red Staining was performed according the manufacturer’s instructions (IHC World).

Hematoxylin-Eosin Staining of Liver Sections and Scoring of Liver Lesions

Sections of paraffin-embedded tissues were stained with hematoxylin and eosin. The pathological readings were scored by a pathologist at our institution. Images were taken Nikon Eclipse Ti microscope. During time course experiments, the lesions were counted under a 200× microscope, where 15 fields per slide were counted. The small lesions typically observed early in the liver consist of hepatocellular degeneration and hepatocyte necrosis along with Kupffer cell hyperplasia attended by a small number of lymphocytes were quantified as previously described(19).

Genotyping

DNA was extracted from ear tissue, and PCR for p53 gene was performed using the primer sequence as previously described by Lang et al (2). The presence of IL27RA gene was confirmed via using forward CAAGAAGAGGTCCCGTGCTG primer sequence and reverse TTGAGCCCAGTCCACCACAT sequence. PCR primers for the absence of IL27RA are as follows: GCTTTCGTCTCCCGTGTGCT (forward) and TGAGCCCAGAAAGCGAAGGA (reverses).

Quantitative real-time PCR

Samples were prepared and processed as previously described (20). Briefly, total RNA was isolated from cells using TRIzol Reagent (Life Technologies). Residual genomic DNA was removed from total RNA using the TURBO DNA-free kit (Life Technologies). Two micrograms of RNA were used for cDNA synthesis using the High-Capacity RNA-to-cDNA Kit (Life Technologies). The relative gene expression levels were determined using RT- qPCR and SYBR Green labeling method in an StepOnePlus™ Real-Time PCR System (Life Technologies). The reaction contained 2 uL cDNA, 12.5 uL SYBR Green PCR Master Mix (Life Technologies), and 200 uM primer in a total volume of 25 uL. The PCR cycling conditions were as follows: 40 cycles of 15 s at 95C and 60 s at 60C. All samples were run in duplicates. The CT value of each sample was acquired, and the relative level of gene expression was calculated by the delta CT method, which was normalized to the endogenous control of GAPDH. Data were expressed as n-fold relative to control. The primer sequences for the mouse GAPDH and P53 are: Gapdh CCAGCCTCGTCCCGTAGAC and CGCCCAATACGGCCAAA; P53 CTCTCCCCCGCAAAAGAAAAA and CGGAACATCTCGAAGCGTTTA.

Statistical Analysis

Student’s T test was used to analyze the differences amongst groups. GraphPad Prism for Windows was used to prepare and analyze the graphs (GraphPad Software).

Flow cytometry analysis

Purified liver immune infiltrates were blocked for non-specific staining in 2% BSA and with anti-CD16 (5 μg/mL) for 20′ at 4°C. Afterwards, cells were stained in 2% BSA with CD4-Violet421 (1:100, Pharminogen), CD8-PE-Cy7 (1:100, Biolegend), F4/80- Fitc (1:100, eBiosciences), CD3-e450 (eBiosciences), NK1.1- PE-Cy7 (eBiosciences), CD19-Violet421 (1:100, Biolegend) and B220-PerCP (1:100, Biolegend), CD11c-Violet 421 (1:100, Biolegend), MHCII-PerCP (1:100, Biolegend). At last, the cells were washed twice in 2% BSA and analyzed by flow cytometry using Attune (Invitrogen), and FlowJo.

Elisa

Elisa for TNF (eBiosciences) and IL6 (Biolegend) was performed according to the manufacturer’s instructions. Briefly, liver tissue was processed as described in western blot procedure and the supernatant was quantified for both total protein amount and the levels of cytokines via ELISA. Cytokine levels were normalized per total protein amount.

RESULTS

To understand the impact of mutant p53 in liver disease, we choose a mouse model of patients with Li Fraumeni-like syndrome. These mice carry a germline mutant p53 allele encoding p53R172H, the mouse equivalent of the human hot spot mutant p53R175H (2). Because these mice do not develop spontaneous liver abnormalities, we crossed mutant p53 carrying either one knock-in allele of p53R172H (p53^H/+) or two knock-in alleles of p53R172H (p53^H/H) to IL27RA^−/− mice since these mice are more susceptible to inducible liver inflammation. Mice of different groups (C57Bl/6 [WT], IL27RA^−/−, p53^H/H, p53^H/+, IL27RA^−/−p53^H/+, and IL27RA^−/−p53^H/H) were monitored over time. Once IL27RA^−/−p53^H/H and IL27RA^−/−p53^H/H mice became moribund, the livers of these mice and those from the control age-matched groups were analyzed and compared. The livers in aged WT, IL27RA^−/−, and p53^H/+ mice did not contain any abnormalities (Figure 1A); however, IL27RA^−/−p53^H/+ mice spontaneously developed liver abnormalities with an increased accumulation of inflammatory cells in livers (Figure 1A). Grossly, mice developed a different phenotype, ranging from enlarged livers to mild or severe steatosis (Figure 1B). Meanwhile, these mice also developed large necrotic areas, extensive fibrosis, hepatocellular degeneration, and both micro- and macrovesicular steatosis, bile accumulation, or hemorrhage (Figure 1C). More than 75% of IL27RA^−/−p53^H/+ mice developed these liver abnormalities, while all of the control mice had normal livers (Figure 1D). Pathological readings for fibrosis and collagen deposition were further confirmed by Sirius Red staining (Figure 1E). Mice with more advanced liver disease developed both fat accumulation and fibrosis in livers, consistent with patients with non-alcoholic steatohepatitis (Supplementary Figure 1). While chronic inflammation was a suspected phenotype, steatosis was an unexpected observation, although studies have implicated kupffer cells in liver steatosis (22).

Figure 1. The combination of mutant p53 and IL27RA^−/− causes spontaneous liver inflammation and steatosis.

(A) Representative HE staining of livers of 12–16 months of age of (i)WT, (ii) IL27RA^{−/ −}, (iii) p53^H/+, or (iv) IL27RA^{−/ −} p53^H/+. White scale bar, 50 μm.(B) Representative photomicrographs of the gross livers of IL27RA^{−/ −}p53^H/+. (C) Representative HE staining for (i) pigment accumulation, (ii) necrotic area surrounded by immune infiltrates, (iii-v) macro and micro- steatosis, and (vi) hemorrhage. White scale bar, 50 μm. Black scale bar, 25 μm. (D) Quantification of the percentage of mice that developed liver-related inflammation: WT, (N=5), IL27RA^{−/ −} (N=8), p53^H/+ (N=8), IL27RA^{−/ −} p53^H/+ (N=18). Lymphomas were excluded as a phenotype. (E) Fibrosis was assessed based on Sirius red staining of (i)WT, (ii) IL27RA^{−/ −}, (iii) p53^H/+, or (iv) IL27RA^{−/ −}p53^H/+ age-matched mice. (F) Percentage of mice IL27RA^{−/ −} p53^H/+ developing liver-related abnormalities based on pathological readings. N= 18.

Different from IL27RA^−/−p53^H/+ mice, IL27RA^−/− p53^H/H mice carrying both mutant p53 alleles developed liver inflammation at a lower frequency than IL27RA^−/−p53^H/+ mice (less than 3% when compared to 75%). Such reduced frequency may be attributed to the shorter lifespan in these mice due to succumbing to tumor burden at an early age (between 4–7 months of age), while IL27RA^−/−p53^H/+ mice survive for 14–18 (referred as “aged mice” throughout this work) months allowing a longer time span for chronic liver inflammation to occur.

Next, the livers of younger mice, aged 2 and 4 months, were analyzed to understand the early events that led to liver fibrosis and steatosis in our model. Indeed, a higher frequency of IL27RA^−/−p53^H/+ mice developed inflammatory foci in the liver when compared to other groups (Figure 2A, B). Liver damage and compensatory proliferation are prerequisites for inflammatory cell infiltration, thus the increase in the number of Ki67-positive cells in the livers of IL27RA^−/−p53^H/+ mice was expected (Figure 2C–D). These results indicate that compensatory proliferation and accumulation of immune cells are early events leading to the phenotype observed in these older mice.

Figure 2. Liver inflammation and compensatory proliferation are early events in liver pathogenesis.

(A, B) Time course study showing liver pathology and quantification in mice as early as 2 and 4 months in IL27RA^{−/ −}p53^H/+ evaluated by HE staining. N=3–4 per group. Pictures takes at 200X. (C, D) Time course study showing Ki67 staining and quantification in the liver of 2 and 4 month old mice of different genotypes. Dots indicate data from the liver of individual mice. Pictures takes at 200X.

Because accumulation of immune cells is an early event in IL27RA^−/−p53^H/+ mice, we wished to investigate which immune cells are most likely involved in this process. Intrahepatic immune cells were profiled the via flow cytometry in mice at 3 months of age (because inflammatory foci were consistently seen in 2 and 4 month old mice). Others have found that natural killer T cells, kupffer cells, CD4 T cells, and natural killer cells play an important role in liver inflammation (23). In the immune cell profiling among different groups at age 3 months, no significant changes were seen in immune cell subgroups amongst all genotypes (Figure 3A). However, as the mice age, the number of F4/80-positive cells were significantly increased in the livers of IL27RA^−/−p53^H/+ mice when compared to the controls (Figure 3B, C).

Figure 3. Increased number of macrophages associated with liver disease phenotype.

(A) No significant changes were seen in mice of different genotypes in CD4, CD8, B, NK, NKT, Macrophages or CD11c cells in the liver of 3 month old mice analyzed by flow cytometry. (B, C) Significantly enhanced number of F4/80-positive in the aged IL27RA^{−/ −}p53^H/+ mice when compared to age-matched control one analyzed via immunohistochemistry. Dots indicate data from the liver of individual mice. Pictures takes at 200X. *, p<0.05

F4/80 is a hallmark of kupffer cells which are involved in liver pathogenesis induced by multiple individual stressors, such as chemicals, toxins, carbon tetrachloride, and endotoxins, mainly through the release of cytokines, superoxides, and nitric oxide. Because F4/80 cells were significantly increased in the inflammatory-prone IL27RA^−/−p53^H/+ mice, we sought to define any signaling changes seen in these cells. Others have previously reported that Endothelin/Endothelin receptor axis to be involved in carbon tetrachloride (CCL4)-mediated liver fibrosis (24). Interestingly, ETAR was highly and significantly upregulated in macrophages in the liver of the aged IL27RA^−/−p53^H/+ when compared with other cohorts, suggesting the involvement of ETAR-positive kupffer cells in sustaining an unresolved chronic inflammation (Figure 4A). In supporting this notion, ET1, the ligand for this receptor, was also elevated in the macrophages of these aged mice (Figure 4B). Similar elevations are seen in cirrhosis patients when compared to healthy ones (25).

Figure 4. ETAR-positive macrophages are responsible for inflammatory response in these mice.

(A) Dual staining for ETAR and F4/80 and ETAR quantification in different genotypes showed that ETAR-positive macrophages are highly accumulated in the liver of aged mice. (B) Dual staining for ET1 and F4/80 and ET1 quantification in different genotypes showed that ET1-positive macrophages are highly accumulated in the liver of aged mice. (C, D) Immunofluorescence and western blot analysis showing that chemical inhibitor of ETAR signaling, ZD4054, repressed the downstream signaling of ETAR, pERK1/2 in macrophages. (E, F) Photomicrograph and quantifications showing reducing in inflammation in steatosis in IL27RA^{−/ −}p53^H/+ mice treated with ZD4054 (N=9), when compared to sham treated mice (N=7). Chi square test was used to determine the differences among these two groups. In vivo data with ZD4054 inhibitor was repeated twice with similar results. (i)WT, (ii) IL27RA^{−/ −}, (iii) p53^H/+, or (iv) IL27RA^{−/ −}p53^H/+. Dots indicate data from the liver of individual mice. Pictures takes at 200X. *, p<0.05.

One interesting question is whether IL27 can directly affect ETAR levels in myeloid cells. To test this hypothesis, myeloid cells were used since these cells contain detectable levels of endogenous ETAR expression. In agreement with the in vivo observation, treatment with recombinant IL27 reduced the level of ETAR expression in these myeloid cells (Supplementary Figure 2). Furthermore, because ETAR levels are not detectable in kupffer cells following treatment with LPS in vivo, the isolated primary kupffer cells from these mice were re-stimulated with LPS in vitro in presence or absence of recombinant IL27 (rIL27) overnight. Indeed, rIL27 inhibited LPS to induce ETAR levels (Supplementary Figure 3). And lastly, to better understand if IL27 can inhibit ETAR levels in vivo under inflammatory conditions, wildtype and IL27RA^−/− mice were treated in vivo with CCL4 and these fibrotic liver specimens were analyzed for ETAR levels (Supplementary Figure 4). As expected, ETAR levels were higher in the liver of IL27RA^−/− mice when compared to the wildtype ones, further suggesting the involvement of IL27 in the regulation of ETAR.

To establish whether these ETAR-macrophages indeed regulate establishment and sustaining of inflammation, ZD4054, a chemical inhibitor of ETAR signaling was used. Once IL27RA^−/−p53^H/+ mice reached 3 months of age, these mice were treated with ZD4054 or sham for 8 consecutive weeks. Because ET1/ETAR signaling induces phosphoERK1/2 activation, we sought to determine if ZD4054 inhibits ET1/ETAR-induced pERK1/2 in macrophages. Indeed, both immunohistochemistry and western blot results confirmed that ZD4054 inhibited ETAR-mediated downstream signaling (Figure 4C, D). Hematoxylin and eosin staining of the livers of IL27RA^−/−p53^H/+ mice treated with ZD4054 inhibitor showed a significant reduction in the number of inflammatory foci and early signs of steatosis in the liver (Figure 4E, F).

The above result led us to hypothesize that the enhanced ETAR signaling sustains an inflammatory phenotype in the macrophages of livers. IL6 and TNF are the two of the most dominant inflammatory cytokines involved in patients with non-alcoholic steatohepatitis (NASH) liver pathogenesis (5, 26). Triple staining for ETAR, F4/80, and TNF showed that the inhibition of ETAR signaling reduced significantly TNF levels in ETAR-positive macrophages (Figure 5A, B) when compared to controls. Similarly, triple staining for ETAR, F4/80, and IL6 showed that the inhibition of ETAR signaling significantly reduced IL6 levels in ETAR-positive macrophages in the inflammatory prone mouse model IL27RA^−/−p53^H/+ (Figure 5C, D) when compared to control treatment. The reduction of IL6 and TNF levels in the liver with ZD4054 inhibitor were also confirmed by ELISA (Figure 5E, F).

Figure 5. ETAR signaling sustains inflammatory phenotype in the liver macrophages.

(A) Triple staining for ETAR, F4/80, and TNF showed that the inhibition of ETAR signaling reduced TNF levels in ETAR-positive macrophages. (B) Quantification of TNF staining. (C) Triple staining for ETAR, F4/80, and IL6 showed that the inhibition of ETAR signaling reduced IL6 levels in ETAR-positive macrophages. (D) Quantification of IL6 staining. (E, F) Quantification of TNF and IL6 levels in the liver via ELISA. Dots indicate data from the liver of individual mice. Pictures takes at 200X. *, p<0.05

The levels of mutant p53 protein, same as the wildtype one, are strictly regulated. The p53R172H protein in p53^H/+ mice is inherently unstable (27, 28). Because of the chronic unresolved inflammation in the liver IL27RA^−/−p53^H/+, one hypothesis is that the level of this p53 mutant protein are involved. In line with this hypothesis, both immunofluorescence and western blot analysis revealed that p53 protein accumulation were raised in IL27RA^−/−p53^H/+ when compared to other control cohorts though the transcript levels remain the same (Figure 6A, B, C). Meanwhile, p21 levels were low and similar between different cohorts, suggesting that wildtype p53 is unchanged in these cells (Supplementary Figure 5). These observations advocate that the involvement of mutant p53 in the chronic liver inflammation/fibrosis.

Figure 6. P53 stability in macrophages is a downstream event of ETAR signaling.

(A,B) Immunofluorescence and RT-PCR analysis showed that p53 protein accumulation but not transcript levels were raised in IL27RA^{−/ −}p53^H/+ when compared to other control cohorts in aged mice. (C) Western blot showing increased p53 protein stability in the aged livers of IL27RA^{−/ −}p53^H/+ when compared to the age-matched genotypes. (D) Triple staining for p53, F4/80, and albumin showed p53 is stabilized specifically in macrophages but not hepatocytes. (E, F) Western blot and immunofluorescence staining showing p53 levels (E) and pERK1/2 (F) levels are reduced in IL27RA^{−/ −}p53^H/+ mice treated with ZD4054 (N=9), when compared to sham treated mice (N=7). In vivo data with ZD4054 inhibitor was repeated twice with similar results. Each lane or dot indicates data from the liver of individual mice. Pictures takes at 200X. *, p<0.05

Since p53 stability is tightly linked to its function, and given the involvement of wildtype p53 in hepatocytes in the liver, next we sought to investigate in which cell type is p53 stabilized in the liver of IL27RA^−/−p53^H/+ mice. Surprisingly, mutant p53 is selectively stabilized in the macrophages rather than in hepatocytes (Figure 6D).

While ETAR signaling in macrophages plays a substantial role in sustaining chronic and unresolved inflammation in the liver of IL27RA^−/−p53^H/+ mice, mechanistically how ETAR-macrophages sustain chronic inflammation in the liver has not yet been investigated. Because p53 protein is accumulated mainly in macrophages but not hepatocytes (Figure 6D), we wanted to further investigate whether expression of mutant p53 in macrophages was a downstream target from ETAR signaling. Indeed, ETAR induces pERK1/2 activation in our model (Figure 4C, D), and others have shown that pERK1/2 can induce phosphorylation of p53 at Ser15, therefore enhancing its stability by preventing its interaction with MDM2 (29). Indeed, both immunohistochemistry and western blot confirmed that inhibition of ETAR results diminished p53 accumulation, ERK1/2 phosphorylation, and phosphorylation of p53 at S15 (Figure 6E, G, and Supplementary Figure 6).

DISCUSSION

Unresolved chronic inflammation in the liver is tightly linked to fibrinogenesis; however, inflammatory pathways alone rarely cause chronic and unresolved inflammation. Indeed, these pathways aid the process of fibrinogenesis when induced by viruses or other inflammatory agents, but additional stimuli are needed to sustain unresolved inflammation. The involvement of oncogenic molecules, such as mutant p53, in this process remains elusive. More interesting, the combination of these two pathways has never been examined prior to this study. This novel investigation revealed that the combination mutant p53^R172H knock-in and an absence of IL27RA causes spontaneous chronic liver inflammation and steatosis in mice as they age while the expression of either genotype alone does not cause any liver inflammation or steatosis. Histologically, the livers of these IL27RA^−/−p53^H/+ mice closely reproduce features observed in NAFLD/NASH patients. Together, this information suggests a multiple-hit model for the initiation and progression of unresolved chronic liver inflammation.

Two schools of thought exist on the driving forces of NAFLD/NASH. The first school suggests the sequential “two-hit” model, where steatosis is the first hit, and inflammation is the second hit necessary to develop NASH (30). Because only 10-20% of patients that have NAFLD (benign disease) develop inflammation and fibrosis, Moschen and colleagues proposed that NASH might be a “separate disease with different pathogenesis” (31). This group proposes that many hits might act in parallel, where inflammation may precede steatosis in NASH (31). Some evidence for this hypothesis exists in that NASH patients present with little steatosis, and drugs targeting the inflammatory TNF pathway improve steatosis (32, 33). Meanwhile, genetic models that portray stepwise this multiple-hit hypothesis in the absence of hypernutrition have been limited. We believe that our genetic model supports the multiple-hit hypothesis: inflammation occurs as early as 2-4 months, whilst the accumulation of fat in the liver happens as a later event. Aside for histopathologically modeling NAFLD and NASH patients, this study further distinguished that ETAR-positive kupffer cells accumulate over time, and play a significant role in this process. Furthermore, chemical inhibition of this ETAR pathway reverses inflammation and improves early signs of steatosis (When the inhibitor studies were concluded at 5 months of age after 8 weeks of treatment, only early signs of steatosis were observed. The authors did not extend the inhibitor studies beyond 8 weeks due to the possibility of liver damage from extended use of the drug). Another novel finding from this study, the phenotype developed in absence of hypernutrition. It is thought-provoking that the genetic mouse model presented in this study might represent a different aspect of the NASH etiology that has not yet been appreciated, thereby adding to already established models and further contributing to the understanding of this disease in humans.

The overall understanding of ETAR-positive macrophages in liver pathogenesis is limited. The results reported here define ETAR-positive macrophages as the major contributor in sustaining liver inflammation in our model. Indeed, these cells were highly enriched in the liver of IL27RA^−/−p53^H/+ mice, and inhibition of ETAR resolved or prevented accumulation of liver inflammatory foci (Figure 4). Others have shown via inhibition that this pathway is indeed involved in liver pathogenesis. More concretely, inhibition of ETAR and ETBR pathways in the liver significantly reverses liver fibrosis induced by CCL4. Similar to our finding that ET1 is increased in diseased livers in IL27RA^−/−p53^H/+ mice, cirrhosis patients have much higher expression of ET1 when compared to healthy individuals, suggestive indeed of the clinical relevance of this pathway in liver pathogenesis (25). Additionally, our study confirmed that the upstream molecular regulators, such as IL27RA^−/− and mutant p53, are needed to sustain the accumulation of these pro-inflammatory macrophages in the liver (Figure 4). Mechanistically, ETAR-positive macrophages sustain inflammation in our model is closely associated with upregulation of TNF and IL6 (Figure 5), two well-characterized cytokines upregulated in the liver of NASH patients (5, 26).

An increase of p53 levels is closely associated with the ability of IL27RA^−/−p53^H/+ mice to develop spontaneous inflammation (Figure 6) which relates well with previously published results. In patient samples, p53 levels were significantly higher in those with steatohepatitis when compared to healthy individuals (9). Most likely, mutant p53 is accumulating in our studies since p21 levels were minimal among all different controls (Supplementary Figure 3). Other evidence exists in the literature coupling the activity of mutant p53 to inflammation. Under inflammatory conditions, intestinal stem cells expressing mutant p53 can proliferate at a faster rate than the wildtype cells (34).Furthermore, the p53R172H protein in p53^H/+ mice, similar to wildtype p53, is kept at basal levels. Multiple stress signals activate mutant p53 in vivo (35). Kupffer cells are the predominant cell that displays p53 protein accumulation in the liver of IL27RA^−/−p53^H/+ mice (Figure 6). ETAR signaling tightly regulated p53 levels in our model (Figure 6). Most likely, ETAR-induced pERK1/2 will stabilize p53, as others have shown that pERK1/2 activation enhances phosphorylation of p53 at serine 15 thereby reducing the affinity of this protein to interact with MDM2, the major ubiquitin ligase of p53 (29, 36). Mutant p53 protein is involved in many steps from cancer initiation to progression, and in our model a similar phenomenon occurs, as mutant p53 sustains kupffer cell numbers which is in turn associated with hepatocyte necrosis and fibrosis, hallmarks of hepatocellular carcinoma development. HCC development did not occur in the livers of these mice in our model, perhaps due to the shortened lifespan related to the rapid development of other tumors, such as sarcomas and lymphomas.

In this study, high p53 levels were observed in the kupffer cells of IL27RA^−/−p53^H/+ mice when compared to other control genotypes; however, the levels of p21 remained basal in p53^H/+, IL27RA^−/−, and IL27RA^−/−p53^H/+ mice, suggesting that mutant p53 rather than the wildtype p53 is stabilized. This selective protein stabilization suggests the significance of mutant p53 protein in regulating the inflammation or ETAR expression in kupffer cells in these mice. Such selective stabilization of mutant p53 is logical because an increased accumulation of wildtype p53 would cause apoptosis and elimination of these macrophages. Perhaps this mechanism exists in the kupffer cells of IL27RA^−/− mice, which is why no increase expression of ETAR was found in the liver of that mouse strain. Meanwhile, kupffer cells from IL27RA^−/−p53^H/+ mice stabilized mutant p53, and most likely these cells survive and accumulate over time due to a well-attributed survival-dependent mechanism of mutant p53.

Our proof-of-principle study revealed that inflammatory macrophages with mitogenic or survival signals are contributors to chronic liver inflammation, steatosis, or steatohepatitis. This study linked mutant p53 to spontaneous and sustained liver inflammation and steatosis when combined with an inflammatory signal, such as the absence of IL27 signaling. Mechanistically, ETAR-positive macrophages are highly accumulated in the inflamed liver, and chemical inhibition of ETAR signaling reverses the observed phenotype and inflammatory cytokine levels, such as TNF and IL6. These results further contribute to our understanding of NAFLD/NASH in humans.

Figure 7.

Schematic overview of events that lead to liver disease in IL27RA^{−/ −}p53^H/+ mice.

Abbreviations

IL27RA

IL27 receptor a

NAFLD

non-alcoholic fatty liver

NASH

non-alcoholic steatohepatitis

GOF

gain-of-function

HCC

hepatocellular carcinoma

CTGF

connective tissue growth factor

IL27

Interleukin 27

ETAR

Endothelin A receptor

CCL4

carbon tetrachloride