If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Type I interferon (IFN-I) has a well-known function in controlling viral infections, but its contribution in hepatocyte proliferation and hepatocellular carcinoma (HCC) formation remains unclear. Mice deficient in IFN-α receptor expression in whole mice or only in hepatocytes (Ifnar−/− and IfnarΔliver) were used to investigate the role of IFN-I signaling in cell proliferation and cancer formation in the liver. Ifnar−/− mice were resistant to chemical-induced HCC formation in the absence of infection. The results show that low grade of IFN-I and interferon-stimulated gene were expressed substantially in naïve mouse liver. The low level of IFN-I activation is constantly present in mouse liver after weaning and negatively modulates forkhead box O hepatic expression. The IFN-I signaling can be partially blocked by the clearance of lipopolysaccharide. Mice lacking IFN-I signaling have lower basal proliferation activity and delayed liver regeneration processes after two-thirds partial hepatectomy. The activation of IFN-I signaling on hepatocyte controls glucose homeostasis and lipid metabolism to support proliferation potency and long-term tumorigenesis. Our results reveal a positive role of low-grade IFN-I singling to hepatocyte proliferation and HCC formation by modulating glucose homeostasis and lipid metabolism.
The liver has both metabolic and immunoregulatory functions. Various nutrients, xenobiotics, and bacterial components are absorbed in the small intestine and pass through the liver via the portal circulation, undergoing metabolic elimination or conversion.
Bacterial translocation, viral infection, and hepatocyte death, caused by xenobiotics, may activate innate immune responses, leading to cytokine and interferon production. After injuries caused by hepatotoxic substances or infection, the liver undergoes compensatory proliferation to maintain homeostasis and tissue mass. Constant inflammation and compensatory regeneration are thought to promote cancer development.
and persistent infection caused by hepatitis B and C viruses is one of the risk factors for human HCC. Type I interferon (IFN-I) is part of the innate immune response that controls virus infection. Although the role of IFN-I in controlling viral infection is well recognized, its impact on hepatocyte proliferation and HCC development is poorly understood.
Expression of IFN-I proteins, such as IFN-α and IFN-β, is triggered by pathogen-associated molecular patterns through activation of pattern-recognition receptors, including RIG-I–like receptors and toll-like receptors (TLRs).
The binding of IFN-I to its receptor activates JAK1 and TYK2 tyrosine kinases, leading to the phosphorylation and dimerization of STAT1 and STAT2, which turn on interferon-stimulated gene (ISG) expression to fight viral infections.
The effect of IFN-I signaling on cell fate might be cell type and context dependent.
In the present study, IFN-α receptor knockout (Ifnar−/−) mice were used to examine the contribution of IFN-I signaling to HCC development, hepatocyte proliferation, and liver regeneration. A low level of IFN-I activation was found to positively regulate compensatory hepatocyte proliferation by modulating glucose and lipid metabolism, which might contribute to HCC development.
Materials and Methods
Wild-type (WT) C57BL/6 and Ifnar−/− male mice (C57BL/6 background)
were maintained at the National Health Research Institutes. IfnarF/F mice were obtained from the Jackson Laboratory (Bar Harbor, ME), and albumin-Cre mice were kindly provided by Dr. Michael Karin. IfnarF/F mice were crossed with albumin-Cre mice to obtain IfnarΔliver mouse line. Unless specified, the mice were weaned at 4 weeks of age and fed a diet of normal mouse chow. All mouse-related experiments were conducted in compliance with the guidelines of the Laboratory Animal Center of the National Health Research Institutes and used approved animal protocols (National Health Research Institutes–Institutional Animal Care and Use Committee 100013, 102126, 103104, 104006, and 106051).
HCC Tumor Model and Two-Thirds PHx
Two-week–old male mice were injected with diethylnitrosamine (DEN; 25 mg/kg; Sigma, St. Louis, MO) intraperitoneally, and HCC tumor development was examined at the age of 8 months.
The blood collected from mouse tail tip was used to measure glucose level with Accu-Chek Instant glucometer (Roche, Basel, Switzerland).
Antibodies, Chemicals, and Reagents
Antibodies for studying forkhead box O (FoxO) 1, FoxO3, epidermal growth factor receptor, Met, cyclin D1, cyclin A2, cyclin B1, p27, p21, and p15 were purchased from Cell Signaling Technology (Danvers, MA). Cyclin E1 was obtained from Santa Cruz Biotechnology (Dallas, TX). Tubulin antibodies were purchased from Sigma-Aldrich (St. Louis, MO). Adipose differentiation-related protein (ADRP) antibody was purchased from PROGEN Biotechnik Gmbh (Heidelberg, Germany). The anti-mouse APOL9 antibody was raised in rabbits (Animal Health Research Institute, Taiwan) immunized with purified His-tagged APOL9 protein. Insulin and intralipid emulsion were purchased from Sigma-Aldrich.
Paraffin-embedded liver sections were subjected to antigen retrieval with Target Retrieval Solution (S1700; DakoCytomation, Carpinteria, CA) at 95°C for 45 minutes and stained with anti–Ki-67 antibody (GeneTX, Hsinchu, Taiwan) by methods as described previously.
The Ki-67 staining images were taken under a microscope and quantified by ImageJ version 1.50i software (NIH, Bethesda, MD; https://imagej.nih.gov/ij, last accessed March 26, 2016). For FOXO1 staining, the antigen retrieval was performed with citrate buffer, pH 6.0, with 0.1% Tween-20 at 80°C for 6 hours and stained with anti-FOXO1 antibody (GeneTX). The hematoxylin and eosin staining for the tissue sections was performed by the pathology core laboratory in National Health Research Institutes.
Quantitative Real-Time PCR Analysis
cDNA synthesis and quantitative real-time PCR were performed as described previously.
Oil Red O Staining and Lipid Droplet Fractionation
Frozen liver sections were fixed with formalin for 10 minutes and washed in tap water. After being rinsed in 60% isopropanol in water, the liver sections were stained with 1.5% oil red O (Sigma) in isopropanol and water mixture (3:2) for 20 minutes and further counterstained with hematoxylin. Lipid droplet fractionation was performed by the method described previously,
with minor modification. In brief, 300 mg mouse livers was lysed in HLM buffer (20 mmol/L Tris-HCl, pH 7.4, 1 mmol/L EDTA, 10 mmol/L sodium fluoride, and 1× Roche protease inhibitor) by grinding with a glass Dounce homogenizer and centrifuged for 10 minutes at 1000 × g, 4°C. The post-nuclear supernatant was mixed with one-third volume of ice-cold HLM containing 60% sucrose and layered on the bottom of the ultracentrifuge tube for SW 41 Ti rotor. A total of 5 mL of HLM containing 5% and 0% sucrose was further layered on top sequentially. Flotation of lipid droplets was performed 30 minutes at 160,000 × g, 4°C. Fractions (1 mL/fraction) were collected from top to bottom and subjected to immunoblotting.
A total of 10 to 100 μg cell or liver lysates, prepared with radioimmunoprecipitation assay buffer (10 mmol/L Tris-HCl, pH 8.0, 140 mmol/L NaCl, 1 mmol/L EDTA, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS), was subjected to SDS-PAGE gel electrophoresis and immunoblotting with specific primary antibodies and horseradish peroxidase–conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA). The blot was developed with SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Scientific, Rockford, IL). All the images used for quantification were within the linear range.
IFN-I Signaling Enhances Liver Tumor Development
As HCC formation is closely associated with repetitive liver injuries and hepatocyte regeneration processes, whether IFN-I signaling has an active role in HCC development was examined. Carcinogen DEN can induce HCC development in young male mice with one injection.
To test the role of IFN-I in HCC development, two-week–old male WT and Ifnar−/− mice were injected with DEN (25 mg/kg) intraperitoneally, and HCC development was examined at the age of 8 months (Figure 1A). Ifnar−/− mice are deficient in IFN-α receptor expression. A total of 100% DEN-injected WT male mice developed approximately 20 liver tumors on average (Figure 1B). In contrast, Ifnar−/− mice had a slightly lower incidence rate (70%) and lower tumor multiplicity (six tumors on average) with a smaller size (Figure 1, B and C), suggesting that basal IFN-I signaling might have an active role in promoting HCC development in the absence of virus infection. Interestingly, IFN-I activation by the transient expression of hepatitis C virus full-length nonstructural protein 5B expression with adenoviral vector
could further enhance HCC development in a DEN-induced tumor model in 4-week–old mice (Supplemental Figure S1).
To examine microscopic alterations in the tumor-harboring livers, the liver tissue of the DEN-treated mice at 8 months old was subjected to hematoxylin and eosin staining (Figure 1D). In addition to the presence of tumor nodules, more lipid vacuoles were shown in some of the WT liver sections than in the Ifnar−/− liver sections (Figure 1D and Supplemental Figure S2). In addition to STAT1-dependent ISG activation, IFN-I signaling also engages PI3K, leading to AKT activation, which inhibits FOXO transcription factor activity via proteasome-dependent degradation.
The liver sections of the DEN-treated mice were therefore stained for FOXO1 by immunohistochemistry. As shown in Figure 1E, the FOXO1 expression was found sporadically in the nontumor region and enhanced in some tumor regions. The FOXO1 expression was significantly stronger in the Ifnar−/− tumor nodules than in the WT tumor nodules.
A Basal Level of IFN-I Signaling Is Present in the Liver
To examine whether IFN-I signaling is activated in a normal physiological state in the liver, IFN-I mRNA expression in adult WT and Ifnar−/− mice was evaluated by quantitative real-time PCR. As shown in Figure 2A, Ifna9 was the main IFN-I expressed in adult mouse liver. Most other IFN-Is, including Ifnb, a1, a2, a4, a5, a6, a7, a8, a11, a12, a13, a15, and a16, were at a low level or not detected in the liver. Ifng (type II IFN) mRNA expression was also low in unchallenged liver tissues. No difference was observed in IFN-I mRNA expression between the WT and Ifnar−/− liver. IFN-I mRNA expression in mouse spleen was also evaluated (Figure 1A). The Ifna6, a9, and a13 mRNA expression levels were detected in the spleen. As lymphocytes are abundant in the spleen, Ifng mRNA expression in the lymphocytes was much higher than in the liver. The results suggest a high level of Ifna9 expression in the naïve mouse liver, which may activate IFN-I signaling constantly.
To evaluate whether IFN-I signaling is active in the liver during postnatal liver development, the expression of an interferon-stimulated gene, Apol9, was examined using quantitative real-time PCR. Apol9 mRNA expression was significantly increased in the 4-week–old WT mouse liver but was much lower in the Ifnar−/− mouse liver (Figure 2B). The APOL9 protein expression was also detected in the 4-week–old WT liver but was undetectable in Ifnar−/− livers (Figure 2C). The results suggest that activation of IFN-I signaling in the WT liver began at a young age. Interferon-γ receptor knockout (Ifngr−/−) adult mice expressed a comparable amount of Apol9 mRNA, as in naïve WT mouse liver (Supplemental Figure S3), suggesting that type II IFN signaling did not contribute to Apol9 mRNA expression in naïve mouse liver.
To examine whether activation of IFN-I signaling in the liver modulates the PI3K/AKT/FOXO pathway, FOXO protein expression in 4-week–old mouse liver was evaluated by immunoblotting. As shown in Figure 2D, both FOXO1 and FOXO3 expression levels were slightly lower in 4-week–old WT liver relative to Ifnar−/− liver, suggesting that IFN-I might fine-tune PI3K/AKT/FOXO pathway activity in naïve liver. The negative regulation of IFN-I signaling on FOXO level was also observed in mouse liver with recombinant adenovirus to expressing hepatitis C virus active nonstructural protein 5B mutant, which can activate IFN-I production (Supplemental Figure S4).
To examine whether LPS translocated from intestine to liver could be the trigger of IFN-I signaling in naïve mice, the subnanometer gold nanocluster with modified methyl motifs nanoparticles that block LPS and LPS-associated inflammation
were used. As shown in Figure 2E, WT livers treated with subnanometer gold nanocluster with modified methyl motifs (75 mg/kg) for 2 or 3 days showed a 40% reduction in APOL9 protein level, suggesting that LPS/TLR4-dependent IFN expression participated in IFN-I activation in the liver.
IFN-I Signaling Enhances Basal Hepatocyte Proliferation in Naïve Mice
To evaluate whether IFN-I signaling modulates hepatocyte proliferation during postnatal development, liver tissue was stained to examine the proliferation marker Ki-67 using immunohistochemistry. As shown in Figure 3, A and B , liver tissue from 2-week–old mice contained high levels of proliferating cells, and these cells were significantly reduced in liver tissue obtained from 4-week–old weaning mice. The number of Ki-67+ cells in young (aged 2 and 4 weeks) mice was not significantly different between the WT and Ifnar−/− mice, but Ifnar−/− mice had less proliferating cells in adult naïve mice (aged 8 or 12 weeks), as quantified in Figure 3B. Hepatocytes are diploid after birth and differentiate into mature hepatocyte cells with multiple chromosome copies (polyploidy) to fulfill the needs for protein synthesis and metabolic functions. Mature hepatocytes, therefore, have a large nuclear size.
Quantification of the nuclear size of proliferating cells (Figure 3C) indicated that Ki-67+ hepatocytes in 2-week–old mouse liver had an average nuclear size of 1000 arbitrary units based on pixel size. The differentiated hepatocytes in adult mice were a mixed population ranging from 1000 to 3000 arbitrary units in size, with an average size of 2000 arbitrary units. Although the average nuclear size of the proliferating hepatocytes was not significantly different between the WT and Ifnar−/− in any age group, in 4-week–old mice, Ki-67+ hepatocytes with a nuclear size >2000 arbitrary units were more abundant in Ifnar−/− livers than in WT livers (Figure 3C), suggesting that the differentiation program in Ifnar−/− liver might start earlier than in WT mice. Assessment of cell cycle–associated cyclins by real-time quantitative PCR in different ages of mice indicated significantly higher levels of E2f1 and Ccnd1 mRNA in 4-week–old WT mice relative to Ifnar−/− mice (Supplemental Figure S6). Decreased proliferation and early onset of hepatocyte differentiation in Ifnar−/− mice were further confirmed in 4-week–old mice challenged with the recombinant adenovirus expressing hepatitis C virus nonstructural protein 5B active mutant (Supplemental Figure S7). Taken together, these data suggest that IFN-I signaling participates in the regulation of the balance of hepatocyte proliferation and differentiation during postnatal liver development.
Ablation of IFN-I Signaling Delays Hepatocyte Regeneration
To evaluate the role of IFN-I signaling in hepatocyte proliferation in the absence of pathogen invasion, the hepatocyte regeneration process was monitored in WT and Ifnar−/− liver after two-thirds PHx. Two-thirds PHx is a surgery that removes two-thirds of the mouse liver to evaluate its regeneration ability.
As shown in Figure 4A, WT mice had extensive proliferating cells in the liver at 48 hours after PHx and few proliferating cells on day 7 after surgery. In contrast, the Ifnar−/− liver had significantly lower numbers of Ki-67+ cells at 48 hours after surgery. The regeneration process was still ongoing on day 7 after liver PHx in the Ifnar−/− liver (Figure 4A), which may be because liver mass had not reached homeostasis. To sustain metabolic homeostasis, the liver size is maintained in proportion to the body size.
The liver weight/body weight ratio was significantly elevated on day 7 in WT mice but not in Ifnar−/− mice (Supplemental Figure S8). In line with the slow proliferation rate, Ifnar−/− liver expressed lower amounts of Ccna2 and Ccnb1 mRNA at 48 hours following PHx (Figure 4B). Reduced cyclin A2 and B1 protein expression at 48 hours after hepatectomy in the Ifnar−/− liver was also observed (Figure 4C). To control cell cycle progression, cell cycle kinases are negatively regulated by members of the Kip/Cip family of cell cycle kinase inhibitors, including p27. p27 expression remained at a higher level in Ifnar−/− liver at 48 hours following PHx (Figure 4C). Interestingly, fewer proliferating cells and more cell cycle kinase inhibitor p27 and p21 accumulation were also observed in Ifnar−/− liver compared with the WT liver during the regeneration time after virus infection (Supplemental Figures S7 and S9). Similarly, compensatory liver regeneration, induced by DEN challenge, was also decreased in Ifnar−/− liver (Supplemental Figure S10). Taken together, the data suggest that hepatic IFN-I signaling positively modulates compensatory regeneration processes in the liver.
IFN-I Signaling Enhances Hepatic Lipid Metabolism during Regeneration
FOXOs are transcription factors involved in the regulation of metabolism and cell fate decisions in responding to various metabolic conditions, growth factor stimulation, and stress.
FOXOs are translocated to the nucleus to activate target gene expression. FOXO1 expression was found predominantly in the nucleus in naïve WT and Ifnar−/− mice (Supplemental Figure S11). Metabolic changes, including hypoglycemia and transient hepatic lipid droplet accumulation, occur before cell proliferation during the liver regeneration process.
As the activation of IFN signaling modulates the FOXO level in the liver (Figure 2D and Supplemental Figure S4), IFN signaling was suspected to modulate metabolic alterations during the regeneration process. Therefore, the change of metabolic genes involved in gluconeogenesis and lipid metabolism after PHx was examined by real-time PCR (Figure 5, A–D ). G6pc gene encodes glucose-6-phosphatase, a catalytic subunit that is an enzyme functioning in gluconeogenesis and glycogenolysis. The baseline G6pc mRNA expression was slightly higher in Ifnar−/− liver compared with the expression level in WT liver. Although both significantly reduced at 6 hours after PHx, the G6pc expression level was kept significantly higher in Ifnar−/− liver than in WT liver (Figure 5A). As lipid metabolism can be regulated at various levels, such as lipid uptake, de novo lipogenesis, fatty acid oxidation, and lipid export, the change of gene expression related to these processes was examined. Genes involved in de novo lipogenesis [acetyl-CoA carboxylase (Acc)] and lipid export [apolipoprotein B (Apob)] were activated in WT liver at 6 hours after PHx but not in Ifnar−/− liver (Figure 5, B and C). The expression level of peroxisome proliferator-activated receptor-α gene (Ppara), which is involved in fatty acid oxidation, was significantly down-regulated in WT liver at 6 hours after PHx compared with the expression level in Ifnar−/− liver (Figure 5D). The results suggest that the IFN signaling modulates hepatic metabolism during the regeneration process.
Whether IFN-I signaling participates in the transient lipid accumulation during liver regeneration was further evaluated. A low amount of small lipid droplets were present in the livers of both WT and Ifnar−/− naïve mice, as indicated by oil red O staining (Figure 5E). High amounts of large lipid droplets accumulated in the WT liver at 24 hours after PHx (Figure 5F). In contrast, only the smaller sized lipid droplets increased in Ifnar−/− liver (Figure 5F). The results suggest a role of IFN-I signaling in lipid accumulation during the liver regeneration process.
ADRP is one of the lipid droplet-associated membrane proteins that are required for stabilization of the lipid droplet structure.
The protein expression of ADRP was highly elevated at 24 hours after PHx in WT livers but to a lesser extent in Ifnar−/− livers (Figure 6A), which is consistent with the oil red O staining result. To examine whether APOL9 functions in lipid accumulation in the liver, lipid droplets in the liver were purified on a discontinuous 0% to 20% sucrose gradient, in which lipid droplets were enriched in the top three fractions (Supplemental Figure S12).
As shown in Figure 6B, both ADRP and APOL9 were present in the lipid-enriched fractions separated from WT mouse liver from 24 hours after PHx, but not in the lipid fractions from naïve WT mouse liver. Lower ADRP amounts were induced in liver lysate and lipid fractions separated from Ifnar−/− mice liver from 24 hours after PHx (Figure 6C). Taken together, the results suggest that IFN-I signaling might participate in the metabolic regulation during the regeneration process.
IFN-I Signaling Modulates Glucose Homeostasis
Glucose and lipid metabolic alterations occur after PHx to initiate the liver regeneration process. As lipid metabolism was dysregulated in Ifnar−/− liver, whether glucose metabolism is also affected by IFN-I signaling was further evaluated. The average fasting blood glucose level in WT mice was 110 mg/dL and was elevated to 140 mg/dL in Ifnar−/− mice (Figure 7A). With the high serum glucose, the insulin level was also slightly elevated in Ifnar−/− mice at baseline (Supplemental Figure S13A). The serum insulin level was not significantly different between WT and Ifnar−/− mice after PHx (Supplemental Figure S13B). Ifnar−/− mice did not show insulin-resistant phenotype in the insulin tolerance test and glucose tolerance test (Supplemental Figure S13C). Transient hypoglycemia occurred in WT mice after PHx surgery, and the blood glucose level returned to normal within 24 hours (Figure 7B). In contrast, the blood glucose level in Ifnar−/− mice decreased at PHx and was kept at a low level at 24 hours after PHx. As dextrose supplementation has been known to have an inhibitory effect on liver regeneration,
the slow regeneration phenotype in Ifnar−/− mice might be due to the high blood glucose level and dysregulation of catabolic metabolism in these mice. Whether lowering blood glucose level could accelerate the regeneration process in Ifnar−/− mice was examined. In Ifnar−/− mice treated with insulin (0.75 U/kg, intraperitoneally) at 2 hours after PHx, the kinetics of blood glucose level in insulin-treated Ifnar−/− mice became similar to the WT mice (Figure 7B). The number of Ki-67+ hepatocytes in the insulin-treated Ifnar−/− mice was also significantly increased (Figure 7C). The transient hepatic lipid droplet accumulation might be triggered by hypoglycemia to maintain glucose homeostasis. Therefore, whether lipid supplement affects blood glucose level and regeneration was examined in Ifnar−/− mice. On the i.p. injection of 20% intralipid emulsion (soybean oil, 100 mg/kg), blood glucose was significantly elevated (Figure 7D) and liver regeneration was enhanced (Figure 7E) in Ifnar−/− mice. Taken together, these results suggest that IFN-I signaling regulates glucose homeostasis to support a timely regeneration process.
Liver-Specific IFN Signaling Regulates Metabolism to Initiate Regeneration
To address whether parenchymal cells in liver are the main cell type receiving IFN signaling during liver regeneration, the conditional knockout mice (IfnarΔliver; IfnarF/F × albumin-Cre+) lacking IFN-I receptor in albumin-expressing parenchymal cells, including hepatocytes and cholangiocytes, were established. As shown in Figure 8A, the Apol9 mRNA expression level in IfnarΔliver liver was significantly lower than in WT (IfnarF/F) liver. IfnarΔliver mice had a slightly higher baseline blood glucose level (Figure 8B). On PHx, the blood glucose level in IfnarΔliver mice decreased at 6 hours after PHx and remained at a low level at 24 hours (Figure 8C). In contrast, the blood glucose level in WT mice decreased at 6 hours after PHx and gradually increased at 24 hours after PHx. Similar to Ifnar−/− mice, IfnarΔliver mice had less amount of lipid droplet accumulation at 24 hours after PHx and less hepatocyte proliferation compared with the WT mice (Figure 8, D and E). Taken together, IFN-I signaling modulates metabolic alterations in parenchymal cells and facilitates liver regeneration.
In the present study, a basal level of IFN-I signaling was identified in the mouse liver, leading to constant activation of ISG expression and FOXO degradation. The low level of IFN-I activation enhanced baseline proliferation and compensatory liver regeneration, which might be dependent on the regulation of glucose and lipid metabolism in hepatocytes. Moreover, the constant hepatic IFN-I activation enhanced HCC development. The study reveals the possible contribution of IFN-I signaling to the regulation of glucose and lipid metabolism, hepatocyte proliferation, and HCC development in the absence of infection.
IFN-I has been used to treat several types of cancer,
and its ability to inhibit cell growth at high concentrations can be easily observed in cell culture. The positive role of IFN-I signaling in hepatocyte proliferation identified in this study might be intensity- and context-dependent, providing the hepatocyte with a capacity to tolerate stress and facilitate subsequent proliferation. In the absence of infection, trace amount of LPS translocated from the intestine is one of the potential stimuli to induce a low level of IFN production to modulate hepatocyte physiology, including metabolic homeostasis.
Signals transduced directly from IFN receptor or downstream ISG expression might participate in shaping liver homeostasis. IFN-I signaling activates STAT1, PI3K, and p38 pathways. Lack of these signaling pathways can all potentially be involved in the slow proliferation and regeneration ability of Ifnar−/− hepatocytes. Genetic ablation of p38α in hepatocytes enhances chemical-induced injury, compensatory proliferation, and HCC development,
therefore, the role of p38 downstream of IFN-I signaling in the regulation of hepatocyte proliferation might not be dominant. FOXO1, a critical effector of the PI3K/Akt pathway, regulates multiple metabolic pathways, including gluconeogenesis and lipogenesis in the liver. The current findings suggest that the basal IFN-I activation in the liver regulates FOXO1, which might, therefore, modulate blood glucose level. The coordinated metabolic status in hepatocytes may provide a higher capacity to proliferate during the regeneration process. On the other hand, disruption of IFN-I signaling leads to an accumulation of FOXO1 and an unregulated blood glucose level. Dysregulation of FOXO1 and blood glucose level then affect subsequent lipid metabolism, which leads to delayed regeneration. In line with this, insulin treatment to lower the glucose and FOXO1 level could accelerate liver regeneration in the Ifnar−/− mice. Insulin binds to the insulin receptor, activating extracellular signal-regulated kinase and PI3/Akt–mechanistic target of rapamycin pathways, which may also contribute to cell proliferation.
Therefore, in addition to the modulation of the FOXO1 and glucose level, the insulin treatment might also potentially contribute to hepatocyte proliferation during liver regeneration in Ifnar−/− mice.
ISG expression downstream of the IFN-I/STAT1 pathway may also play a role to engage IFN-I signaling to glucose and lipid metabolism during liver regeneration. Although the exact role of the APOL family is not clear yet, APOL9, which is highly induced by IFN-I activation in the liver, may be involved in lipid droplet accumulation during hepatocyte regeneration. APOL9 expression is maintained at a high level in mouse liver after weaning and is diminished in Ifnar−/− mice, making APOL9 a good marker for tracing IFN-I signaling activity in the liver. Mice, aged 3 to 4 weeks, begin eating solid food during weaning and are exposed to various food-born substances, some of which may lead to direct IFN-I induction in the liver or through effects on the microbiota at the mucosal immune system. For instance, LPS reaching the liver through the portal vein activates TLR4 and induces IFN-I expression. Interestingly, Tlr4−/− mice also show delayed liver regeneration after partial hepatectomy,
Whether other ISGs downstream of IFN-I signaling modulate various metabolic homeostasis levels in the liver needs further investigation.
Transient hypoglycemia and lipid droplet accumulation in the liver are critical events to initiate liver regeneration after PHx, but the interplay between these physiological events remains unclear. Blood glucose level reduces right after PHx, which may be a trigger for lipid metabolism, including de novo lipogenesis. Transient lipid accumulation, induced by metabolic disturbance after PHx, may be used as an energy resource for liver regeneration. Lipid supplementation is also known to enhance liver regeneration. The role of IFN-I signaling in modulating glucose and lipid metabolism during the regeneration process may contribute to HCC development and will be worthy of further exploration.
We thank Dr. Michael Karin (University of California, San Diego) for providing Ifnar−/− mice and for critical suggestions.
M.-S.W., Y.-P.K., Y.-C.L., and D.-J.T. performed most of the experiments. C.-Y.L. and T.-H.C. conducted apolipoprotein L9 antiserum production. S.-Y.L. provided nanoparticles. W.-T.T. and P.-J.L. participated in animal work. M.-S.W., S.-Y.L., and T.-H.C. participated in experimental design, data interpretation, and manuscript preparation with G.-Y.Y.
Virus infection enhances hepatocellular carcinoma (HCC) formation. A: Four-week–old wild-type (WT) mice were challenged with Ad–nonstructural protein 5B full-length (Ad-NS5B-FL; 1 × 109 plaque-forming units/mouse) to induce transient expression of hepatitis C virus NS5B protein in the liver. Control mice were infected with Ad–green fluorescent protein (GFP). The mice were injected with diethylnitrosamine (DEN; 100 mg/kg, intraperitoneally) 2 days after the adenovirus challenge. Beginning when the mice were 2 months old, they received 0.07% phenobarbital in their drinking water until the end of the experiment. HCC incidence rate, tumor number, and tumor size were evaluated when the mice were 9 months old. B: Four-week–old WT mice were administered DEN (100 mg/kg, intraperitoneally) followed by the adenovirus challenge (Ad-GFP or Ad-NS5B-FL; 1 × 109 plaque-forming units/mouse) 2 days after the DEN injection. Hepatic tumor formation was examined when the mice were 9 months old. n = 8 to 10 (). ∗P < 0.05.
Lipid accumulation in tumor-harboring mouse livers. Wild-type and Ifnar−/− liver sections from the diethylnitrosamine-treated mice at 8 months old were subjected to hematoxylin and eosin staining. Lipid vacuoles in the liver sections were evaluated and scored. Representative images of scores 0 to 5 are shown. Scale bar = 100 μm.
Apolipoprotein L9 (ApoL9) expression is controlled by type I interferon signaling. ApoL9 mRNA expression levels in adult wild-type (WT), Ifnar−/−, and interferon-γ receptor knockout (Ifngr−/−) mice were examined with quantitative real-time PCR. n = 3. ∗P < 0.05.
Forkhead box O FOXO) degradation is activated by virus infection. Wild-type mice were challenged with the recombinant adenoviruses Ad–green fluorescent protein (GFP; control virus) or Ad–nonstructural protein 5B (NS5B)–Del to express the hepatitis C virus NS5B active mutant (1 × 109 plaque-forming units/mouse). Liver tissues were collected at 24, 48, and 72 hours after virus infection for immunoblotting. APOL9, apolipoprotein L9.
Interferon (IFN)-β treatment promotes forkhead box O (FOXO) degradation. Human hepatoma Huh7 cells were treated with various doses of IFN-β for 48 or 72 hours, and the cell lysates were subjected to immunoblotting.
Type I interferon signaling regulates cell cycle–related gene expression during postnatal liver development. Cell cycle–related gene expression in the liver obtained from wild-type (WT) or Ifnar−/− mice of different ages was examined with quantitative real-time PCR. n = 3. ∗P < 0.05.
Type I interferon signaling promotes the compensatory hepatocyte proliferation after virus infection. Four-week–old wild-type (WT) and Ifnar−/− mice were challenged with Ad–green fluorescent protein and Ad–nonstructural protein 5B–Del (1 × 109 plaque-forming units/mouse). A: The infected liver tissues were collected 72 hours after virus infection for hematoxylin and eosin (H&E) and Ki-67 staining. B: Ki-67+ cell counts and nuclear size were quantified. ∗∗P < 0.01. Scale bar = 10 μm (A). AU, arbitrary unit.
The liver regeneration was assessed by liver weight/body weight ratio. Wild-type (WT) and Ifnar−/− mice were subjected to two-thirds partial hepatectomy. The remaining liver lobes were collected and weighted at 48 hours and 7 days after surgery. The liver weight/mouse body weight ratio was calculated. ∗P < 0.05.
Type I interferon signaling affects cell cycle–related protein expression after virus infection. Four-week–old wild-type (WT) and Ifnar−/− mice were challenged with Ad–green fluorescent protein and Ad–nonstructural protein 5B–Del (1 × 109 plaque-forming units/mouse). Cell cycle–related gene expression in the infected liver tissue was also analyzed by immunoblotting. EGFR, epidermal growth factor receptor.
Type I interferon signaling accelerates the compensatory hepatocyte proliferation caused by diethylnitrosamine (DEN) treatment. Wild-type (WT) and Ifnar−/− mice were subjected to a DEN challenge (100 mg/kg). Liver tissues were collected 72 hours after the challenge for Ki-67 staining (A) and cyclin mRNA detection by quantitative real-time PCR (B). n = 3 (). ∗∗P < 0.01. Scale bar = 20 μm (A). AU, arbitrary unit.
Examination of hepatic forkhead box O 1 (FOXO1) expression and activation status by immunostaining. A: The frozen liver sections collected from naïve wild-type (WT) and Ifnar−/− mice were subjected to immunostaining with rabbit anti-FOXO1 antibody (GeneTX) at 4°C overnight and then incubated with a secondary antibody with Alexa Fluor 488 conjugate (Invitrogen, Carlsbad, CA) at room temperature for 1 hour. The tissue section was then further counterstained with DAPI for nuclear DNA. B: FOXO1 activation was quantified with the FOXO1 staining intensity [integrated density (IntDen)/nuclear]. ∗∗P < 0.01. Scale bar = 20 μm (A).
Lipid fractionation by a discontinuous sucrose gradient. Gradient fractions were subjected to Western blot analysis with adipose differentiation-related protein (ADRP) antibody (lipid droplet marker), PDI (endoplasmic reticulum marker), and EEA1 (endosome marker) antibody detection. Arrow indicates full-length ADRP.
Blood insulin level, insulin tolerance test (ITT), and glucose tolerance test (GTT). A: Insulin level in naïve mouse serum was measured by Insulin Mouse ELISA Kit (Invitrogen, Carlsbad, CA) by following the manufacturer's instructions. B: Serum insulin level collected from mice at various times after partial hepatectomy was measured by enzyme-linked immunosorbent assay. C: Starved mice were intraperitoneally injected with insulin (1.2 U/kg) for ITT or glucose (1.5 g/kg) for GTT. Blood glucose level was monitored every 15 minutes. n = 4 to 5 mice (C). ∗∗P < 0.01. WT, wild type.
Antigen-presenting cell function in the tolerogenic liver environment.