Potential conflict of interest: Nothing to report.
The liver has a remarkable regenerative capacity, allowing recovery following injury. Regeneration after injury is contingent on maintenance of healthy residual liver mass, otherwise fulminant hepatic failure (FHF) may arise. Understanding the protective mechanisms safeguarding hepatocytes and promoting their proliferation is critical for devising therapeutic strategies for FHF. We demonstrate that A20 is part of the physiological response of hepatocytes to injury. In particular, A20 is significantly upregulated in the liver following partial hepatectomy. A20 protects hepatocytes from apoptosis and ongoing inflammation by inhibiting NF-κB. Hepatic expression of A20 in BALB/c mice dramatically improves survival following extended and radical lethal hepatectomy. A20 expression in the liver limits hepatocellular damage hence maintains bilirubin clearance and the liver synthetic function. In addition, A20 confers a proliferative advantage to hepatocytes via decreased expression of the cyclin-dependent kinase inhibitor p21waf1. In conclusion, A20 provides a proliferative advantage to hepatocytes. By combining anti-inflammatory, antiapoptotic and pro-proliferative functions, A20-based therapies could be beneficial in prevention and treatment of FHF. (HEPATOLOGY 2005;42:156–164.)
Necrosis and apoptosis of hepatocytes are critical pathological features associated with liver injury. Hepatocyte apoptosis is a feature of viral hepatitis, ischemic liver injury, sepsis, cholestasis, extensive surgical resection and of exposure to hepatotoxic substances.1 Massive hepatocyte loss results in fulminant hepatic failure (FHF).2 Only 14% of patients diagnosed with FHF recover with medical therapy. Orthotopic liver transplantation (OLT) has dramatically improved the fate of these patients (49% undergo OLT), yet 37% die while awaiting OLT.3 This gloomy picture is balanced by the unique capacity of the liver to regenerate. Hepatocyte proliferation leads to the full recovery of liver function and mass 1 to 2 weeks following surgical, viral, or chemical hepatic loss, provided a critical and healthy liver mass is maintained.4, 5 We propose that protecting hepatocytes from apoptosis and promoting their proliferation are two strategies that could beneficially have an impact on FHF.
A20 is a key element of the cellular response to injury and inflammation. A20 is a cytoplasmic seven zinc finger protein that is upregulated by the proinflammatory transcription factor NF-κB in most cell types, including hepatocytes.6–9 In endothelial cells, A20 has a dual cytoprotective function. It is anti-inflammatory through inhibition of NF-κB in a negative feedback loop and is antiapoptotic through inhibition of the caspase cascade at the level of initiator caspase 8.10, 11 In hepatocytes, direct and indirect evidence suggest that A20 is dominant in the hierarchy of anti-inflammatory and antiapoptotic defense mechanisms. A20 knockout mice are born cachectic and die within 3 weeks of birth as a consequence of unfettered liver inflammation.12 A20 protects mice from lethal FHF following treatment with D-galactosamine (D-gal) and lipopolysaccharide (LPS) by decreasing hepatocyte loss and safeguarding liver function.9 The impact of A20 on hepatocyte proliferation and liver regeneration is unknown.
A model of liver failure and necrosis following incrementally more extensive liver resections has been established in rats.13 Our goal was to adapt this progressive liver resection model to mice engineered to express A20 in the liver and analyze the impact of heightened A20 expression on liver regeneration, liver function, and survival.
We generated recombinant adenovirus encoding A20 (rAd.A20) or the control gene, β-galactosidase, (rAd.βgal) as described.10 We purchased mouse hepatocytes (NMuli, CRL-1638) from ATCC (Manassas, VA). These hepatocytes were infected with rAd. at a multiplicity of infection (MOI) of 25 plaque-forming units (pfu) per cell to achieve high expression of the transgene in more than 98% of cells 48 hours after infection with minimal toxicity.9 All experiments used two control groups, a noninfected (NI) group and a rAd.β-gal–infected group to control for adenoviral toxicity.
Experimental Liver Resection Model.
A20 expression in the liver was achieved by penile vein injection of 1×109 pfu of rAd. in 100 μL of normal saline. Optimal transgene expression was observed 5 days after injection in 30% to 40% of hepatocytes. Liver resections were conducted on day 5 following rAd. administration in 8-week-old BALB/c mice weighing 25 to 30 grams (Taconic, Germantown, NY).9 A standard two-third hepatectomy comprised removal of the lateral left, medial left, and medial right lobes. Our experiments employed a more extensive series of resections. An “extended” hepatectomy comprised 78% of the total liver mass by resecting the lateral right lobe in addition to a two-third resection. A “radical” hepatectomy comprised 87% of the total liver mass by resecting the quadrate lobe and the pyriform process, sparing only the caudate lobe. We took extreme care to avoid blood loss and maintain euvolemia and normal body temperature in these animals. Animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals. All procedures were performed according to the recommendations of the institutional animal care committee.
Serum Determinations From Experimental Animals.
We measured circulating levels of interleukin (IL)-6 and tumor necrosis factor (TNF) from mouse sera using commercially available ELISA (Endogen Inc., Woburn, MA). We measured aspartate aminotransferase (AST) levels in mouse sera with the Infinity Reagent System (Sigma, St. Louis, MO). We measured total bilirubin using a colorimetric assay (Sigma). Citrated plasma was separated, frozen and the prothrombin time (PT) determined at the comparative coagulation diagnostic laboratory of the Cornell University College of Veterinary Medicine (Ithaca, NY).
Histology and Immunochemistry.
Liver tissue samples were recovered and fixed in 10% buffered formalin. Five-μm sections were stained with hematoxylin and eosin for morphological examination and proliferating cell nuclear antigen (PCNA; Santa Cruz Biotechnology, Santa Cruz, CA), Ki67 (Dako, Glostrup, Denmark) and p21waf1 (BD Biosciences, San Jose, CA) staining were performed with monoclonal antibodies. Apoptosis was evaluated using the ApopTag assay (Oncor, Gaithersberg, MD). We used appropriate secondary antibodies and isotype specific negative controls for each antibody. Two investigators independently read 3 high-power fields (HPF) per slide in a blinded fashion.
Western Blot Analysis of IκBα Degradation.
We prepared protein extracts from NI, rAd.β-gal– and rAd.A20-infected NMuli hepatocytes and expression of cytoplasmic IκBα was detected by Western blot analysis using a polyclonal rabbit anti-IκBα antibody (Santa Cruz Biotechnology) as described.10
Analysis of Apoptosis.
Apoptosis of NMuli hepatocytes was induced by pretreatment with 333 nmol/L of Actinomycin D (Sigma) followed by 100 U/mL of recombinant human TNFα 30 minutes later (R&D Systems Inc. Minneapolis, MN). Cells are harvested 7 hours following TNF stimulation and DNA content analyzed on a FACScan using CELLQuest software (Beckton Dickson Immunocytometric Systems, San Jose, CA) as described.10
Cell Counts and S-phase Fraction Determination.
NI, rAd.β-gal–infected, and rAd.A20-infected NMuli hepatocytes were serum starved for 48 hours to synchronize them in the G0/G1 phase of the cell cycle. We performed cell counts before and 24 hours after feeding with 10% fetal calf serum and 10 ng/mL of human hepatocyte growth factor (HGF; R&D). The S-phase fraction is that portion of a FACScan generated histogram lying between the G1 (2N) and G2 (4N) peaks and represents cells synthesizing DNA.
Gene Chip Arrays.
We performed gene microarray analysis at the BIDMC Genomic Center using the Affymetrix GeneChipR Mouse Genome 430.2.0 Array from Affymetrix (Santa Clara, CA). This chip provides a comprehensive mouse genome expression covering over 39,000 transcripts. The scanned array images were analyzed by dChip.14 Samples were compared using the lower confidence bound (LCB) of the fold change (FC). If the 90% LCB of the FC between samples was above 2, the corresponding gene was considered differentially expressed. We extracted total mRNA from livers infected with rAd.β-gal and rAd.A20 using the RNAeasy extraction kit (Qiagen Inc., Valencia, CA). We included 6 mice per group and per timepoint. Two GeneChipR performed with RNA pooled from three animals were performed per group and per timepoint.
Analysis of A20 mRNA by Real-Time PCR.
Quantitative analysis of A20 messenger RNA (mRNA) before and 24 hours following extended liver resection was performed in NI and rAd.A20-infected mice by real-time polymerase chain reaction (PCR) using the ABI Prism 7900HT sequence detector system (Applied Biosystems, Foster City, CA). Human (to detect the rAd.A20 transgene) and mouse (to detect endogenous murine A20) A20 primers and probes (TaqMan MGB probes, FAM dye-labeled) were purchased from Applied Biosystems, Assays-on-Demand. Expression of the target gene was normalized to that of the housekeeping gene, 18S ribosomal RNA.
Quantitative data were expressed as mean ± SEM. If standard deviations (SD) were equal within the populations, an unpaired two-tailed Student t test was used to compare treatment groups. If SD values differed significantly, a nonparametric Mann-Whitney test was used. A P value of less than .05 was considered to be statistically significant.
Adenoviral-Mediated Expression of A20 in the Mouse Hepatocyte Cell Line NMuli Blocks TNF-Induced NF-κB Activation and Apoptosis.
Activation of the proinflammatory transcription factor NF-κB requires phosphorylation, ubiquination, and degradation of its inhibitor, IκBα.15 To evaluate the effect of A20 on TNF-induced NF-κB activation, we evaluated by Western blot analysis expression of IκBα in NI, rAd.β-gal and rAd.A20-infected NMuli hepatocytes before, 15 minutes and 120 minutes following addition of 100 U/mL of TNF. Data showed that A20 overexpression prevents IκBα degradation 15 minutes after addition of TNF (Fig. 1A, n = 3 experiments). Expression of the housekeeping protein β-tubulin was used to control for equal loading. These results are consistent with the known anti-inflammatory function of A20 via blockade of NF-κB.7–9
To assess the effect of A20 on hepatocyte apoptosis, we treated NI, rAd.β-gal– and rAd.A20-infected NMuli hepatocytes with 333 nmol of Actinomycin D to block transcription, followed by 100 U/mL of TNF and measured apoptosis 7 hours later by FACS analysis of DNA content. A20 overexpression significantly protected hepatocytes from TNF-mediated apoptosis. Only 5% ± 0.4% of rAd.A20-infected hepatocytes underwent apoptosis, as compared to 21.5% ± 0.8% of NI and 27% ± 0.5% of rAd.β-gal–infected hepatocytes (P < .0001; n = 4 experiments; Fig. 1B).
Mice Engineered to Express A20 Are Protected From Lethality Following Extended and Radical Hepatectomy.
To determine how A20 would affect hepatocytes in vivo, we engineered mice livers to transiently express A20 by rAd-mediated gene transfer. Then, we established two-third (67%), extended (78% of total liver mass), and radical (87%) hepatectomy models that resulted in successively poorer outcomes. Whereas 100% of mice survived following 67% hepatectomy in all groups (data not shown), we observed 50% (6 of 12) and 8% (1 of 12) survival in NI and rAd.β-gal–infected mice, respectively, following extended hepatectomy (Fig. 2). In marked contrast, rAd.A20-infected mice were significantly protected with 84% of mice surviving the procedure (2 of 12) (P = .009 vs. NI and P < .0001 vs. rAd.β-gal; n = 12 mice/group; Fig. 2).
A20 mRNA was significantly upregulated in the liver of surviving NI mice as determined by real time PCR. Relative A20 mRNA expression in the mouse liver increased from 1.03 ± 0.44 to 13 ± 7.2 at 24 hours following extended hepatectomy (P = .04; n = 6 mice/group; Fig. 3), indicating that A20 is part of the liver protective response to hepatectomy. Relative human A20 mRNA levels in mouse livers engineered to express A20 was 50- to 60-fold higher than those detected in NI mice following hepatectomy, reaching 733 ± 223 (P = .01; n = 6 mice; Fig. 3).
The protective effect of A20 was maintained following radical hepatectomy. We observed 0% (0 of 12) and 8% (1 of 12) survival in NI and rAd.β-gal–infected mice, respectively, following radical hepatectomy, whereas 42% of mice engineered to express A20 in the liver survived the procedure (5 of 12) (P = .01 vs. NI and P = .06 vs. rAd.β-gal; n = 12 mice/group; Fig. 2). All surviving mice were monitored for 8 weeks. No lethality, hepatomegaly, or liver tumors were observed irrespective of the experimental group.
A20 Does Not Alter Release of Priming Cytokines Following Extended Hepatic Resection.
TNF and IL-6 serum levels peak 1.5 hours and 6 hours following hepatectomy and are required for priming hepatocytes to proliferate in response to growth factors.16 We observed no significant difference in circulating IL-6 or TNF elevations among NI, rAd.β-gal– and rAd.A20-infected mice following extended hepatectomy (n = 3 mice/timepoint/group; Fig. 4A-B). These data demonstrate that expression of A20 in the liver does not affect the immediate levels of the two major priming cytokines following partial hepatectomy.
A20 Expression in Mice Suppresses Aminotransferase and Bilirubin Elevation Following Extended Hepatectomy and Maintains Hepatocytes Synthetic Capacity as Evaluated by Prothrombin Time.
We obtained serum for aminotransferase determination before and 6 and 24 hours following extended hepatectomy. We did not observe a statistical difference in the serum levels of AST amongst NI, rAd.β-gal– and rAd.A20-infected mice before hepatectomy. Six hours following extended hepatectomy, AST levels increased but still with no statistical difference among groups. There was a net tendency for higher AST levels in rAd.β-gal–infected mice (2251 ± 406 IU/mL vs. 1258 ± 81 in NI mice and 1043 ± 170 in rAd.A20-treated mice; n = 3 mice/timepoint/group; Fig. 5A). This finding is a likely a reflection of adenoviral toxicity. However, mice treated with rAd.A20 had significantly lower AST levels (478 ± 55 IU/mL) in comparison with NI (1661 ± 149 IU/mL; P = .003) and rAd. β-gal–infected mice (3555 ± 357 IU/mL; P = .001) 24 hours following extended hepatectomy, reflecting protection of hepatocytes from ongoing cellular damage (n = 3 mice/timepoint/group; Fig. 5A).
Mice infected with rAd.A20 also had significantly lower total serum bilirubin levels (2.1 ± 0.2 mg/dL) in comparison with NI (5.2 ± 0.4 mg/dL) and rAd.β-gal–infected mice (6.1 ± 0.7) 48 hours following extended hepatectomy, indicating preservation of bilirubin conjugation and excretion by hepatocytes (P< .001; n = 3 mice/timepoint/group; Fig. 5B).
Moreover, A20 expressing mice maintained good liver synthetic function and were protected against the coagulopathy associated with FHF. Mice expressing A20 had significantly lower PT (22.7 ± 2.5 seconds) in comparison to NI (71 ± 10 seconds) and rAd.β-gal–infected (>90 seconds) mice 48 hours following extended hepatectomy (P = .001; n = 3 mice/timepoint/group; Table 1).17 PT is dependent upon hepatocyte synthesis of the coagulation factors II, V, VII, X, and fibrinogen.18 As a further demonstration of unaltered liver function, mice infected with rAd.A20 remained euglycemic in the first 24 hours after resection compared with NI and rAd.β-gal–infected mice that suffered from hypoglycemia (data not shown). Hypoglycemia is often a premorbid observation.
A20-Mediated Protection Is Associated With Preservation of Liver Morphology and Increased Hepatocyte Proliferation.
We assessed liver architecture and morphology by histopathology before and 36 hours following extended hepatectomy in NI, rAdβ-gal– and rAd.A20-infected mice. This analysis was impossible following radical hepatectomy given the high mortality rate observed in control groups. All slides were reviewed in a blinded fashion by two investigators and a minimum of 3 HPF/slide, were analyzed. There was no significant difference among experimental groups prehepatectomy. A20-expressing livers had significantly less hepatic steatosis, sinusoidal congestion and hemorrhage 36 hours following extended hepatectomy (Fig. 6A-C). A20-expressing livers also had a significant advantage in proliferation, as evidenced by increased staining for proliferating cell nuclear antigen (PCNA), an endogenous marker of cellular proliferation, 36 hours following extended hepatectomy. We observed 63 ± 4 PCNA positive hepatocytes/HPF in A20-expressing livers compared with 5 ± 2/HPF in NI and 9 ± 3/HPF in rAd.β-gal–infected livers (P< .0001; n = 3 mice/group; Fig. 6D-F). Apoptosis was minimal and did not significantly differ among experimental groups 36 hours following extended hepatectomy (n = 3 mice/group; Fig. 6G-I).
Of note is that rAd.β-gal–infected mice faired worse for all measured parameters compared with NI mice, likely as a result of additional adenoviral toxicity.19 This result indicates that A20 overexpression not only protected mice from the untoward effects of extended and radical hepatectomy but also blunted the inflammatory insult of the adenoviral vector.
A20 Provides a Proliferative Advantage to Mouse Hepatocytes In Vitro.
We next investigated the molecular basis of the proliferative advantage provided by A20 in hepatocytes. NMuli hepatocytes were NI or infected with rAd.A20 or rAd.β-gal, then synchronized in the G0 phase of the cell cycle by 48 hours of serum starvation. We induced progression through the cell cycle by adding 10% fetal calf serum and 10ng/mL of recombinant human HGF. Cell counts before and 24 hours following addition of FCS/HGF demonstrated a significant proliferative advantage for A20 expressing hepatocytes. The number of rAd.A20-infected hepatocytes/well increased from 6.3 ± 0.6 × 105 to 20 ± 0.7 × 105 24 hours after addition of growth factors compared with an increase from 6.5 ± 0.1 × ± 105 to 13.2 ± 0.4 × × 105 in NI hepatocytes and from 5.5 ± 0.2 × 105 to 11 ± 0.7 × 105 in rAd.β-gal–infected hepatocytes (P < .001; n = 3 experiments done in triplicate; Fig. 7A).
In support of this, study of cell cycle progression by DNA content analysis one hour following addition of growth factors demonstrated a similar significant advantage for A20-expressing hepatocytes with regard to entry into the S-phase of the cell cycle. The percentage of NMuli hepatocytes in S-phase during starvation did not vary among experimental groups. One hour following treatment with 10 ng/mL of HGF and 10% fetal calf serum, 17.4% ± 2.1% of rAd.A20-infected hepatocytes had entered S-phase compared with 4.7% ± 0.5% and 7% ± 1.2% of NI and rAd.β-gal–infected hepatocytes respectively (P = .0014; n = 4 experiments done in duplicate; Fig. 7B).
A20 Promotes Proliferation In Vivo and Downregulates the Expression of the Cyclin-Dependent Kinase Inhibitor p21waf1.
We next analyzed gene expression in liver tissue taken from rAd.A20- and rAd.β-gal–infected mice before and 24 hours following extended hepatectomy. Duplicate samples of total RNA were probed using the Affymetrix GeneChip Mouse Genome 430 2.0 Array. We were interested in observing differences among genes regulating cell cycle progression. The most notable finding was a 2.4-fold lower expression of the cyclin-dependent kinase inhibitor (CDKI) p21waf1 in livers expressing A20 compared with livers expressing β-gal prior to resection (n = 2 gene microarrays/timepoint/group). This difference persisted in tissues retrieved 24 hours following liver resection with 3.1-fold lower p21waf1 expression in A20 compared with β-gal–expressing livers. This demonstrates that A20-treated animals are primed to enter the cell cycle even before resection via mitigation of the p21waf1 cell cycle brake.20
To confirm changes in gene expression, we performed immunohistochemical analysis on liver tissue taken from rAd.A20- and rAd.β-gal–infected mice before and 24 hours following extended hepatectomy (n = 6 mice/timepoint/group). Prior to resection, livers expressing A20 showed significantly fewer p21waf1-positive hepatocytes (1.2 ± 0.55/HPF) compared with β-gal–expressing livers (55 ± 8.1/HPF; P = .004; Fig. 8A). The pattern of p21waf1 expression inversely correlated with that of Ki67, 24 hours after resection. Ki67 is a confirmed marker of cell proliferation in the liver and is expressed during the G1, S, G2, and M phases of the cell cycle. Twenty-four hours following extended hepatectomy, mice infected with rAd.A20 had significantly more (107 ± 7.9/HPF) Ki67 positive hepatocytes compared with those infected with rAd.β-gal (7.7 ± 5.64/HPF; P < .0001; Fig. 8B). p21waf1 and Ki67 were detected in the nucleus and expression was mainly in hepatocytes.
Hepatocyte proliferation is an adaptive response to acute liver injury and is important in maintaining liver function in many chronic liver diseases.5 Following injury, surviving hepatocytes must meet increased metabolic demands while proliferating to restore liver mass. Liver regeneration is a complex process including priming of hepatocytes to enable them to respond to growth factors and enter the cell cycle, progression beyond the restriction point in the G1 phase of the cycle and apoptosis of excessive hepatocytes to restore optimal liver mass to body mass ratio.5 A strategy designed to augment hepatocyte priming and accelerate cell cycle progression while promoting physiological function could prove crucial to improving survival following acute liver injury. This study demonstrates that the NF-κB–dependent, cytoprotective gene A20 is well poised to achieve each of these benefits.
Our data demonstrate that A20 has potent anti-inflammatory and antiapoptotic properties in hepatocytes treated with TNF, which should improve hepatocyte survival and function following extensive hepatectomy. TNF is a critical contributor to inflammation and hepatocyte loss in acute liver injury including after liver resection.21 TNF is also an important proximal co-mitogen required for priming and proliferation of hepatocytes during liver regeneration.22 Lack of signaling through the TNF receptor 1 (TNF-R1) in TNF-R1 knockout mice inhibits DNA replication and results in significant mortality following partial hepatectomy.22 The hepatotrophic functions of TNF following partial hepatectomy relate, at least in part, to activation of NF-κB, which orchestrates the expression of a large number of genes that are essential for liver growth, differentiation, regeneration and protection from apoptosis.23–27 Blockade of NF-κB by overexpression of its inhibitor IκBα in hepatocytes, results in increased apoptosis and decreased mitotic index.28 Targeted gene disruption of the dominant transcription member p65(RelA) or of the IκBα kinase (IKKβ), leads to embryonic lethality with severe liver degeneration.25, 29 Our results demonstrate that expression of A20 by itself, despite its potent NF-κB inhibitory effect, substitutes for NF-κB activation in protecting hepatocytes and promoting liver regeneration. Overexpression of A20 in hepatocytes is sufficient to enhance liver regeneration while maintaining liver function following extended and radical hepatectomy.
Importantly, expression of A20 in hepatocytes does not alter expression of TNF or IL-6 following liver resection. IL-6 is a downstream hepatotrophic effector cytokine induced by TNF. Lethality following partial hepatectomy in TNF-R1 knock out mice is reversed by administration of IL-6 and associated activation of STAT3 (signal transduction and activation in T cells).22 Mice with targeted disruption of IL-6 suffer liver failure and defective regeneration.30, 31 Liver regeneration is enhanced by a designer IL-6/soluble IL-6 receptor fusion protein with superagonistic IL-6 properties or upon overexpression of STAT3.32, 33 In addition, IL-6 prevents mortality following fatty liver transplants in rats.34
These results indicate that A20 is either “the” NF-κB–dependent gene required for liver regeneration or that other NF-κB–independent proliferative pathways triggered by TNF and IL-6 are enabled in hepatocytes expressing A20. TNF and mainly IL-6–independent pathways play a minor role in hepatocyte proliferation.35 This pro-proliferative function for A20 in hepatocytes seems independent from its antiapoptotic properties. Analysis of cell cycle progression in vitro in nonapoptotic conditions show that hepatocytes expressing A20 demonstrate increased response to growth factors with higher number of proliferating cells and earlier entry in the cell cycle when compared to control hepatocytes.
We demonstrate that the pro-proliferative effect of A20 is associated with decreased expression of the CDKI p21waf1 and is likely related to it. p21waf1 is a universal inhibitor of cyclins/cyclin-dependent kinase (CDK) activity, limiting progression beyond the G1 phase of the cell cycle.36 DNA synthesis, cyclins/CDK activation and S phase gene expression occur earlier in transgenic p21waf1 knockout mice than in wild type mice following partial hepatectomy.37 In contrast, transgenic mice with hepatic overexpression of p21waf1 demonstrate diminished liver size and markedly impaired hepatocyte proliferation following resection.38
Little is known about the mechanisms regulating p21waf1 expression in the liver. In compensatory hepatocyte replication after partial hepatectomy or carbon tetrachloride intoxication, p21waf1 is upregulated in a biphasic manner through a p53-independent mechanism.39, 40 We believe that p21waf1 expression may also be p53-independent in our model, as p53 levels in microarray analysis were neither modified by liver resection nor by A20 expression (data not shown). Other factors are involved in modulating p21waf1 expression including the transcription factors p150, CEBPα, and CEBPβ and altered mRNA and protein stability.36, 39, 41, 42 Whether the novel ubiquitin ligase and deubiquinating enzymatic activities of A20 are involved in modifying any of these components remains to be determined.43
Importantly, decreased p21waf1 mRNA levels and protein expression in livers engineered to express A20, as analyzed by gene chip arrays and immunohistochemistry, is observed prior to liver resection, which suggests that the proliferative advantage provided by A20 to hepatocytes is effective prior to any injury hence is likely independent from the antiapoptotic effect of A20. Indeed, we demonstrate a reciprocal relationship between low p21waf1-positive hepatocytes prior to resection and high Ki67-positive hepatocytes after resection.
In summary, we demonstrate that A20 is part of the regenerative response in hepatocytes. A20 promotes hepatocyte proliferation likely through decreased expression of the cell cycle “brake” p21waf1. This novel function is independent from its antiapoptotic effect and occurs despite inhibition of NF-κB activation. This, added to the antiapoptotic and anti-inflammatory effects of A20 in hepatocytes, accounts for metabolic and synthetic advantages as well as improved survival of A20-treated mice following extended and radical hepatectomy. Indeed, it is widely appreciated that recovery of liver mass following hepatectomy requires a metabolic compromise between differentiated function and organ regrowth and that hepatic failure after resection in clinical settings is more common when the organ is diseased.44 We recognize that expression of an antiapoptotic gene such as A20 might disturb the regulatory apoptosis required for involution of excessive liver mass or promote tumor formation. These problems can be avoided with limited expression of A20 as achieved by rAd. (maximal by 5 to 7 days and substantially decreased by 14-21 days).9 We did not observe excessive hepatomegaly or liver neoplasia during the long term follow-up of mice treated with A20.
We propose that short term A20-based therapies may be highly beneficial for patients presenting with severe liver damage but who still have a fraction of viable hepatocytes. Protecting this reduced functional liver mass in the face of ongoing inflammation would meet metabolic demands and allow enough time for regeneration. Expression of A20 is particularly promising for reducing the donor graft size necessary for living donor liver transplantation (LDLT) and hepatocellular transplants.45–47 LDLT has emerged as a solution to ease the shortage of available organs for transplantation and has achieved remarkable success in the pediatric population.48, 49 However, success is limited in adults by the size of the graft that can be safely harvested.47 Our results clearly demonstrate that A20-based therapy is a legitimate strategy to help ease this shortage and allow for safer extended liver resection in the treatment of neoplasia.
We wish to acknowledge Drs. Hasan Otu and Towia Lieberman from the Genomics Center at BIDMC for their help with Gene Array analysis, Mrs. Christina Mottley from the Division of Immunology for her help with real time PCR, and Pr. Fritz H. Bach from the BIDMC for his continuous support and for critical review of this manuscript.