HIP/PAP accelerates liver regeneration and protects against acetaminophen injury in mice

Authors

  • Hanh-Tu Lieu,

    1. Institut National de la Santé et de la Recherche Médicale U-370; Faculté de Médecine, Université Paris 5; Hôpital Necker, Institut Pasteur de Paris, Paris, France
    Search for more papers by this author
    • Hanh-Tu Lieu and Frédèric Batteux contributed equally to the study.

  • Frédéric Batteux,

    1. Laboratoire d'Immunologie, Hôpital Cochin, Faculté de Médecine, Université Paris 5, Paris, France
    Search for more papers by this author
    • Hanh-Tu Lieu and Frédèric Batteux contributed equally to the study.

  • Marie-Thérèse Simon,

    1. Institut National de la Santé et de la Recherche Médicale U-370; Faculté de Médecine, Université Paris 5; Hôpital Necker, Institut Pasteur de Paris, Paris, France
    Search for more papers by this author
  • Alexandre Cortes,

    1. Institut National de la Santé et de la Recherche Médicale U-370; Faculté de Médecine, Université Paris 5; Hôpital Necker, Institut Pasteur de Paris, Paris, France
    Search for more papers by this author
  • Carole Nicco,

    1. Laboratoire d'Immunologie, Hôpital Cochin, Faculté de Médecine, Université Paris 5, Paris, France
    Search for more papers by this author
  • Flora Zavala,

    1. Institut National de la Santé et de la Recherche Médicale U-580; Institut Necker, Hôpital Necker, Paris, France
    Search for more papers by this author
  • Alain Pauloin,

    1. Biologie Cellulaire et Moléculaire, Institut National de la Recherche Agronomique, Jouy-en-Josas, France
    Search for more papers by this author
  • José Guilherme Tralhao,

    1. Institut National de la Santé et de la Recherche Médicale U-370; Faculté de Médecine, Université Paris 5; Hôpital Necker, Institut Pasteur de Paris, Paris, France
    Search for more papers by this author
  • Olivier Soubrane,

    1. Laboratoire de Recherches Chirurgicales, Hôpital Cochin, Faculté de Médecine, Université Paris 5, Paris, France
    Search for more papers by this author
  • Bernard Weill,

    1. Laboratoire d'Immunologie, Hôpital Cochin, Faculté de Médecine, Université Paris 5, Paris, France
    Search for more papers by this author
  • Christian Bréchot,

    1. Institut National de la Santé et de la Recherche Médicale U-370; Faculté de Médecine, Université Paris 5; Hôpital Necker, Institut Pasteur de Paris, Paris, France
    Search for more papers by this author
  • Laurence Christa

    Corresponding author
    1. Institut National de la Santé et de la Recherche Médicale U-370; Faculté de Médecine, Université Paris 5; Hôpital Necker, Institut Pasteur de Paris, Paris, France
    • Institut National de la Santé et de la Recherche Médicale U-370, Faculté de Médecine, Université Paris 5, Hôpital Necker, Institut Pasteur de Paris, 156 rue de Vaugirard, 75742 Paris cedex 15, France
    Search for more papers by this author
    • fax: (33) 1 40 61 55 81


  • Potential conflict of interest: Nothing to report.

Abstract

Human hepatocarcinoma-intestine-pancreas/pancreatic-associated protein HIP/PAP is a secreted C-type lectin belonging to group VII, according to Drickamer's classification. HIP/PAP is overexpressed in liver carcinoma; however, its functional role remains unclear. In this study, we demonstrate that HIP/PAP is a paracrine hepatic growth factor promoting both proliferation and viability of liver cells in vivo. First, a low number of implanted hepatocytes deriving from HIP/PAP-transgenic mice (<1:1,000) was sufficient to stimulate overall recipient severe combined immunodeficiency liver regeneration after partial hepatectomy. After a single injection of HIP/PAP protein, the percentages of bromodeoxyuridine-positive nuclei and mitosis were statistically higher than after saline injection, indicating that HIP/PAP acts as a paracrine mitogenic growth factor for the liver. Comparison of the early events posthepatectomy in control and transgenic mice indicated that HIP/PAP accelerates the accumulation/degradation of nuclear phospho–signal transducer activator transcription factor 3 and tumor necrosis factor α level, thus reflecting that HIP/PAP accelerates liver regeneration. Second, we showed that 80% of the HIP/PAP-transgenic mice versus 25% of the control mice were protected against lethal acetaminophen-induced fulminate hepatitis. A single injection of recombinant HIP/PAP induced a similar cytoprotective effect, demonstrating the antiapoptotic effect of HIP/PAP. Comparison of Cu/Zn superoxide dismutase activity and glutathione reductase-like effects in control and transgenic liver mice indicated that HIP/PAP exerts an antioxidant activity and prevents reactive oxygen species-induced mitochondrial damage by acetaminophen overdose. In conclusion, the present data offer new insights into the biological functions of C-type lectins. In addition, HIP/PAP is a promising candidate for the prevention and treatment of liver failure. (HEPATOLOGY 2005;42:618–626.)

C-type lectins are cation-dependent carbohydrate-binding proteins with a common carbohydrate recognition domain of 115–130 amino acid residues.1 These proteins exhibit various functions such as complement activation,2 endocytosis,3 cell recognition,4 and morphogenesis.5 It has been recently reported that the viral EP153R protein is endowed with antiapoptotic properties in vitro, which is unexpected for a C-type lectin.6

The hepatocarcinoma-intestine-pancreas/pancreatic-associated protein (HIP/PAP) is a human 16-kd secreted protein whose complementary DNA has been cloned in humans7, 8 as well as rats, mice, and hamsters under the names PAPI or peptide 23, Reg-2 or RegIIIβ, and INGAP,9 respectively. The HIP/PAP and the lithostatine/RegI-related genes are classified in group VII of the C-type lectin gene superfamily and encode small secreted proteins with a single carbohydrate recognition domain linked to a signal peptide.1, 10 Because the physiological functions of group VII lectins remain unclear, structure–function relationships are difficult to establish. The human HIP/PAP protein has been shown to bind lactose,11 promote rat hepatocyte adhesion in primary culture, and interact with the proteins of the extracellular matrix.12In vitro, the rat HIP/PAP has been reported to prevent cell death in neuronal primary cultures13 and pancreatic cells,14 promote the growth of epithelial intestinal cells,15 and stimulate DNA synthesis in Schwann cells.16 Recently, INGAP peptide (the sequence of which is derived from the hamster homologue to HIP/PAP) has been shown to enhance nerve outgrowth from explanted dorsal root ganglia.9In vivo, HIP/PAP can function as an acute phase reactant in human pancreatitis.8 We have discovered HIP/PAP by differential screening of a human hepatocarcinoma,7 and we have shown its frequent overexpression in liver tumors and regenerative hepatic/ductular diseases.17 These results also suggest that HIP/PAP could regulate both viability and proliferation in hepatocytes. Thus we have recently shown that human HIP/PAP expressed in hepatocytes exhibits mitogenic and an antiapoptotic properties in primary cultures and that liver regeneration after partial hepatectomy (PHX) is stimulated in HIP/PAP transgenic mice.18

We investigated the clinical applications of C-type lectins using two complementary approaches: (1) stimulation of the hepatic regeneration in severe combined immunodeficiency (SCID) mice transplanted with HIP/PAP-expressing hepatocytes or in C57Bl/6 mice injected with HIP/PAP protein, and (2) prevention of acetaminophen (APAP)-lethal acute liver failure (ALF) in HIP/PAP-transgenic mice or in C57Bl/6 mice preventively injected with HIP/PAP protein. Liver regeneration after PHX and APAP-induced ALF are well-known models for investigating the proliferation and apoptosis of liver cells in vivo. Both models are relevant to human liver failure resulting from chronic, acute clinical conditions and surgery. Indeed, the main causes of ALF are viral infections, drugs (generally APAP with suicidal intent) or alcohol intoxication, and major surgical resection.

Abbreviations

HIP/PAP, hepatocarcinoma-intestine-pancreas/pancreatic-associated protein; PHX, partial hepatectomy; SCID, severe combined immunodeficiency; APAP, acetaminophen; ALF, acute liver failure; BrdU, bromodesoxyuridine; STAT3, signal transducer activator transcription factor 3; TNF-α, tumor necrosis factor α; PBS, phosphate-buffered saline; IL-6, interleukin 6; TGF, transforming growth factor; LT-β, lymphotoxin β.

Materials and Methods

Animals.

HIP/PAP transgenic mice previously described,18 C57BL/6J mice (wild-type of HIP/PAP transgenic mice), and SCID mice (Janvier, L'Arbresles, France) (8 to 10 weeks old; 20–22 g) were used after an adaptation of 1 week to their environment. The animals received human care in compliance with institutional guidelines, and experiments were approved by the appropriate board of our institution. Mice were anesthetized with xylazine (Bayer, Leverkusen, Germany) and ketamine (Biomérieux, Lyon, France) dissolved in saline (0.9% NaCl).

Hepatocyte Implantation and Detection of HIP Expression.

To minimize cell rejection, 7-week-old female SCID mice were used as recipients for hepatocytes isolated from male HIP/PAP transgenic or C57BL/6 mice as previously described.19 Recipient SCID mice were held for 30 days to allow proliferation and reorganization of donor hepatocytes into the liver parenchyma. PHX was then performed. HIP/PAP expression was evaluated via immunohistochemistry using anti-HIP/PAP antibodies and via reverse-transcriptase polymerase chain reaction as previously described18 with the following primers: HIP/PAP 19 sense, 5′-cgccccgggatgctgcctcccatggccctg-3′; HIP/PAP 101 antisense, 5′-cgcgaatccgcccatgatgagttgcacaccaaac-3′. Housekeeping mouse β globine gene expression was evaluated with the following primers: β globine sense: 5′-tgaccggcttgtatgctatc-3′; β globine antisense, 5′-cagtgtgagccaggatatag-3′.

Liver Regeneration After PHX.

Mice were starved for 12 hours and hepatectomized at 8 A.M. as previously described.18 Six hundred nanograms of purified recombinant HIP/PAP protein diluted into 100 μL of saline20 was injected into the spleens of SCID mice 36 hours after PHX and into the spleens of C57Bl/6 mice during surgery. Overall, liver regeneration was evaluated by weighing the livers and assessing bromodeoxyuridine (BrdU) incorporation into nuclei following intraperitoneal injection of 60 mg kg−1 body of BrdU 2 hours before dissection. BrdU-labeled nuclei and mitosis were scored as previously described.18 Western blottings of whole regenerative liver extracts were hybridized with cyclin A antibody (sc-596, Santa Cruz Biotechnology, Santa Cruz, CA) at 1:1,000 as previously described.18

APAP Intoxication.

Drug-induced ALF was achieved via intraperitoneal injection of a lethal dose of 1,000 mg/kg (APAP1,000) or a sublethal dose of 500 mg/kg (APAP500) of APAP as previously described.21 C57BL/6J female mice were treated intravenously with either 600 ng or 1,200 ng of purified HIP/PAP protein or saline per mouse 1 hour before the administration of APAP. In sublethal experiments (APAP500), animals were sacrificed 12 hours after APAP intoxication. Blood was collected for measurement of serum transaminase activities. Cytosolic (S100) and mitochondrial (HM) fractions of mouse liver were isolated as described by Yang et al.22 Protein concentration was measured using the BSA microbiuret assay (Pierce, Bezons, France).

The level of cytokine expression was evaluated via RNase protection assay as previously described.19 Signal transducer activator transcription factor 3 (STAT3) activation on nuclear extracts was assessed either using the TramsAM kit (Active Motif, Rixensart, Belgium) according to the manufacturer's instructions or via Western blotting18 with anti–phospho-STAT3 antibodies (Cell Signaling Technology, Beverly, MA) on nuclear extracts prepared as previously described23 with the following modifications: liver tissues (10 mg) were homogenized in 800 μL of buffer A (10 mmol/L HEPES [pH 7.6], 25 mmol/L KCl, 0.15 mmol/L spermine, 0.5 mmol/L spermidine, 1 mmol/L EDTA, 2 mol/L sucrose, 10% glycerol, 1 mmol/L dithiothreitol (DTT), 0.1 mmol/L phenylmethylsulfonyl fluoride (PMSF), 0.2% Nonidet P 40, 0.5 mol/L NaF, 2 mmol/L sodium orthovanadate, 1 mmol/L sodium glycerophosphate, 3 μg/mL aprotinin). The homogenate was filtered through a 70-μm membrane, loaded to 1 mL of buffer A, and centrifuged for 30 minutes at 38,000 rpm to pellet nuclei (rotor TLA55, Beckman Coulter, optima Max ultracentrifuge, Fullerton, CA). The nuclei were lysed in 80 μL of buffer B (400 mmol/L NaCl, 10 mmol/L HEPES [pH 7.9], 1.5 mmol/L MgCl2, 0.1 mmol/L EGTA, 0.5 mmol/L DTT, 5% glycerol, 0.5 mmol/L PMSF, 0.5 mmol/L NaF, 2 mmol/L sodium orthovanadate, 2 mmol/L sodium glycerophosphate, 3 μg/mL aprotinin). After 30 minutes at 4°C, the resuspended nuclei were centrifuged for 30 minutes at 16,000g. Ten micrograms of soluble nuclear protein (Biorad protein assay) were used for the TramsAM assay and the Western blotting.

Tumor Necrosis Factor α Immunoassay.

Liver remnants were removed and homogenized in extraction buffer (Tris 50 mmol/L [pH 7.2], NaCl 150 mmol/L, EDTA 5 mol/L, Triton-X-100 1%, and a protease inhibitor cocktail) (complete mini; Roche, Basel, Switzerland). The homogenate was shaken on ice for 30 minutes and centrifuged at 3,000g and 15°C for 15 minutes. Tumor necrosis factor α (TNF-α) was detected using a commercial ELISA kit (Quantikine M, TNF-α; R&D Systems, Abingdon, Oxon, UK).

Measurement of Liver Glutathione Content and Superoxide Dismutase Activities.

Glutathione levels in liver cytosolic (S100) or mitochondrial (HM) fractions were measured using the method of Baker et al.24 Superoxide dismutase activities in liver cytosolic or mitochondrial fractions were measured by the method of Beauchamp and Fridovich.25

Lipid Peroxidation Assay.

The concentrations of 4-hydroxyalkenals and malondialdehyde were measured in cytosolic fractions using a lipid peroxidation assay from Calbiochem (Paris, France). The level of lipid peroxydation was expressed as the amount of 4-hydroxyalkenals + malondialdehyde per milligram of protein.

Cytochrome c Determination.

Cytochrome c was determined on the cytosolic fractions or on the enriched mitochondrial fractions of the livers. Levels of cytochrome c in each fraction were determined via ELISA (R&D, Oxon, UK). Results were expressed as the ratio of cytochrome c (in picograms) per microgram of protein in mitochondrial fraction versus cytochrome c (in picograms) per microgram of protein in cytosolic fraction.

Caspase-3 Activity.

Caspase-3 activity was measured in liver cytosolic fractions as previously described.21 Results are expressed as pmol/min/mg protein.

Serum Biochemical Analysis.

Alanine aminotransferase and aspartate aminotransferase were quantified using a standard clinical automatic analyzer (Hitachi, type 747).

Histological Examination.

Liver tissues were fixed overnight at 4°C in 4% paraformaldehyde in phosphate-buffered saline (PBS), then dehydrated and embedded in paraffin. Sections (4 μm) of liver tissue were stained with hematoxylin-eosin and red picrosirius for histological examination. Percentages of lysed hepatocytes per field were blind-evaluated with a reticle.

Statistical Analysis.

The statistical significance of variations in liver weights, BrdU incorporation, mitotic indexes, densitometric cyclin A levels, and parameters in APAP intoxication was determined using the Mann-Whitney U test. The statistical significance of differences in the survival rates following APAP intoxication was calculated using the Kaplan-Meier method. A Student t test was used for comparisons of STAT3, and TNF-α ELISA tests. Quantitative variations were considered positive when the P value was less than .05.

Results

HIP/PAP Stimulates the Proliferation of Liver Cells and Accelerates Liver Regeneration.

HIP/PAP-expressing or wild-type hepatocytes were implanted into the spleens of SCID mice that were hepatectomized 1 month later (Fig. 1A). Thirty days after the implantation, reverse-transcriptase polymerase chain reaction was used to confirm human HIP/PAP gene expression in recipient SCID livers (Fig. 1B). A number of 0.75 × 106 implanted hepatocytes represented less than 1:1,000 hepatocytes in the recipient livers. HIP/PAP expression was detected in a restricted number of liver cells without preferential distribution in the liver sections (portal or centrolobular area). Macroscopic evaluation of the livers of SCID mice implanted with HIP/PAP hepatocytes or wild-type hepatocytes showed that there is no alteration in liver growth, and liver weights were similar before partial PHX (Fig. 1C). The HIP/PAP gene was still expressed in liver sections obtained 46 hours and 7 days after hepatectomy, indicating that implanted cells persisted and retained gene expression during liver regeneration (data not shown). Seven days after PHX, livers implanted with hepatocytes derived from HIP/PAP transgenic mice showed a marked increase in liver weight (P < .0008) (Fig. 1C). Thus a low number of implanted HIP/PAP-expressing hepatocytes (<1:1,000) was sufficient to stimulate overall recipient SCID liver regeneration. We then investigated the paracrine effect of HIP/PAP by injecting HIP/PAP into the spleens of SCID mice 36 hours after PHX and by evaluating the liver weights 7 days after PHX. The results showed a 10% increase in liver weight in mice injected with 600 ng of HIP/PAP versus saline (P = .0022) (Fig. 1D). Thus the HIP/PAP lectin acts as a paracrine growth factor on overall liver regeneration. The HIP/PAP paracrine mitogenic effect was then tested by comparing BrdU incorporation and mitosis after intrasplenic injection of HIP/PAP (600 ng/mouse; 30 μg/kg) or saline (100 μL) and simultaneous surgery. The C57Bl/6 mice were sacrificed 46 hours after PHX. At this time, medians for BrdU incorporations were 23% (minimal 2.5; maximal 37) and 7.3% (minimal 1; maximal 46) for HIP/PAP and saline injection, respectively (P = .0168). Medians for mitotic indexes were 0.7% (minimal 0; maximal 6.3) and 0% (minimal 0; maximal 5.7) for HIP/PAP and saline injection, respectively (P = .0362) (Fig. 2A). We have been able to further confirm our observations using cyclin A, a reliable marker of the S phase entry. Indeed, we observed that cyclin A levels were statistically higher in HIP/PAP-injected mice than in saline-injected mice as evaluated via Western blot analysis (P < .05) (Fig. 2B). Thus the percentages of BrdU-positive nuclei, the number of mitoses, and the levels of cyclin A were statistically higher after injection of HIP/PAP than after saline injection, indicating that HIP/PAP acts as a paracrine mitogenic growth factor for liver. Furthermore, a single postsurgery injection is enough to stimulate overall regeneration.

Figure 1.

HIP/PAP is a paracrine hepatic growth factor. We took advantage of the human HIP/PAP transgene expression to monitor the fate of implanted liver cells via (A) immunohistochemical staining of liver sections with anti-HIP/PAP antibodies (original magnification ×400) and (B) reverse-transcriptase polymerase chain reaction with human HIP/PAP primers. (C) Box plots of the SCID liver weights 1 month after implantation with HIP/PAP or wild-type hepatocytes at surgery (P value not significant) and 7 days after PHX (P < .0008). (D) Box plots of the SCID liver weights injected with 600 ng HIP/PAP protein or saline (P = .0022). HIP/PAP, hepatocarcinoma-intestine-pancreas/pancreatic-associated protein; PHX, partial hepatectomy.

Figure 2.

HIP/PAP is a paracrine mitogenic factor, activating liver regeneration. (A) Immunohistochemical detection of S phase and mitotic hepatocytes (arrows) in liver sections of C57Bl/6 mice injected with saline or HIP/PAP (600 ng) (original magnification ×40). Box plots of the distribution of BrdU incorporation (*P = .0168), and mitotic index (**P = .0362) in saline (white bars) and HIP/PAP (shaded bars) injected mice 46 hours after PHX are shown, as is a graph defining each mouse by its BrdU incorporation (% on Y axis) and mitotic index (% on X axis). There were more HIP/PAP-injected mice than saline-injected mice that traverse the S and mitosis phases at 46 hours after PHX. (B) Western blot analysis of the hepatic cyclin A levels in regenerative liver extracts prepared from C57Bl/6 mice injected with saline or HIP/PAP (600 ng). Box plots of the distribution of cyclin A levels (***P < .05) in mice injected with saline (white bars) and HIP/PAP (shaded bars) 46 hours after PHX are shown. BrdU, bromodesoxyuridine; HIP/PAP, hepatocarcinoma-intestine-pancreas/pancreatic-associated protein.

To investigate the molecular mechanisms underlying the improvement of liver regeneration, we examined whether HIP/PAP might modulate early signaling events, including the kinetics of activation of the downstream interleukin (IL-6) transcription factor STAT3 and liver cytokine expressions.26, 27 Using the model of hepatocyte implantation, we compared cytokine gene expression in SCID livers implanted with either HIP/PAP or wild-type hepatocytes at surgery or 46 hours later (Fig. 3A). Consistent with the known regulated expression of transforming growth factor (TGF)-β during liver regeneration, a similar increase in TGF-β expression was observed in SCID mice implanted with either HIP/PAP or wild-type hepatocytes 46 hours after PHX. Thus HIP/PAP has no effect on TGF-β expression. In contrast, expression of lymphotoxin β (LT-β) and TNF-α, which belong to the same functional family of cytokines, was lower in the SCID livers implanted with HIP/PAP hepatocytes than in livers implanted with wild-type hepatocytes. This result suggests that TNF-α and LT-β expressions are repressed or degraded earlier in SCID livers expressing HIP/PAP.

Figure 3.

(A) RNase protection assay. LT-β, TNF-α and TGF-β were simultaneously evaluated in a pool of four liver extracts of SCID mice implanted with HIP/PAP-expressing (lanes 1 and 2) or wild-type (lanes 3 and 4) hepatocytes during PHX (lanes 1 and 3) and 46 hours after PHX (lanes 2 and 4). Bands were detected via phosphoimaging, and the results were calculated as a ratio of the band of interest to the mean of the two bands corresponding to housekeeping genes encoding the large ribosomal protein and GAPDH. Two RNase protection assays were performed. (B) STAT3 activation was monitored in liver nuclear extracts of C57BL/6 (○) and HIP/PAP transgenic (▪) mice by using the TramsAM kit during the indicated time course of regeneration. Results are expressed as percentages of the positive control provided in the kit (nuclear extract prepared from HepG2 cells treated with IL-6). Results corresponded to means obtained in 4 mice (*P = .029, ** P = .04). Phospho-STAT3 antibodies were used in a Western blot analysis on liver nuclear extracts of wild-type and HIP/PAP transgenic mice. Normalization was visualized using a nonspecific coloration of the membrane with the Red Ponceau dye. (C) Effect of PHX on hepatic TNF-α levels. TNF-α production occurred earlier in HIP/PAP transgenic than in C57Bl6/ wild-type mice, with a bulk at 1 hour versus 3 hours after PHX, respectively (***P < .001 at 1 hour; n = 4 in each group). (D) Western blot analysis of the time course of hepatic cyclin A levels in a pool (n = 3 mice) of regenerative liver extracts prepared from HIP/PAP or C57Bl/6 mice. SCID, severe combined immunodeficiency; HIP/PAP, hepatocarcinoma-intestine-pancreas/pancreatic-associated protein; PHX, partial hepatectomy; LTβ, lymphotoxin β; TNFα, tumor necrosis factor α; IFNγ, interferon γ; TGFβ, transforming growth factor β; L32, large ribosomal protein; STAT3, signal transducer activator transcription factor 3.

We further observed HIP/PAP transgenic mice and C57Bl/6 mice to investigate the TNF-α and IL-6 downstream transcription factor STAT3 and comfort the effects of HIP/PAP on TNF-α/LT-β that were obtained in SCID livers. Activation of the transcription factor STAT3 was quantified using an ELISA test in which the DNA oligonucleotide consensus to STAT3 had been immobilized to the 96-well plate (Fig. 3B). After its direct phosphorylation, activated STAT3 was also evaluated via Western blotting on nuclear extracts with anti–phospho-STAT3 antibodies. Both approaches provided identical results: STAT3 was activated and inactivated earlier in HIP/PAP transgenic mice than in wild-type mice (P = .029 at 1 hour and P = .04 at 12 hours after PHX). Thus the accumulation/degradation time course of nuclear phospho-STAT3 was accelerated in HIP/PAP transgenic mice. Next, we investigated whether HIP/PAP influenced the time course of the production of TNF-α after PHX by evaluating TNF-α hepatic levels in both types of mice (Fig. 3C). TNF-α production occured earlier in HIP/PAP transgenic mice than in C57Bl/6 wild-type mice, with a bulk at 1 hour versus 3 hours after PHX (P < .001 at 1 hour).

TNF-α and IL-6 are two known regulators of the initial phase of liver regeneration that prime hepatocytes to enter the G1 phase of the cell cycle.26 However, the persistence of TNF-α/IL-6 expression and STAT3 activation are deleterious and delay the time course of regeneration.28 We recently showed that HIP/PAP counteracted the apoptotic effect of TNF-α combined with actinomycine D on hepatocytes in primary culture.18 Using either the model of HIP/PAP-expressing hepatocytes implanted into SCID mice or the model of HIP/PAP-transgenic mice, we showed that HIP/PAP modulates the effects of TNF-α, IL-6, and STAT3 during liver regeneration, suggesting a shortening of G1 responsible for the acceleration of liver regeneration. This conclusion was supported by the graph gained from BrdU incorporation and mitotic index for each mouse (Fig. 2C). Indeed, there are more HIP/PAP-injected mice than saline-injected mice that traverse the S and mitosis phases at 46 hours after PHX. Moreover, we further confirmed our observations with cyclin A levels evaluated via Western blotting during the time course of liver regeneration (Fig. 3D). We observed that cyclin A expression, a reliable marker of the S phase entry, occurred earlier in HIP/PAP-expressing livers than in wild-type livers, with a bulk at 46 versus 60 hours, respectively (Fig. 3D). In conclusion, using three different experimental models based on hepatocyte implantation, transgenesis, and injection of recombinant protein, we showed that HIP/PAP accelerates the time course of liver regeneration.

HIP/PAP Protects From APAP-Induced ALF and Exerts an Antioxidant Activity.

The second part of our work was to examine the in vivo antiapoptotic properties of HIP/PAP. For that purpose, we investigated the preventive effect of HIP/PAP protein in a murine model of APAP-induced ALF. In the mouse, as in humans, a large single dose of APAP causes massive centrolobular parenchymal destruction and hepatocyte death.29 Therefore, we tested the resistance of HIP/PAP transgenic mice against a lethal intraperitoneal dose of APAP. Eighty percent of HIP/PAP transgenic mice (males or females) survived for more than 24 hours versus 25% in the wild-type group (P = .0003) (Fig. 4A). Moreover, the intravenous administration of HIP/PAP protein into wild-type mice 1 hour before lethal APAP intoxication dose-dependently increased the survival of mice by 45% versus 25% after the administration of HIP/PAP 600 ng and saline, respectively, and 70% versus 20% after the administration of HIP/PAP 1,200 ng and saline, respectively (600 ng, P value not significant; 1,200 ng, P = .02). Thus, by using two different experimental models, based on transgenesis and injection of dose-dependent recombinant protein, we showed that HIP/PAP protects the mice against a lethal dose of APAP.

Figure 4.

HIP/PAP protein protects mice against APAP-induced acute liver failure and exerts an antioxidant activity. (A) Injection of a lethal dose of 1,000 mg/kg APAP. Left panel: the protection conferred by HIP/PAP on APAP intoxication was evaluated on male and female HIP/PAP transgenic mice (♦), and in C57BL/6 mice (▴). Negative controls are represented by PBS-injected C57BL/6 mice (▪). Middle and right panels: the protection conferred by recombinant HIP/PAP protein on APAP intoxication was evaluated in C57BL/6J female mice treated intravenously with either 600 ng (middle panel) or 1,200 ng (right panel) of purified HIP/PAP protein per mouse (▵) 1 hour before the induction of ALF via APAP. Negative controls are represented by PBS-injected C57BL/6 mice (▵) 1 hour before the induction of ALF via APAP. Survival rates were blindly evaluated for 24 hours. (B) Male and female C57BL/6 mice transgenic or not for HIP/PAP were intoxicated with a sublethal dose of APAP (APAP500). A control group of 10 mice transgenic or not for HIP/PAP was injected with saline. Twelve hours later, the animals were sacrificed, blood was collected for measurement of serum aminotransferase activities, and livers were removed for analysis of mitochondrial and cytosolic concentrations of glutathione, superoxide dismutase, and cytochrome c and for determination of cytosolic activities of caspase-3 and lipid peroxidation. (C) Microscopic views of livers 12 hours following the injection of a sublethal dose of APAP (hematoxylin-eosin staining). Upper left panel: nonintoxicated C57BL/6 mice. Upper right panel: nonintoxicated HIP/PAP transgenic mice. Lower left panel: C57BL/6 mice intoxicated with APAP. Lower right panel: HIP/PAP transgenic mice intoxicated with APAP. Apoptosis/necrosis scores appear in the lower left corner of each panel (original magnification ×200). HIP/PAP, hepatocarcinoma-intestine-pancreas/pancreatic-associated protein; PBS, phosphate-buffered saline; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GSH, glutathione; mito, mitochondrial; cyto, cytosolic; SOD, superoxide dismutase; 4-HNE, 4-hydroxyalkenals; MDA, malondialdehyde; APAP, acetaminophen.

A sublethal dose of APAP (500 mg/kg) caused massive hepatocyte injury 12 hours after intoxication in wild-type and transgenic mice, as evidenced by the significant increases in serum alanine aminotransferase and aspartate aminotransferase (P < .001). However, we have shown that the rises in alanine aminotransferase and aspartate aminotransferase levels were twofold lower (both P < .01) in HIP/PAP transgenic mice than in wild–type mice, reflecting the cytoprotective effect of HIP/PAP against acute liver injury (Fig. 4B). Moreover, histological examination of livers showed a threefold decrease in hepatocyte apoptosis/necrosis in HIP/PAP transgenic mice compared with wild-type mice, confirming the protective role of HIP/PAP against APAP-induced liver damages (Fig. 4C).

APAP overdose induces hepatocyte injuries through the increased production of reactive oxygen species associated with mitochondrial damages, such as the collapse of mitochondrial transmembrane potential, the release of cytochrome c into the cytoplasm, and caspase activation that leads to hepatocyte apoptosis.30 In addition, N-acetyl-p-benzoquinone-imine, a highly reactive molecule generated during APAP metabolism, depletes the intracellular pool of glutathione, a nonprotein thiol with both oxidant scavenger– and redox-regulating capacities.31 Consequently, during APAP intoxication, the imbalance between reactive oxygen species overproduction and antioxidant depletion leads to ALF.32 In wild-type and HIP/PAP transgenic mice, the sublethal dose of APAP induced significant increases in the release of cytochrome c (P < .001), caspase-3 activity (P < .001), and the mitochondrial lipid peroxidation products 4-hydroxyalkenals and malondialdehyde (P < .02). Significant decreases were also observed in cytosolic and mitochondrial glutathione (P < .001) and in cytosolic superoxide dismutase (P < .001). However, HIP/PAP transgenic mice intoxicated with APAP showed fewer oxidative alterations than intoxicated wild-type mice, as demonstrated by the higher levels of both cytosolic and mitochondrial glutathione content (both P < .001) and by the lower concentrations of 4-hydroxyalkenals plus malondialdehyde (P < .05). Cytochrome c release and caspase-3 activation were also reduced (P < .01 and P < .05, respectively) in the livers of transgenic versus nontransgenic mice. Finally, we observed a lower decrease in cytosolique Cu/Zn superoxide dismutase activity in HIP/PAP transgenic mice compared with wild-type mice during APAP intoxication (52.5 ± 7.8 vs. 12.1 ± 2.9, P < .001), whereas the decrease in mitochondrial Mn superoxide dismutase was equally observed in both types of mice. Interestingly, we observed that the cytosolique Cu/Zn superoxide dismutase activity is also higher in PBS-treated HIP/PAP transgenic than in PBS-treated C57Bl/6 mice (71.6 ± 8.9 vs. 43.7 ± 6.1; P < .05). Considering the structure of HIP/PAP protein, a C-type lectin, we hypothesize that this molecule could exert an antioxidant activity. The crystal structures of the Reg/lithostatine33 and HIP/PAP,34 both belonging to group VII, reveal changes in the Ca2+-binding sites observed in the other C-type lectins (e.g., mannose-binding protein), and no Ca2+ is found in the structure. Instead of Ca2+, we showed that HIP/PAP binds Zn2+ (allowing its dimerisation34) as well as Cu2+ (unpublished data). In addition, HIP/PAP contains seven cysteins, six of which are included in three disulfide bonds, instead of two in the other C-type lectins. Those structural differences could account for some of HIP/PAP's biological functions. HIP/PAP may exhibit a superoxide dismutase–like and a glutathione reductase–like activity, explaining its protective effect against reactive oxygen species–induced mitochondrial damages, subsequent cytochrome c release, and caspase-3 activation induced by APAP overdoses.30 Similar beneficial effects have been obtained using other antioxidant molecules such as N-acetyl-L-cysteine35 and the superoxide dismutase mimics MnTBAP30 and Mangafodipir21 during ALF induced by APAP.

Discussion

We have demonstrated that the human HIP/PAP C-type lectin is a hepatic growth factor acting through paracrine dual mitogenic and antiapoptotic properties. Upon a single injection, HIP/PAP markedly stimulates liver regeneration and protects against APAP-induced ALF. Most of the multifunctional C-type lectins are transmembrane proteins that have not yet been implicated in intracellular signaling pathways. In contrast, we have recently shown that HIP/PAP is able to interact with the regulatory RIIα subunit of protein kinase A,36 indicating that a fraction of the cellular pool of HIP/PAP escaped the secretory pathway. HIP/PAP stimulates the protein kinase A–dependent phosphorylation of the proapoptotic Bad factor in hepatocytes,18 and Reg-2 (the homologue of HIP/PAP in rats) also stimulates Akt in rat motoneurons.13 In this study, we show that HIP/PAP (1) accelerates the accumulation/degradation of nuclear phospho-STAT3 and hepatic TNF-α levels during the time course of liver regeneration and (2) exerts an antioxidant activity, preventing reactive oxygen species–induced mitochondrial damages, probably as a result of its superoxide dismutase–like and glutathione reductase–like activity.

Medical treatment of ALF is difficult, because few molecules are clinically available. The antioxidant molecule N-acetyl-L-cysteine is used because of its antiapoptotic properties, but it is not a mitogenic factor. Eventually, it has a preventive but not curative effectiveness in vivo. Both antiapoptotic and mitogenic properties are crucial to treat all types of ALF, because the lack of hepatocyte regeneration in the presence of an ongoing loss of hepatocytes is ultimately responsible for the severe prognosis of acute liver disease. It has been proposed that cytokines (e.g., TNF-α and IL-6) and growth factors (hepatocyte growth factor, TGF-α, and epidermal growth factor) stimulate liver regeneration. However, none of them has proved its efficacy without toxicity and pleiotropic side effects.37–39 Moreover, beneficial effects of hepatocyte growth factor on liver regeneration,27 acute liver injury,40, 41 and chronic liver failure42 require either iterative injections at short intervals, continuous intraportal infusion, or a gene transfer strategy43 to allow persistence of significant levels of expression. Furthermore, hepatocyte growth factor needs to be processed from an inactive precursor and is unstable in peripheral blood, with a half-life of only 3 to 5 minutes.44 HIP/PAP fulfills the requirements to treat ALF: it protects hepatocytes from TNF-α–induced apoptosis in vitro18 and prevents ALF in vivo. On the other hand, it also acts as a mitogen for hepatocytes in primary culture,18 accelerating the time course of STAT3 activation and of TNF-α–repression, and, as we now know, it is also a mitogen in vivo.

We have previously shown that HIP/PAP-expressing mice have a normal life span and do not develop tumors or macroscopic alterations during a 2-year follow-up, suggesting that HIP/PAP could be devoid of adverse effects.18 Moreover, we have demonstrated that a single injection of human HIP/PAP recombinant protein was sufficient to stimulate liver regeneration and to protect against acute liver failure, despite a plasma half-life of 4.8 ± 1.4 minutes reported for rat HIP/PAP (Reg-2, peptide 23/PAPI).45 Therefore, the present results highlight HIP/PAP as a promising candidate to treat liver failure through its dual action on the viability of hepatocytes and parenchymal regeneration. Finally, in ALF of viral or toxic origin, HIP/PAP is a potential substitute to liver transplantation that is presently the only effective treatment available. On the other hand, the lack of cadaveric grafts has led to the development of split-liver grafts from living donors.46, 47 Consequently, the small-for-size syndrome has become more frequent in recent years, and therapeutic compounds that can be clinically used to stimulate liver regeneration and/or counteract massive hepatocyte necrosis are greatly needed.

Acknowledgements

We would like to thank Anne Sophie Chaplaut for taking care of HIP/PAP transgenic mice and genotypic screening, Evelyne Souil for helping with the analysis of the histopathological study (from the anatomo-pathological platform of Université Paris 5), and Marie Mianowski for reviewing the manuscript.

Ancillary

Advertisement