| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institut National de la Santé et de la Recherche Médicale U-370, Institutes Necker-Pasteur, Université Paris V, 75742 Paris Cedex 15, France;
* Biologie Cellulaire et Moléculaire, Institut National de la Recherche Agronomique, Jouy-en-Josas, France;
Service d’Anatomo-pathologie, Hôpital Necker-Enfants Malades, Paris, France; and
Service d’Anatomo-pathologie, Hôpital Européen Georges Pompidou, Paris, France
1Correspondence: Institut National de la Santé et de la Recherche Médicale U-370, Necker-Pasteur Institutes, Universite Paris V, 156 rue de Vaugirard, 75742 Paris Cedex 15, France. E-mail: christa{at}necker.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
plus actinomycin-D-induced apoptosis. HIP/PAP anti-apoptotic effects against TNF- were clearly demonstrated when protein kinase A activity was specifically inhibited by KT5720, and HIP/PAP stimulated PKA-dependent phosphorylation of the proapoptotic Bad protein at Ser-112, suggesting that HIP/PAP may compete with cAMP to stimulate PKA activity. Overall, our results led us to propose a new role for a C-type lectin, HIP/PAP, as a hepatic cytokine that combines mitogenic and anti-apoptotic functions regarding hepatocytes, and consequently acts as a growth factor in vivo to enhance liver regeneration.—Simon, M.-T., Pauloin, A., Normand, G., Lieu, H.-T., Mouly, H., Pivert, G., Carnot, F., Tralhao, J. G., Bréchot, C., Christa, L. HIP/PAP stimulates liver regeneration after partial hepatectomy and combines mitogenic and anti-apoptotic functions through the PKA signaling pathway.
Key Words: HIP/PAP/Reg-2 • C-type lectin • TNF- • PKA
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2
). The second regenerative pathway is triggered after toxic injury, upon massive necrosis or carcinogenesis, when the proliferation of hepatocytes or biliary cells is impaired or slowed by the injury (3
4
5)
. Under these conditions, it has been proposed that "stem-like" cells proliferate and differentiate into hepatocytes and biliary epithelial cells, then repopulate the liver. In rodents, these so-called oval cells represent a heterogeneous cellular compartment in which well-defined subpopulations have yet to be isolated. In humans, the oval cell compartment may participate in repopulating the liver after acute massive necrosis and has been identified in chronic liver diseases (6
, 7)
. Finally, whether oval cells act as tumor precursors remains controversial, and the identification of specific markers may help to clarify this point (8
9
10
11)
.
The differential screening of a human hepatocellular carcinoma cDNA library using subtracted probes enabled us to identify abundant expression of the HIP/PAP gene in tumor but not in nontumor or normal livers (12)
. HIP/PAP protein expression was undetectable in normal mature hepatocytes, but certain ductular cells located at the interface between the portal tracts and the parenchyma were HIP/PAP immunoreactive in normal liver (13)
. Since it has been suggested that these ductular cells may represent potential progenitor cells with the characteristics described above, they may constitute potential precursors for a subset of liver cancers with hepatocytes or cholangiolar phenotypes. This observation, seen alongside the expression of HIP/PAP in tumor cells with hepatocyte or cholangiolar phenotypes, supports the suggestion that HIP/PAP may be implicated in hepatocytes or cholangiolar differentiation and proliferation.
We produced and purified human HIP/PAP (also called WAP-HIP) from the milk of lactating transgenic mice. We crystallized and resolved its structure as a C-type lectin, belonging to group VII of the C-type lectin family according to Drickamer’s classification (14
15
16)
. We showed that this lectin exhibits lactose binding activity and is capable of promoting rat hepatocyte adhesion and the binding of elements in the extracellular matrix. This HIP/PAP C-type lectin contains one carbohydrate recognition domain linked to an amino-terminal peptide signal driving secretion of the protein. Specific expression of the HIP/PAP gene in hepatocellular carcinoma causes the secretion of high levels of HIP/PAP protein into the serum in a large percentage of patients with primary liver cancers (75%). Because there is no correlation between HIP/PAP and -fetoprotein levels (13
, 17)
and because HIP/PAP protein levels are elevated in a smaller group of patients with cirrhosis but without detectable cancer (39%), HIP/PAP may represent a predictive marker for the early diagnosis of hepatocellular carcinoma (13)
.
The aim of this study was to understand whether the biological functions of HIP/PAP are related to its specific expression in tumor hepatocytes and potential progenitor ductular cells. We therefore examined the consequences of human HIP/PAP expression in transgenic mouse hepatocytes on the regulation of proliferation/apoptosis, both in vivo after partial hepatectomy and in vitro in hepatocytes in primary culture.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-) was obtained from R&D Systems (Minneapolis, MN, USA). Bovine growth hormone poly A fragment (nt 1021 to nt 1235) was from pcDNA 3.1 Invitrogen (Groningen, The Netherlands). Human HIP/PAP protein was purified from the milk of lactating transgenic mice as described previously (16)
.
Transgenic mice
The regulatory region of the mouse albumin gene (18)
was cloned upstream of the HIP/PAP gene fragment to drive human HIP/PAP gene expression specifically in the liver, as described in Fig. 1
. The entire NotI/KpnI linearized construct was microinjected into single-cell mouse zygotes of C57Bl/6xCBA hybrid strains in the Experimentation and Transgenesis Department (Villejuif, France). The 24 and 27 homozygous transgenic lines were developed from independent founders on genetic C57Bl/6 background. Animal welfare, conditions for animal handling before slaughter, and all experimental procedures were in line with French Ministry of Agriculture guidelines (dated 19 April 1988).
|
Partial Hepatectomy and BrdU incorporation in vivo
Liver resection represents 70% of the total liver mass, as described by Higgins and Anderson (19)
in 2-month-old mice. Animals received one intraperitoneal injection of 60 mg kg–1 body weight BrdU in 0.9% NaCl 2 h before dissection. They were killed 24, 36, 46, and 55 h posthepatectomy. Animals and livers were weighed, BrdU-labeled nuclei were scored after incubation with anti-BrdU antibody (clone Bu 20A Dako), and revelation was performed using the Universal LSAB2 horseradish peroxidase kit (Dako) with at least 20 low magnification (x10) microscope fields for each liver slide (Olympus BX60). More than 1600 nuclei were screened per slide.
Hepatocytes in primary culture
Primary mouse hepatocytes were isolated from 2- to 3-month-old mice, as described by Klaunig (20)
and Renton (21)
, with Liberase Blendzyme. Viable hepatocytes were purified using a low-speed, iso-density Percoll centrifugation method as described by Kreamer (22)
. Cells were resuspended in 199-medium containing penicillin, streptomycin, Fungizone, bovine serum albumin (0.1%), and fetal calf serum (10%) at densities of 2 x 105 and 4 x 105 for proliferation and apoptotic experiments, respectively, in Primaria plates. Cells were maintained at 37°C in a humidified atmosphere and the medium was changed after attachment to the plates for 2–3 h.
After attachment, the cells were rinsed once and cultured with the same medium containing no serum, then exposed to actinomycin D (ActD; 0.05 µg mL–1) plus TNF- at concentrations ranging from 0.2 to 40 ng mL–1 for 17–18 h, unless otherwise specified in the figure legends. For proliferation experiments, the medium was supplemented with 3,5,3'-triiodothyronine 5 10–8M, dexamethasone 10-7 M, Insulin 10 µg mL–1 (2 10-6 M), transferrin 5.5 µg mL–1, selenium 7 ng mL–1, pyruvate 20 mM, and fetal calf serum 5%.
DNA synthesis in primary culture hepatocytes
To measure DNA synthesis, BrdU (20 mM) was added for the last 16 h before evaluation. The hepatocytes were washed with PBS, fixed, and rendered permeable in 30:70 acetic acid/ethanol solution at -20°C for 30 min. Incorporated BrdU was localized using the BrdU Labeling and Detection kit II. Replicative DNA synthesis was measured by scoring the percentage of BrdU-labeled cells in at least 10 low magnification (x20) microscope fields for each sample (Olympus CK2). More than 1000 hepatocytes were screened per well.
Cell viability and evaluation of apoptosis in primary culture hepatocytes
Seventeen hours after the addition of TNF-, the monolayer was fixed with 4% paraformaldehyde for 20 min at room temperature and stained with Hoechst 33258 (0.5 µg mL–1). Apoptotic cells were examined at wavelengths between 350 and 460 nm using an Olympus BX60 inverted fluorescence microscope (Olympus America Inc.). Loss of cell viability was quantified using the MTT assay: 30,000 cells/well in a 96-well microtiter plate were treated with 100 µL (0.5 mg mL–) MTT solution, freshly dissolved in medium for 1 h at 37°C. The medium was then aspirated and 100 µL DMSO was added to solubilize the dye. Absorbance was measured at 570 nm using a Dynex MRX 96-well microplate reader (Dynex Technologies, France). Each measurement was performed in quadruplicate for HIP/PAP and wild-type hepatocytes dispensed on the same plate. Percentage cell survival was calculated by taking the optical density reading of cells receiving a particular treatment, dividing that number by the OD reading for untreated control cells, then multiplying by 100. Comparison of the results with the number of apoptotic cells visualized using Hoechst 33258 validated the accuracy of the MTT assay.
HIP/PAP assays: Western blot analysis, immunohistochemistry, and ELISA test
HIP/PAP protein was produced and purified from the milk of lactating transgenic mice (16)
. Western blot analysis and immunohistochemistry were performed with pre-HIP antibodies, as already described (13)
. Serum HIP/PAP levels were assayed using a sandwich ELISA test in accordance with the manufacturer’s instructions (Dynabio, La Gaude, France).
Immunodetection of phosphorylated and nonphosphorylated Bad
Freshly isolated hepatocyte-derived from wild-type or transgenic mice were seeded in 35 mm dishes at 400,000 cells per dish. After adhesion, they were rinsed in serum-free 199- medium and incubated with purified HIP/PAP (100 ng mL–1), forskolin 20 µM, or KT 5720 5 µM for 30 min. The cells were rinsed with ice-cold PBS, then lysed in 30 µL of an extraction buffer including 5 mM PO4H2K, pH7.5, 75 mM NaCl, 5 mM EDTA, 1% Triton X-100, 0.1 mM sodium fluoride, 1 mM sodium ß-glycerophosphate, 2 mM Na3VO4, supplemented with a protease inhibitor cocktail tablet. The entire cell lysate was collected by centrifugation at 13,000 x g for 30 min to remove cell debris and the amount of total soluble protein was determined using the Bio-Rad protein assay. One hundred micrograms of protein were loaded onto 12% polyacrylamide gels. After transfer to nitrocellulose (Amersham), hybridizations were performed according to the manufacturer’s instructions (Phospho Ser-112 1:500 and total anti-Bad antibodies 1:500 came from Upstate biotechnologies; anti-actin 1:3000 came from Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at +4°C. The secondary antibody was incubated at 1:2000 dilution for 1 h and the protein band was visualized with ECL+ plus Western Blotting Detection System according to the manufacturer’s instructions (Amersham). Phosphorylated Bad levels were quantified by NIH 1.62 software after the scanning of autoradiographies, using actin normalization.
Statistical analyses
Results for hepatocytes in primary culture were expressed as mean ± SD, and statistical significance (P<0.05) was determined using an unpaired Student’s test. In vivo liver regeneration was represented by the percentages of nuclei incorporating BrdU using the box and whiskers representation; the statistical significance of differences between HIP/PAP transgenic and wild-type mice was determined by the Mann-Whitney U test (P<0.05) because the data distribution was not normal (Statview 5', Abacus Concepts, Berkeley, CA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Immunohistolocalization analysis detected HIP/PAP protein in the liver of transgenic mice as diffuse intrahepatocyte immunostaining, occupying most of the cytoplasm of the hepatocytes (Fig. 2
A, 1 and 2). Staining was heterogeneous and positive regions were located either in centrolobular or portal areas of the liver acinus. This heterogeneous distribution likely reflects HIP/PAP secretion; thus hepatocytes could be either positive or negative before or after HIP/PAP secretion, respectively. Although different zonal expression of the HIP/PAP transgene cannot be excluded, there is no evidence to support this hypothesis. HIP/PAP protein was secreted into the serum (250 ng mL–1 to 700 ng mL–1) in homozygote transgenic lines 24 and 27 and into the culture medium of derived primary hepatocytes (30–120 ng mL–1 per 2 105 cells). No difference in morphology and ploidy was detected between HIP/PAP-expressing and control hepatocytes by histological examination (mouse hepatocytes were 80% binuclear after adhesion; see ref 23
). HIP/PAP immunohistochemistry views of hepatocytes in culture showed that >50% of the hepatocytes were HIP/PAP labeled (Fig. 2A
, 3 and 4).
|
Western blot analysis detected HIP/PAP as a 16 kDa protein in liver extracts and primary culture hepatocytes from HIP/PAP transgenic mice (Fig. 2B
). HIP/PAP protein was not detected in wild-type mice. Actin hybridization allowed an accurate estimation of the 50 µg protein loaded for livers and hepatocytes (50 µg corresponded to 
50,000 hepatocytes).
Liver regeneration is stimulated in mice expressing the human HIP/PAP gene
To test in vivo the HIP/PAP effect on liver cell proliferation, we examined liver regeneration induced by partial hepatectomy. We present low magnification (x20) views for 24, 36, 46, and 55 h postpartial hepatectomy (Fig. 3
A). At the indicated times, percentages of positive BrdU cells were higher in HIP/PAP transgenic than in wild-type livers despite the low overall frequency of nuclei that had incorporated BrdU in both groups. The percentages of nuclei incorporating BrdU were significantly higher in HIP/PAP transgenic mice (median 33%; range 20–42%) than wild-type (median 18%; range 11–27%) (P=0.0014) 46 h after partial hepatectomy (Fig. 3B
).
|
To reinforce the hypothesis that HIP/PAP protein may act as growth factor during liver regeneration, we established the time course of the hepatic mass restoration in wild-type and transgenic mice after hepatectomy (Fig. 3C
). Animal and liver weights were measured in normal nonhepatectomized mice. The liver/body ratio of weight was calculated and expressed as the average percentage ± SD. There was no difference in this ratio between the two groups: 0.0460 ± 0.0064, n = 12 and 0.0489 ± 0.0035 n = 16 for wild-type and HIP/PAP transgenic mice, respectively. Liver recovery was higher in the HIP/PAP transgenic than in wild-type mice and the difference was statistically significant at 48 h (P<0.001), 60 h (P<0.003), and 96 h (P<0.002). At 120 h, the liver weight recovered to the same percentage in wild-type and HIP/PAP transgenic mice.
HIP/PAP mitogenic effect in primary culture hepatocytes
To further investigate the enhanced liver regeneration observed in vivo after hepatectomy in HIP/PAP transgenic mice, primary cultures of hepatocytes were used to evaluate the HIP/PAP mitogenic effect. Hepatocytes derived from HIP/PAP transgenic and wild-type mice exhibited two peaks in DNA synthesis, 60 and 84 h after plating, when stimulated by EGF (Fig. 4
A, B). At 60 h, mean percentages of BrdU-positive hepatocytes were 31±7% (n=19) and 16 ± 4% (n=20) in transgenic and wild-type mice, respectively (P<0.0001). When cells were stimulated by HGF, DNA synthesis was also higher in HIP/PAP than in wild-type hepatocytes (41±4% n=4, vs. 31±11%, n=4, respectively, after 60 h), although this difference did not attain significance. When hepatocytes were not stimulated by growth factor, BrdU incorporation were 11% ± 3 (n=8) and 6% ± 3 (n=7) in transgenic and wild-type hepatocytes, respectively, and the difference was statistically significant (P=0.0146). HIP/PAP is a secreted protein, and we therefore tested whether it might act as a paracrine mitogenic factor. When HIP/PAP protein (40 ng mL–1) was added to wild-type hepatocytes, EGF-induced DNA synthesis increased from 16 ± 4% to 24 ± 7% (P=0.0168; n=8; Fig. 4C
). These results showed that HIP/PAP was a mitogenic factor for hepatocytes in primary culture. The mitogenic effect of HIP/PAP on hepatocyte proliferation was thus demonstrated both in vivo and in vitro.
|
HIP/PAP anti-apoptotic effect against apoptosis induced by TNF- + ActD in primary culture hepatocytes
We next examined whether the HIP/PAP mitogenic effect was associated with the HIP/PAP anti-apoptotic effect. Rat hepatocytes in primary cultures were not sensitive to cell death caused by TNF- treatment alone. Instead, they died by apoptosis after exposure to TNF- combined with a low dose of ActD (22)
. In our present study, mouse hepatocyte cell death was induced by TNF- combined with an ActD dose as low as 0.05 µg mL–1 although ActD (0.05 µg mL–1) alone did not induce any loss of viability (Fig. 5
B). We showed that hepatocytes expressing HIP/PAP resisted TNF- + ActD–induced apoptosis after a 16–17 h of treatment (Fig. 5A
). Cell survival reached 75% vs. 43% (P<0.0001) for 2 ng mL–1 TNF-, and 60% vs. 27% for 20 ng mL–1TNF- (P<0.0001). The LD50 for TNF- was >40 ng mL–1 and 1 ng mL–1 in HIP/PAP and wild-type hepatocytes, respectively. Pretreatment of cells with pan-anticaspase z-VAD-fmk (50 µM) prevented TNF--induced cell death, indicating that this process occurs via hepatocyte apoptosis. We also examined whether dying cells exhibited the typical features of apoptosis. When stained with Hoechst 33258, nonviable cells displayed condensed chromatin, fragmented nuclei, and apoptotic bodies whereas viable cells did not. The features of apoptotic bodies were organized in "rosettes" characteristic of the hepatocyte apoptosis induced by TNF- (Fig. 5C
). When HIP/PAP protein (40 ng mL–1) was added to wild-type hepatocytes, protection against 20 ng mL–1 TNF- + ActD rose from 27% to 47% (P<0.0001). These data demonstrate that HIP/PAP partly abrogated TNF--induced apoptosis in primary hepatocytes.
|
HIP/PAP stimulated the cAMP/PKA pathway
The underlying mechanisms of HIP/PAP mitogenic and anti-apoptotic effect are unknown. Thus, we tested whether cyclic AMP nucleotide-elevating agents might interfere with HIP/PAP effects during EGF-stimulated proliferation and TNF--induced apoptosis. Under our experimental conditions, forskolin (20 µM) had no effect on hepatocyte DNA synthesis without growth factor (data not shown). In contrast, forskolin (20 µM) enhanced EGF-stimulated DNA synthesis in HIP/PAP transgenic (31±7%, n=19 vs. 39±6%, n=8, P=0.02) but not in wild-type hepatocytes (16±4%, n=20 vs. 17±3%, n=8). Thus, the data suggested a cooperative effect of cAMP and HIP/PAP to induce hepatocyte proliferation (Fig. 4C
).
We next showed that Br-cAMP was able to protect wild-type mouse hepatocytes against apoptosis induced by a low dose of TNF- 2 ng mL–1, cell viability increasing from 43% to 57% and 65% with Br-cAMP 200 µM (P=0.0016) and 800 µM (P=0.0004), respectively. Br-cAMP also protected hepatocytes against a high dose of TNF- 20 ng mL–1, cell viability increasing from 27% to 39% and 46% with Br-cAMP 200 µM (P=0.0025) and 800 µM (P=0.0004), respectively (Fig. 6
A). In contrast, Br-cAMP enabled only a small and nonsignificant increase in the viability of HIP/PAP transgenic hepatocytes after treatment with low and high doses of TNF- (Fig. 6A
). In an attempt to understand why HIP/PAP transgenic hepatocytes did not respond to the protective effect of Br-cAMP, we inhibited the protein kinase A (PKA) by the cell-permeable inhibitor KT5720. Under these experimental conditions, we showed that TNF- (without ActD) induced apoptosis in hepatocytes and that HIP/PAP hepatocytes resisted better than wild-type hepatocytes. Viability was 67% and 87% (P<0.025) in wild-type and transgenic hepatocytes, respectively, after treatment with TNF- (2 ng mL–1) + KT5720 (5 µM). Viability was 51% and 76% (P<0.0001) in wild-type and transgenic hepatocytes, respectively, after treatment with TNF- (20 ng mL–1) + KT5720 (5 µM). KT5720 alone did not affect baseline hepatocyte viability (data not shown). Collectively, these results demonstrate that KT5720-dependent PKA inhibition induces TNF- apoptosis and that HIP/PAP expression enhances PKA biological effects. Moreover, the absence of cooperative effects between HIP/PAP and Br-cAMP suggest that they might compete for PKA stimulation.
|
The protective effect of activated PKA against TNF--induced apoptosis and the role of HIP/PAP in stimulating PKA were tested by treating hepatocytes with TNF- combined with KT5720 and ActD (Fig. 6A
). Cell viability decreased dramatically in wild-type hepatocytes when they were preincubated with KT5720 before the addition of TNF- + ActD. After a treatment with 2 ng mL–1 TNF- + ActD, viability fell from 43% to 20% and 13% with 2.5 µM KT5720 (P<0.0001) and 5 µM KT5720 (P<0.0001), respectively. After a treatment with 20 ng mL–1 TNF- + ActD, viability also fell from 27% to 14% and 13% in the presence of 2.5 µM KT5720 (P=0.0013) and 5 µM (P=0.0003) respectively. In contrast, HIP/PAP-expressing hepatocytes resisted KT5720 preincubation before the addition of TNF- + ActD. Viability fell from 75% to 54% and 43% in the presence of 2.5 µM KT5720 (P=0.0081) and 5 µM KT7520 (P<0.0001), respectively, after treatment with TNF- (2 ng mL–1). Viability fell from 60% to 40% and 24% in the presence of 2.5 µM KT5720 (P=0.0042) and 5 µM KT 5720 (P=0.0001), respectively, after treatment with 20 ng mL–1 TNF-. Thus, KT5720 preincubation increased toxicity induced by TNF- + ActD in both type of hepatocytes, but less in HIP/PAP than in wild-type hepatocytes. These results further confirm that HIP/PAP may stimulate PKA.
HIP/PAP-dependent PKA activation induces the phosphorylation of Bad at Ser-112
To reinforce the hypothesis that HIP/PAP may stimulate PKA, we examined the consequence of HIP/PAP-dependent activation of PKA. Among PKA substrates, Bad has been identified as a proapoptotic target of the PKA signaling pathway (24)
. Activated PKA can directly phosphorylate and inactivate Bad at Ser-112. Consistent with this result, we verified that forskolin increased and KT5720 decreased Bad phosphorylation in hepatocytes, further demonstrating the validity of our experimental conditions in this model (Fig. 6B
). Densitometric analyses of Western blotting showed that forskolin (20 µM) increased the P=Bad/actin ratio of intensity by 2.5 ± 0.1 and 3.2 ± 0.9 times in wild-type and HIP/PAP hepatocytes respectively in 3 independent experiments. KT5720 (5 µM) decreased the P=Bad/actin ratio of intensity by 0.6 ± 0.2 and 0.6 ± 0.3 in wild-type and HIP/PAP hepatocytes respectively. In the absence of drug, basal phosphorylation of Bad at Ser-112 was identical in both types of hepatocytes. To test whether HIP/PAP activated PKA-dependent phosphorylation, we examined the PKA-dependent phosphorylation of Bad at Ser-112 after addition of recombinant HIP/PAP protein to wild-type and HIP/PAP hepatocytes: the P=Bad/actin ratio of intensity increased by 2.1 ± 0.5 and 2.3 ± 0.9 times in wild-type and HIP/PAP hepatocytes, respectively. Overall, these results show that HIP/PAP mimics the forskolin effects and confirm that HIP/PAP activates PKA and, consequently, induces the PKA-dependent phosphorylation of Bad at Ser-112.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
HIP/PAP has been identified using several independent approaches and exhibits marked tissue specificity: liver, pancreas, intestine, and nervous system. In the exocrine pancreas, the pattern of HIP/PAP expression is consistent with that of an acute phase reactant, and HIP/PAP protein has been associated with pancreatic acinar cell protection from oxidative stress (25)
and TNF--induced pancreatic stress (26)
. HIP/PAP is 50% homologous with the Reg-1 protein, which is a predominant ß cell replication factor and improves surgically induced diabetes in rats (27)
.
We generated transgenic mice secreting into their serum high levels of human HIP/PAP protein from hepatocytes. In vivo, HIP/PAP transgenic mice were more likely to regenerate than controls after partial hepatectomy, as judged by stimulation of DNA synthesis and an increase in the recovery of liver weight in the HIP/PAP transgenic mice. Hepatocytes need to be primed before they fully respond to HGF, TGF-, or EGF. Two known regulators of the priming phase of liver regeneration are the TNF- and IL-6 cytokines (28)
. TNF- signaling has been reported to mediate EGF-dependent DNA replication. The levels of these cytokines rise after partial hepatectomy, with the peak of circulating TNF- levels preceding that of IL-6. The rat HIP/PAP-I promoter contains sequences similar to the IL-6 response element, suggesting that HIP/PAP gene expression may be activated by IL-6 (29)
. Thus, HIP/PAP may be a candidate for mediating the positive effect of TNF- /IL-6 during liver regeneration.
Consistent with the positive effect of HIP/PAP upon regeneration in vivo, we show that under primary culture conditions transgenic HIP/PAP-derived hepatocytes proliferate more and resist TNF--induced apoptosis better than wild-type hepatocytes. The proportion of hepatocytes passing through the S phase was higher in HIP/PAP transgenic hepatocytes than in wild-type hepatocytes in the absence or presence of growth factors. However, HIP/PAP alone only exhibits a low mitogenic capacity and requires growth factors such as EGF to enhance DNA proliferation, indicating that HIP/PAP cooperates with mitogenic stimuli to increase the cycling fraction of hepatocytes.
Although TNF- can initiate liver regeneration in vivo, it also induces apoptosis in vitro if given in conjunction with drugs such as ActD or cycloheximide, which block respectively gene transcription or translation, resulting in the production of excess reactive oxygen species in the absence of NF
B activation (22)
. We show here that HIP/PAP protects hepatocytes against TNF- + ActD-induced apoptosis.
Links between the cAMP transduction network and the cell cycle machinery have already been described in primary hepatocytes, with cAMP exerting a bidirectional effect depending on the concentration of the cAMP-elevating agent and the time of addition. In our model, wild-type hepatocytes did not respond to forskolin, an activator of adenylate cyclase, whenever the mitogenic effect of HIP/PAP was increased in the presence of forskolin. Thus, HIP/PAP and cAMP cooperate to increase the hepatocyte proliferation induced by EGF, and HIP/PAP may stimulate entry into the S phase of cultured mouse hepatocytes via a cAMP pathway. The HIP/PAP gene has been shown to be markedly up-regulated after axotomy in rat motoneurons (30)
and acts as a weak mitogen for Schwann cells in vitro. Moreover, in this model HIP/PAP combined to forskolin has demonstrated the synergistic mitogenic effect of their association.
It has been shown that agents elevating cAMP levels can protect rat hepatocytes from apoptosis induced by Fas, TNF-, and biliary salts (31
32
33)
. Our results show that the cell-permeable cAMP analog Br-cAMP partially prevented TNF--induced apoptosis in mouse wild-type hepatocytes, but not in HIP/PAP transgenic hepatocytes, suggesting that HIP/PAP may compete with cAMP to stimulate PKA activity and consequently partially prevent TNF--induced apoptosis in hepatocytes. We also showed that KT 5720 was able to promote apoptosis by TNF- without ActD, demonstrating for the first time that the specific inhibition of PKA induces TNF- apoptosis and therefore that PKA protects hepatocytes against TNF--induced apoptosis. HIP/PAP transgenic hepatocytes resisted better than wild-type cells to apoptosis induced by TNF- combined with KT5720, confirming the hypothesis that HIP/PAP or cAMP stimulates PKA activity.
A similar interpretation was possible after hepatocytes were treated by the combination of TNF- + ActD + KT 5720. Under these conditions, HIP/PAP transgenic hepatocytes also resisted better than wild-type hepatocytes, although apoptosis was dramatically enhanced overall, suggesting the coexistence of a multiple survival signaling pathway in hepatocytes, i.e., both PKA-dependent and -independent mechanisms. Our present results do not exclude the possibility that HIP/PAP may participate in different signaling pathways. HIP/PAP (=Reg-2) was able to keep rat motoneurons alive by activating both the PI3K/Akt kinase pathway and NF-B (34)
. Indeed, active phosphorylated-Akt kinase (=PKB) protects also mouse hepatocytes from TNF--mediated apoptosis (35)
and cyclic nucleotides rapidly activate endogenous Akt in a PI3K-dependent manner in hepatocytes (32)
. However, complete inhibition of cyclic nucleotide-dependent Akt activation has little effect on cyclic nucleotide-mediated hepatocyte survival. This result thus involves cyclic nucleotide-dependent PKA activation to mediate the apoptotic signaling cascade after TNF--induced apoptosis (32)
. In the present study, we confirm that PKA mediates TNF--dependent apoptosis in hepatocytes. Moreover, we show that the inhibition and the stimulation of PKA mediate the Bad phosphorylation at Ser-112. Finally, we show that HIP/PAP protein increases the PKA-dependent phosphorylation of Bad, indicating that HIP/PAP, as for cAMP, stimulates PKA activity. Knowing that cAMP binds to the PKA-RII regulatory subunit to activate the PKA catalytic enzyme subunit, we have also found an interaction between HIP/PAP and the PKA-RII regulatory subunit (Demaugre et al., unpublished results).
Despite the HIP/PAP-dependent PKA activation and consequent Bad phosphorylation, we cannot relate the HIP/PAP anti-apoptotic effect to the Bad phosphorylation in hepatocytes with the present data. Harada et al. (24)
showed, however, that PKA-dependent Bad phosphorylation inactivates the proapoptotic Bad protein, which cannot interact with Bcl-2. Consequently, Bad bound to 14-3-3 is sequestered in the cytosol, freeing BCL-Xl or BCL-2 to promote FL5.12 cells survival.
Even though Bcl-2 could be involved in apoptosis of certain hepatoma cells, no information is now available regarding the role of Bad in hepatocarcinoma. However, our present work confirms that of Li et al. (32)
, who showed that PKA activation could partially protect hepatocytes against apoptosis induced by combining TNF-+ActD. Finally, our present work shows that HIP/PAP may stimulate the PKA activity.
The addition of purified human HIP/PAP protein to wild-type hepatocytes in primary culture mimicked the results observed in transgenic hepatocytes: elevation of DNA synthesis and protection against TNF- + ActD-induced apoptosis. It has been reported that Reg-2 (rat HIP/PAP) may act in both autocrine and paracrine ways to prevent neuronal cell death in rats (35)
and promote the cellular growth of epithelial intestinal cells (36)
as well as in a paracrine way to stimulate DNA synthesis in Schwann cells (30)
or promote rat hepatocyte adhesion in primary cultures (37)
. These results therefore indicate that the biological functions of HIP/PAP may be dependent on the autocrine and paracrine pathways. A paracrine effect was anticipated for a protein secreted in hepatocarcinoma patient’s sera, in transgenic mice, and in the culture medium of primary HIP/PAP transgenic hepatocytes. These results suggest the existence of an unknown HIP/PAP receptor on the external membrane of hepatocytes.
In summary, the polypeptides known to enhance liver regeneration in vitro and in vivo are limited in number, and, as a C-type lectin, HIP/PAP is unrelated structurally with any of those so far described. Lectins play a role in biological recognition events and Galectins are involved in cellular migration during development and metastasis. Galectins may be proapoptotic (38)
or anti-apoptotic molecules (39
, 40)
. To date, no C-type lectin had been ascribed the role of a growth factor and no C-type lectin combines both mitogenic and anti-apoptotic properties.
During liver carcinogenesis (hepatocarcinoma and cholangiocarcinoma), HIP/PAP gene activation does not result from the reexpression of a fetal marker, as has been described for the -fetoprotein gene. Indeed, HIP/PAP is not expressed in mouse adult liver or during liver development (41)
. In contrast, a moderate expression of the mouse HIP/PAP gene was detected in liver after partial hepatectomy (data not shown). In human adult liver, HIP/PAP protein is not detected in normal differentiated hepatocytes, but is present during active cirrhosis and chronic active hepatitis in reactive ductular structures. Moreover, in normal liver HIP/PAP protein is detected in some ductular cells located at the interface of the portal tracts with the parenchyma, suggesting that they may represent progenitor cells with a dual potential to differentiate into bile duct cells and hepatocytes (13)
. These observations together with our present report further support the possibility that HIP/PAP is implicated in certain stages of liver carcinogenesis, since HIP/PAP reduces the local control of both survival and proliferation.
|
ACKNOWLEDGMENTS |
|---|
Received for publication October 23, 2002. Accepted for publication April 13, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Gironella, E. Folch-Puy, A. LeGoffic, S. Garcia, L. Christa, A. Smith, L. Tebar, S. P Hunt, R. Bayne, A. J H Smith, et al. Experimental acute pancreatitis in PAP/HIP knock-out mice Gut, August 1, 2007; 56(8): 1091 - 1097. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Gurr, M. Shaw, Y. Li, and R. Sherwin RegII Is a {beta}-Cell Protein and Autoantigen in Diabetes of NOD Mice Diabetes, January 1, 2007; 56(1): 34 - 40. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Namikawa, T. Okamoto, A. Suzuki, H. Konishi, and H. Kiyama Pancreatitis-associated protein-III is a novel macrophage chemoattractant implicated in nerve regeneration. J. Neurosci., July 12, 2006; 26(28): 7460 - 7467. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Lu, A. Ponton, H. Okamoto, S. Takasawa, P. L. Herrera, and J.-L. Liu Activation of the Reg family genes by pancreatic-specific IGF-I gene deficiency and after streptozotocin-induced diabetes in mouse pancreas Am J Physiol Endocrinol Metab, July 1, 2006; 291(1): E50 - E58. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Folch-Puy, S. Granell, J. C. Dagorn, J. L. Iovanna, and D. Closa Pancreatitis-Associated Protein I Suppresses NF-{kappa}B Activation through a JAK/STAT-Mediated Mechanism in Epithelial Cells J. Immunol., March 15, 2006; 176(6): 3774 - 3779. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. D. De Leon, C. Farzad, M. F. Crutchlow, J. Brestelli, J. Tobias, K. H. Kaestner, and D. A. Stoffers Identification of transcriptional targets during pancreatic growth after partial pancreatectomy and exendin-4 treatment Physiol Genomics, January 12, 2006; 24(2): 133 - 143. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R.-H. Yuan, Y.-M. Jeng, H.-L. Chen, F.-J. Hsieh, C.-Y. Yang, P.-H. Lee, and H.-C. Hsu Opposite Roles of Human Pancreatitis-Associated Protein and REG1A Expression in Hepatocellular Carcinoma: Association of Pancreatitis-Associated Protein Expression with Low-Stage Hepatocellular Carcinoma, {beta}-Catenin Mutation, and Favorable Prognosis Clin. Cancer Res., April 1, 2005; 11(7): 2568 - 2575. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Ostrowski, A. A. Reilly, D. N. Collins, and A. I. Ramsingh Progression or Resolution of Coxsackievirus B4-Induced Pancreatitis: a Genomic Analysis J. Virol., August 1, 2004; 78(15): 8229 - 8237. [Abstract] [Full Text] [PDF]
|
|
| |||||||||
) and HIP/PAP mice (
) at various time points after partial hepatectomy shows stimulated recovery in the HIP/PAP transgenic mice (5 to 9 mice were hepatectomized at each time for each group). The difference was statistically significant at 48 h (P<0.001), 60 h (P<0.003) and 96 h (P<0.002).
) wild-type (hatched) HIP/PAP.
). The histograms represent the mean values ± SE of 5 separate experiments with four replicates, and 3 separate experiments with KT5720 but without ActD. B) HIP/PAP activates the cAMP/PKA signaling pathway inducing the phosphorylation of Bad at Ser-112. Freshly isolated hepatocytes-derived from wild-type or transgenic mice were incubated without drugs (1, 5), with 5 µM KT 5720 (2, 6), 20 µM forskolin (3, 7) or 100 ng mL–1 recombinant HIP/PAP (4, 8) for 30 min.