Abstract
The mRNA expression of two activin growth factor subunits (βA- and βC-activin), activin receptor subunits (ActRIIA, ActRIIB) and the activin-binding protein follistatin, and peptide expression of βA-activin and βC-activin subunits, were examined in regenerating rat liver after partial hepatectomy (PHx). Liver samples were collected from adult, male Sprague–Dawley rats, 12–240 h (n=3–5 rats per time point) after PHx or from sham-operated controls at the same time points. Hepatocyte mitosis and apoptosis were assessed histologically and by in situ cell death detection. RT and PCR were used to assess relative gene expression. βA- and βC-activin peptide immunoreactivity was assessed in liver and serum samples by western blotting, whereas cellular expression was investigated by immunohistochemistry, using specific monoclonal antibodies. βA- and βC-activin mRNA dropped to < 50% of sham control values 12 h after PHx and remained at this level until 168 h post-PHx, when βA-activin expression increased to three times sham control values and βC-activin mRNA returned to pre-PHx levels. A peak in follistatin expression was observed 24–48 h post-PHx, coincident with an increase in hepatocyte mitosis. No changes were observed in ActRIIA mRNA, whereas ActRIIB expression paralleled that of βA-activin mRNA. βC-activin immunoreactive homo- and heterodimers were observed in regenerating liver and serum. Mitotic hepatocytes frequently contained βC-activin immunoreactivity, whereas apoptotic hepatocytes were often immunoreactive for βA-activin. We conclude that βA- and βC-activin subunit proteins are autocrine growth regulators in regenerating liver and when expressed independently lead to hepatocyte apoptosis or mitosis in a subset of hepatocytes.
Introduction
Cell proliferation and cell death are subject to control by stimulatory and inhibitory factors. Simplistically, tropic factors will stimulate mitosis and at the same time inhibit cell death, while negative growth signals will inhibit mitosis and cause apoptosis. In contrast to the many heterotropic factors that have been identified, very few growth inhibitors or apoptosis inducers have been characterised in the liver (Fausto et al. 1995). The activins belong to the transforming growth factor-(TGF-β) gene family. Members of this family share structural similarities, but are diverse in their biological activities (Kingsley 1994). Activins are hetero- or homodimers of inhibin β-subunits held together by disulphide bonds (Ying 1988). Activins control cell proliferation and programmed cell death in a variety of tissues including the liver (Schwall et al. 1993, Yasuda et al. 1993, Hully et al. 1994). Activin A is a homodimer of the inhibin βA-activin subunit. It is expressed in hepatocytes and has been shown to be a negative regulator of hepatocyte cell growth (Schwall et al. 1993, Yasuda et al. 1993, Hully et al. 1994). In the normal liver the expression of activin A is relatively low (Yasuda et al. 1993). Activin A is involved in the maintenance of constant liver mass, since blocking activin A action by administering the activin-binding protein follistatin leads to hepatocyte proliferation (Kogure et al. 1995, 1996, 1998, 2000, Kozian et al. 1997, Zhang et al. 1997a). Activin A is also up-regulated in the development of liver cirrhosis concomitant with an increase in hepatocyte apoptosis (Gold et al. 2003).
βC-activin mRNA appears to be predominately expressed in the liver (Hotten et al. 1995, Lau et al. 1996, 2000, Schmitt et al. 1996, Gold et al. 2004) and is down-regulated 12 h after partial hepatectomy (PHx) in the rat, suggesting that it too may be a negative regulator of liver cell growth (Esquela et al. 1997, Zhang et al. 1997b). Two recent studies show that βC-activin inhibits DNA synthesis of hepatic cells in vivo (Chabicovsky et al. 2003) and induces apoptosis in hepatoma cells in vitro (Vejda et al. 2003). Stimulation of hepatocyte proliferation by administration of di-n-butyl phthalate was accompanied by a rise in βB- and βC-activin mRNA and a decrease in βA-activin mRNA (Kobayashi et al. 2002). The βC-activin subunit can form heterodimers with the βA-, βB- and βE-activin subunits in vitro (Mellor et al. 2000, 2003, Vejda et al. 2002) and acts at the intra- cellular level to modulate the synthesis of activin A or activin B via the formation of activin AC or BC heterodimers (Mellor et al. 2000, 2003, Kobayashi et al. 2002). Recent research demonstrates that βC-activin homodimer can be growth promoting and confirms that intracellular AC dimerisation prevents the formation of activin A (Wada et al. 2004). Thus βC-activin has been shown to be either growth inhibitory (Esquela et al. 1997, Zhang et al. 1997b) or growth promoting (Kobayashi et al. 2002, Wada et al. 2004) in liver. A balance between cell proliferation and apoptosis is crucial for regulating normal liver function (Chen et al. 2000), since abnormalities in liver regeneration may contribute to chronic hepatitis, cirrhosis and cancer (Leevy 1998). Cellular control of activin expression appears to occur at many levels, including the amount of growth factor, dimer combinations formed, the presence of binding proteins, receptor recruitment and presence of other peptides (Phillips 2000). The presence of β-activin subunits, follistatin and activin receptor mRNAs, and the ability of these proteins to affect hepatic cell division, implies that these growth factors may be involved in the liver’s regenerative response. In order to examine the ability of the βC-activin subunit to antagonise activin A action in vivo, we examined hepatocyte cell death and division after PHx in the liver of male rats and assessed mRNA expression of βA-activin, βC-activin, follistatin, and activin receptor subunits ActRIIA and ActRIIB; and also βA- and βC-activin subunit dimer formation. We hypothesised that βC-activin expression is associated both cellularly and temporally with βA-activin expression, at times of liver growth. If this hypothesis is correct we would predict co-localisation of the two activin subunits in mitotic or resting hepatocytes. Furthermore we predict βA-activin expression to be high, and the expression of antagonists of activin A (βC-activin and follistatin) to be decreased at times of maximal hepatocyte apoptosis.
Methods
Ethical approval
All experiments using animals were performed in accordance with the 1999 Animal Welfare Act regulations of New Zealand after approval was granted by the University of Otago Medical School Animal Ethics Committee.
Surgery
Male (220–250 g) Sprague–Dawley rats were randomly allocated to a PHx or control group. PHx was performed under ether anaesthesia according to a previously described method (Higgins & Anderson 1932). Control sham operations consisted of opening the abdominal cavity and briefly handling the liver. Two sham-operated control animals were killed at 12, 24, 48, 72, 96, 168 and 240 h post-sham PHx. Five PHx rats were killed at 12, 48 and 168 h; while three PHx rats were killed at 24, 72, 96 and 240 h post-PHx. Samples of caudate lobe from regenerating or sham control liver samples were removed and either immediately stored at −80 °C or prepared for histology. The caudate lobe is not affected by the surgery, but does participate in the restoration of liver mass (Higgins & Anderson 1932).
Histology
Samples were fixed in 4% paraformaldehyde and processed to 4 μm thick sections. Serial sections were used for immunohistochemistry, cell death detection, and haematoxylin and eosin staining for histology. Hepatocyte cell death was assessed with an in situ cell death detection kit (TUNEL, Roche) and was confirmed by counting apoptotic bodies in serial sections, as previously described (Gold et al. 2003). Mitotic hepatocytes were identified based on histology (prophase, metaphase, anaphase or telophase). Within sections, the percentage of mitotic hepatocytes or TUNEL-positive hepatocytes were counted. Five randomly selected fields were assessed in duplicate sections, in three to five animals per time point and the results are presented as means ± s.d.
Molecular biology
Total RNA was extracted from liver using a RNeasy Mini Kit (Qiagen) according to manufacturer directions. RNA concentration was determined by spectrophotometry at 260 nm. An equal amount of RNA from each tissue was used for generation of first-strand cDNA. Primers for PCR were designed based on published rat sequences, or from consensus sequence of human and mouse βC-activin cDNA (Gold et al. 2003). All primers were designed so that the amplicon was derived from two or more exons. Preliminary PCR assays were performed to determine the optimum annealing temperature, MgCl2 concentration and range of cycles over which the amplified cDNA samples remained in the exponential phase of amplification (Gold et al. 2003). Reactions without template, genomic DNA and RNA controls, were run alongside all experimental samples. PCR products were separated on agarose gel, photographed and semi-quantified by densitometry as previously described (Gold et al. 2003). The density of amplified product for each sample was normalised with respect to the density of β-actin mRNA. To confirm amplicon identity, PCR products from two separate reactions were sequenced in both directions using the direct-cycle method, by the University of Otago Centre for Gene Research sequencing facility.
Immunoblotting
The identity of activin dimers present in tissue extracts and serum was determined by western immunoblotting. Protein samples were prepared for electrophoresis by homogenising 50 mg tissue in 500 μl non-reducing sample buffer (4% (w/v) SDS, 2 mM EDTA, 50 mM Tris–HCl), heating to 100 °C for 5 min and centrifuging at 10 000 g for 10 min. Total protein concentration was determined with a Sigma BCA kit (Sigma–Aldrich). Equal amounts of protein (2 mg) were loaded onto 15% polyacrylamide gels and separated by electrophoresis under non-reducing conditions. The proteins were transferred to polyvinylidene diflouride membrane with a reducing transfer buffer (700 mM glycine, 300 mM Tris, 15.6% ethanol, 15 mM dithiothreitol). Antibodies were diluted [βA-activin 1:2000 (R&D Systems, Inc., Minneapolis, MN, USA) and βC-activin 1:3000 (Mellor et al. 2000)], incubated overnight at 4 °C in Tris-buffered saline (TBS)–5% milk powder. Biotinylated goat anti-mouse secondary antibody (Amersham) was diluted 1:3000 in TBS and blots incubated for 1 h at room temperature, followed by streptavidin biotinylated horseradish peroxidase (Amersham) at a dilution of 1:1500. Blots were developed with the ECL chemiluminescent detection system (Amersham) according to manufacturer’s directions. Three representative liver and serum samples per time point were assessed in duplicate and results from each were consistent. Duplicate samples were run at the same time on two separate gels. Gels were incubated with either βA- or βC-activin antibodies, exposed to X-ray film, stripped and re-probed with the alternative activin antibody. When films were overlaid, each activin antibody was shown to detect different-sized monomer peptides. This also confirmed the presence or absence of a βAβC activin dimer.
Immunohistochemistry and staining indices
Paraffin-embedded tissue blocks were sectioned and deparaffinised by standard techniques. Sections underwent microwave heat antigen retrieval submerged in 0.01 M glycine buffer, pH 4.4 (Mellor et al. 2000). After cooling to room temperature, endogenous peroxidase activity was eliminated with hydrogen peroxidase. βC-activin antibody, diluted to a final concentration of 5 μg/ml in PBS, was added. Sections were incubated overnight at 4 °C, washed in PBS containing 0.1% (v/v) Tween-20 (PBS-T) and incubated with biotinylated goat anti-mouse secondary antibody (Amersham). Sections were washed with PBS-T and incubated with streptavidin biotinylated horseradish peroxidase (Amersham). Peroxidase activity was detected using 3′3′ diaminobenzidine tetrahydrochloride (DAB, Vector Laboratories Inc, Peterborough, UK), before counter-staining with Gill’s haematoxylin (Mellor et al. 2000, Gold et al. 2003). The βC-activin subunit antibody has previously been shown not to cross-react with βA-or βB-activin subunits or the closely related βE-activin subunit (Mellor et al. 2000). βC-activin staining patterns were confirmed using a DAKO autostainer (DAKO Corporation, Carpinteria, CA, USA), with 0.01 M glycine buffer, pH 4.4, for antigen retrieval, a 90 min primary antibody incubation at room temperature, a 20-min secondary antibody incubation and a 15 min tertiary incubation. Antibodies were diluted in TBS containing 1% BSA (Sigma) and sections washed with TBS containing 0.1% Tween-20. Negative controls were included in every experiment: sections were incubated with primary antibody only, secondary antibody only or mouse monoclonal IgG substituted for primary antibody (Sigma–Aldrich). Specificity of staining was also assessed by pre-absorption of the βA- or βC-activin primary antibodies with 500 μg/ml antigenic peptide as described previously (Mellor et al. 2000). The percentages of intensely stained βA- and βC-activin hepatocytes undergoing mitosis or apoptosis were assessed in liver sections. Slides from regenerating liver were analysed for positive β-activin subunit staining in apoptotic (12 h, 7 day and 10 day sections) or mitotic hepatocytes (24 h, 48 h and 72 h sections). Apoptotic hepatocytes in βA- or βC-activin-stained sections were identified by their histological appearance (Gold et al. 2003) and by positive TUNEL staining in the adjacent section. Mitotic hepatocytes in prophase, metaphase, anaphase or telophase were identified by histological appearance. These time points were chosen as they contained increased evidence of apoptosis or mitosis compared with the other time points (see Fig. 1). The frequencies of apoptotic or mitotic hepatocytes that exhibited intense activin subunit immunoreactivity were presented as a staining index (stained hepatocytes/total hepatocytes in mitosis or apoptosis × 100).
Statistics
The normalised densities of PCR products (n=41) were log transformed and variation with time assessed by ANOVA (SPSS for Windows 6.1). In figures, normalised densities are expressed as a percentage of sham-operated controls at each time point. Gene expression results are presented as means ± s.d. of 3 independent PCR reactions with different cDNAs, n=3–5 PHx and 2 sham controls per time point.
Mitotic/apoptotic staining percentages were analysed by Mann–Whitney U-test (SPSS for Windows 6.1).
Results
Body weight
Rats lost 4% of their body weight post-PHx, returning to pre-operation weight after 5 days (120 h). At 7 days (168 h) post-PHx, by which time liver mass had been restored, rat body weight was steadily increasing (14% increase compared with pre-operation weight). Control rats did not lose weight after the sham operation and their body weight increased throughout the study (25% increase by day 7).
Histology
The percentages of mitotic and apoptotic hepatocytes in regenerating male rat liver are presented in Fig. 1. Mitosis was observed rarely in sham-operated control liver. An increase in the percentage of mitotic hepatocytes was evident in the caudate lobe 48 h (P < 0.001) and 72 h (P < 0.02) post-PHx. Cell death was low in sham-operated control liver, but an increased rate of TUNEL staining was evident at 12 h and from 96 h on, in regenerating liver (P < 0.01).
Gene expression
Concentrations of β-actin mRNA did not vary significantly after control sham operations or during the course of liver regeneration (P=0.537, ANOVA). Control sham operations did not lead to significant changes in βA- or βC-activin, ActRIIA, ActRIIB or follistatin mRNA (Fig. 2). Results are mean ± range in n=2 control rats at each time point. Follistatin mRNA expression was very low in control livers. Consistently more βC-activin than βA-activin mRNA and equivalent amounts of ActRIIA and ActRIIB mRNA were detectable.
Results from the caudate lobe of regenerating rat liver are presented graphically as a percentage of sham controls. Significantly lower amounts of βA-activin mRNA were measured in the caudate lobe of regenerating liver at 12 and 24 h following PHx, compared with sham-operated controls (P < 0.05, Fig. 3A). βA-activin mRNA increased at 168 h, when liver mass appeared fully restored (P < 0.001). Amounts of βC-activin mRNA decreased significantly compared with controls levels at 12, 24 (P < 0.05) and 72 h (P < 0.001) post-PHx (Fig. 3A). Follistatin mRNA increased from very low values to a peak at 24 and 48 h in regenerating liver (P < 0.001), decreasing to barely measurable values at 72, 168 and 240 h compared with controls (P < 0.001, Fig. 3B). The expression of ActRIIA mRNA showed no significant changes with restoration of liver mass, whereas ActRIIB receptor mRNA increased markedly at 168 h (P < 0.001, Fig. 3C), compared with controls.
Activin peptides in liver
Activin subunit immunoreactivity in control and restoring rat liver are shown in Fig. 4. Immunoblotting revealed that the caudate lobe of sham-operated male rat liver contained moderate amounts of activin pro-proteins (above 32 kDa), small amounts of a βA-activin homodimer, no evidence of the βAβC (23 kDa) band and moderate levels of the βC-activin homodimer at 21 kDa. βA-activin immunoreactive bands corresponding to pro-peptides (above 36 kDa), βAβA-activin (26 kDa) and βAβC-activin (23 kDa) were evident in the regenerating caudate lobe (Fig. 4A). There were βC-activin immunoreactive bands above 36 kDa (likely to be βC-activin subunit pro-peptides), at 23 kDa (βAβC-activin) and at 21 kDa (βCβC-activin) in regenerating liver (Fig. 4B).
Activin dimer combinations appeared to alter in restoring rat liver (Fig. 4), but no attempt was made to quantify amounts of immunoreactivity on the blots. The 12-h time point showed little evidence of dimers of either subunit. The proportion of βAβA (26 kDa) and βAβC (23 kDa) activin dimers appeared to increase at 24 h. There was proportionally more βAβC activin (23 kDa) at 48 h than either βAβA (26 kDa) or βCβC (21 kDa) and at 96 h there appeared to be more βC-activin than βA-activin immunoreactivity compared with other time points. At 168 and 240 h the male rat restoring caudate lobe showed evidence of more βAβA activin (26 kDa) and βAβC activin (23 kDa), than βCβC activin (21 kDa).
Activin peptides in serum
In the serum of PHx rats the anti-βA-activin antibody detected a double band at 23–26 kDa; this is likely to be activin AC (βAβC dimer), activin AB (βAβB dimer) and/or activin A (βAβA, Fig. 5A). The βC-activin antibody also detected two circulating peptides in regenerating and control rat serum (Fig. 5B). These bands may be activin AC (23 kDa), activin BC and/or activin C (βCβC dimer, 21 kDa). The unidentified intensely immunoreactive band evident at around 30 kDa for both β-activin antibodies is likely to be an artefact caused by excess loading of serum onto the gel.
Activin subunit immunohistochemistry
Representative views of activin subunit immunoreactivity are displayed in Fig. 6. Mild βA-activin subunit immunoreactivity was present in hepatocyte cytoplasm in control and regenerating male rat liver. At 48 h post-PHx, staining was mainly associated with granular or vesicular structures, rather than hepatocyte cytoplasm (Fig. 6A). In the later stages of restoration of liver mass, there was increased βA-activin subunit immunoreactivity associated with apoptotic hepatocytes (Fig. 6B).
Moderate βC-activin subunit immunoreactivity was observed in the majority of hepatocytes in control male liver. At 48 h after PHx, there was intense βC-activin subunit immunoreactivity associated with mitotic cells (Fig. 6C). βC-activin immunoreactivity was evident in approximately 65% of hepatocyte nuclei, as well as in the cytoplasm, at 168 and 240 h post-PHx (Fig. 6D).
βA-activin subunit staining was assessed in 355 mitotic hepatocytes and 29 were strongly positive. βC-activin subunit immunoreactivity was assessed in 243 mitotic hepatocytes and 218 were intensely immunoreactive. Of 195 apparently apoptotic hepatocytes assessed for βA-activin subunit staining, 144 were markedly positive, whereas 32 of 191 apoptotic hepatocytes assessed for βC-activin subunit immunoreactivity were positive. These data are shown in Fig. 7. Thus an increase in βA-activin subunit peptide was associated with hepatocyte cell death (P < 0.01), whereas an increase in βC-activin subunit immunoreactivity was significantly associated with hepatocyte cell division (P < 0.001).
Discussion
Temporal changes in mRNA expression of βA- and βC-activin subunits, follistatin and activin receptor RIIB were observed in the regenerating caudate lobe of male rat liver.
The ratio of βA- and βC-activin subunit dimers appeared to alter in restoring rat liver. An increase in βA-activin subunit protein was associated with hepatocyte apoptosis, while increased βC-activin subunit protein was associated with hepatocyte mitosis.
Activin A and activin receptor subunit ActRIIA have previously been shown to be associated with apoptosis (Schwall et al. 1993, Yasuda et al. 1993, Hully et al. 1994, Coerver et al. 1996, Draper et al. 1997). Our results indicate that locally produced activin A acts as an autocrine factor to regulate hepatocyte cell proliferation, however we failed to demonstrate changes in ActRIIA mRNA. Activin-induced apoptosis of hepatoma cells is blocked by dominant negative forms of ActRIIB or Smad 2, whereas over-expression of either ActRIB, ActRIIB, Smad 2 or Smad 4 is sufficient to stimulate apoptosis in the absence of activin A (Chen et al. 2000, 2002). This suggests that activin A signals through ActRIB/ActRIIB receptors and the receptor Smad pathway plays a central role in determining cellular fate. Hubner et al.(1999) and Rosendahl et al.(2001) have examined activin receptor type II mRNA in wound healing and lung inflammation. Neither study demonstrates changes in these mRNAs in their in vivo models. To the best of our knowledge the data presented here represent the first report of dynamic regulation of activin type IIB receptor subunit mRNA.
The mechanisms by which activin A regulates hepatocyte growth arrest are largely unknown. In HepG2 cells, retinoblastoma hypophosphorylation – via modulation of p53, p21 and the cyclin-dependent kinase Cdk4 – is involved in activin-mediated cell growth inhibition (Zauberman et al. 1997). Activin has also been shown to induce hepatocyte growth arrest through induction of the cyclin-dependent kinase inhibitor p15 and transcription factor Sp1 (Ho et al. 2004). Examining the expression of genes shown to be associated with activin-mediated cell cycle arrest in vitro in an in vivo model of liver regrowth may provide further evidence as to the mechanisms by which activins control hepatocyte cell growth. Follistatin has previously been shown to increase DNA synthesis and to antagonise the inhibitory effects of activin A in regenerating liver (Kogure et al. 1995, 1996, 1998, 2000, Zhang et al. 1997a). The increase in follistatin demonstrated in our study may therefore serve to protect proliferating hepatocytes from the apoptotic effects of activin A. Mitosis was evident at 48–72 h when follistatin mRNA expression was high. This may indicate that increased follistatin expression promotes hepatocyte cell division, either independently or by blocking the negative growth effects of activin A.
βC-activin has been shown to be growth inhibitory in liver (Esquela et al. 1997, Zhang et al. 1997b) or growth promoting (Kobayashi et al. 2002, Wada et al. 2004). Lau et al.(2000) demonstrated, by RNase protection assay with a probe from the 3′ untranslated region of C-activin cDNA, an increase in βC-activin mRNA 12–24 h after PHx; this differs both from our results and from those of Esquela et al.(1997) and Zhang et al. (1997b). Discrepancies may be related to the region of βC-activin cDNA examined or the part of the liver used in this study. In our study, βC-activin mRNA was low in extracts of regenerating liver, but significantly associated histologically with mitotic hepatocytes. Thus the gene expression data presented support the theory that βC-activin may act as a liver chalone, whereas the immunohistochemical data suggest a local role for βC-activin in hepatocyte proliferation. In addition, our data suggest that activin mRNA may be labile and translated peptide stored in hepatocytes, as mRNA levels did not accurately reflect the amount of activin subunit proteins evident by immunohistochemistry or immunoblotting. These discrepancies highlight the importance of examining activin at both the mRNA and protein level.
It has been proposed that high levels of βC-activin subunit peptide may protect dividing cells from the apoptotic effects of activin A by forming intracellular heterodimers (Mellor et al. 2000). There was, however, very little evidence of βA-activin subunit immunoreactivity in mitotic hepatocytes, indicating that the heterotrophic effects of βC-activin subunit protein were independent of βA-activin subunit protein within mitotic hepatocytes. Mellor et al.(2000, 2003) and Vejda et al.(2002) reported the formation of βC-activin subunit peptide homodimers and heterodimers in vitro. Our study confirms that βC-activin homo- and heterodimers also form in vivo. When non-reduced samples were assessed, the β-activin subunit antibodies detected a range of activin dimers. The presence of a βAβC activin dimer was confirmed with both the βA-and βC-activin subunit antibodies. The exact identity of other in vivo βC-activin peptides needs to be confirmed by immunoprecipitation studies. To the best of our knowledge our study is the first evidence of in vivo βC-activin subunit peptide dimers in regenerating rat liver. Wada et al.(2004) propose that activin C homodimer, rather than activin A, is formed in normal liver and PHx may lead to increased activin A, terminating liver regeneration (Wada et al. 2004). We show activin homodimers are present at low to moderate levels in control liver while increased activin AC was evident in regenerating liver. Therefore the formation of activin heterodimers may antagonise the formation of homodimers and allow restoration of liver mass. βE-activin is also expressed in the liver and can dimerise with βA- and βC-activin subunits, and is therefore likely to be an additional player in the restoration of liver mass (Lau et al. 2000, Vejda et al. 2002).
Significant amounts of βC-activin subunit peptide were detected in the serum of rats by western blotting. During regrowth the liver is able to carry out most normal physiological functions; however, functional reserve is very low. Therefore some peptides, rather than being processed in the liver, are secreted (Fausto 1990). The presence of circulating βC-activin subunit dimers in control rat serum suggests these dimers are not merely a by-product of a liver with low functional reserve. However the local effects of liver activins are difficult to reconcile with evidence of secreted activins. These results may indicate that βC-activin subunit proteins, like activin A, have endocrine, autocrine and paracrine roles in vivo. These findings also raise the possibility that serum levels of βC-activin subunit peptide may be able to be used as a diagnostic marker of liver health.
The nuclear staining observed with the βC-activin subunit antibody in the later stages of regeneration of male rat liver mass (168 and 240 h) was unexpected. Nuclear localisation of βA-activin subunit in rat seminiferous epithelium was presented by Blauer and colleagues in 1999. This group showed the βA-activin subunit precursor contained a functional nuclear localisation sequence (Blauer et al. 1999). A search of protein databases (Psort database) indicates that human βC-activin contains a leucine zipper domain at amino acids 148–169 (http://psort.nibb.ac.jp). Despite the specific times post-hepatectomy at which nuclear staining was observed and our observation of nuclear βC-activin staining with a polyclonal antibody, in mouse ovarian cysts (Richardson et al. 2005), we cannot entirely discount the possibility that nuclear staining may be a tissue fixation artefact. In order to determine if nuclear localisation of βC-activin subunit in the later stages of regeneration of male rat liver mass indicates a novel signalling pathway, future studies will perform western blotting on nuclear extracts of regenerating rat liver.
Lau et al.(2000) proposed that βC-activin performs no essential in vivo function, as mice deficient in βC-activin showed no obvious phenotype, even when female mice were subjected to PHx. However, a redundancy of roles, with other TGF-βs substituting for the lack of the liver activins cannot be excluded. Furthermore, we consider it is highly unlikely that a protein with no physiological function in the liver would show in vivo temporal and cellular responses to liver insult.
We have demonstrated concurrent changes in the expression of βA-activin, βC-activin, follistatin and activin receptor subunit mRNA in the caudate lobe of regenerating rat liver, over 240 h post-PHx. βC-activin appears to be regulated in parallel to βA-activin in vivo, however the two subunits are not always localised in the same cell types.
We conclude that βA- and βC-activin subunit proteins are autocrine growth regulators in regenerating liver and when expressed independently lead to hepatocyte apoptosis (βA-activin) or mitosis (βC-activin) in a subset of hepatocytes. In the liver, apoptosis is an important part of the tightly controlled homeostatic mechanisms regulating liver function (Chen et al. 2000), perhaps indicating the importance of locally produced, autocrine growth regulators such as the activins in the liver. The mechanism by which βC-activin promotes hepatocyte mitosis and the role of the closely related βE-activin subunit in liver regeneration will form the basis of future studies.
Hepatocyte cell death (%) assessed by in situ TUNEL staining (filled columns) and cell division (open columns) –prophase, metaphase, anaphase or telophase – assessed histologically, in rat regenerating liver following PHx. Results are means±s.d., n=3 male rats per time point. Results were analysed by Mann–Whitney U-test: *P < 0.02, **P < 0.01, ***P < 0.001. Mitosis was evident at 48 and 72 h, whereas increased cell death occurred at 12 h and from 96 h on in regenerating male rat liver.
Citation: Journal of Molecular Endocrinology 34, 2; 10.1677/jme.1.01657
RT-PCR band densities, relative to β-actin densities, in the caudate lobe of livers from sham-operated control rats. Results are means±range of triplicate RNA extracts from male control rats (n=2 at each time point). Sham operation did not lead to significant differences in expression of mRNA for any of the genes analysed.
Citation: Journal of Molecular Endocrinology 34, 2; 10.1677/jme.1.01657
(A) βA-activin (•) and βC-activin (○) mRNA expression after PHx, assessed by RT-PCR, in triplicate RNA extracts from 5–7 male rats (2 controls and 3–5 PHx) at each time point. Data are presented after normalisation against the β-actin density, as the percentage of sham-operated control liver expression at each time point (means±s.d.). Data were log transformed and analysed by ANOVA for group differences (sham vs PHx) and time differences (12–240 h). βA- and βC-activin densities were significantly different from sham control values (P values between 0.001 and 0.05), 24 h post-PHx. βC-activin densities were significantly lower than sham control and other time points (P values between 0.001 and 0.05) at 72 h, whereas βA-activin densities were significantly higher than sham control and other time points (P values between 0.001 and 0.05) at 168 h post-PHx. (B) Expression of follistatin mRNA in regenerating rat liver expressed as the percentage of sham-operated control liver expression at each time point (means±s.d.). Follistatin densities were significantly higher than sham control values and other time points (P values between 0.001 and 0.05), 24 and 48 h post-PHx and significantly lower than sham control values and other time points (P < 0.05) at 72, 168 and 240 h post-PHx. (C) ActRIIA (•) and ActRIIB (○) mRNA expression as a percentage of sham control. ActRIIB densities were significantly higher than sham control values or other time points at 168 h post-PHx (P values between 0.001 and 0.05).
Citation: Journal of Molecular Endocrinology 34, 2; 10.1677/jme.1.01657
βA-activin (R&D, monoclonal antibody) (A) and βC-activin (clone 1, monoclonal antibody) (B) immunoreactive proteins in the caudate lobe of sham-operated control and regenerating rat liver were assessed by SDS–PAGE and non-reducing Western blot, at various times after PHx. Results are representative of duplicate, independent tissue extracts of caudate lobe samples from 3 rats at each time point. Bands representing pro-βA-or βC-activin were present at 43–46 kDa; activin A, activin AC, 23 kDa; activin C, 21 kDa; βA-activin, 13 kDa; and βC-activin, 12 kDa (Mellor et al. 2000). The control sample, run on a different gel to the PHx samples, contained more activin C homodimer (*) and less activin AC than the PHx samples.
Citation: Journal of Molecular Endocrinology 34, 2; 10.1677/jme.1.01657
Circulating βA-activin (A) and βC-activin (B) peptide subunit in serum of sham-operated control and PHx male rats during the restoration of liver mass. βA-activin (R&D, monoclonal antibody) and βC-activin (clone 1, monoclonal antibody) subunit peptides were assessed by SDS–PAGE and non-reducing Western blot in the serum of male rats at various times after PHx. Results were repeated in duplicate, independent serum samples (2 mg) from 3 rats at each time point. Immunoblots showed bands representing: activin A, 26 kDa; activin AC, 23 kDa; and activin C, 21 kDa (Mellor et al. 2000).
Citation: Journal of Molecular Endocrinology 34, 2; 10.1677/jme.1.01657
Selected views of βA-activin and βC-activin subunit peptide immunoreactivity in paraffin-embedded, paraformaldehyde-fixed sections of the caudate lobe of male rat regenerating liver. (A) Granular βA-activin subunit immunoreactivity in hepatocytes. Adjacent hepatocytes undergoing mitosis show no positive staining with the βA-activin antibody. (B) βA-activin subunit immunoreactivity in a cell undergoing apoptosis. (C) Cytoplasmic βC-activin subunit immunoreactivity in association with mitosis. (D) βC-activin subunit immunoreactivity within hepatocyte nuclei, 10 days after PHx. (E and F) βA-and βC-activin control sections incubated with mouse IgG. Final magnifications: A, B and C, ×400; D, ×40; E and F, ×200. No staining was observed in control sections incubated with primary antibody pre-absorbed with 100 times excess βA-or βC-activin subunit synthetic peptide (not shown, see Gold et al. 2003).
Citation: Journal of Molecular Endocrinology 34, 2; 10.1677/jme.1.01657
Mitotic (filled columns) and apoptotic (open columns) labelling indices for βA-activin and βC-activin subunit immunoreactivity. Slides from regenerating liver were assessed for markedly positive β-activin subunit staining in apoptotic (12 h, 7 days and 10 days PHx) or mitotic hepatocytes (24 h, 48 h and 72 h PHx). Hepatocytes undergoing cell death were identified by histological appearance and apoptosis confirmed on an adjacent TUNEL-stained section (not shown). Mitotic hepatocytes were identified based on histological appearance. Apoptotic or mitotic cells that were markedly positive for activin subunit immunoreactivity were counted and divided by the total number of cells assessed; results are presented as a staining index (%).
Citation: Journal of Molecular Endocrinology 34, 2; 10.1677/jme.1.01657
We thank the University of Otago Histology Unit for technical advice, Dr Peter Hurst (Department of Anatomy and Structural Biology, University of Otago) for guidance with assessing apoptotic and mitotic staining, Professor David Jans (Monash University) for comments on nuclear staining and Dr David Phillips (Monash IRD) for comments on the manuscript.
Funding
This research was supported by a University of Otago Research Grant (E J G, J S F) and NH&MRC programme grant number 143786 (G P R).
References
Blauer M, Husgafvel S, Syvala H, Tuohimaa P & Ylikomi T 1999 Identification of a nuclear localization signal in activin/inhibin betaA subunit; intranuclear betaA in rat spermatogenic cells. Biology of Reproduction 60 588–593.
Chabicovsky M, Herkner K & Rossmanith W 2003 Overexpression of activin beta(C) or activin beta(E) in the mouse liver inhibits regenerative deoxyribonucleic acid synthesis of hepatic cells. Endocrinology 144 3497–3504.
Chen W, Woodruff TK & Mayo KE 2000 Activin A-induced HepG2 liver cell apoptosis: Involvement of activin receptors and Smad proteins. Endocrinology 141 1263–1272.
Chen YG, Lui HM, Lin SL, Lee JM & Ying SY 2002 Regulation of cell proliferation, apoptosis, and carcinogenesis by activin. Experimental Biology and Medicine (Maywood, NJ)227 75–87.
Coerver KA, Woodruff TK, Finegold MJ, Mather J, Bradley A & Matzuk MM 1996 Activin signaling through activin receptor type II causes the cachexia-like symptoms in inhibin- deficient mice. Molecular Endocrinology 10 534–543.
Draper LB, Chong H, Wang E & Woodruff TK 1997 The uterine myometrium is a target for increased levels of activin A during pregnancy. Endocrinology 138 3042–3046.
Esquela A, Zimmers T, Koniaris L, Sitzmann J & Lee S-J 1997 Transient down regulation of inhibin-bC expression following partial hepatectomy. Biochemical and Biophysical Research Communications 235 553–556.
Fausto N 1990 Regulation of liver cell mass. In Oxford Textbook of Clinical Hepatology, pp 189–212. Eds J Bircher, J-P Benhamou, N McIntrye, M Rizzetto & J Rodes. Oxford: Oxford University Press.
Fausto N, Laird AD & Webber EM 1995 Liver regeneration. 2. Role of growth factors and cytokines in hepatic regeneration. FASEB Journal 9 1527–1536.
Gold EJ, Francis RJ, Zimmermann, A, Mellor SL, Cranfield M, Risbridger GP, Groome NP, Wheatley AM & Fleming JS 2003 Changes in activin and activin receptor subunit expression in rat liver during the development of CCl4-induced cirrhosis. Molecular and Cellular Endocrinology 201 143–153.
Gold EJ, O’Bryan MK, Mellor SL, Cranfield M, Risbridger GP, Groome NP & Flemming JS 2004 Cell-specific expression of βC-activin in rat ovary, testis, pituitary, uterus and adrenal, as well as in rat liver. Molecular and Cellular Endocrinology 222 60–69.
Higgins G & Anderson R 1932 Experimental pathology of the liver 1. Restoration of the liver of the white rat following partial surgical removal. Archives of Pathology 12 186–202.
Ho J, de Guise C, Kim C, Lemay S, Wang XF & Lebrun JJ 2004 Activin induces hepatocyte cell growth arrest through induction of the cyclin-dependent kinase inhibitor p15ink4b and sp1. Cellular Signalling 16 693–701.
Hotten G, Neidhardt H, Schneider C & Pohl J 1995 Cloning of a new member of the TGF18 beta family: a putative new activin beta C chain. Biochemical and Biophysical Research Communications 206 608–613.
Hubner G, Alzheimer C & Werner S 1999 Activin: a novel player in tissue repair processes. Histology and Histopathology 14 295–304.
Hully J, Chang L, Schwall R, Windmer H, Terrell T & Gillett N 1994 Induction of apoptosis in the murine liver with recombinant human activin A. Hepatology 20 854–861.
Kingsley DM 1994 The TGF-beta superfamily – new members, new receptors, and new genetic tests of function in different organisms. Genes Development 8 133–146.
Kobayashi T, Niimi S, Fukuoka M & Hayakawa T 2002 Regulation of inhibin β chains and follistatin mRNA during rat hepatocyte growth induced by peroxisome proliferator di-n-butyl phthalate. Biological and Pharmaceutical Bulletin 25 1214–1216.
Kogure K, Omata W, Kanzaki M, Zhang YQ, Yasuda H, Mine T & Kojima I 1995 A single intraportal administration of follistatin accelerates liver regeneration in partially hepatectomized rats. Gastroenterology 108 1136–1142.
Kogure K, Zhang YQ, Kanzaki M, Omata W, Mine T & Kojima I 1996 Intravenous administration of follistatin: delivery to the liver and effect on liver regeneration after partial hepatectomy. Hepatology 24 361–366.
Kogure K, Zhang YQ, Shibata H & Kojima I 1998 Immediate onset of DNA synthesis in remnant rat liver after 90% hepatectomy by an administration of follistatin. Journal of Hepatology 29 977–984.
Kogure K, Zhang YQ, Maeshima A, Suzuki K, Kuwano H & Kojima I 2000 The role of activin and transforming growth factor-beta in the regulation of organ mass in the rat liver. Hepatology 31 916–921.
Kozian DH, Ziche M & Augustin HG 1997 The activin-binding protein follistatin regulates autocrine endothelial cell activity and induces angiogenesis. Laboratory Investigation 76 267–276.
Lau AL, Nishimori K & Matzuk MM 1996 Structural analysis of the mouse activin beta C. Biochimica et Biophysica Acta 1307 145–148.
Lau AL, Kumar TR, Nishimori K, Bonadio J & Matzuk MM 2000 Activin betaC and betaE genes are not essential for mouse liver growth, differentiation, and regeneration. Molecular and Cellular Biology 20 6127–6137.
Leevy CB 1998 Abnormalities of liver regeneration: a review. Digestive Diseases 16 88–98.
Mellor SL, Cranfield M, Ries R, Pedersen J, Cancilla B, de Kretser D, Groome NP, Mason AJ & Risbridger GP 2000 Localization of activin beta(A)-, beta(B)-, and beta(C)-subunits in human prostate and evidence for formation of new activin heterodimers of beta(C)-subunit. Journal of ClinicalEndocrinology and Metabolism 85 4851–4858.
Mellor SL, Ball EM, O’Connor AE, Ethier JF, Cranfield M, Schmitt JF, Phillips DJ, Groome NP & Risbridger GP 2003 Activin betaC-subunit heterodimers provide new mechanism of regulating activin levels in the prostate. Endocrinology 144 4410–4419.
Phillips DJ 2000 Regulation of activin’s access to the cell: why is mother nature such a control freak? Bioessays 22 689–696.
Richardson MJ, Beaugié CR, McQuillan HJ & Fleming JS 2005 Nuclear expression of βC activin in ovarian cysts from mice with varying ovulation number. Proceedings of the Endocrine Society of Australia vol 47 (2005), NZ Society of Endocrinology Supplement (abstract B19).
Rosendahl A, Checchin D, Fehniger TE, ten Dijke P, Heldin C-H & Sideras P 2001 Activation of the TGF-ββ/activin-Smad 2 pathway during allergic airway inflammation. American Journal of Respiratory Cell and Molecular Biology 25 60–68.
Schmitt J, Hotten G, Jenkins NA, Gilbert DJ, Copeland NG, Pohl J & Schrewe H 1996 Structure, chromosomal localization, and expression analysis of the mouse inhibin/activin beta C (InhbC) gene. Genomics 32 358–366.
Schwall RH, Robbins K, Jardieu P, Chang L, Lai C & Terrell TG 1993 Activin induces cell death in hepatocytes in vivo and in vitro. Hepatology 18 347–356.
Vejda S, Cranfield M, Peter B, Mellor SL, Groome N, Schulte-Hermann R & Rossmanith W 2002 Expression and dimerization of the rat activin subunits βC and βE: evidence for the formation of novel activin dimers. Journal of Molecular Endocrinology 28 137–148.
Vejda S, Erlach N, Peter B, Drucker C, Rossmanith W, Pohl J, Schulte-Hermann R, Grusch M 2003 Expression of activins C and E induces apoptosis in human and rat hepatoma cells. Carcinogenesis 24 1801–1809.
Wada W, Maeshima A, Zhang YQ, Hasegawa Y, Kuwano H, Kojima I 2004 Assessment of the function of the betaC-subunit of activin in cultured hepatocytes. American Journal of Physiology -Endocrinology and Metabolism 287 E247–54.
Yasuda H, Mine T, Shibata H, Eto Y, Hasegawa Y, Takeuchi T, Asano S & Kojima I 1993 Activin A: an autocrine inhibitor of initiation of DNA synthesis in rat hepatocytes. Journal of Clinical Investigation 92 1491–1496.
Ying S-H 1988 Inhibins, activins and follistatins: gonadal proteins modulating the secretion of follicle-stimulating hormone. Endocrine Reviews 9 267–293.
Zauberman A, Oren M & Zipori D 1997 Involvement of p21 (waf1/cip1), cdk4 and rb in activin a mediated signaling leading to hepatoma cell growth inhibition. Oncogene 15 1705–1711.
Zhang YQ, Kanzaki M, Shibata H & Kojima I 1997a Regulation of the expression of follistatin in rat hepatocytes. Biochimica et Biophysica Acta 1354 204–210.
Zhang YQ, Shibata H, Schrewe H & Kojima I 1997b Reciprocal expression of mRNA for inhibin betaC and betaA subunits in hepatocytes. Endocrine Journal 44 759–764.
