HMG Box Transcriptional Repressor HBP1 Maintains a Proliferation Barrier in Differentiated Liver Tissue

Construction of HBP1 transgenic mice.

The entire coding region of the rat HBP1 DNA (GenBank accession number U09551) with an N-terminal hemagglutinin (HA) epitope tag was blunt ended and inserted into the StuI site of the pTTR1ExV3 vector (Fig. 1A), resulting in clone TTR-HBP. The pTTR1ExV3 vector contains transcriptional regulatory sequences from the murine transthyretin (TTR; a generous gift from Terry Van Dyke [48]). A ClaI-SmaI fragment comprising the TTR promoter with the HA-tagged HBP1 cDNA was microinjected into single-cell mouse oocytes derived from male B2D6F1 mice crossed with female C57BL/6 mice. Founder mice and their offspring were crossed back into the C57BL/6 background to maintain genetic homogeneity. Genotyping of transgenic mice was carried out using Southern blotting or PCR with tail DNA. The PCR assay was transgene specific using a 5′ primer (5′-AAAGTCCTGGATGCTGTCCGAG-3′) hybridizing to the second TTR exon (48) and a 3′ primer (5′-CACTTTGAACAGCCTGAAG-3′) hybridizing to rat HBP1 cDNA. A similar strategy was used to construct the transgenic vectors for the N-terminal deletion of HBP1.

Partial hepatectomy and tritiated thymidine incorporation assay.

The [³H]thymidine incorporation assay provides a quantitative measurement for the kinetics of DNA synthesis during the first synchronous cell cycle after partial hepatectomy (0 to 54 h). Thus, subtle differences are reliably detected. Three-month-old HBP1 transgenic mice and nontransgenic littermates matched for weight, age, and sex were used in this assay. Each animal was anesthetized with 0.015 ml of avertin/g of body weight. Two-thirds partial hepatectomy was performed by ligating and then resecting the left lateral and median lobes of the liver. Sham-operated animals received laparatomy without liver damage. Animals were sacrificed at indicated time points. One hour before death, each mouse was injected with 50 μCi of [³H]thymidine intraperitoneally. Mice were sacrificed; the livers were frozen in liquid nitrogen. Approximately 80 mg of liver was homogenized in cold 10% trichloroacetic acid (TCA). Homogenates were centrifuged (1,500 × g), and the pellets were washed in warm 95% ethanol to remove fatty acids. The precipitates were subsequently boiled in 5% TCA to solubilize DNA. The [³H]thymidine incorporation into DNA was determined by liquid scintillation. The DNA content was determined using a diphenylamine colorimetric assay with calf thymus DNA as a standard (37). The data were expressed as counts per minute of [³H]thymidine incorporated per microgram of DNA. Generally, the average and standard error of at least three measurements were shown with the indicated P values. The latter were determined by standard Student t test.

RNase protection assay.

Total RNA was isolated from frozen liver tissues using Trizol reagent (Sigma) according to the manufacturer's directions. RNase protection assays were performed using the RPAIII kit (Ambion) according to the manufacturer's directions. The RNA probes were derived from murine HBP1, rat HBP1, murine c-Jun, murine c-Fos, murine 18S, and murine H3.2. The specific probes are readily available upon request.

Tissue analysis.

Freshly isolated tissues were dounced in WCE buffer (25 mM HEPES [pH 7.5], 300 mM NaCl, 1.5 mM MgCl₂, 0.2 mM EDTA, 0.1% Triton X-100, 20 mM β-glycerophosphate, 0.1 mM Na₃VO₄, 0.5 mM dithiothreitol (DTT) 0.2 mM phenylmethylsulfonyl fluoride (PMSF), 1 μg of leupeptin per ml, 1 μg of pepstatin per ml) and extracted by rocking for 30 min at 4°C. The homogenates were centrifuged at 100,000 × g for 30 min at 4°C. The supernatant was collected and stored at −80°C. To detect the expression of HA-HBP1 protein, a total of 2 mg of extract was rocked with 30 μl of protein A beads (Repligen) at 4°C for 30 min. The beads were centrifuged and discarded. The supernatants were incubated with rabbit anti-HBP1 antiserum at 4°C for 1 h. A volume of 50 μl of protein A beads was added to each sample, and the sample was rocked at 4°C for 1 h. The beads were collected by centrifugation, washed three times with WCE buffer, and boiled in sodium dodecyl sulfate (SDS) sample buffer. The immune complexes were analyzed by Western blotting after SDS–7.5% polyacrylamide gel electrophoresis (PAGE). The presence of HA-HBP1 transgene was visualized by enhanced chemiluminescence detection with a mouse anti-HA monoclonal antibody, 12CA5, and goat anti-mouse horseradish peroxidase antibody (diluted 1:5,000; Jackson). The endogenous HBP1 protein was detected by immunoblot with a rabbit antibody made against the HMG box region of rat HBP1. The expression of cell cycle genes was examined by direct Western blotting. Frozen livers were dounced and extracted in TLB (20 mM Tris [pH 7.4], 137 mM NaCl, 2 mM EDTA[pH 8.0], 1% Triton X-100, 25 mM β-glycerophosphate, 1 mM Na₃VO₄, 2 mM sodium pyrophosphate, 10% glycerol, 0.2 mM PMSF, 1 μg of leupeptin per ml, 1 μg of pepstatin per ml) in the same way as described above. A total of 200 μg of tissue extract was used for Western analysis. The primary antibodies for cyclin E, cyclin A, PCNA, and p38 were purchased from Santa Cruz Biotechnology, Inc.

Immunoprecipitation kinase assay.

Fresh liver was homogenized in TLB buffer as described above. A total of 500 μg of liver extract was precleared with 30 μl of protein A Sepharose beads (Amersham) with gentle rocking at 4°C for 30 min. A volume of 30 μl of protein A Sepharose beads was incubated with 1 μg of anti-cyclin E antibody (Santa Cruz) at room temperature for 30 min, and the beads-antibody conjugate was washed once with TLB buffer. The precleared extract was briefly centrifuged. The supernatant was incubated with the beadantibody conjugate at 4°C for 2 h with gentle rocking. The beads were then collected by centrifugation and washed twice with TLB buffer and twice with kinase buffer (25 mM HEPES [pH 7.5], 20 mM β-glycerophosphate, 2.5 mM MgCl₂, 0.5 mM Na₃VO₄, 2 mM DTT). For kinase reactions, the beads were resuspended in 20 μl of kinase buffer containing 10 μCi of [γ-³²P]ATP (3,000 Ci/mmol; ICN), 20 μM ATP (Boehringer Mannheim), and 1 μg of histone H1 (Sigma). The reaction was performed at room temperature for 30 min and was stopped by adding 20 μl of SDS sample buffer. The ³²P-labeled proteins were visualized by SDS–10% PAGE and autoradiography.

Liver-specific expression of HBP1 in transgenic mice.

We established five pedigrees of transgenic mice expressing wild-type HBP1 under control of the TTR 5′ flanking sequence (Fig. 1A). For specific detection of the transgene, an influenza virus HA epitope tag was incorporated at the amino terminus of the HBP1 coding region (Fig. 1A). The TTR promoter was selected because transgene expression is restricted to the postnatal liver, thus avoiding potentially deleterious effects of overexpressing a potential cell cycle regulatory protein during murine development (48). As shown below, all experiments are conducted under modest overexpression of HBP1, despite the use of the powerful TTR promoter. We used RNase protection assays with gene-specific RNA probes that could discriminate between the endogenous (mouse) and the transgenic (rat) HBP1 RNAs. The expression of transgenic HBP1 RNA was approximately twofold higher than that of the endogenous murine gene in the highest expressing pedigree, HBP 1Z1 (see Fig. 1B). Similarly, the transgenic rat HBP1 protein was at most two fold greater that the endogenous mouse HBP1 protein (see Fig. 5).

We used the same RNase protection assay to determine the tissue profile of transgenic HBP1 expression. Unlike the ubiquitously expressed endogenous HBP1 mRNA (data not shown), transgenic HBP1 mRNA was expressed only in the liver. To confirm this, we used immunoblotting of HBP1 immunoprecipitates with an anti-HA antibody to demonstrate tissue specificity, since no transgenic HBP1 protein expression was detected in other tissues (Fig. 1C). Results from the highest expressing line, HBP 1Z1, are shown in subsequent figures, but essentially identical results were noted with pedigree HBP4, the second highest HBP1 expressing line. The pattern of endogenous and transgenic HBP1 protein expression during liver regeneration will be described in Fig. 5 in the light of cell cycle changes.

Importantly, animals from all pedigrees appeared to develop and reproduce normally. The modest overexpression of HBP-1 did not affect liver growth and development, since transgenic mice showed a normal liver-to-body mass ratio. Examination of multiple livers from the HBP 1Z1 and HBP4 lines revealed no discernible changes in liver structure or hepatocyte morphology. Serum liver function tests, which included alanine aminotransferase, aspartate aminotransferase, albumin, glucose, cholesterol, and insulin-like growth factor 1, were identical in transgenic mice and nontransgenic littermates (data not shown). These results suggested that HBP1 overexpression did not induce any gross or cellular liver dysfunction.

HBP1 transgenic mice show delayed onset of S phase during liver regeneration following partial hepatectomy.

To determine whether HBP1 regulates cell cycle activity in the liver, we examined the effects of HBP1 overexpression on the synchronous cell cycle reentry of normally quiescent hepatocytes that occurs during liver regeneration following partial hepatectomy. We first examined [³H]thymidine incorporation into DNA following two-thirds partial hepatectomy to determine the timing of S phase in both transgenic and nontransgenic animals. In nontransgenic mice, a peak in [³H] thymidine incorporation occurred 38 h after resection (Fig. 2A). This correlates well with published murine S-phase kinetics in response to partial hepatectomy (16, 24). In contrast, transgenic mice overexpressing HBP1 showed a significant delay (P < 0.05) in S phase, with peak DNA synthesis occurring at 48 h (Fig. 2A). As expected, control sham surgeries did not elicit a proliferative response in either mouse set and is represented as the zero time point. To confirm the [³H]thymidine labeling studies, we examined expression of histone 3.2 (H3.2) RNA as an independent marker which is highly induced in S phase (Fig. 2B and C). The expression of H3.2 in the HBP1 transgenic liver occurred at 48 h after partial hepatectomy. In contrast, the peak expression of H3.2 in nontransgenic mice occurred at 36 h. Thus, two independent assays demonstrate a delay in S phase for HBP1 transgenic mice following a two-third partial hepatectomy.

HBP1 does not affect the onset of immediate-early gene induction following partial hepatectomy.

The induction of immediate-early genes by cytokine and growth factor signaling pathways marks the G₀-to-G₁ transition during liver regeneration (reviewed in reference 16). A delay in activation of these signaling pathways in mice overexpressing HBP1 represents one possible mechanism for the observed delay in S phase. The analysis of immediate-early gene induction revealed rapid and identical induction of c-Fos and c-Jun at 30 min after partial hepatectomy in both non-transgenic and transgenic mice. This result indicated that the onset of upstream signaling pathways was unaffected by the overexpression of HBP1 (Fig. 3). The levels of c-Fos and c-Jun fell rapidly in nontransgenic mice to low levels in 2 h, consistent with the transient nature of the immediate-early response and a characteristic of G₁. In contrast, the decline in the immediately-early response was dysregulated in the transgenic mice overexpressing HBP1, with persistent expression of c-Fos and c-Jun at 2 and 8 h after partial hepatectomy. A second wave of immediate-early gene expression characteristic of late G₁ at 24 h was seen in both nontransgenic and HBP1 transgenic mice. However, the immediate-early gene response was blunted in the HBP1 transgenic mice, further indicating dysregulation of the cell cycle. By 48 h after partial hepatectomy, expression of c-Fos or c-Jun was undetectable (data not shown). The identical onset of the immediate-early response in both HBP1 transgenic and nontransgenic mice suggested that the delay in S phase seen in transgenic mice does not result from a delay in the G₀-to-G₁ transition. Furthermore, the normal onset indicates that transgenic HBP1 expression apparently does not affect the upstream signaling pathways that initiate proliferation and G₁ entry. However, the prolonged immediate-early response observed in HBP1 transgenic animals suggested the possible alteration of normal G₁ progression.

Relative to nontransgenic mice, the expression of the p21 cyclin-dependent kinase (CDK) inhibitor was also increased in the HBP1 transgenic mice after partial hepatectomy and suggested a possible delay in G₁ progression. As shown in Fig. 3, the normal pattern of p21 expression was transient in G₁, with a strong maximum at 8 h. By contrast, p21 expression was dysregulated in the HBP1 transgenic mice, with elevated p21 expression extending to 24 h. By 48 h after partial hepatectomy in both nontransgenic and HBP1 transgenic lines, there was no detectable p21 RNA expression (data not shown). Nonetheless, the increase in p21 expression in the transgenic mice was consistent with a delayed S phase and will be interpreted further below.

Extended G₁ progression results in an S-phase delay in HBP1 transgenic liver.

To determine if delayed G₁ progression was the basis of the delayed S-phase onset in the HBP1 transgenic mice, the expression of several cell cycle markers for the G₁ and S phases was examined. We selected only those markers that had been characterized for the liver regenerative cell cycle. Induction of cyclin D1 mRNA, a marker of early G₁, occurred by 16 h in both nontransgenic and HBP1 transgenic mice, indicating that the early transition into early G₁ was unaltered (Fig. 4A). The kinetic profile of D1 mRNA was consistent with a recently published study on ERK activation in liver regeneration (46).

In contrast, the expression of the late G₁ marker, cyclin E, was delayed in the regenerating livers from HBP1 transgenic mice after partial hepatectomy (Fig. 4A and B). In nontransgenic mice, cyclin E protein was induced at 16 h after partial hepatectomy, whereas induction was delayed by 8 h in the HBP1 transgenic animals. The onset of cyclin E-associated CDK activity was delayed by 12 h in the HBP1 transgenic mice (Fig. 4B). Because assembly of the active kinase is required, cyclin E-associated kinase activity temporally follows the induction of cyclin E protein and is an important factor in G₁ progression (22). Thus, a delay in G₁ progression in the HBP1 transgenic mice was supported by a delay in both cyclin E protein and its associated kinase activity.

Extending the observation of S-phase delay in HBP1 transgenic mice (Fig. 2), the expression of the S-phase-associated markers cyclin A and PCNA was also delayed in the HBP1 transgenic mice (Fig. 4C). In non transgenic liver, the cyclin A protein was detectable at 24 h and peaked at 38 h after partial hepatectomy. By contrast, in the HBP1 transgenic liver the initial induction of cyclin A was delayed to 32 h, with a peak at 48 h after partial hepatectomy. PCNA exhibited a similar delay in expression to that observed with cyclin A in the transgenic mice. The results with cyclin A and PCNA further confirm the delay in S phase in mice overexpressing HBP1. The kinetics of cyclin E and p21 gene expression suggest that delays in S-phase entry during liver regeneration in the transgenic mice result from prolongation of mid-G₁ phase following partial hepatectomy.

HBP1 protein levels are regulated during liver regeneration.

In Fig. 5, we compared the expression of endogenous HBP1 across the G₁ and S phases of the liver cell cycle using extracts similar to those used for Fig. 2 through 4. The specific HBP1 antibody was generated against the HMG box of the rat HBP1. The HBP1-specific bands are denoted with arrows in Fig. 5. The criterion for specificity is that the signal is abolished with antiserum that is preadsorbed with the glutathione S-transferase (GST)-HMG box antigen but not with GST protein (data not shown). In nontransgenic mice, a modest twofold increase in HBP1 protein was consistently observed at 8 h (early G₁). These results are similar to the data previously reported for HBP1 mRNA (25 and unpublished data). In addition, we consistently observed that the expression of the endogenous HBP1 protein declined in late G₁ (24-h point; lane 4) but returned to control levels by 48 h (compare the first and fifth lanes). As a control, p38 mitogen-activated protein (MAP) kinase protein was unchanged, demonstrating specificity to the changes in HBP1 protein levels. Interestingly, the decline in HBP1 protein at 24 h correlates with G₁ progression, suggesting that HBP1 transgene expression may differ from endogenous expression, resulting in the altered G₁ phenotype.

In the HBP1 transgenic mice, two bands corresponding to endogenous and transgenic HBP1 were detected with the anti-HBP1 antibody. In the HBP1 transgenic mice, the endogenous HBP1 protein declined at 24 h, consistent with the endogenous expression in nontransgenic mice. Yet the transgenic HBP1 protein was still expressed at 24 h. Thus, the overall level of HBP1 remained elevated compared to that nontransgenic mice at 24 h after partial hepatectomy. Finally, in the HBP1 transgenic mice the expression of both transgenic and endogenous HBP1 declined at 48 h. These results are consistent with the observed S-phase delay in HBP1 transgenic mice. Thus, the delayed G₁ progression seen in Fig. 2 to 4 upon transgenic HBP1 expression could be attributed to both overall increased HBP1 expression and a delay in the decline of HBP1 protein levels. Because endogenous and transgenic HBP1 were expressed from different promoters, this analysis indicates that the overall HBP1 protein levels may be a result of both transcription and regulation of protein stability during the cell cycle.

Mutant HBP1 increases the magnitude of S phase in liver regeneration.

If the function of HBP1 could be blocked by a dominant-interfering mutant of HBP1, the predicted phenotype would be an enhanced cell cycle response. We had previously characterized a mutant HBP1 (ΔPst) in which the N-terminal 50 amino acids are deleted, thus removing the high-affinity binding site for RB and p130. As expected, this mutant has decreased interaction with RB and p130 and has reduced potency as a transcriptional repressor (46). This deleted 50-amino-acid region of HBP1 is highly conserved in HBP1 protein from human, mouse, rat, and zebra fish (see GenBank). The prolonged G₁ phase and delayed S phase following two-thirds partial hepatectomy in the HBP transgenic mice implied that HBP1 regulated G₁ progression in the differentiated liver. We reasoned that this HBP1 mutant may have reduced RB regulation and might exhibit a dominant-interfering phenotype with respect to the cell cycle. If the role of HBP1 is to control cell G₁ progression, we asked if this mutant HBP1 might enhance S phase in response to partial hepatectomy. Therefore, we examined the consequences of expressing the mutant HBP1 in transgenic mice on the cell cycle during liver regeneration.

Three lines of transgenic mice expressing the mutant HBP1 under the control of the TTR promoter were identified (Fig. 6A). All lines appeared to develop and reproduce normally and had normal liver function and morphology using the criteria discussed previously. The HA tag was inadvertently omitted, and the predicted size of the mutant HBP1 protein migrated with background bands in the Western blots (not shown). Thus, we relied on the RNA expression of the mutant HBP1 transgene in our analysis. Two lines (denoted 1U3 and 6) expressed the highest levels of mutant HBP1 RNA expression (Fig. 6A) and were selected for detailed studies.

The cell cycle analysis of the mutant HBP1 transgenic mice had different considerations than those for Fig. 2 to 4. We reasoned that enhancement of the liver cell cycle by mutant HBP1 expression might be optimally detected with a less robust stimulus and chose a one-third partial hepatectomy. It is well established that this mitogenic stimulus results in an attenuated proliferative response and lower [³H]thymidine incorporation than two-thirds partial hepatectomy (9) (compare Fig. 2 and 6). In the nontransgenic mice, the peak of S phase occurred with different kinetics—48 h after one-third partial hepatectomy compared to 38 h with a two-thirds partial hepatectomy.

Mice expressing wild-type and mutant HBP1 exhibited different proliferative responses to partial hepatectomy. As shown in Fig. 6B, mice expressing the mutant HBP1 exhibited an increased S-phase response to one-third partial hepatectomy. As expected, the S-phase response to a one-third partial hepatectomy of wild-type HBP1 transgenic mice was consistently less robust than the response to the stronger two-thirds partial hepatectomy (compare Fig. 2 and 6). In the wild-type HBP1 transgenic mice, the weaker response to one-third partial hepatectomy hampered a more detailed kinetic and gene expression analysis (data not shown). The previously observed kinetic differences were probably not detectable due to reduced signal strength. This observation illustrated both the flexibility and the importance of signal strength for inducing different levels of proliferation in the liver model.

By contrast, expression of the mutant HBP1 mice gave an increased S-phase response, with consistent changes in expression of the immediate-early and cell cycle genes upon the one-third partial hepatectomy signal. The expression patterns of immediately-early and cell cycle genes were analyzed in the nontransgenic and mutant HBP1 transgenic mice. Expression of two independent markers of S phase, H3.2 and cyclin A, were both increased at 48 h in transgenic mice expressing the mutant HBP1, again confirming increased S-phase magnitude (Fig. 6C and D). We observed a comparable enhancement of DNA synthesis in another line expressing the mutant HBP1 at lower levels, suggesting this reduced expression was sufficient for a near maximal response. However, no differences were observed when a maximal stimuli was delivered by a two-thirds partial hepatectomy, suggesting that expression of mutant HBP could not augment a maximal proliferative response (data not shown). Taken together, the mutant HBP1 exhibited some functional characteristics of a dominant-interfering phenotype with an augmentation of proliferation in response to a one-third partial hepatectomy.

We examined the immediate-early response following one-third partial hepatectomy to determine whether the mutant HBP1 altered the earliest signaling events. c-Jun and c-Fos expression rose rapidly 30 min after partial hepatectomy in both nontransgenic mice and mice expressing the HBP1 mutant and indicated that the mutant protein did not alter the initial signaling response. In nontransgenic mice, the c-Jun and c-Fos signals are sustained at 2 h and then decline at 8 h. Again, the kinetics are similar to those described previously for mice that have undergone a one-third partial hepatectomy (9). However, at 2 and 8 h, the abundance of c-Jun and c-Fos transcripts was increased in mice expressing the HBP1 mutant (Fig. 7). By 24 and 48 h after a one-third partial hepatectomy, the c-Jun and c-Fos expression was reduced to background levels in both transgenic and nontransgenic mice, reflecting the normal transient expression profile in G₁. Relative to the nontransgenic mice, the p21 CDK inhibitor declined at least 6 h earlier in the mutant HBP1 transgenic mice. While an increase in overall p21 expression was observed in the mutant HBP1 transgenic mice, the basis remains unclear. The rapid decline of p21 was consistent with a more robust S phase. The results shown in Fig. 6 and 7 are consistent with the notion that expression of mutant HBP1 enhances the proliferative response during liver regeneration.

Summary of results.

Our results with transgenic mice expressing HBP1 in the liver support a role for HBP1 as a cell cycle inhibitor in differentiated tissues. Furthermore, our data are consistent with a model in which HBP1 regulates the progression through the G₁ phase of the cell cycle (Fig. 8). Specifically, the mid-late G₁ expression of cyclin E and associated kinase activity in the HBP1 transgenic mice was delayed. The decline of immediate-early gene expression, an early-to-mid G₁ event, was inhibited in the HBP1 transgenic mice. Because of an extended G₁ phase, S-phase entry was detectably delayed in HBP1 transgenic mice. Interestingly, expression of a defective HBP1 protein in the liver led to an increased S-phase magnitude. This suggested that the mutant HBP1 may interfere with the normal HBP1 functions. Finally, the effects of HBP1 are likely limited to cell cycle regulation and not the upstream signaling cascades initiated by the plethora of cytokines and growth factors described for liver regeneration. In both wild-type HBP1 and mutant HBP1 transgenic mice, the onset of immediate-early gene expression was normal, suggesting that signaling cascades and the G₀-to-G₁ transition were intact. By contrast, the expression pattern of p21 CDK inhibitor correlated with the different S-phase responses in mice expressing either wild-type or mutant HBP1 and suggested that p21 may mediate the cell cycle changes (see below for further discussion). Lastly, a complexity is that the HBP1 protein itself may be subject to cell cycle regulation, regardless of expression from the endogenous or transgenic promoters. Strikingly, when S phase was delayed in the transgenic HBP1 mice, the decline in HBP1 protein was also delayed. The data depicted in Fig. 5 suggest that both transcriptional and posttranscriptional regulation may contribute to the overall HBP1 levels. Taken together, we conclude that expression of HBP1 is inhibitory to G₁ progression in liver regeneration. Because the level of transgenic overexpression was modest, HBP1 protein levels must be subject to tight control to ensure orderly cell cycle progression.

Merits of the liver model.

Understanding cell proliferation controls in normal tissues requires the use of an experimental model where cell cycle kinetics can be manipulated reproducibly. For this reason, we have focused on the model of liver regeneration. It is unlikely that the questions addressed in the present work could have been addressed well in primary hepatocyte culture where differentiation is lost. For example, TNF-α is a mitogen for cultured hepatocytes (39) and is required for liver regeneration (1). However, the true function of TNF-α, which appears to prime hepatocytes for growth, required the use of animal models (16, 47, 49). Several other reports demonstrate the difficulty in using cultured hepatocytes to recapitulate gene regulation of the differentiated liver (20, 21, 29, 36, 44). By focusing on the reproliferation of the differentiated liver, our data suggest that HBP1 is a component of the growth suppression machinery in normal tissues that restrain proliferation. Specifically, the primary effect of HBP1 expression is unlikely to regulate the direct action on the signaling pathways activated by cytokines and/or growth factors. Rather, our data suggest that HBP1 may be involved in regulating the proliferative response to these factors by controlling G₁ progression.

Several studies have now highlighted the utility of liver regeneration for probing cell cycle regulation. In this paper and in other works, the goal was to examine the responding proteins that may govern G₀-to-G₁-to-S transitions. The end result has been the discovery of some expected and unexpected G₁ regulators in liver regeneration. Of the expected G₁ regulators, Van Dyke and coworkers established that p21 was a relevant regulator by using the TTR transgenic model (48). Of particular relevance, overexpression of p21 resulted in a block in liver regeneration. Three different studies have all used the liver regeneration models and have identified some unexpected G₁ regulators. Mice that are deficient in CREM have a similar liver regeneration and G₁ phenotype compared to that of HBP1 (42). CREM is a regulator of the CREB transcription factor and of the cyclic AMP pathway. Second, the partial hepatectomy model has been used to dissect a requirement for ERK in G₁ progression during regeneration (45). Finally, expression of the winged helix transcription factor HFH-11 in the liver using the TTR transgenic expression system resulted in increased S phase during regeneration. The authors suggested that HFH-11 is a positive regulator of hepatocyte proliferation (50), and their observations are grossly similar to ours with mice expressing the mutant HBP1.

While there is some similarity in the studies of these very different signalling networks, the potential relationship between CREM-, ERK-, HFH-11-, and/or HBP1-mediated cell cycle pathways is unclear. In normal growth homeostasis, the regulation of HBP1 and other proteins would be expected to restrain proliferation of the differentiated liver. The nearest link is with p21, for which Van Dyke showed that transgenic overexpression gave a regeneration defect (48). In our case, increased p21 expression in the HBP1 transgenic mice correlated with delayed G₁ progression. An attractive idea is that the increased p21 levels could be responsible for the S-phase delay in the wild-type HBP1 transgenic mice. How HBP1 expression in the mouse liver leads to increased p21 RNA expression is unclear. To add complexity, we have shown that HBP1 repressed the p21 promoter in cell lines (18). However, it is well known that p21 gene expression involves multiple transcriptional regulators and posttranscriptional stabilization in liver regeneration (2, 3). Furthermore, the factors that control p21 expression in liver tissue and cell lines may differ. In the liver studies shown in Fig. 2 to 4, the sustained p21 levels correlate well with the cell cycle delay in the HBP1 transgenic mice.

Possible pathways for HBP1 involvement.

What might be the pathway(s) affected by HBP1 expression? An earlier work with cell culture showed that transcriptional repression by HBP1 contributed to cell cycle regulation. It showed that RB and p130 augment intrinsic transcriptional repression by HBP1 (46). The loss of proliferation inhibition was consistent with loss of RB regulation, since the higher affinity RB interaction motif was deleted in the HBP1 mutant, which resulted in a heightened cell cycle response (Fig. 5 and 6). Recent work has shown that an HBP1 site may regulate chromatin configuration in position-effect variegation (58). Mechanistically, interaction of RB or p130 with HBP1 may be significant in chromatin regulation. Since one aspect of RB and p130 repression occurs through recruitment of histone deacetylases and DNA methylases (reviewed in references 8 and 19), a new twist might be that HBP1 (and RB) might provide repression by regulating chromatin configuration. Whether the loss of the major RB and p130 binding site is the sole reason behind the observed mutant HBP1 phenotype awaits further investigation of the liver. Intriguingly, the increased S-phase response of the mutant HBP1 is limited to the animal model and has not yet been recapitulated in cell culture models.

While we have focused on the proliferation properties of HBP1 in this work, HBP1 is a unique member of the HMG box transcription factor family and remains one of few dedicated transcriptional repressors of this class (46). HMG box transcription factor family members LEF and TCF have a major function as the nuclear effectors of the Wnt/β-catenin oncogenic pathway. Several components (e.g., adenomatous polypoisis coli [APC], β-catenin, and axin) are mutated in cancer. The Wnt/β-catenin oncogenic pathway signals to the LEF and TCF HMG box transcription factors to activate transcription of oncogenes (e.g., cyclin D1 and c-MYC) and other growth and/or tumor regulatory genes (7, 35). The Wnt/β-catenin pathway has special significance in the liver. Hepatocellular carcinoma (HCC) is now associated with mutations in the Wnt/β-catenin pathways. The mutations of axin and β-catenin further underscore liver as a relevant tissue to address the functions of the Wnt/β-catenin pathway (13, 40) (reviewed in references 12 and 35). The APC tumor suppressor protein, which is an important negative factor in Wnt signaling, is also a documented inhibitor of G₁ progression (4). An output of β-catenin signaling is expression of c-MYC, which resides in a parallel and RB-independent pathway of G₁ control (27, 38). In other work, the transgenic expression of c-MYC has been used extensively as a model for HCC in mice (for examples see reference 13). Finally, recent work demonstrates that regulation of β-catenin levels is part of the early events of liver regeneration (32).

Future perspectives.

Recent data indicates that HBP1 is an efficient repressor of Wnt/β-catenin signaling. Thus, this new work expands the possibilities for negative regulation of proliferation by inhibiting an oncogenic pathway (37a). To add complexity, suppression of Wnt signaling occurs by a physical inhibition of the LEF and TCF factors and does not require DNA binding by HBP1. Together with published data on the N-MYC promoter (46), this defines HBP1 as a complex repressor with both sequence-independent and sequence dependent modes of regulation.

The significance of the present work is the demonstration that HBP1 can regulate G₁ in proliferation of a differentiated tissue, and this finding provides a foundation for future work. The liver regeneration model in this study represents one of few differentiated tissue models in which cell cycle issues can be addressed. The results of this work place HBP1 in G₁ regulation of an animal model of proliferation. Thus, a possibility is that HBP1 may block the RB, Wnt/β-catenin, and/or c-MYC pathways to regulate G₁ progression in the liver. Future work will address HBP1 in the RB and Wnt pathways in liver and other tissues.

A major question for tumor suppression is how normal proliferation controls are lost in the early stages of certain cancers, in which there is a reproliferation of differentiated cells. While liver regeneration is clearly not HCC, a shared similarity is the reproliferation of a differentiated tissue that is normally quiescent. The inhibition of proliferation suggests that HBP1 could have tumor suppressor function. Consistent with the notion of tumor suppression, the HBP1 gene (UniGene Hs.10882) resides in a region that is mutated in many cancers (human chromosome 7q31) (6, 14, 23, 28, 54–57). A loss-of-function mutation in HBP1 may have an increased tumor onset or an enhanced cell cycle response in the liver. The latter phenotype is akin to transgenic expression of the HBP1 mutant described in this work. Future work will be directed at the importance of HBP1 in tumor suppression and in signaling networks that govern reproliferation of differentiated tissues.