Nikolai A. Timchenko, Department of Pathology & Immunology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA. Tel.: 713 798 1567; fax: 713 798 4161; e-mail: firstname.lastname@example.org
CCAAT/Enhancer Binding Proteins family proteins are important regulators of liver functions. Here, we show the critical role of C/EBPα-mediated chromatin remodeling in the age-associated dysfunctions of the liver and in the maintenance of physiological homeostasis. Because ph-S193 isoform of C/EBPα is increased in livers of old mice, we have generated C/EBPα-S193D knockin mice, which mimic the ph-S193 isoform of C/EBPα. Analyses of these mice showed that the S193D mutation causes chromatin remodeling leading to histological appearance of ‘foci-like’ nodules, which are also observed in livers of old mice. These ‘foci-like’ structures contain K9 trimethylated histone H3, a marker of heterochromatin. The increase of heterochromatin regions in S193D mice correlates with the elevation of S193D-C/EBPα-HDAC1 complexes and with dys-regulation of gene expression including epigenetic silencing of cyclin D1 and D2 promoters and the inhibition of liver proliferation. The elimination of C/EBPα-HDAC1 complexes in S193D mice by inhibition of HDAC1 corrects chromatin structure and normalizes expression of cyclin D1 and D2. We found that epigenetic dys-regulation is also associated with the elevation of C/EBPβ and with the increase of C/EBPα/β heterodimers in S193D mice. The C/EBPα/β heterodimers activate transcription of Glut4 and increase the levels of Glut4. As the result, S193D livers have accelerated uptake of glucose and accumulation of glycogen in the liver. Thus, this study demonstrates that the phosphorylation of C/EBPα at S193 leads to the appearance of heterochromatin regions, which correlates with the development of age-related dysfunctions of the liver.
Senescent liver is characterized by many alterations in biological processes leading to the decline of liver functions and to the alterations in body homeostasis (Schmucker, 2005; Timchenko, 2009). The most dramatic alteration in liver functions is the impaired liver regeneration after partial hepatectomy (PH) and after surgical resections (Bucher et al., 1964; Fry et al., 1984; Iakova et al., 2003; Conboy et al., 2005; Gielchinsky et al., 2010). The decline of regenerative capacity of old livers is mediated by the appearance of C/EBPα-Brm-HDAC1 complex (Iakova et al., 2003; Conboy et al., 2005; Wang et al., 2006) and following epigenetic silencing of E2F-dependent promoters (Wang et al., 2007, 2008a; Timchenko, 2009). Recent reports have shown that the reduction of GSK3β in aged livers causes the age-associated elevation of C/EBPα-HDAC1 complexes (Jin et al., 2009a,b,c). In agreement with these observations, Seo et al. (2008) have found that the inhibition of GSK3β leads to development of senescence phenotype in human liver-derived Chang cells. The aging liver is also characterized by changes in the size of hepatocytes. Chipchase et al. (2003)have found that livers of 2- year-old mice have increased number of hepatocytes containing enlarged nuclei with increased ploidy, while livers of 3- month--old mice do not contain enlarged hepatocytes. The premature liver polyploidy has been also described in DNA repair (Ercc1) deficient mice (Chipchase et al., 2003). One of the characteristics of the old livers is a lipid accumulation that leads to the development of steatosis (Kuk et al., 2009). It has been also shown that the activities of alanine transaminase (ALT) and aspartate transaminase (AST) and concentration of triglycerides are increased in the blood of old animals as the result of liver dysfunctions (Schmucker, 2005). Molecular basis for these age-associated dysfunctions of the liver has been not elucidated.
The liver is able to regenerate itself after PH (Fausto et al., 2006). The regeneration of liver after PH is a very complex process that includes reorganizations of the networks of several signal transduction pathways (Michalopoulos, 2007). Despite the global re-organization of these pathways, the remaining portion of liver proliferates and it is able to completely support its differentiation functions and body homeostasis. The molecular basis for these enormous capabilities of the small remaining portion of the liver is not known. Two members of C/EBP family, C/EBPα and C/EBPβ, are expressed in the liver and play critical role in the regulation of liver biology (Johnson, 2005; Timchenko, 2009). Transcription factor C/EBPα is an important component of network because C/EBPα knockout mice die shortly after birth because of impaired energy homeostasis (Wang et al., 1995). C/EBPα displays many of its functions through direct interactions with proteins. C/EBPα interacts with the SWI/SNF chromatin remodeling complex and activates expression of genes involved in adipogenesis (Pedersen et al., 2001). The interaction of C/EBPα with Brm is involved in the inhibition of cell proliferation (Iakova et al., 2003; Muller et al., 2004; Conboy et al., 2005). In aging liver, C/EBPα is hyperphosphorylated at S193 by cyclin D3-cdk4 and is associated with histone deacetylase 1, HDAC1, and with heterochromatin protein 1α, HP1α (Wang et al., 2008a,b). In agreement with these observations, Demarco et al. (2006) have identified a direct interaction of C/EBPα with HP1α in the regions of centromeric heterochromatin.
We have recently generated S193D-C/EBPα knockin mice and have shown that liver proliferation after PH is completely inhibited in these mice (Wang et al., 2010). Here, we examined the role of phosphorylation of C/EBPα at S193 in the age-associated decline of liver functions. We found that the age-related mutation of C/EBPα leads to the significant increase of heterochromatin regions and that these alterations are similar to those observed in old mice that have abundant ph-S193 isoform of C/EBPα. The elevation of heterochromatin regions correlates with the development of age-like dys-functions such as inhibition of liver proliferation, accumulation of glycogen in the liver and the elevation of ALT/AST and triglycerides in the blood.
The C/EBPα-S193D increases size of hepatocytes and alters chromatin structure
In these studies, we have examined liver functions in C/EBPα-S193D knockin mice, which have been recently generated (Wang et al., 2010). The Fig. 1A shows an example of genotyping of mice used for these studies. We have mutated serine to aspartate by TC-GA substitution which also created new BamHI site and destroyed MluI site (Fig. 1A). In the course of examination of S193D mice, we found that S193D isoform significantly changes liver morphology. The H&E staining showed that the size of hepatocytes and size of nuclei are significantly increased in livers of C/EBPα-S193D mice compared to WT mice (Fig. 1B). The size of nuclei in S193D mice is around twofold bigger than that in WT mice (see bottom part of the Fig. 1B). Because the liver weight/body weight ratio is not changed in S193D mice, the increase in the size of hepatocytes is consistent with a reduction in the number of hepatocytes in the liver down to 50% (Fig. 1C). Because further studies showed that liver proliferation is inhibited in S193D mice during postnatal liver development (Fig. 6), these observations suggested that normal size of the S193D livers is supported by increased cell growth leading to the increased size of hepatocytes. To determine if the alterations of size of nuclei in S193D mice change chromatin structure, we have performed 4'-6-diamidino-2-phenylindole staining of livers of 2- month-old mice. Consistent with H&E staining, the DAPI staining showed that hepatocytes in C/EBPα-S193D livers have a larger size of nucleus. In addition to that, we found that hepatocytes in S193D livers have a significant increase of ‘foci-like’ nodules (Fig. 1D). Because livers of old mice are characterized by increased size of nuclei (Chipchase et al., 2003), we have next performed H&E and DAPI staining of livers of WT young (2 month) and old (24 month) mice (Fig. 1E–G). We found that up to 35–40% of hepatocytes are enlarged in livers of old mice (Fig. 1F). Examination of chromatin structure by DAPI staining revealed the abundance of ‘foci-like’ nodules in nuclei of livers from old mice, especially in cells with enlarged nuclei (Fig. 1G and see additional images in Fig. S1).
‘Foci-like’ structures in young S193D mice and in old wild-type mice are heterochromatin regions
We have next examined if the ‘foci-like’ structures in S193D and old WT hepatocytes might be heterochromatic regions. Livers of young WT, young S193D and old WT mice were stained with antibodies to the marker of heterochromatin, histone H3 trimethylated at K9 (H3K9Me). Figure 2 shows typical pictures of the staining. Nuclei of hepatocytes of young WT mice show mainly diffused staining of H3K9Me with a rare punctuated location of H3K9Me. On the contrary, nuclei in hepatocytes of young S193D mice and nuclei in large hepatocytes of old WT mice contain a much bigger number of H3K9Me located in ‘foci-like’ structures. In S193D livers, around 80–90% of hepatocytes contain H3K9Me in ‘foci-like’ structures. It is interesting to note that, in livers of old mice, ‘foci-like’ structures determined by DAPI staining and H3K9me are mainly co-localized in enlarged hepatocytes. Thus, these studies show that the S193D mutation of C/EBPα leads to the elevation of heterochromatin regions. Similar alterations are observed in large hepatocytes of aged livers.
Livers of S193D mice contain abundant C/EBPα-HDAC1 complexes and elevated levels of C/EBPβ
We have next examined pathways by which C/EBPα-S193D mutant causes alteration of chromatin structure. Our previous studies have shown that the S193-ph isoform of C/EBPα form complexes with a chromatin remodeling protein histone deacetylase 1, HDAC1, in livers of old mice (Wang et al., 2008a). Therefore, we determined if the C/EBPα-S193D might affect chromatin structure via interactions with HDAC1. Staining of livers of 2- month-old mice with Abs to HDAC1 and C/EBPα showed that livers of WT mice have a diffuse nuclear staining for both HDAC1 and C/EBPα, while HDAC1 and C/EBPα are observed in punctuate ‘foci-like’ structures of the C/EBPα-S193D livers (Fig. 3A). We have next examined interactions of C/EBPα with HDAC1 and with heterochromatin protein α, HP1α, using Co-IP approach and HPLC-based size-exclusion chromatography. HP1α has been included in the studies because this protein cooperates with HDAC1 in the repression of promoters and because this protein interacts with C/EBPα (Demarco et al., 2006; Wang et al., 2008a). Co-IP approach showed that larger amounts of HDAC1 and HP1α are observed in the complexes with the mutant C/EBPα, while C/EBPα-HDAC1 complex is detected in WT mice only after longer exposure (Fig. 3B). Previous studies have shown that the livers of old mice contain the C/EBPα-Brm complex (Iakova et al., 2003; Conboy et al., 2005) and that this complex is up-regulated by phosphorylation of C/EBPα at S193 (Wang et al., 2007). Therefore, we have examined if Brm might be in the complexes with C/EBPα in young S193D mice. Under sensitivity of our assay, we could not detect Brm in C/EBPα IPs (Fig. 3B). The lack of Brm in C/EBPα complexes is perhaps associated with the very low levels of Brm in livers of 2 -month-old mice. To confirm results of Co-IP and to determine the size of C/EBPα-HDAC1 complex, we have utilized HPLC-based techniques. These studies showed that the complex is significantly larger (around 680 kD) than the sum of masses of C/EBPα, HDAC1 and HP1α, suggesting that it might contain additional components (Fig. 3C).
Given the elevation of heterochromatin regions in hepatocytes of S193D mice, we suggested that S193D mice might have changes in the transcription of many genes. Therefore, we have performed micro array studies with livers of 2- month-old WT and S193D mice. These studies identified a number of genes with reduced or increased expression on livers of S193D mice (data not shown). In this paper, we have investigated in detail a portion of these genes including C/EBP family proteins, genes which are involved in control of liver growth and in the control of glucose metabolism. The alteration in the expression of these genes was confirmed by Q-RT–PCR and by Western blotting. These data are presented throughout the manuscript (see Figs 3, 5 and 8). C/EBP family is one of the groups of genes with altered mRNA levels in livers of S193D mice. Figure 3D shows differences in the levels of C/EBP family mRNAs, which were further confirmed by Q-RT–PCR. We found that the levels of C/EBPβ and C/EBPε mRNAs are increased in the livers of S193D mice, while levels of C/EBPα and C/EBPγ are slightly reduced. To examine if protein levels of C/EBP proteins are changed in S193D mice, we have performed Western blotting analyses. These studies have shown that, although C/EBPε and C/EBPγ mRNAs are detected by RT–PCR, these proteins are not detectable in livers of both WT and S193D mice by Western blotting (data not shown). Examination of C/EBPα showed that it is expressed as three isoform with MW 42, 40 and 30 kD. In agreement with a slight reduction of C/EBPα mRNA, the protein levels of these three isoforms of C/EBPα are slightly reduced in livers S193D mice (Fig. 3E). However, this reduction is minor and represents <20% (Fig. 3E, middle bar graphs). Quite different result was observed for C/EBPβ expression. C/EBPβ is also expressed in the liver as three isoforms: full-lengths protein (FL), C/EBPβ-LAP and C/EBPβ-LIP. We have found that protein levels of C/EBPβ-LAP are elevated in livers of S193D mice, while levels of a truncated isoform C/EBPβ-LIP are not changed. Calculation of C/EBPβ-LAP as ratios to β-actin showed around 2.5-fold increase in livers of S193D mice (Fig. 3E, bottom). Taken together, these studies showed that the alterations of chromatin structure in S193D mice are associated with elevation of C/EBPα-HDAC1 complexes and with elevation of C/EBPβ.
The inhibition of HDAC1 in S193D mice corrects size and chromatin structure of nuclei in hepatocytes
Given the abundance of C/EBPα-HDAC1 complexes in livers of S193D mice, we suggested that these complexes might be involved in the alterations of chromatin structure. To test this hypothesis, we have inhibited the expression of HDAC1 by siRNA in S193D mice and examined chromatin structure of the nuclei. We have previously established conditions for the efficient delivery of siRNA HDAC1 and for a strong inhibition of HDAC1 in the liver (Wang et al., 2008b). The animals were injected with siRNA to HDAC1 for 3 days, and livers were harvested and analyzed. Using these conditions, we have inhibited the expression of HDAC1 to <15% in livers of C/EBPα-S193D mice (Fig. 4A). Examination of C/EBPα-HDAC1 complexes by Co-IP approach showed that siRNA-mediated inhibition of HDAC1 eliminates the C/EBPα-HDAC1 complexes in livers of S193D mice (Fig. 4B). H&E staining and DAPI staining showed that the elimination of C/EBPα-HDAC1 complexes correlates with the correction of size and chromatin structure of the nuclei in hepatocytes (Fig. 4C). Staining of the livers with antibodies to C/EBPα and to HDAC1 showed that the reduction of HDAC1 by siRNA redistributes C/EBPα from foci-like structures to the diffuse location within nucleus. In addition to the correction of chromatin structure, the inhibition of HDAC1 reduces the size of hepatocytes. We suggest that the enlarged size of hepatocytes in S193D mice might be the result of epigenetic alterations and that the elimination of foci-like structures also corrects the size of existing hepatocytes.
Previous studies have shown that C/EBPα-HDAC1 complexes are also abundant in livers of old mice (Wang et al., 2008a) suggesting that these complexes might be involved in the age-associated alterations of chromatin structure in old animals. To determine the role of C/EBPα-HDAC1 complexes in these age-associated alterations, we have initially tried to reduce the complexes by inhibiting HDAC1; however, the siRNA-mediated inhibition of HDAC1 in old mice caused a massive cell death shortly after injection of siRNA. Therefore, we have used an alternative approach to eliminate the complex, which is the treatment of mice with growth hormone (GH). It has been shown that GH activates pathways that de-phosphorylate C/EBPα at S193 and reduce amounts of C/EBPα-HDAC1-Brm complex (Wang et al., 2008a). In agreement with previous reports, the treatments of old mice with GH eliminated the C/EBPα-HDAC1-Brm complex (Fig. 4E). H&E staining and examination of nuclei by DAPI staining revealed that the elimination of C/EBPα-HDAC1-Brm complex also corrects the size and chromatin structure of the nuclei. The Fig. 4F shows typical pictures of H&E and DAPI staining. We found that the treatment of old mice with GH reduced the amounts of enlarged hepatocytes down to 10–15% (Fig. 4G).
Previous studies showed that GH reduces the C/EBPα-HDAC1 complex by activation of pathways that de-phosphorylate C/EBPα at S193. Because S193D mutation prevents elimination of the negative charge on S193, we examined if S193D mutation will prevent GH-mediated correction of the chromatin structure in S193D mice. We found that the treatment of S193D mice with GH is not sufficient to alter chromatin structure and size of hepatocytes in S193D mice (Fig. 4H). This result suggested that the S193D mutation took C/EBPα away from the GH-dependent control. Taken together, the studies of S193D mice and examinations of old mice showed that the elevation of C/EBPα-HDAC1 complex correlates with alterations in chromatin structure and that elimination of these complexes in S193D and in old mice contributes to the correction of chromatin structure.
C/EBPα-HDAC1 complexes reduce levels of cyclin D2 in C/EBPα-S193D livers via repression of the cyclin D2 promoter
We have next examined biological consequences of the change of chromatin structure in S193D mice. The micro array analysis has determined a group of mRNAs that have reduced levels in livers of S193D mice. This group represents genes that are involved in the regulation of liver growth including cyclins D1 and D2 (Fig. 5A). Because D-type cyclins play a critical role in the liver proliferation, we focused our further studies on the promoters of cyclin D1 and D2. Western blotting confirmed that the expression of cyclin D2 is reduced in C/EBPα-S193D livers (Fig. 5B), while cyclin D1 protein was not detected in quiescent livers. Therefore, we first examined if C/EBPα-HDAC1 complexes down-regulate cyclin D2 promoter in quiescent livers. Examination of the cyclin D2 promoter identified four consensuses for C/EBPα (Fig. 5C). To determine if C/EBPα binds to these sites, we have examined the interaction of C/EBPα with DNA probes covering C/EBPα sites within the cyclin D2 promoter using electrophoretic mobility shift assay (EMSA). These studies showed that C/EBPα binds to the cyclin D2 promoter in vitro (Fig. 5D).
To test if C/EBPα-HDAC1 complexes down-regulate the cyclin D2 in S193D mice, we have examined the expression of cyclin D2 in livers of S193D mice treated with siRNA to HDAC1. Western blotting has shown that the levels of cyclin D2 are normalized in S193D mice by siRNA to HDAC1 and that this normalization correlates with the elimination of C/EBPα-HDAC1 complexes (Fig. 5E). To determine the causal role C/EBPα-HDAC1 complexes in the repression of cyclin D2 promoter in the livers of S193D mice, we have performed chromatin immunoprecipitation (ChIP) analyses. Chromatin immunoprecipitation studies showed that C/EBPα-HDAC1 complex is abundant on the cyclin D2 promoter in the C/EBPα-S193D livers, while only C/EBPα is observed on this promoter in WT mice (Fig. 5F). The elimination of HDAC1-C/EBPα complexes leads to the removal of HDAC1 from the cyclin D2 promoter (Fig. 5F). To determine if the cyclin D2 promoter is repressed in S193D livers, we examined the modifications of histone H3 associated with the cyclin D2 promoter. K9 acetylated histone H3 occupies the cyclin D2 promoter in WT mice, while tri-methylated H3K9 is not detected. In contrast, the Ac-K9-histone H3 is dramatically reduced on the D2 promoter in the livers of C/EBPα-S193D mice, but amounts of trimethylated K9H3 are increased (Fig. 5F). The inhibition of C/EBPα-HDAC1 complexes reversed the patterns of modifications of histone H3 on the cyclin D2 promoter. These patterns of histone H3 modifications show that C/EBPα-HDAC1 complex inhibits the cyclin D2 promoter.
C/EBPα-HDAC1 complexes repress cyclin D1 promoter in the liver during postnatal development
Because the expression of cyclin D1 protein is not detectable in quiescent livers of 2- month-old mice, we have examined the role of C/EBPα-HDAC1 complexes in the regulation of cyclin D1 using mice during early postnatal development when liver proliferates. We have first asked if the C/EBPα-S193D inhibits liver proliferation after birth. Proliferation of the livers during postnatal development was examined by measuring BrdU uptake in 1-, 3-, 7-, 15- and 60- day-old mice. Figure 6A shows that the rate of liver proliferation in WT mice is relatively high at day 1 and then declines with age. We have found that the phosphomimetic C/EBPα-S193D reduces liver proliferation at all stages of postnatal liver development. We have next examined mechanisms by which C/EBPα-S193D mutant and ph-S193 isoform inhibit liver proliferation during postnatal liver development. Figure 6B shows that total protein levels of both WT and S193D-C/EBPα are increased with age, and the phosphorylation of WT C/EBPα at S193 is detectable at days 7, 15 and 60 in WT mice. Because antibodies to S193-ph isoform of C/EBPα do not recognize S193D mutant (Wang et al., 2006, 2007), no specific signals were observed with protein extracts from the livers of C/EBPα-S193D mice. We have next examined the expression of cyclin D1 during postnatal development. In addition to cyclin D1, the expression of FoxM1 was examined because this protein is expressed only in proliferating livers and serves as an additional marker of liver proliferation. Figure 6C shows that the levels of cyclin D1 and FoxM1B are significantly lower in C/EBPα-S193D mutant livers during whole course of the studies. Calculations of cyclin D1 levels as ratios to β-actin revealed twofold to threefold higher levels of cyclin D1 protein at days 1–7 in WT mice compared to the levels in S193D mice. The expression of cyclin D1 is inhibited and is not detectable at days 15 and 60 in both WT and S193D mice.
We next asked if the inhibition of cyclin D1 in S193D livers is mediated by C/EBPα-HDAC1 complexes. Examination of the cyclin D1 promoter identified a consensus for C/EBPα to which C/EBPα binds (Fig. S2). To determine if the C/EBPα-HDAC1 complexes inhibit the cyclin D1 promoter through interactions with the identified site, we have performed ChIP assay with primers covering this site. For the ChIP studies, 7- day-old animals were used because the differences in the expression of cyclin D1 between WT and S193D mice are significant at that time point. Chromatin immunoprecipitation studies showed that the cyclin D1 promoter is occupied by C/EBPα in WT mice, while HDAC1 is not detectable on the promoter and histone H3 is acetylated at K9 (Fig. 6D). This pattern of histone H3 modifications shows that the promoter is active. On the contrary, the cyclin D1 promoter is occupied by C/EBPα-HDAC1 complex in S193D mice leading to partial de-acetylation of histone H3 and following trimethylation at K9. These alterations in the histone H3 modifications revealed that the cyclin D1 promoter is partially repressed in the livers of S193D mice by C/EBPα-HDAC1 complex.
C/EBPα has several pathways of inhibition of cell proliferation. Previous studies showed that C/EBPα arrests liver proliferation in young mice through binding to and inhibition of cdk2 (Wang et al., 2001, 2004; Tan et al., 2005); therefore, we have examined if this pathway might be also involved in the inhibition of liver proliferation during postnatal liver development. Figure 6E shows that cdk2 is associated with the mutant S193D-C/EBPα at all tested stages of postnatal liver development and that this association is proportional to amounts of C/EBPα. On the contrary, the association of cdk2 with C/EBPα in WT mice is observed only in livers of 15- and 60- day-old mice. Examination of cdk2 IPs in in vitro kinase assay with histone H1 substrate shows that, in WT livers, cdk2 activity is high at days 1–7 and is reduced at later stages of liver development. However, cdk2 activity is inhibited in C/EBPα-S193D livers at all stages of postnatal development (Fig. 6F). Taken together, these data demonstrate that the mutant S193D-C/EBPα and ph-S193 isoform of C/EBPα arrest liver proliferation during postnatal development through repression of the cyclin D1 promoter and via inhibition of cdk2.
Livers of C/EBPα-S193D mice have increased uptake of the glucose and accumulation of glycogen
Examination of blood parameters showed that the levels of glucose are reduced in S193D mice (Wang et al., 2010 and see Fig. 7A). Therefore, we have investigated pathways by which S193D mutation affects glucose levels. The glucose tolerance test revealed that the C/EBPα-S193D animals remove glucose from the blood much faster than WT mice (Fig. 7B). To determine if the liver is involved in the increased removing glucose from the blood, we have injected radioactive glucose and have examined the uptake of H3-glucose by livers in WT and in S193D mice. These studies revealed that the livers of S193D mice utilize glucose much more efficiently than livers of WT mice (Fig. 7C). These data suggest that the reduction of glucose in the blood of C/EBPα-S193D mice is because of an increased uptake of glucose by the liver. Because one of the pathways of utilization of the glucose in the liver is the conversion of glucose into glycogen, we have isolated glycogen from WT and S193D livers after injection with H3-glucose. Measurements of total and H3-glycogen showed that the utilized H3-glucose is converted into glycogen (Fig. 7D). In agreement with these data, staining of WT and S193D livers for glycogen detected a significant increase of glycogen in livers of C/EBPα-S193D mice (Fig. 7E).
The C/EBPα-S193D increases expression of genes involved in utilization of glucose
We have next examined molecular mechanisms by which the S193D mutant increases liver-mediated uptake of glucose in C/EBPα-S193D mice. Examination of expression of genes involved in glucose metabolism by Q-RT–PCR showed that levels of mRNAs coding for the glucose transporters Glut2 and Glut4, GcK and GyS2 are increased in livers of C/EBPα-S193D mice (Fig. 8A,B). Because Glut4 is also expressed in adipose tissues, we have examined the levels of Glut4 mRNA in adipocytes and found around 15-fold elevation of Glut4 mRNA in adipose tissue of S193D mice. Because levels of Glut4 mRNA are dramatically activated in both liver and adipocytes, we determined mechanisms of regulation of Glut4 by C/EBPα-S193D. We have focused the studies on the livers of S193D mice, because the S193D liver has increased the uptake of glucose. Western blotting confirmed that livers of C/EBPα-S193D mice contain higher levels of Glut4 protein (Fig. 8C). We have next asked if there is a difference in the ability of WT C/EBPα and the mutant S193D-C/EBPα to activate the Glut4 promoter. The Glut4 promoter was cloned into pGL3 vector and co-transfected with WT C/EBPα and with S193D-C/EBPα into Hep3B2 cells. In addition to these constructs, C/EBPα-R290A mutant was used because it does not bind to DNA (Miller et al., 2003; Wang et al., 2004). Co-transfection studies showed that the S193D mutation significantly increases the ability of C/EBPα to activate the promoter (Fig. 8D). Our further studies in tissue culture systems showed that the cyclin D3-cdk4-mediated phosphorylation of endogenous C/EBPα at S193 enhances its ability to activate the Glut4 promoter (Fig. S3). Taking together these data, we conclude that the S193D mutation and phosphorylation of C/EBPα at S193 enhance the ability of C/EBPα to activate the Glut4 promoter.
C/EBPβ forms heterodimers with C/EBPα-S193D and co-operates with C/EBPα-S193D in the activation of Glut4 promoter
Because protein levels of C/EBPβ are increased in livers of S193D mice (Fig. 3), we have examined if this elevation might be involved in the activation of Glut4 promoter. We first performed EMSA with the probe covering C/EBP site within the Glut4 promoter. Specific antibodies were incorporated to distinguish C/EBPα and C/EBPβ isoforms. Consistent with elevation of protein levels of C/EBPβ, significantly higher amounts of C/EBPβ are bound to the Glut4 promoter in livers of S193D mice (Fig. 8E). This is especially clear after supershift of the C/EBPα (Fig. 8E, lane 8). Most important, the incorporation of antibodies to C/EBP proteins showed that the majority of S193D-C/EBPα binds to the Glut4 promoter as a heterodimer with C/EBPβ and that the binding of C/EBPα homodimers is not detectable after supershift with antibodies to C/EBPβ (Fig. 8E, lane 9). It is important to note that antibodies to C/EBPβ specifically interact with C/EBPβ and do not cross-react and do not supershift C/EBPα in livers of WT mice (Fig. 8E, lane 4). To determine if the elevation of C/EBPβ and C/EBPα/β heterodimers up-regulate Glut4 in livers of S193D mice, we have examined interactions of C/EBP proteins with the Glut4 promoter and modifications of histone H3 on the Glut4 promoter using ChIP analysis. HDAC1 protein was also included in these studies because it forms complexes with C/EBP proteins in the liver (Wang et al., 2008a,b). These studies showed that the Glut4 promoter is occupied and partially repressed in WT mice by C/EBPβ-HDAC1 and C/EBPα-HDAC1 complexes because histone H3 is trimethylated at K9 on the Glut4 promoter (Fig. 8F). On the contrary, HDAC1 is not detectable on the Glut4 promoter in S193D livers. Our data show that the Glut4 promoter is activated by S193D-C/EBPα and C/EBPβ because histone H3 is acetylated at K9 on this promoter.
C/EBPβ is required for the increased ability of C/EBPα-S193D to activate the Glut4 promoter
Electrophoretic mobility shift assay and ChIP assays suggested that C/EBPα-S193D activates the Glut4 promoter via the formation of heterodimers with C/EBPβ. To directly examine the role of C/EBPα:C/EBPβ heterodimers, we have performed experiments in cultured HEK293 cells. These cells were chosen for these studies because they do not express endogenous C/EBPα; therefore, effects of the transfected S193D mutant will not be influenced by endogenous C/EBPα. The expression of C/EBPβ was inhibited in these cells by siRNA, and the WT and S193D mutant C/EBPα were co-transfected with the Glut4 promoter. As the control, the activity of Glut4 promoter was examined in cells transfected with C/EBPβ siRNA alone. Figure 8G shows that the inhibition of C/EBPβ in cells transfected with WT C/EBPα does not change significantly the activation of Glut4 promoter. However, the inhibition of C/EBPβ in cells transfected with S193D mutant reduces the activity of Glut4 promoter to the levels observed for activation by WT C/EBPα. These studies clearly demonstrated that the increased ability of C/EBPα-S193D to activate the Glut4 promoter is mediated via interactions with C/EBPβ, perhaps through the formation of heterodimers with C/EBPβ. Thus, these studies showed that the formation of C/EBPβ:C/EBPα-S193D heterodimers is involved in the activation of Glut4 promoter and following increase of Glut4 protein in S193D mice.
Age-related dysfunctions of the liver are mediated by the ph-S193 isoform of C/EBPα
Senescent liver is characterized by the decline of several functions and alterations in liver morphology (Schmucker, 2005; Timchenko, 2009). In this paper, we have examined the role of age-specific S193-ph isoform of C/EBPα in the age-related alterations using recently generated C/EBPα-S193D mice. In our previous paper, we have shown that C/EBPα-S193D mutant inhibits liver proliferation after PH (Wang et al., 2010). The impaired liver proliferation in S193D mice is similar to that observed in livers of old WT mice (Bucher et al., 1964; Timchenko, 2009). Examination of liver morphology and liver functions in young C/EBPα-S193D mice showed that hepatocytes of S193D mice have enlarged size and have abundant heterochromatin regions, which are similar to those observed in livers of old WT mice. These data are in agreement with previous reports showing that mammalian aging is associated with remodeling of chromatin structure (Narita et al., 2003; Bandyopadhyay et al., 2007; Sedivy et al., 2008). Our data suggest that alterations of chromatin structure might be the cause of the increase of size of hepatocytes of old mice; however, it is also possible that the increased ploidy in old hepatocytes (Chipchase et al., 2003) explains the appearance of DAPI-bright heterochromatic foci, because the increased number of foci could reflect the heterochromatic centromeres of additional chromosomes.
We have performed a detailed investigation of pathways by which S193D-C/EBPα and ph-S193 isoform of C/EBPα might change chromatin structure. Our data show that alterations in chromatin structure of hepatocytes correlate with the elevation of C/EBPα-HDAC1-HP1α complexes. These complexes are abundant in livers of old mice, but contain some additional proteins such as Brm and E2F4 (Iakova et al., 2003; Conboy et al., 2005; Wang et al., 2008a). The elevation of C/EBPα-HDAC1 complex seems to be responsible for the alterations of chromatin structure because the inhibition of complex by siRNA to HDAC1 or by treatment of old mice with GH corrects chromatin structure in livers of S193D mice and in livers of old mice, respectively. In this paper, we have shown that the alterations of chromatin structure are involved in the inhibition of liver proliferation and in dys-regulation of gluconeogenesis. It is interesting to note that our previous paper has identified an additional age-associated and S193-ph-dependent alteration which is a high frequency of development of liver cancer in old mice. We have shown that the development of liver cancer in old mice is associated with the specific elimination of ph-S193 isoform of C/EBPα by gankyrin-mediated degradation through ubiquitin proteasome system (Wang et al., 2010). Taken together, the investigations of S193D mice show that the expression of age-specific C/EBPα-S193D isoform in the liver of young mice leads to alterations that mimic liver dysfunctions observed in old animals (Schmucker, 2005; Timchenko, 2009). Alterations in S193D mice include enlarged hepatocytes, inhibition of liver proliferation, steatosis, accumulation of AST/ALT and TG in the blood and accumulation of glycogen in the liver (Fig. 9).
It is interesting to note that Seo et al. (2008) have shown that the accumulation of glycogen itself led to senescence. Therefore, the accumulation of glycogen in livers of S193D mice might contribute to the age-associated changes in S193D mice. Our findings also suggest that the correction of chromatin structure might be sufficient to improve liver functions in elderly. One of the possible tools for the correction of chromatin structure is a treatment of mice with growth hormone. Consistent with this hypothesis, Krupczak-Hollis et al. (2003) have shown that the treatment of old mice with GH corrects liver regeneration via stimulation of proliferation of hepatocytes. A recent paper has shown another pathway that might correct liver regeneration in aged mice. Gielchinsky et al. (2010) found that activation of Akt/mTORC1 pathway in aged livers restores liver regeneration by activation of the process of cell growth.
Examination of the S193D mice has provided an answer to the critical question of liver biology. Liver supports body homeostasis and, during liver regeneration, the small portion of remaining liver completely maintains its functions of the differentiated tissue. The majority of studies of liver regeneration were performed with 70% model of PH; however, there are many reports of successful liver regeneration after 90% PH (Moser et al., 2001; Corbin et al., 2002; Benko et al., 2010; Eipel et al., 2010). How does the remaining, very small portion of the liver control the multiple functions? Data in this paper and our published observations (Wang et al., 2010) present an example of pathways which liver uses to control proliferation and differentiation functions at the same time. Our data suggest that the enormous capabilities of liver to control multiple functions are associated with a proper regulation of C/EBPα activities by timely phosphorylation/de-phosphorylation at a single amino acid residue and that disruption of this regulation causes pathological alterations in the liver and in blood.
Multiple pathways of C/EBPα-mediated arrest of cell proliferation
C/EBPα inhibits cell proliferation mainly through the interactions with cell cycle proteins (Wang et al., 2001; Tan et al., 2005; Timchenko, 2009). The analysis of S193D mice allowed us to identify additional pathways by which C/EBPα inhibits liver proliferation and additional pathways that regulate the biological activities of C/EBPα. The additional pathway of growth arrest involves silencing the cyclin D1 and D2 promoters by C/EBPα-HDAC1 complexes through direct interactions with these promoters. Although this pathway requires protein–protein interactions with HDAC1, it differs from previously described pathways and also requires DNA binding activity of C/EBPα. This is a first observation showing that direct interaction of C/EBPα with the promoters of D-type cyclin genes is involved in the arrest of proliferation. Examination of the S193D mice has also identified additional levels of complexity of the regulation of C/EBPα activities. Data in our paper show that S193D-C/EBPα activates the Glut4 promoter, but it inhibits cyclin D2 and cyclin D1 promoters. How are these opposite activities of C/EBPα regulated? One possible pathway is supported by the findings that S193D mice have the increased amounts of S193D-C/EBPα/β heterodimers. Our data show that the C/EBPα/β heterodimers bind to the Glut4 promoter and activate the promoter stronger than homodimers of each of these proteins. We have also shown that the interactions of C/EBPα with chromatin remodeling proteins might be one of the mechanisms which control these opposite transcriptional functions of C/EBPα. We have found that C/EBPα is observed on the cyclin D1 and cyclin D2 promoters in the complexes with HDAC1, while C/EBPα binds to the Glut4 promoter independently on HDAC1, mainly in the complexes with C/EBPβ. Thus, our data show that the transcriptional activities of C/EBPα are also controlled, at least in part, by phosphorylation at S193.
Antibodies and reagents
Antibodies to cyclin D3 (C-16), C/EBPα (14AA and N19), Glut4, C/EBPβ (C19), cyclin D1, cyclin D2 HDAC1, HP1α, cdk4 and cdc2 are from Santa Cruz Biotechnology (Richmond, CA, USA). Monoclonal anti-β-actin antibodies were from Sigma-Aldrich (Saint Louis, MO, USA).
C/EBPα-S193D knockin mice
The generation of C/EBPα-S193D knockin mice was described in our previous paper (Wang et al., 2010). Briefly, the C/EBPα-S193D mice were generated by replacement of the endogenous C/EBPα gene with knockin construct which contained a substitution of TC to GA in the position of Ser193 (See Fig. 1A). This substitution leads to the mutation of Ser to Asp and to alterations in restriction sites for MluI and BamH1. Therefore, the genotyping of these mice included the examination of 858- bp PCR product by restriction enzymes MluI and BamHI. Our previous paper contains more details for the generation and genotyping of these mice (Wang et al., 2010). For all experiments presented in this paper, we have bred heterozygous mice and have examined WT, heterozygous and homozygous mice from the same littermates. Animal experiments were approved by the Institutional Animal Care and Use Committee at Baylor College of Medicine (protocol AN-1439).
Experiments with GH and with siRNA to HDAC1
For GH treatment of mice, recombinant mouse growth hormone (rmGH) provided by the National Hormone and Peptide Program was injected into mice (2 mg kg−1) subcutaneously for 3 days. The physiological saline (PS, 0.9% NaCl) was used as vehicle for rmGH. The control animals were injected with PS. Liver tissues were collected and used for the staining with H&E and DAPI and for the examination of protein expression as described in the following paragraphs. The inhibition of HDAC1 by siRNA was performed as described in our previous paper (Wang et al., 2008b). The control animals were injected with a control siRNA containing random composition of nucleotides (Wang et al., 2008b).
Blood test, H&E and DAPI staining
Examination of the blood parameters, H&E staining and DAPI staining were performed in Baylor College of Medicine Facilities.
Immunostaining of livers with antibodies to C/EBPα, HDAC1 and K9 trimethylated histone H3
The immunostaining of the WT, S193D and WT old livers was performed as described in our previous papers (Wang et al., 2008a,b). The antibodies to C/EBPα (14AA) and HDAC1 were from Santa Cruz Biotechnology. Antibodies to H3K9-trimethyl were from Abcam (Cambridge, MA, USA).
Examination of the Glut4 promoter
The proximal region of the mouse Glut4 promoter (containing C/EBPα site) was cloned into pGl3 vector. The reporter construct was co-transfected with WT C/EBPα and with C/EBPα mutants (S193D and R290A) into HEK293 cells. The activity of the promoter was calculated as a ratio to Renilla luc control.
Nuclear extracts were isolated from livers as described earlier (Timchenko et al., 1997) and fractionated by size-exclusion column SEC400 (HPLC, BioLogic HR; Bio-Rad Laboratories Inc, Hercules, CA, USA). The detailed procedure for the analysis of C/EBPα complexes is described in our previous papers (Wang et al., 2001, 2004). Briefly, gel filtration fractions were loaded on denaturing gradient (4–20%) PAAG, blotted onto membrane and probed with antibodies to C/EBPα (14AA), HDAC1, HP1α and Brm (Santa Cruz Biotechnology). To detect C/EBPα complexes, C/EBPα was immunoprecipitated from each fraction, and IPs were probed with antibodies to HDAC1.
Protein isolation and Western blotting
Nuclear extracts were isolated from cultured cells and from livers as described in previous papers (Wang et al., 2001, 2004). Briefly, tissues were homogenized with buffer A (20 mm Tris–HCL pH 7.5, 30 mm KCl, 10% glycerol and inhibitors of phosphatases) and spun down at 12 000 rpm for 10 min at 4°C. Supernatant (cytoplasm) was frozen, and the pellet was treated with high salt buffer (20 mm Tris–HCl pH 7.5, 0.42 m NaCl, 25% sucrose, 5 mm DTT and inhibitors of phosphatases). After centrifugation, the supernatant (NE) was used for the Western blotting and for Co-IP or frozen in −80°C freezer. Nuclear extracts were isolated from mouse livers of WT and S193D mice and examined by Western blotting as described below.
C/EBPα was immunoprecipitated from nuclear extracts or from gel filtration fractions with polyclonal antibodies (14AA; Santa Cruz Biotechnology), and the presence of HDAC1 and HP1α in C/EBPα IPs was examined by Western blotting with monoclonal antibodies to the mentioned proteins.
Electrophoretic mobility shift assay was performed with the Glut4 promoter and with promoters of cyclin D1 and cyclin D2. The sequence of the Glut4 probe is as follows. 5′-CTGCACTCAATTCTTTCAGAAATTTCGCAGT-3′ (C/EBP site is shown in bold). The sequences of the probes with C/EBP sites within cyclin D1 and cyclin D2 promoters are shown in Figs 5C and S2. The double stranded probes were gel-purified and labeled by Klenow ‘fill in’ reaction with radioactive P32-dCTP. The DNA probes were incubated with nuclear extracts isolated from WT and from S193D livers. The binding buffer contained 20 mm Tris–HCl pH 7.5, 100 mm KCl, 5 mm MgCl2, 5 mm DDT, 1 μg/10 μL poly (dI-dC) and 10% glycerol. Antibodies were added to the binding reactions before probe addition. The reaction mixtures were separated by 5% native gel electrophoresis; the gel was dried and exposed with X-ray film.
Chromatin immunoprecipitation assay was performed using Chip-It kit according to the instruction. Briefly, chromatin solutions were prepared from livers of WT and S193D mice. C/EBPα, C/EBPβ, HDAC1, histone H3K9 and histone H3-trimethyl K9 were immunoprecipitated from the solutions. DNA was isolated and used for the PCRs with primers covering C/EBPα sites within the Glut4, cyclin D1 and cyclin D2 promoters (see Figs 5 and S2). The sequences of these primers for the Glut4 promoter are as follows: 5′-ACACACACACACACACACACAC-3′ (forward) and 5′-TAAGGTTCCCGCCTGCTTCTGAGTT-3′ (reverse). PCR mixtures were amplified for one cycle of 95°C (5 min), 60°C (5 min) and 72°C (2 min). Then PCR mixtures were amplified for 31 cycles of 95°C (1min), 60°C (2 min) and 72°C (1.5min). PCR products were separated by 8% PAGE.
Micro array studies
Total RNA was isolated from three animals of each genotype and used for micro array assay. The micro array was performed as described in our previous paper (Wang et al., 2008c). Briefly, 200 ng of total RNA was amplified using Illumina TotalPrep RNA Amplification Kit (Cat# IL1791; Ambion, Austin, TX, USA) following kit instructions. In vitro transcription was performed, and biotinylated cRNA was synthesized by 14-h amplification with dNTP mix containing biotin-dUTP and T7 RNA polymerase. Amplified cRNA was subsequently purified and concentration was measured by NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The amplified products were used for analysis of Illumina Sentrix Beadchip Array Mouse-6 arrays. The arrays were scanned with BeadArray Reader (Illumina, San Diego, CA, USA). Data were analyzed using BeadStudio software (Illumina). Clustering and pathway analysis were performed with BeadStudio and Ingenuity Pathway Analysis (Ingenuity Systems, Inc., Third Floor Redwood City, CA, USA) softwares, respectively.
We thank Estela Medrano and Gretchen Darlington for the discussion of this work and for useful suggestions. This work is supported by NIH grants GM55188, CA100070 and AG20752 (NAT).
Jingling Jin has performed experiments with the examination of heterochromatin regions in S193D and old mice, ChIP assay, treatments of S193D and old mice with growth hormone, examination of glucose metabolism in S193D mice and genotyping. Guo-Li Wang has performed Q-RT–PCR, immunostaining and a portion of Western blotting studies. Polina Iakova has performed immunostaining, ChIP and Western blotting experiments. Xiurong Shi was responsible for genotyping and maintenance of mice. Simon Haefliger and Milton Finegold have provided their expertise for the analyses of blood parameters and for examination of liver tissues. Nikolai A. Timchenko performed Co-IP- and HPLC-based examination of protein–protein complexes, was overall supervisor of the research and contributed to the writing and editing the manuscript.