Chromatin remodeling plays an important role in transcriptional regulation. Evidence suggest that acetylation and deacetylation of histones and/or nonhistone proteins play significant roles in chromatin remodeling and thus in the regulation of transcription. 1, 2 Consequently, histone acetyl transferases (HATs) and histone deacetylases (HDACs) are emerging as important molecules in transcriptional regulation. Histone deacetylases inhibitors (HDIs) are promising agents for anticancer therapy as they exhibit strong antitumor activities in vivo with low toxicity in preclinical studies. HDIs belong to a heterogenous class of compounds that includes derivatives of short chain fatty acids, hydroxamic acids, cyclic tetrapetides and benzamides.
Sodium butyrate (NaB), a short chain fatty acid, is a histone deacetylase inhibitor and is produced in the colonic lumen as a consequence of microbial degradation of dietary fibers. 3, 4 Sodium butyrate and other deacetylase inhibitors are not growth inhibitory to normal cells. 5, 6 The growth inhibitory effect of NaB against cancer cells has been attributed to its ability to induce cell cycle arrest, differentiation and apoptosis. 7, 8, 9, 10 NaB induced cell cycle arrest and differentiation has been correlated with its ability to induce p21WAF1/CIP1 in a p53-independent manner 11, 12, 13, 14, 15 and modulate levels of cyclin D1, 12, 16, 17 cdc2, 17 cdk2 12, 18 and proliferating cell nuclear antigen (PCNA). 19 However, the mechanism of apoptosis induction by NaB is not clearly understood. Acetylation of nontranscriptional targets such as nonhistone proteins like tubulin and microtubules has been correlated with apoptosis induction by TSA. 20 Although mitochondrial pathway of apoptosis seems to get activated in sodium butyrate-treated cells, the exact mechanism by which signals are passed on is less understood.
The tumor suppressor p53 is stabilized and activated in response to cellular stress through posttranslational modifications including acetylation. MDM2 has been shown to promote p53 deacetylation by recruiting a complex containing histone deacetylase, HDAC1. 21 This suggests an important role for p53 pathway in the growth inhibition by deacetylase inhibitors. However, the reports in the literature about the role of p53 in the sodium butyrate-mediated growth inhibition of cells are unclear. While some studies suggest a possible role for p53 in HDIs-mediated growth inhibition, 21, 22, 23, 24, 25 others suggest p53 is not required. 26, 27, 28
In our study, we provide evidence for an important role for p53 pathway in sodium butyrate-mediated growth suppression. Addition of NaB to cells resulted in increased levels of functional p53 by a p14ARF-dependent posttranscriptional mechanism. We also show that NaB suppresses the growth of wild-type (WT) p53-containing cells more efficiently. While DNA synthesis inhibition by NaB was not affected by p53 status, the cell cycle phase where majority of cells get arrested was determined by the p53 status. Moreover, the apoptosis induction by NaB was greatly reduced in the absence of p53. These results suggest that p53 pathway of growth suppression involving apoptosis induction has an important role to play in NaB-mediated growth inhibition of cancer cells.
Human cancer cell lines H460, HCT116 WT and MCF7 were described previously. 29, 30, 31 HCT116 p53−/−, which lacks both copies of p53, was kindly provided by Dr. B. Vogelstein, Johns Hopkins University. 32 H460 NEO, H460 E6, HCT116 NEO and HCT116 E6 cell lines were kindly provided by Dr. W.S. El-Deiry, University of Pennsylvania. 33, 34, 35 NEO or E6 stable cell lines were made by transfecting pCMV-NEO or E6 (derived from HPV 16) construct followed by selection for G418 resistance. Transfections into HCT116 WT cells were carried out by using a transfection reagent ESCORT 1 (Sigma Chemical Co., St. Louis, MO) as described previously. 29 Luciferase assays were done as described previously. 36
Western blot analysis
Western blot analysis and immunohistochemical staining were performed as described before 30 with mouse anti-human p21WAF1/CIP1 (Ab-; Oncogene Science, Manhasset, NY), mouse anti-human p53 monoclonal (Ab-2; Oncogene, USA), mouse anti-human Rb monoclonal (LM95.1 clone; Ab5; Oncogene Science), mouse anti-human PARP antibody (Oncogene Science) and goat anti-human actin polyclonal (I-19; Santa Cruz Biotechnology, Santa Cruz, CA) antibodies. Cells were harvested after 24 hr or appropriate time point as described after drug addition and subjected to analysis.
MTT assay was carried out as described previously. 37 1.5 × 103 cells per well were plated in a 96-well plate. After 24 hr of plating, the cells were treated with indicated amounts of NaB. Twenty microliters (5 mg/ml) of MTT was added to each well 48 hr after addition of drug. MTT is a tetrazolium salt that is converted by living cells into blue formazan crystals. The medium was removed from the wells 3 hr after MTT addition and 200 μl of DMSO was added to dissolve the formazan crystals and then the absorbance was measured at 550 nm in an ELISA reader.
RT-PCR was carried out using a 2-step strategy: 37 cDNA was generated using Reverse Transcription kit (Promega, Madison, WI) in the first step and then using gene specific primer sets, PCR was carried out with cDNA as templates. GAPDH (Glyceraldehyde-3-phosphate dehydrogenase) is used for normalization.
DNA synthesis measurement
DNA synthesis was measured by monitoring BrdU incorporation as described earlier. 38 BrdU (20 μM) was added 20 hr after NaB or Adriamycin addition. The experiment was terminated 4 hr after addition of BrdU and the DNA synthesis was measured by using anti-BrdU antibody (Ab-3; Oncogene Science).
Fluorescence-activated cell sorting was carried out as described before. 38 HCT116 WT and HCT116 p53−/− cells were treated with NaB (1 mM). BrdU (20 μM) was added to cells 4 hr before the time point at which they are harvested. The cells were washed with PBS twice and harvested by trypsinization. The cells were washed again with PBS and fixed with cold 70% ethanol for 1 hr. Subsequently, the cells were centrifuged and the pellet was resuspended in 2 N HCl and incubated at room temperature for 30 min. Cells were washed thrice with PBS and incubated with 1 μg of anti-BrdU antibody (Ab-3; Oncogene Science) for 30 min. Cells were washed again twice with PBS and incubated with 2 μg of FITC-conjugated goat anti-mouse antibody (Oncogene Science) for 30 min at room temperature. Cells were washed with PBS once and then incubated with 4 μg of ribonuclease A (Roche, City, State) for 30 min at room temperature. Propidium iodide was added to the cell suspension at a final concentration of 20 μg/ml and incubated for 30 min. Cells were then analyzed by flow cytometry using FACScan (Becton Dickinson, Mountain View, CA). The results were quantified by using the WinList software (Verity Software House, Inc., City, State).
NaB induces p53-mediated transactivation
In order to find out the effect of sodium butyrate on p53 function, we measured the p53-mediated transcription levels in a wild-type p53-containing cell line, HCT116, in the absence and presence of sodium butyrate. Adriamycin, a known activator of p53, was used as positive control. We used a p53-specific reporter PG13-Luc to assay p53-mediated transcription. PG13-Luc contains a synthetic promoter with 13 copies of p53-binding sequence cloned in tandem upstream of a basal polyoma promoter; 39 transcription from this promoter is dependent upon the presence of wild-type p53 protein. 40 Treatment of HCT116 cells with NaB (1 mM) resulted in several-fold increase in transcription from PG13-Luc reporter (Fig. 1a, compare bar 3 with 1). As expected, adriamycin also activated PG13-Luc reporter activity (Fig. 1a, compare bar 2 with 1). From the above finding that p53-mediated transcription is induced in NaB-treated cells, it is evident that p53 pathway is activated by NaB.
NaB induces p53 by a post-transcriptional p14ARF-dependent mechanism
We further investigated the mechanism by which NaB activates p53-dependent transcription. NaB induced p53-mediated transactivation could be due to increased transcription of p53 gene, increased stability of p53 protein or both; p300/CBP-mediated acetylation of p53 leads to increase in sequence-specific DNA- binding. 41 Similarly, mdm-2-mediated recruitment of HDAC1 to p53 has been shown to promote degradation of p53 through ubiquitin system. 21 In order to investigate the mechanism of NaB induced p53-mediated transactivation, we studied the p53 protein levels in the absence and presence of NaB. NaB treatment resulted in a many-fold increase in p53 protein levels in HCT116 cells (Fig. 1b, compare lane 6 with 4) and H460 cells (Fig. 1b, compare lane 3 with 1). As expected, adriamycin treatment also resulted in increase in p53 levels in both cell lines (Fig. 1b, compare lane 2 with 1 and compare lane 5 with 4). Adriamycin seems to induce p53 more strongly than NaB in H460 cells while NaB induced p53 more strongly than adriamycin in HCT116 cells, which could be a cell type-dependent phenomenon. These results suggest that NaB treatment leads to increased p53 protein levels. To verify whether increased p53 level upon NaB treatment is due to increased protein stability or increased transcription rate of p53 gene, we measured the levels of p53 transcripts in the absence and presence of NaB by semiquantitative RT-PCR. NaB treatment did not result in any significant change in p53 messenger levels in H460 cells (Fig. 1c, compare lane 2 with 1). As expected, adriamycin treatment also did not show any significant change in p53 messenger levels (data not shown). These results suggest that NaB induced increase in p53 protein level involves a posttranscriptional mechanism.
p14ARF (p19ARF in mouse) has been implicated as a critical mediator of p53 stabilization or checkpoint activation in response to aberrant expression of a number of positive growth regulatory proteins including c-myc, ras, E1A and E2F 42, 43, 44, 45, 46 and overexpression of BRCA1. 31 To identify the role of p14ARF in p53 stabilization by NaB, we compared the ability of NaB to stabilize p53 in human cell lines differing in their p14ARF status. The human breast carcinoma cell line, MCF7, lacks p14ARF expression. 46 Human lung carcinoma cell line H460 and colon carcinoma cell line HCT116 both carry p14ARF. 31 Adriamycin, a DNA damaging agent, was used as control because it does not require p14ARF to stabilize p53. 31 Both H460 and HCT116 cells stabilized p53 in response to addition of NaB (Fig. 2, compare lane 3 with 1 and compare lane 6 with 4, respectively). However, MCF7 failed to stabilize p53 upon addition of NaB (Fig. 2, compare lane 9 with 7). On the contrary, adriamycin treatment resulted in the stabilization of p53 in all the 3 cell lines tested (Fig. 2, compare lane 2 with 1, lane 5 with 4 and lane 8 with 7). These results suggest that p14ARF may be required for p53 stabilization by NaB. Next we analyzed the steady-state levels of p14ARF mRNA in NaB-treated cells. H460 cells treated with NaB showed a 2-fold increase in p14ARF mRNA levels as compared to the untreated cells (Fig. 1c, compare lane 2 with 1). These results suggest that p53 stabilization by NaB may in part result from increased p14ARF expression.
NaB-induced p53 is functional
In order to verify whether NaB-induced p53 is functional, we studied the ability of p53 to activate the target gene p21WAF1/CIP1, inhibit cellular DNA synthesis and induce apoptosis. p53 stabilized by NaB increased the expression of p21WAF1/CIP1 in H460 (Fig. 2, compare lane 3 with 1; Fig. 1c, compare lane 2 with 1) and HCT116 (Fig. 2, compare lane 6 with 4). In MCF7 cells, NaB treatment, although it failed to induce p53, increased p21WAF1/CIP1 levels (Fig. 2, compare lane 9 with 7). This could be attributed to the ability of NaB to induce p21WAF1/CIP1 using p53-independent pathways. 9, 10, 11, 12, 13 NaB-induced p53 also inhibited cellular DNA synthesis (Fig. 3a, compare f and hwith b; Fig. 3b, compare bars 3 and 4 with bar 1). Adriamycin, which was used as a positive control, also inhibited cellular DNA synthesis (Fig. 3a, compare d with b; Fig. 3b, compare bar 2 with 1). p53, stabilized upon treatment with NaB, also lead to apoptosis induction (Fig. 3c, compare d with b). These experiments suggest that p53 stabilized in response to NaB treatment is functional as it activates p21WAF1/CIP1, inhibits cellular DNA synthesis and induces apoptosis.
Role of p53 in NaB-mediated cytotoxicity
Since above experiments suggest that p53 pathway is activated in NaB-treated cells, we hypothesized that p53 may play an important role in NaB-mediated growth inhibition. To test this hypothesis, we studied the cytotoxicity profile of NaB-treated HCT116 p53−/− cell line (derived from HCT116 cell line) and compared it with that of HCT116 (which carries wild-type p53) cells. HCT116 p53−/− cells carry deletion of both alleles of p53. 31, 32 NaB inhibited the growth of HCT116 p53−/− cells less efficiently than that of HCT116 WT cells (Fig. 4a), suggesting that the growth inhibition by NaB is partially p53-dependent. To further confirm this observation, we used HPV 16 E6 stable transfectants of HCT116 and H460 cell lines. HPV 16 E6 protein forms complex with p53 and promotes its degradation via ubiquitin-mediated proteolysis in an E6-associated protein (E6AP)-mediated reaction. 47, 48, 49 The same cell lines transfected with empty vector (denoted as NEO) were used as control cells. NaB inhibited the growth of HCT116 NEO and H460 NEO cells more efficiently than that of HCT116 E6 and H460 E6 cells (Fig. 4b,c). These results indicate that NaB suppresses the growth of cells carrying wild-type p53 more efficiently, suggestive of an important role for p53 pathway in NaB-mediated cytotoxicity.
Role of p53 on NaB-mediated cell cycle arrest and apoptosis
To study mechanistically the contribution by p53 in NaB-mediated growth suppression, we monitored cellular DNA synthesis and induction of apoptosis in NaB-treated HCT116 and HCT116 p53−/− cells by both BrdU incorporation assay and Fluorescence-activated cell sorting (FACS). NaB treatment resulted in a many-fold reduction in percentage of cells incorporating BrdU in HCT116 (Fig. 5a) as well as HCT116 p53−/− (Fig. 5c) cells. The results obtained in Figure 5a,c were quantified and % BrdU incorporation under different conditions is shown in Figure 5b,d. This result suggests that cellular DNA synthesis inhibition by NaB is p53 independent, which is in good correlation with earlier observations. 11, 12, 13, 14, 15 To further confirm this observation as well as to find out the role of p53 in NaB-mediated apoptosis, we carried out fluorescence-activated cell sorting (FACS). The proportion of cells actively replicating DNA, which represent S phase cells, decreased drastically to a similar extent in both HCT116 WT cells (Fig. 6a; 32.0 to 8.0) and HCT116 p53−/− cells (Fig. 6b; 28.0 to 8.0%) by 24 hr after NaB treatment. Although NaB-inhibited cellular DNA synthesis in both cell lines to similar extent, we found a marked difference in the cell cycle phase where the cells got arrested. By 24 hr of NaB treatment, HCT116 cells arrested mainly at G2/M phase (35%) rather than G1 phase (21%), while HCT116 p53−/− cells arrested mainly in G1 phase (47%) than at G2/M phase (20%). A similar pattern is also seen by 48 hr after NaB treatment. In good correlation, hypophosphorylated form of Rb is seen as early as 24 hr in NaB-treated HCT116 p53−/− cells unlike HCT116 WT cells, where no visible hypophosphorylated form of Rb is seen even after 48 hr of NaB addition (Fig. 7). In the presence of p53, NaB-treated cells arrested mainly in G2/M phase of the cell cycle. This result suggests that p53 is needed for NaB-mediated G2/M arrest and in the absence of p53, NaB-treated cells mainly arrest in G1 phase of the cell cycle.
Cells with less than 2 N content of DNA, which represents apoptotic cells, appeared in high proportion by 24 hr in NaB-treated HCT 116 p53 WT (Fig. 6a; 36.0%). However, there was a substantial difference in the level of apoptosis induced by NaB in HCT116 p53 WT vs. HCT116 p53−/− cells (Fig. 6a,b; 36% and 25%, respectively) by 24 hr. The difference in apoptosis induction by NaB between HCT116 p53 WT and HCT116 p53−/− cells is further increased by 48 hr of NaB treatment (Fig. 6a,b; 51.0% and 31.0%, respectively). The above result was confirmed by PARP cleavage assay as well. Poly (ADP-ribose) polymerase (PARP), an enzyme involved in DNA repair mechanisms, is cleaved by the caspase-3. Therefore, the cleavage of PARP from the native 115 to 89 kDa is considered as a hallmark of apoptosis. Although PARP cleavage is seen in both cell lines, there was a marked difference in PARP cleavage levels. By 24 hr after addition of NaB, about 18% of PARP is cleaved in HCT116 p53 WT cells, while there is hardly any PARP cleavage in HCT116 p53−/− cells (Fig. 7). The PARP cleavage level increased to 29% in HCT116 p53 WT cells as against 20% in HCT116 p53−/− cells by 48 hr of NaB addition (Fig. 7). Marked difference in the PARP cleavage between these 2 cell lines supports the results obtained from flow cytometry analysis. This result suggests that NaB induces apoptosis both by p53-dependent and -independent pathways. Together from these experiments, we conclude that NaB induces apoptosis more efficiently in the presence of p53, which explains the fact that NaB inhibited the growth of wild-type p53 containing cells more efficiently.
In our report, we show evidence for activation of functional p53 involving a post-transcriptional p14ARF-dependent mechanism upon NaB treatment. NaB inhibited the growth of wild-type p53, containing cells more efficiently. Although we found that NaB-mediated inhibition of cellular DNA synthesis is independent of p53, NaB-treated cells mainly arrested in the G2/M stage in the presence of p53, while they get arrested in the G1 phase of the cell cycle in the absence of p53. Furthermore, apoptosis induction by NaB was greatly reduced in the absence of p53, suggesting that NaB can induce apoptosis by both p53-dependent and p53-independent mechanisms.
p300/CBP-dependent acetylation of C-terminus lysine residues of p53 increases its sequence-specific DNA-binding. 41 Evidence also suggests acetylation of lysine residues in the C-terminus of p53 controls the p53 stability by interfering with MDM2-mediated ubiquitination. MDM2 promotes p53 deacetylation by recruiting a complex containing HDAC1, which promotes ubiquitin-mediated degradation of p53. 21 HDAC1 interaction with p53 results in reduced acetylation and downregulation of p53-dependent transcription. 23 Considering the fact that acetylases and deacetylases play an important role in regulation of p53 activity, histone deacetylase inhibitors would be expected to activate p53 functions. In our experiments, NaB addition activated transcription from p53-specific reporter PG13LUC, which suggested that the p53 pathway is activated upon NaB treatment. NaB-treated cells showed increased p53 levels involving a p14ARF dependent posttranscriptional mechanism. Moreover, NaB-induced p53 is found to be functional as it activated its target gene expression, inhibited cellular DNA synthesis and induced apoptosis. These results suggest that the p53 pathway is activated upon NaB addition.
p53 has been implicated as a determinant of chemosensitivity and radiosensitivity. 34, 50 Low concentrations of paclitaxel has been shown to induce cell type-dependent p53, p21 and G1/G2 arrest instead of mitotic arrest. 51 Since our results show that p53 is activated and fully functional in NaB-treated cells, we hypothesized that NaB may inhibit the growth of wild-type p53 containing cells more efficiently. Our result shows that NaB inhibits HCT116 wild-type cells more efficiently than that of HCT116 p53−/− cells. Similarly, we found the HPV 16 E6 stable transfectants of HCT116 and H460 cells are relatively resistant to NaB than the NEO versions of the same cell lines, which carry wild-type p53. These experiments suggest that the p53 pathway induced upon addition of NaB may play an important role in NaB-mediated cytotoxicity.
Suppression of cancer cell growth upon activation of p53 is mediated by two mechanisms, cell cycle arrest and apoptosis, occurring either individually or in combination. 52 In our experiments, NaB induced a potent cellular DNA synthesis inhibition in both HCT116 p53 WT as well as HCT116 p53−/− cells. However, in the presence of p53, NaB-treated cells arrested mainly in G2/M phase of cell cycle, while in the absence of p53 the arrest was mainly in G1 phase of the cell cycle. Induction of G2/M arrest in NaB-treated cells has been reported previously. Many studies have linked the ability of NaB to induces G2/M arrest to repression of cyclin B1, reduction of cdc2 kinase activity 53, 54, 55 and induction of GADD45. 56 However, these studies did not analyze the role of p53 in NaB-mediated G2/M arrest. p53 has been shown induce G2/M arrest by repressing cyclin B1, 57, 58 inducing GADD45 59 and inducing 14-3-3σ. 60 Because our results show that NaB failed to induce G2/M arrest in the absence of p53, we conclude that p53 is essential for NaB to induce G2/M arrest.
Caspase activation plays a central role in the execution of apoptosis. The two best-studied pathways of caspase activation are the cell surface death receptor pathway and the mitochondrion-initiated pathway. 61 NaB has been shown to induce apoptosis involving caspase-8- and caspase-9-dependent mechanisms. 62 However, the role of p53 in NaB-mediated apoptosis is not clear. p53 induces apoptosis by both cell surface death receptor pathway and the mitochondrion-initiated pathway. 63 In our experiments, NaB treatment resulted in induction of apoptosis. However, apoptosis induction by NaB is less efficient in absence of p53, which suggests that p53 also plays an important role in NaB-mediated apoptosis. One can conclude that NaB induces apoptosis by both p53-dependent and independent mechanisms.
Taken together, our results suggest that p53 status is an important determinant of chemosensitivity of cancer cells treated with NaB. Lack of G2/M arrest and reduced apoptosis in the absence of p53 in NaB-treated cells could be correlated with the fact that NaB is more cytotoxic to wild-type p53 containing cells. Perhaps it is important to find out the p53 status in deacetylase inhibitor-based cancer chemotherapy. In addition, combined p53 gene therapy and deacetylase inhibitor chemotherapy could be considered for treatment of p53 mutant or null tumors.
We thank Dr. O. Joy for technical assistance in FACS analyses. KS is a Wellcome Trust International Senior Research Fellow. JJ is supported by a postdoctoral fellowship from Department of Biotechnology (DBT), Government of India. NW is supported by a fellowship from University Grants Commission (UGC), Government of India.