Differential response of cancer cells to HDAC inhibitors trichostatin A and depsipeptide

Background:

Over the last decade, several drugs that inhibit class I and/or class II histone deacetylases (HDACs) have been identified, including trichostatin A, the cyclic depsipeptide FR901228 and the antibiotic apicidin. These compounds have had immediate application in cancer research because of their ability to reactivate aberrantly silenced tumour suppressor genes and/or block tumour cell growth. Although a number of HDAC inhibitors are being evaluated in preclinical cancer models and in clinical trials, little is known about the differences in their specific mechanism of action and about the unique determinants of cancer cell sensitivity to each of these inhibitors.

Methods:

Using a combination of cell viability assays, HDAC enzyme activity measurements, western blots for histone modifications, microarray gene expression analysis and qRT–PCR, we have characterised differences in trichostatin A vs depsipeptide-induced phenotypes in lung cancer, breast cancer and skin cancer cells and in normal cells and have then expanded these studies to other HDAC inhibitors.

Results:

Cell viability profiles across panels of lung cancer, breast cancer and melanoma cell lines showed distinct sensitivities to the pan-inhibitor TSA compared with the class 1 selective inhibitor depsipeptide. In several instances, the cell lines most sensitive to one inhibitor were most resistant to the other inhibitor, demonstrating these drugs act on at least some non-overlapping cellular targets. These differences were not explained by the HDAC selectivity of these inhibitors alone since apicidin, which is a class 1 selective compound similar to depsipeptide, also showed a unique drug sensitivity profile of its own. TSA had greater specificity for cancer vs normal cells compared with other HDAC inhibitors. In addition, at concentrations that blocked cancer cell viability, TSA effectively inhibited purified recombinant HDACs 1, 2 and 5 and moderately inhibited HDAC8, while depsipeptide did not inhibit the activity of purified HDACs in vitro but did in cellular extracts, suggesting a potentially indirect action of this drug. Although both depsipeptide and TSA increased levels of histone acetylation in cancer cells, only depsipeptide decreased global levels of transcriptionally repressive histone methylation marks. Analysis of gene expression profiles of an isogenic cell line pair that showed discrepant sensitivity to depsipeptide, suggested that resistance to this inhibitor may be mediated by increased expression of multidrug resistance genes triggered by exposure to chemotherapy as was confirmed by verapamil studies.

Conclusion:

Although generally thought to have similar activities, the HDAC modulators trichostatin A and depsipeptide demonstrated distinct phenotypes in the inhibition of cancer cell viability and of HDAC activity, in their selectivity for cancer vs normal cells, and in their effects on histone modifications. These differences in mode of action may bear on the future therapeutic and research application of these inhibitors.

Class I histone deacetylases sequentially interact with MyoD and pRb during skeletal myogenesis

We describe a functional and biochemical link between the myogenic activator MyoD, the deacetylase HDAC1, and the tumor suppressor pRb. Interaction of MyoD with HDAC1 in undifferentiated myoblasts mediates repression of muscle-specific gene expression. Prodifferentiation cues, mimicked by serum removal, induce both downregulation of HDAC1 protein and pRb hypophosphorylation. Dephosphorylation of pRb promotes the formation of pRb-HDAC1 complex in differentiated myotubes. pRb-HDAC1 association coincides with disassembling of MyoD-HDAC1 complex, transcriptional activation of muscle-restricted genes, and cellular differentiation of skeletal myoblasts. A single point mutation introduced in the HDAC1 binding domain of pRb compromises its ability to disrupt MyoD-HDAC1 interaction and to promote muscle gene expression. These results suggest that reduced expression of HDAC1 accompanied by its redistribution in alternative nuclear protein complexes is critical for terminal differentiation of skeletal muscle cells.

Regulation of the homeodomain CCAAT displacement/cut protein function by histone acetyltransferases p300/CREB-binding protein (CBP)-associated factor and CBP

The CCAAT displacement protein/cut homologue (CDP/cut) is a divergent homeodomain protein that is highly conserved through evolution and has properties of a potent transcriptional repressor. CDP/cut contains three conserved cut-repeat domains and a conserved homeobox, each involved in directing binding specificity to unique nucleotide sequence elements. Furthermore, CDP/cut may play a role as a structural component of chromatin through its direct interaction with nucleosomal DNA and association with nuclear matrix attachment regions. CDP/cut is cell-cycle regulated through interactions with Rb, p107, specific kinases and phosphatases directing the transcriptional activity of CDP/cut on such genes encoding p21(WAF1,CIP1), c-myc, thymidine kinase, and histones. Our previous studies indicate that CDP/cut is associated with histone deacetylase activity and is associated with a corepressor complex through interactions with histone deacetylases. Here, we report the interaction of CDP/cut with CBP and p300/CREB-binding protein-associated factor (PCAF) along with the modification of CDP/cut by the histone acetyltransferase PCAF. Acetylation of CDP/cut by PCAF is directed at conserved lysine residues near the homeodomain region and regulates CDP/cut function. These observations are consistent with the ability of CDP/cut to regulate genes as a transcriptional repressor, suggesting acetylation as a mechanism that regulates CDP/cut function.

The accessibility of histone H3 tails in chromatin modulates their acetylation by P300/CBP-associated factor

P300/CBP-associated factor (PCAF) is a transcriptional coactivator with intrinsic histone acetylase activity. Reversible acetylation of the core histone tails in chromatin has been linked to transcriptional regulation. Here we investigate the mechanism whereby PCAF acetylates its target in chromatin. We demonstrate that recombinant PACF preferentially acetylates the H3 tail in oligonucleosomes, as compared with nucleosome core particles. The rate of acetylation is directly related to the length of the oligonucleosomal substrate. Using a trypsin accessibility assay, we demonstrate that the rate of acetylation is also related to the accessibility of the H3 tail in uncondensed oligonucleosomes. We suggest that PCAF, and perhaps other histone acetyltransferases, acetylate chromatin templates more efficiently than core particle subunits and that this preference arises from an increased accessibility of the H3 tail in either condensed or uncondensed oligonucleosomes. Acetylation of the H3 tails by the native PCAF complex is not affected by the length of the oligonucleosomal substrate. Our results suggest that the accessibility of the H3 tail in chromatin is a major factor affecting their rate of acetylation and that component(s) in the native PCAF complex function to modify the organization of these tails in chromatin thereby enhancing their accessibility to PCAF.

Histone H1 is a specific repressor of core histone acetylation in chromatin

Although a link between histone acetylation and transcription has been established, it is not clear how acetylases function in the nucleus of the cell and how they access their targets in a chromatin fiber containing H1 and folded into a highly condensed structure. Here we show that the histone acetyltransferase (HAT) p300/CBP-associated factor (PCAF), either alone or in a nuclear complex, can readily acetylate oligonucleosomal substrates. The linker histones, H1 and H5, specifically inhibit the acetylation of mono- and oligonucleosomes and not that of free histones or histone-DNA mixtures. We demonstrate that the inhibition is due mainly to steric hindrance of H3 by the tails of linker histones and not to condensation of the chromatin fiber. Cellular PCAF, which is complexed with accessory proteins in a multiprotein complex, can overcome the linker histone repression. We suggest that linker histones hinder access of PCAF, and perhaps other HATs, to their target acetylation sites and that perturbation of the linker histone organization in chromatin is a prerequisite for efficient acetylation of the histone tails in nucleosomes.

PCAF interacts with tax and stimulates tax transactivation in a histone acetyltransferase-independent manner

Recent studies have shown that the p300/CREB binding protein (CBP)-associated factor (PCAF) is involved in transcriptional activation. PCAF activity has been shown strongly associated with histone acetyltransferase (HAT) activity. In this report, we present evidence for a HAT-independent transcription function that is activated in the presence of the human T-cell leukemia virus type 1 (HTLV-1) Tax protein. In vitro and in vivo GST-Tax pull-down and coimmunoprecipitation experiments demonstrate that there is a direct interaction between Tax and PCAF, independent of p300/CBP. PCAF can be recruited to the HTLV-1 Tax responsive element in the presence of Tax, and PCAF cooperates with Tax in vivo to activate transcription from the HTLV-1 LTR over 10-fold. Point mutations at Tax amino acid 318 (TaxS318A) or 319 to 320 (Tax M47), which have decreased or no activity on the HTLV-1 promoter, are defective for PCAF binding. Strikingly, the ability of PCAF to stimulate Tax transactivation is not solely dependent on the PCAF HAT domain. Two independent PCAF HAT mutants, which knock out acetyltransferase enzyme activity, activate Tax transactivation to approximately the same level as wild-type PCAF. In contrast, p300 stimulation of Tax transactivation is HAT dependent. These studies provide experimental evidence that PCAF contains a coactivator transcription function independent of the HAT activity on the viral long terminal repeat.

Overlapping but distinct patterns of histone acetylation by the human coactivators p300 and PCAF within nucleosomal substrates

A number of transcriptional coactivators possess intrinsic histone acetylase activity, providing a direct link between hyperacetylated chromatin and transcriptional activation. We have determined the core histone residues acetylated in vitro by recombinant p300 and PCAF within mononucleosomes. p300 specifically acetylates all sites of histones H2A and H2B known to be acetylated in bulk chromatin in vivo but preferentially acetylates lysines 14 and 18 of histone H3 and lysines 5 and 8 of histone H4. PCAF primarily acetylates lysine 14 of H3 but also less efficiently acetylates lysine 8 of H4. PCAF in its native form, which is present in a stable multimeric protein complex lacking p300/CBP, primarily acetylates H3 to a monoacetylated form, suggesting that PCAF-associated polypeptides do not alter the substrate specificity. These distinct patterns of acetylation by the p300 and PCAF may contribute to their differential roles in transcriptional regulation.

Histone-like TAFs within the PCAF histone acetylase complex

PCAF histone acetylase plays a role in regulation of transcription, cell cycle progression, and differentiation. Here, we show that PCAF is found in a complex consisting of more than 20 distinct polypeptides. Strikingly, some polypeptides are identical to TBP-associated factors (TAFs), which are subunits of TFIID. Like TFIID, histone fold-containing factors are present within the PCAF complex. The histone H3- and H2B-like subunits within the PCAF complex are identical to those within TFIID, namely, hTAF(II)31 and hTAF(II)20/15, respectively. The PCAF complex has a novel histone H4-like subunit with similarity to hTAF(II)80 that interacts with the histone H3-like domain of hTAF(II)31. Moreover, the PCAF complex has a novel subunit with WD40 repeats having a similarity to hTAF(II)100.

Cloning of Drosophila GCN5: conserved features among metazoan GCN5 family members

PCAF and hGCN5 are distinct human genes that encode proteins related to the yeast histone acetyltransferase and transcriptional adapter GCN5. The PCAF protein shares extensive similarity with the 439 amino acids of yGCN5, but it has an approximately 350 amino acid N-terminal extension that interacts with the transcriptional co-activator p300/CBP. Adenoviral protein E1a can disrupt PCAF-CBP interactions and prevent PCAF-dependent cellular differentiation. In this report, we describe the cloning and initial characterization of a Drosophila homolog of yGCN5. In addition to the homology to yGCN5, the Drosophila protein shares sequencesimilarity with the N-terminal portion of human PCAF that is involved in binding to CBP. In the course of characterizing dGCN5, we have discovered that hGCN5 also contains an N-terminal extension with significant similarity to PCAF. Interestingly, in the case of the h GCN5 gene, alternative splicing may regulate the production of full-length hGCN5. The presence of the N-terminal domain in a Drosophila GCN5 homolog and both human homologs suggests that it was part of the ancestral form of metazoan GCN5.

Nuclear receptor coactivator ACTR is a novel histone acetyltransferase and forms a multimeric activation complex with P/CAF and CBP/p300

We report here the identification of a novel cofactor, ACTR, that directly binds nuclear receptors and stimulates their transcriptional activities in a hormone-dependent fashion. ACTR also recruits two other nuclear factors, CBP and P/CAF, and thus plays a central role in creating a multisubunit coactivator complex. In addition, and unexpectedly, we show that purified ACTR is a potent histone acetyltransferase and appears to define a distinct evolutionary branch to this recently described family. Thus, hormonal activation by nuclear receptors involves the mutual recruitment of at least three classes of histone acetyltransferases that may act cooperatively as an enzymatic unit to reverse the effects of histone deacetylase shown to be part of the nuclear receptor corepressor complex.

The transcriptional coactivators p300 and CBP are histone acetyltransferases

p300/CBP is a transcriptional adaptor that integrates signals from many sequence-specific activators via direct interactions. Various cellular and viral factors target p300/CBP to modulate transcription and/or cell cycle progression. One such factor, the cellular p300/CBP associated factor (PCAF), possesses intrinsic histone acetyltransferase activity. Here, we demonstrate that p300/CBP is not only a transcriptional adaptor but also a histone acetyltransferase. p300/CBP represents a novel class of acetyltransferases in that it does not have the conserved motif found among various other acetyltransferases. p300/CBP acetylates all four core histones in nucleosomes. These observations suggest that p300/CBP acetylates nucleosomes in concert with PCAF.

Equine infectious anemia virus gene expression: characterization of the RNA splicing pattern and the protein products encoded by open reading frames S1 and S2

The utilization of predicted splice donor and acceptor sites in generating equine infectious anemia virus (EIAV) transcripts in fetal donkey dermal cells (FDD) was examined. A single splice donor site identified immediately upstream of the gag coding region joins the viral leader sequence to all downstream exons of spliced EIAV transcripts. The predominant 3.5-kb transcript synthesized in EIAV-infected FDD cells appears to be generated by a single splicing event which links the leader sequence to the first of two functional splice acceptor sites near the 5' end of the S1 open reading frame (ORF). The translation products encoded by the 3.5-kb transcript were examined by producing in vitro transcripts from a cDNA corresponding to this RNA followed by in vitro translation in wheat germ extracts. These transcripts directed the synthesis of three proteins: the virus trans-activator protein (EIAV Tat) encoded by ORF S1, a protein of unknown function encoded by ORF S2, and the virus envelope glycoprotein. When transfected into FDD cells, this cDNA also directed expression of EIAV Tat. Amino-terminal sequence analysis of the in vitro-synthesized S1 protein supports the suggestion that translation of EIAV Tat is initiated at a CUG codon within the virus leader region. Both in vitro-synthesized S2 protein and synthetic peptides corresponding to S2 are shown to react positively with sera obtained from EIAV-infected horses, providing the first direct evidence of expression of this protein in infected animals.

Proviral DNA integration and transcriptional patterns of equine infectious anemia virus during persistent and cytopathic infections

The structure and integration patterns of equine infectious anemia virus (EIAV) proviral DNA and the patterns of viral transcription were examined in persistent and cytopathic infections of cultured cells. The results of Southern blot analyses indicated that, in persistently infected cells, about 30% of the EIAV provirus exists as randomly integrated DNA, while the remaining 70% is equally divided between unintegrated linear and closed circular forms. The cytopathic infection, in contrast, is characterized by levels of integrated provirus ranging from 65 to more than 90% of the total proviral DNA, depending on the extent of cytopathology exhibited by the virus strain employed. In both persistent and cytopathic infections, extensive Northern (RNA) blot analyses have revealed the presence of two major virus-specific transcripts, an 8.2-kilobase (kb) full-length genomic mRNA and a 3.5-kb single-spliced mRNA. A low-abundance 1.5-kb mRNA, presumably formed by a double-splicing event of the full-length RNA, was also detected in the cytopathic EIAV infection. The two major viral transcripts are present in approximately equal quantities in persistently infected cells, while the cytopathic infection reveals nearly a 30-fold higher level of viral transcripts in which the 3.5-kb species constitutes over 75% of the total viral mRNA. The relatively high proportion of proviral DNA integration and the simple pattern of viral transcription observed during EIAV infections appeared to be different from the generally observed patterns of predominantly unintegrated proviral DNA and multi-spliced viral mRNAs in cells infected with other lentiviruses such as visna virus or human immunodeficiency virus type 1. Moreover, the data suggested that the cytopathology of EIAV may be correlated in part with the degree of proviral DNA integration and levels of viral mRNA in infected cells, particularly that of the spliced 3.5-kb mRNA.