Elevated expression of ornithine decarboxylase (ODC) and increased synthesis of polyamines are hallmarks of epithelial tumorigenesis. The skin and tumors of K6/ODC and ODC/Ras transgenic mice, in which overexpression of ODC has been targeted to hair follicles, were found to exhibit intrinsically high histone acetyltransferase (HAT) activity. We identified Tip60 as a candidate enzyme for contributing significantly to this abnormally high HAT activity. Compared with normal littermate controls, the levels of Tip60 protein and an alternative splice variant Tip53 were found to be greater in K6/ODC mouse skin. Furthermore, skin tumors that spontaneously develop in ODC/Ras bigenic mice typically have substantially more Tip60 protein than adjacent non–tumor-bearing skin and exhibit a unique pattern of Tip60 size variants and chemically modified protein isoforms. Steady-state Tip60 and Tip53 mRNA levels were not affected in ODC-overexpressing skin and tumors, implying novel posttranscriptional regulation by polyamines. Given the diverse roles of Tip60, the overabundance of Tip60 protein is predicted to have biological consequences. Compared with normal littermate skin, we detected altered association of Tip60 with E2F1 and a subset of newly identified Tip60-interacting transcription factors in ODC transgenic mouse skin and tumors. E2F1 was shown to be bound in greater amounts to up-regulated target genes in ODC-overexpressing skin. Thus, up-regulation of Tip60 protein, influencing the expression of Tip60-regulated genes, could play a contributing role in polyamine-mediated tumor promotion. (Cancer Res 2006; 66(16): 8116-22)
Overexpression of ornithine decarboxylase (ODC), catalyzing increased polyamine metabolism, is a common feature of epithelial tumors in humans. The role that polyamines play in promoting tumor development is complex due to their ability to interact with a variety of macromolecules to influence diverse cellular mechanisms. It is clear, however, that polyamines can have a major effect on gene expression that likely contributes to promoting and sustaining tumorigenesis (1–4). One way polyamines might affect gene expression in a non-global manner is to modulate multiprotein complexes instrumental in recruiting the transcriptional machinery to genes. Such protein complexes comprised transcription factors that bind to specific DNA sequences in gene regulatory regions as well as various accessory and coregulatory factors, including those that alter chromatin structure by modifying nearby histones. Generally speaking, recruitment of enzymes that acetylate histones precedes gene activation, whereas the recruitment of enzymes that deacetylate histones is associated with gene silencing.
In an effort to better understand how elevated levels of polyamines promote the development of skin tumors, we have explored the possibility that polyamines may modulate histone-modifying enzymes that play a central role in regulating gene transcription. In previous studies using transgenic mice that constitutively overexpress ODC in the hair follicles of skin (K6/ODC), we found that the total histone acetyltransferase (HAT) activity in skin is elevated compared with normal littermate skin (5). Moreover, HAT activity is exceptionally high in the skin tumors that develop spontaneously in ODC/Ras double transgenic mice. This intrinsically high HAT activity acetylates all four core histones provided as substrate in in vitro assay reactions but exhibits a preference for histone H4. Although we found that p300/CBP HAT activity is higher in ODC-overexpressing skin (5), p300/CBP readily acetylates both histone H3 and histone H4 (6–8). This suggests that at least one other HAT enzyme, having a specificity preference for histone H4, may make a major contribution to the intrinsically high HAT activity in ODC transgenic skin and tumors.
Tip60, a member of the MYST family of HAT enzymes, functions as part of a multiprotein complex that consists of at least 14 subunits. In addition to acetyltransferase activity, this highly conserved complex displays ATPase, DNA helicase, and structural DNA-binding activities (9–12). The Tip60 complex predominately acetylates histone H4 in core histones and, to a lesser extent, H2A in the context of nucleosomes (9, 10). Several splice variants of Tip60 have been detected. Removal of exon 5 results in an enzymatically active 53-kDa HAT protein (Tip53/Tip60b/PLIP; refs. 13–15). In another case, intron 1 is not excised, producing a protein slightly larger than full-length Tip60 (16). A third variant contains both in-frame deletions of exon 5 and the HAT domain (17). In addition to these sequence variants, posttranslational modification results in phosphorylated and ubiquitinated forms of Tip60 (18, 19). The functional roles of these different splice variants and chemically modified isoforms remain to be elucidated.
Originally identified as a coactivator for the HIV-1 Tat protein (20), Tip60 is now known to interact with a variety of transcriptional regulators to influence gene expression, including that of nuclear hormone receptor–regulated genes (21, 22), nuclear factorκB (NF-κB)/β-amyloid precursor protein target genes (23, 24), and ribosomal genes (25). Tip60 can also negatively regulate gene expression through its binding to transcription factors (17, 26–28). Besides gene regulation, multiple roles for Tip60 have been shown in the cellular response to DNA damage. For example, Tip60 has been implicated in double-strand break repair and apoptosis (9, 29), selective histone exchange at DNA lesions (30), activation of ATM kinase activity (31), and the p53 response to UV irradiation (10, 18, 32). Accumulating evidence supports additional intranuclear and extranuclear roles for Tip60. For example, Tip60 interactions with the endothelin receptor A and the interleukin-9 (IL-9) and IL-4 α receptors have been implicated in the activation of mitogen-activated protein kinase and cytokine signaling pathways (33, 34). Tip60 and its splice variant Tip53/PLIP also interact with cytosolic phospholipase A2, enhancing apoptosis and prostaglandin production (15, 35).
We have identified Tip60 as a candidate enzyme for significantly contributing to the high overall HAT activity in ODC transgenic mouse skin and tumors. Compared with normal littermate controls, the levels of Tip60 protein isoforms were found to be greater in K6/ODC transgenic mouse skin and even more so in skin tumors that spontaneously develop in ODC/Ras double transgenic mice. Recognizing that Tip60 associates with multiple transcriptional protein complexes, we have shown that interactions between Tip60 and several transcription factors are influenced by polyamines in mouse skin and tumors. Our results provide the first example of Tip60 modulation by alterations in a biosynthetic pathway.
Materials and Methods
Transgenic animals and cell lines. The generation of K6/ODC transgenic mice was described previously (36). K6/ODC mice on C57BL/65 genetic backgrounds were bred with TG.AC transgenic mice on a FVB background to produce ODC/Ras double transgenic progeny (37). Tumors begin to develop spontaneously on ODC/Ras mice at ∼5 weeks of age. Littermate mice were used for individual experiments. The results provided are representative of analyses done using tissue samples from multiple independent litters of mice.
Protein extraction and immunoblot analysis. Mouse epidermis, total skin, and tumor tissues were harvested as described previously (5). Ground tissue was homogenized in Laemmli buffer [62.5 mmol/L Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 2.5% β-mercaptoethanol] containing 2 μg/mL each of aprotinin, leupeptin, and pepstatin; 1 mmol/L sodium fluoride; 1 mmol/L sodium orthovanadate; and 1 mmol/L Pefabloc by passing through a syringe needle and boiling for 10 minutes. Ground tumor and skin tissues used in activity assays were homogenized in Tween 20 lysis buffer [50 mmol/L HEPES (pH 7.5), 150 mmol/L NaCl, 2.5 mmol/L EDTA, 0.01% Tween 20, 1 mmol/L DTT, 10 mmol/L β-glycerophosphate] containing the protease and phosphatase inhibitors. Debris was removed by centrifugation, and the tissue extracts were stored at −80°C. Protein in tissue extracts was separated by SDS-PAGE, transferred to nitrocellulose or polyvinylidene difluoride membrane, briefly stained with Ponceau S to verify efficiency of transfer and equal protein loading, and subjected to immunoblot analyses using antibodies directed against Tip60. The antigen for one of these antibodies is amino acids 494 to 513 at the COOH terminus of Tip60 (Upstate Biotechnology, Lake Placid, NY). The epitopes for the K-17 and N-17 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) reside at the COOH and NH2 termini, respectively, of Tip60. Antibody binding was detected by enhanced chemiluminescence.
Gel activity assay. Skin and tumor tissue extracts (50 μg) were separated on 10% PAGE gels containing 1 mg/mL core histones or bovine serum albumin (BSA) control substrate polymerized in the gel matrix. The running buffer for the histone substrate gel contained 0.1 μg/mL histone. Following electrophoresis, gels were washed four times for 15 minutes in buffer 1 [50 mmol/L Tris-HCl (pH 8), 20% isopropanol, 0.1 mmol/L EDTA, 1 mmol/L DTT] and four times for 15 minutes in buffer 2 [50 mmol/L Tris-HCl (pH 8), 8 mol/L urea, 0.1 mmol/L EDTA, 1 mmol/L DTT], incubated in buffer 3 [50 mmol/L Tris-HCl (pH 8), 0.005% Tween 40, 0.1 mmol/L EDTA, 1 mmol/L DTT] for 15 minutes at 4°C without agitation, and then again in buffer 3 overnight at 4°C without agitation. The following day, gels were incubated in fresh buffer 3 at 4°C and then at room temperature for 30 minutes without agitation before incubation in buffer 4 [50 mmol/L Tris-HCl (pH 8), 10% glycerol, 0.1 mmol/L EDTA, 1 mmol/L DTT] at room temperature for 15 minutes with gentle agitation. The HAT reaction was done by sealing the gels in bags with 10 mL of buffer 4 containing 16.6 μCi [3H]acetyl CoA and 10 mg histones or BSA and incubating them at 35°C for 1 hour. The gels were then washed thrice in fixative (30% methanol, 10% acetic acid) at room temperature with gentle agitation for 20 minutes, developed in Enhance solution containing 3% glycerol for 1 hour, washed in H2O for 30 minutes, and then dried down at 64°C for 3 hours and subjected to autoradiography.
RNA isolation and reverse transcription-PCR. Total RNA was isolated from skin tissue using TriReagent (Molecular Research Center, Cincinnati, OH). Total RNA (2 μg) was reverse transcribed using random hexamer primers at 42°C for 1 hour. Duplicate PCR amplification reactions contained 200 μmol/L deoxynucleotide triphosphates, 0.04 to 0.5 μmol/L forward and reverse primers, 1 mol/L betaine, and 1 unit of polymerase. A constitutively expressed gene, glyceradehyde 3-phosphate dehydrogenase (GAPDH), was simultaneously amplified as an internal control. For quantitation, gels were scanned using a Bio-Rad Fluor-S MultiImager, or 32P-labeled primer (0.5-1 × 106 cpm) was included in the PCR reactions, and gels were quantified using a PhosphorImager. The primers used for PCR were Tip60/Tip53 forward primer A, 5′-CCAAGACACCTACCAAGAACGGACTT-3′; forward primer B, 5′-AAGACCTTGCCAATCCCGGTCCAGAT-3′; and reverse primer C, 5′-TACGGGCTGACCCATTCTGAGGGAAAA-3′; GAPDH forward primer, 5′-TGCTGAGTATGTCGTGGAGTC-3′ and reverse primer, 5′-AGTGGGAGTTGCTGTTGAAGT-3′; plasminogen activator inhibitor-1 (PAI-1) forward primer, 5′-AGGGCTTCATGCCCCACTTCTTCA-3′ and reverse primer, 5′-AGTAGAGGGCATTCACCAGCACCA-3′; cdc2 forward primer, 5′-TCAGACTTGAAAGCGAGGAA-3′ and reverse primer, 5′-TCGTCCAGGTTCTTGACGTG-3′.
Chromatin immunoprecipitation assay. The chromatin immunoprecipitation experiments were done using the Chromatin Immunoprecipitation Assay kit (Upstate Biotechnology) with some modification. Total skin was cross-linked by incubation on ice in 1% paraformaldehyde for 15 minutes. Cross-linking was stopped by rinsing the skin tissue in cold PBS containing 0.125 mol/L glycine. The tissue was subsequently rinsed twice in ice-cold PBS containing 1 μg/mL each of aprotinin, leupeptin, and pepstatin and 1 mmol/L Pefabloc. The tissue was pulverized in liquid nitrogen and homogenized briefly in homogenization buffer [5 mmol/L PIPES (pH 8), 85 mmol/L KCl, 0.5% NP40] containing protease inhibitors as above. Following a brief spin at 4°C, the pelleted material was resuspended in nuclei lysis buffer [50 mmol/L Tris-HCl (pH 8.1), 10 mmol/L EDTA, 1% SDS] containing protease inhibitors, incubated 20 minutes on ice, spun 5 minutes at maxspeed, resuspended in 500 to 800 μL of SDS lysis buffer containing protease inhibitors, and then sonicated to generate 200- to 800-bp DNA fragments. After clearance by centrifugation, the supernatant was diluted 10-fold in chromatin immunoprecipitation dilution buffer containing protease inhibitors and incubated with a monoclonal antibody against E2F1 (Santa Cruz KH95) or with mouse IgG, at 4°C overnight. The immunocomplexes were captured by salmon sperm DNA adsorbed/protein A agarose beads, washed, eluted, and DNA-protein cross-links reversed by incubating 6 hours at 65°C. DNA was purified by phenol/chloroform extraction and ethanol precipitation. The following primers were used to analyze the binding of E2F1 to target promoters (38): carboxylesterase gene forward primer, 5′-CCCAGCAGAGAAACACTGGG-3′ and reverse primer, 5′-AGAGACAGGGAGGTGTCTGC-3′; PAI-1 gene forward primer, 5′-CAGACCAAGGCTCGAGGAAG-3′ and reverse primer, 5′-ATCCAGCTGTGCTCCGTTCC-3′; cdc2 gene forward primer, 5′- GTGGACTGTCACTTTGGTGGCTGGC-3′ and reverse primer, 5′- GGTAAAGCTCCCGGGATCCGCCAAT-3′. The following β-actin gene primers were used as a specificity control: forward primer, 5′-AGCACAGCTTCTTTGCAGCTCCTT-3′ and reverse primer, 5′-TTTGCACATGCCGGAGCCGTTGT-3′. PCR products were resolved by agarose gel electrophoresis and visualized by ethidium bromide staining.
Protein-protein interaction screen. Interactions between Tip60 and a panel of 54 transcription factors were assayed using the TranSignal TF-TF Interaction Array I kit (Panomics, Inc., Fremont, CA). Tissue was homogenized in radioimmunoprecipitation assay buffer [50 mmol/L Tris-HCl (pH 7.5), 1% NP40, 0.25% sodium deoxycholate, 0.25% SDS, 150 mmol/L NaCl, 1 mmol/L EGTA] containing protease inhibitors as described above and 0.1 mmol/L DTT. Ten micrograms of tissue homogenate were used in each immunoprecipitation reaction using 1.8 μg of an antibody directed against Tip60 (Santa Cruz H-93) or rabbit IgG. Binding of transcription factors to cis-element oligonucleotide probes, immunoprecipitation of Tip60-containing DNA/protein complexes, and hybridization of eluted cis-element probes to the TranSignal Array membrane were done according to the manufacturer's protocol. Bound probes were detected using enhanced chemiluminescence and quantitated by densitometry.
There are numerous HAT enzymes that could potentially contribute to the overall elevated HAT activity measured in ODC mouse skin and tumor extracts. As a first step towards identifying which HAT enzymes are influenced by increased levels of polyamines in mouse skin, an in situ activity assay was used to directly detect the presence of HAT enzymes embedded in the polyacrylamide gel matrix following the separation of proteins by electrophoresis. The gel activity assay detected two proteins between 64 and 50 kDa in tissue extracts prepared from ODC/Ras skin and tumors that exhibit HAT activity (Fig. 1). A third band at ∼48 kDa was also readily detectable in the tumor samples.
The approximate molecular weight of the uppermost band detected in the gel activity assay was consistent with the presence of the 60-kDa Tip60 HAT enzyme. Likewise, a lower band was consistent with the possible detection of the 53-kDa isoform of the Tip60 protein, known as Tip53. Therefore, protein extracts prepared from K6/ODC transgenic and normal littermate mouse epidermis were separated by SDS-PAGE and subjected to immunoblot analyses using an antibody directed against the COOH terminus of Tip60 and Tip53. Indeed, the levels of both the Tip60 and Tip53 proteins were found to be higher (∼1.3-fold and ∼1.9-fold, respectively) in K6/ODC skin compared with littermate controls (Fig. 2A). Interestingly, compared with both normal littermate and adjacent non–tumor-bearing total skin, tumors from ODC/Ras bigenic mice typically have substantially greater amounts of Tip60 and, often, Tip53 (Fig. 2B). Tip53 was not readily detected in the normal skin and only barely detectable in the ODC/Ras skin, consistent with our observation that considerably less Tip60, and even moreso, Tip53, is found in total skin and in dermis compared with epidermis (data not shown). An extremely abundant protein and several less abundant proteins in the ∼48-kDa range were also detected in the tumor tissue (denoted by the bracket and double asterisk in Fig. 2B). Although highly reproducible with independent lots of one particular antibody, it is not yet known whether the very prominent band in tumors represents a novel Tip60 isoform, a consistent degradation product, or a nonrelated protein. Using an antibody directed against a different Tip60 epitope, a series of bands having greater molecular weight than Tip60 is better resolved in ODC/Ras tumor extracts (brackets in Fig. 2C). This banding pattern is remarkably similar to the ladder of bands previously shown to reflect differentially ubiquitinated isoforms of exogenously expressed Tip60 in U2OS cells (18). Simultaneous probing with an antibody against ubiquitin produces a partially overlapping banding pattern (data not shown). The antibody used in Fig. 2C also readily detects a prominent band that migrates slightly slower than Tip60 (single asterisks), that likely represents the phosphorylated form of Tip60 (19). Interestingly, this Tip60 isoform seems to be less abundant in ODC/Ras skin tumors compared with adjacent nontumorigenic skin. The inefficiency of available antibodies to immunoprecipitate Tip60 from tissue extracts and the lack of antibodies specific for chemically modified Tip60 isoforms preclude the absolute identification of the various immunoreactive proteins at this time. Nevertheless, the spectrum of Tip60 protein isoforms seems to differ among normal skin, ODC-overexpressing skin, and skin tumors that develop following overexpression of both ODC and activated Ha-Ras.
The increased levels of Tip60 and Tip53 protein in ODC-overexpressing skin and tumors could result from polyamine-induced increased transcription and/or decreased turnover of Tip60/Tip53 mRNA. Alternatively, more Tip60 and Tip53 protein could result from polyamine modulation of protein synthesis or stability. To evaluate these possibilities, the steady-state levels of Tip60 and Tip53 mRNAs in K6/ODC and ODC/Ras skin and tumors were compared with that in skin of normal mice controls. There was no significant difference in the level of either the Tip60 or Tip53 mRNA isoforms in ODC transgenic skin (Fig. 3). Moreover, normalization against GAPDH mRNA suggests a decrease in Tip60 mRNA in ODC/Ras tumors. These results indicate that elevated levels of polyamines exert posttranscriptional regulation of Tip60. Cycloheximide studies have not revealed definitive differences in the half-life of Tip60 protein isoforms in an inducible ODC cell line or K6/ODC primary keratinocytes. Because we have not detected increased levels of Tip60 or Tip53 protein in either of these ODC-overexpressing cell culture models (data not shown), it is possible that experiments involving cultured cells will not adequately mimic the polyamine regulation of Tip60/Tip53 protein.
Because Tip60 is now known to be a component of multiple protein complexes recruited to genes via interactions with transcriptional regulators (21, 23, 24, 28, 39, 40), we reasoned that there would be more Tip60 protein available in ODC-overexpressing tissues to interact with transcription factors. To test this, protein extracted from K6/ODC and normal littermate skin and ODC/Ras tumors was incubated with a pool of biotinylated oligonucleotides containing consensus binding sequences for 54 different transcription factors. Using an antibody against Tip60, protein complexes containing Tip60 that were bound to the oligonucleotides were immunoprecipitated and washed free of nonspecific binding proteins. The biotin-labeled oligo probes were then eluted from the immunoprecipitated protein complexes and hybridized to a TranSignal array membrane spotted with different transcription factor consensus sequences (Fig. 4). Thus, a positive signal on the array as a result of binding of an oligo probe serves as an indirect measurement of the interaction between Tip60 and a specific transcription factor. Notably, a large subset of transcription factors, including E2F1, which is known to recruit Tip60 complexes to chromatin (40), was coimmunoprecipitated with Tip60 as detected by this assay. Importantly, E2F1 and several other transcription factors, including MEF-1, NF-1, NFATc, Pbx-1, Smad 3/4, and USF-1, were found to be associated with Tip60 to a greater degree in K6/ODC skin compared with normal littermate skin. An even greater amount of each of these factors, except NF-1, was associated with Tip60 in ODC/Ras tumors, correlating with the greater abundance of Tip60 protein in tumor tissue. In fact, the interactions between Tip60 and several transcription factors (e.g., glucocorticoid response element, NF-κB, retinoic acid receptor, retinoid X receptor, thyroid hormone receptor, and signal transducers and activators of transcription 4) were only detectable in the tumor tissue extract. Although Tip60 is also known to associate with c-Myc complexes (39, 41), no Myc protein was detected in this assay for any of the extracts, suggesting that the amount of Myc present in mouse skin and tumor tissue is below the limits of detection for this system.
Interestingly, there was one case in which a transcription factor (Brn-3) was found to be highly associated with Tip60 in normal skin but less so in K6/ODC skin (Fig. 4B and C). Moreover, a further reduction in the interaction between Tip60 and Brn-3 was observed in ODC/Ras tumors. This latter result suggests that increased levels of polyamines may sometimes negatively influence the interaction between Tip60 and components of transcription factor protein complexes.
Although more E2F1 was detected associated with Tip60 in ODC-overexpressing skin and tumors by the TranSignal screen, we were unable to verify binding between Tip60 and E2F1 by conventional immunoprecipitation and Western blotting. Instead, chromatin immunoprecipitation analysis of several E2F1 target genes (38) was done. Sonicated chromatin prepared from K6/ODC and normal littermate mouse skin was immunoprecipitated with an antibody specific for E2F1 or without antibody as a specificity control. The immunoprecipitated chromatin fragments were then analyzed by PCR amplification. We detected more binding of E2F1 to relevant promoter regions of the PAI-1, carboxylesterase, and Cdc-2 genes (38) in K6/ODC transgenic mouse skin compared with the skin of normal littermates (Fig. 5A). This was not the case for a region of the β-actin gene, to which E2F1 would not be expected to bind. We then examined the expression of two of these E2F1 target genes in mouse skin (Fig. 5B). There was a small (1.6-fold) increase in PAI-1 mRNA in K6/ODC skin compared with normal mouse skin. This is consistent with the ∼1.5-fold increase in PAI-1 mRNA measured by microarray analyses.1
L. Lan, B. Paul, and S. Gilmour. unpublished results.
We have shown previously that overexpression of ODC, resulting in increased synthesis of polyamines, leads to an abnormally high intrinsic HAT activity in the skin and tumors of transgenic mice; this elevated HAT activity has a pronounced specificity preference for histone H4 (5). Our new studies have revealed high levels of the Tip60 HAT enzyme and its splicing variant Tip53 in K6/ODC and ODC/Ras transgenic mouse skin and tumors. Tip60 functions as part of a multiprotein complex that also contains TRRAP, p400, Tip48, Tip49, BAF53, and β-actin (9–12). Perhaps not coincidentally, this complex predominately acetylates histone H4 in nucleosomes. Besides several splicing variants (15–17), phosphorylated and ubiquitinated forms of Tip60 have been reported (18, 19). We observed that the relative pattern of Tip60 isoforms varies significantly among normal skin, K6/ODC skin, and ODC/Ras tumors. Because the functional significance of the splicing variants and different chemically modified isoforms has not been well defined, it is difficult to predict the consequences of polyamine-mediated alterations in the overall amount and relative ratios of these different Tip60 isoforms on normal cell function.
Nonetheless, given the normal tight regulation of Tip60 (18, 19) and the integral involvement of this HAT enzyme in gene regulation, DNA repair, apoptosis, and cellular signaling, it seems certain that an overabundance of Tip60, Tip53, and any of their posttranslationally modified isoforms would have biological ramifications. Although still to be individually verified and characterized, the significant number of novel Tip60 interactions, direct or indirect, revealed in our transcription factor screen, fuels speculation that greater Tip60 levels affect gene expression. Moreover, seven factors showed increased association with Tip60 in K6/ODC skin extracts and even greater association in ODC/Ras tumor extracts. This subset of transcription factors includes E2F1, which has recently been reported to recruit Tip60 complexes to E2F1 target genes and promote histone hyperacetylation following mitogenic stimulation (40). Although bound Tip60 was not detectable in our chromatin immunoprecipitation assays, we were able to show greater recruitment of E2F1 to several of its target genes in K6/ODC skin. It stands to reason that having more Tip60 protein molecules available in the nuclear compartment to interact with available transcription factor complexes might augment the kinetics of expression of some genes and/or extend the transcriptional regulatory function of Tip60 to additional genes. This may involve increased acetylation of nearby histones (39, 40), and/or other regulatory proteins as occurs for Myc (41), the androgen receptor (22), and UBF (25). Indeed, we measured increased levels of nucleolin (2.8-fold and 2.3-fold) and nucleophosmin (3.4-fold and 2.9-fold) mRNAs in ODC/Ras skin and tumors, respectively (see Supplementary Data), suggesting that the greater amount of Tip60 protein in these tissues may promote increased recruitment of Myc/Tip60 protein complexes to these target genes (39), leading to their increased transcription. Tip60 HAT activity has been shown to stabilize Myc protein as well as its binding to chromatin and contribute to histone acetylation in response to mitogenic stimuli (39, 41). In some instances, Tip60 can negatively influence gene transcription (17, 26–28). In that context, it is intriguing that our transcription factor screen also identified the POU transcription factor Brn-3 as a factor whose interaction with Tip60 is negatively affected in K6/ODC skin and to an even greater extent in ODC/Ras tumors. The Brn-3 family of transcription factors positively and negatively regulates many genes, including some critical for the somatosensory, visual, and auditory/vestibular systems (42), as well as genes relevant to cancer (43). Because the different Brn-3 family members can exert opposite and antagonistic effects on target promoters (44), characterization of Tip60 interactions with the individual Brn-3 homologues will be necessary to predict the consequences of polyamine-mediated modulation of Tip60-Brn-3 function in mouse skin.
Increased ODC activity and concomitant high levels of polyamines are characteristic of human epithelial tumors (45–49). The ODC inhibitor α-difluoromethylornithine exhibits chemopreventive effects for prostate adenocarcinoma in transgenic mice (50). Therefore, it may not be coincidental that increased levels and altered cellular distribution of Tip60 protein positively correlate with the malignant potential of human prostate tumors (51). Our studies have revealed polyamine-mediated accumulation of Tip60 and enhanced association of this coactivator protein with a subset of interacting transcription factors in mouse skin. These changes likely influence the expression of some genes, which could play a contributing role in polyamine-mediated tumor promotion.
Grant support: National Cancer Institute grant CA95592 (S.K. Gilmour).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Karen Knudsen for critical review of this article and Loretta Rossino for editorial assistance.