Abstract
Background/Aim: Alterations of global histone modification levels have been identified in various tumor entities, including bladder cancer (BCA). Our study was designed to investigate the value of global histone acetylation levels as diagnostic and prognostic biomarker for BCA patients. Materials and Methods: A tissue microarray with formalin-fixed paraffin-embedded tissues (271 BCA and 29 normal urothelial samples) was used to determine global histone acetylation levels (histone H3 acetylation (H3Ac); histone H3 lysine 18 acetylation (H3K18Ac); histone H4 acetylation (H4Ac)). Results: Global H3Ac levels were decreased in BCA patients, whereas H3K18Ac and H4Ac levels were similar in both groups. All studied histone acetylation markers were lower in muscle-invasive BCA compared to non-muscle invasive BCA and normal urothelial tissue, thereby indicating a possible prognostic relevance. Conclusion: Global histone acetylation levels undergo quantitative alterations during bladder cancer progression and could be helpful to identify patients at risk for early cancer recurrence.
Bladder cancer (BCA) is among of the ten most common human malignancies (1). The clinical course of patients with similar tumor characteristics often differs considerably and sub-classification of BCA patients based on parameters additionally to TNM staging would be desirable to predict the outcome more accurately and to improve treatment modalities (2).
Epigenetic alterations, including histone modifications, play an important role in bladder carcinogenesis (3, 4, 5). The N-terminal tails of histones undergo various post-translational modifications (e.g., acetylation, methylation), which affect chromatin contacts and recruitment of non-histone proteins to the chromatin. Thereby, histone modifications are involved in the regulation of fundamental cell processes like transcription, replication, DNA repair and chromosome condensation (6). Earlier studies indicated that global levels of histone acetylation are suitable biomarkers for patients with urological malignancies: histone H3 acetylation (H3Ac) and histone H4 acetylation (H4Ac) levels were decreased in prostate cancer compared to normal prostate tissue (9). H3Ac and H4Ac levels were correlated with unfavorable pathological staging and histone H3 lysine 18 acetylation (H3K18Ac) was predictive for progression-free survival (PFS) in renal cell carcinoma patients (7). Former studies also indicated a relevance for the analysis of histone modifications in BCA: Schneider et al. demonstrated altered global levels of lysine 4 of histone 3 (H3K4) and histone H4 lysine 20 (H4K20) methylation in non-muscle invasive BCA (NMIBC), muscle invasive BCA (MIBC) compared to normal urothelial tissue; histone H4 containing the trimethylated lysine 20 (H4K20me3) levels were furthermore increased in MIBC patients with a shorter period of survival after radical cystectomy (8). Phosphorylation of histone H2AX at serine 139 was correlated with a reduced risk of BCA recurrence following transurethral resection (9). So far, global histone acetylation levels have not been investigated in BCA. Given the known deregulation of class I histone deacetylases (HDCA) in BCA (10, 11, 12), we hypothesized that global histone H3 and H4 acetylation levels are decreased in BCA tissue. We, therefore, studied global histone H3Ac, H4 Ac, as well as a specific histone marker (H3K18Ac), using immunohistochemistry of a BCA tissue microarray.
Materials and Methods
Patients. A tissue microarray with samples of urothelial bladder cancer (NMIBC, n=150; MIBC, n=121) and normal urothelial tissue (CTRL); n=29; derived from cystectomy specimens) was used for immunohistochemical analysis. The tissue microarray included three representative cores per patient to ensure representativeness of the samples for the tumor. The tissue samples from BCA patients were derived from transurethral resection (TURB), as well as cystectomy specimens, as available. Patients with NMIBC underwent secondary TURB in case of high grade or pT1 tumor to confirm staging. The tissue microarray was described earlier more in detail (13). The clinicopathological parameters of the study cohort are provided in Table I. The study was approved by the local ethics committee (ethic vote 289/08).
Immunohistochemistry. Immunohistochemical analysis of H3Ac, H4Ac and H3K18Ac was performed as described earlier (14): Paraffin sections (5 μm) were cut from the tissue microarray block, deparaffinized and rehydrated. Slides were placed in target retrieval solution (citrate buffer, pH 6.0) and heated for 10 min at boiling temperature. After cooling for 15 min, endogenous peroxidase activity was blocked (3% hydrogen peroxide, 10 min) and the sections were washed. After a 15-min protein block with normal swine serum, the primary antibodies against histone H3 pan-acetylation (H3Ac; dilution 1:2500; #06-599), histone H3 lysine 18 acetylation (H3K18Ac; 1:2500; #07-354) and histone H4 pan-acetylation (H4Ac; 1:2500; #06-866) were incubated overnight at 4°C. All rabbit polyclonal antibodies were purchased from Millipore (Lake Placid, NY, USA). Immunohistochemical staining was performed using the streptavidin-biotin-peroxidase complex technique (LSAB+; DAKO Cytomation, Glostrup, Denmark). The biotin-conjugated secondary antibody was incubated (30 minutes at room temperature) and the avidin biotin enzyme reagent alike. The peroxidase was developed with the automatic exposure control (AEC) system (DAKO). The sections were counterstained with hematoxylin and mounted. Negative (identical array sections without the primary antibody) and positive controls (liver, colon and breast cancer samples) were included on the array.
Immunostaining results were recorded semi-quantitatively. The number of urothelial cells showing nuclear staining was estimated per core (0, no positive cells; 1, 1-25% positive cells; 2, 26-50% positive cells; 3, 51-75% positive cells; and 4, 76-100% positive cells) and multiplied with an intensity scale (0, negative; 1, weak; 2, moderate; and 3, intensive staining). Mean staining for a patient was calculated; odd values were rounded to nearest whole integer. The immunohistochemical staining was evaluated without knowledge of the specimen identity.
Statistics. Clinicopathological parameters were correlated with global histone acetylation levels using the Mann-Whitney test and Kruskal-Wallis test, as appropriate. The Cox proportional hazard regression analysis was used to correlate the period of PFS and cancer-specific survival (CSS) following surgery with histone acetylation levels. Statistical analyses were performed using SPSS Statistics v21 (SPSS Inc., Chicago, IL, USA); significance was concluded at p<0.05.
Results
We first analyzed whether histone acetylation levels were different in BCA and CTRL tissue: H3Ac levels were significantly lower in BCA than CTRL (p<0.001), whereas H4Ac (p=0.635) and H3K18Ac (p=0.053) levels were similar. However, we observed that global acetylation levels were decreasing during progression from NMIBC to MIBC: H3Ac (p<0.001) and H3K18Ac (p=0.003) levels were also lower in MIBC than CTRL. H4Ac (p=0.448) and H3K18Ac (p=0.316) levels were similar in NMIBC and CTRL, while H3Ac (p<0.001) was significantly lower in NMIBC. All investigated histone modifications were significant lower in MIBC than NMIBC tissue (all p≤0.001; Table II). The distribution of global histone acetylation levels in bladder tissues levels is shown in Figure 1.
We next studied whether histone acetylation levels were correlated to clinicopathological parameters: neither staging nor grading was correlated to histone acetylation levels in patients with NMIBC. However, pT-stage was correlated inversely with histone acetylation in MIBC patients (all p≤0.001). In addition, lymph node metastasis (LNM) was correlated with low H3A (p<0.001) and H4Ac (p=0.001) levels. Grading and cM-stage (all p>0.05) were not correlated with histone acetylation levels; see Table III for details.
We also performed Cox proportional hazard models to analyze the predictive value of histone acetylation levels on patients' outcome. Follow-up information was available from all patients with NMIBC (median follow-up period 60.5 to 73.6 months) and MIBC (median follow-up period 27 months). We did not observe any significant correlation of histone acetylation levels and PFS or CSS in patients with NMIBC or MIBC (all p>0.1).
Discussion
The analysis of global histone modification levels has been recognized as potential diagnostic and prognostic tool in uro-oncology. Histone methylation levels are deregulated in BCA (8, 13) and may indicate poor prognosis (8). The analysis of H3K9Ac in a small cohort of patients with urothelial neoplasm of low malignant potential (PUNLMP) indicated high levels in patients without cancer recurrence (15). In the present work, we observed a decrease of global histone acetylation levels (H3Ac, H4Ac, H3K18Ac) in advanced BCA, namely MIBC, compared to NMIBC. In addition, H3Ac was also decreased in NMIBC compared to normal urothelial tissue. It should be noted that normal urothelial tissues in the present study were derived from cystectomy specimens from patients undergoing surgery for BCA; thus, the so-called field effect could also lead to epigenetic changes in the morphologically normal tissue.
H3 and H4 are deacetylated by histone deacetylases (HDACs). HDACs are overexpressed in various tumor entities, including BCA, whose overexpression is linked to poor prognosis: HDAC-1 and HDAC-2 are highly expressed in BCA, particularly in high-grade BCA (12). HDACs may be inhibited by specific HDAC inhibitors, e.g. vorinostat or trichostatin A, resulting in decreased global histone acetylation levels (16). Treatment of BCA cell lines with HDAC inhibitors inhibited cellular proliferation (17), tumor cell invasion (18), cell cycle progression (19), as well as suppression of cancer cell growth in vitro and in vivo (11, 20). Consequently, a clinical Phase I trial demonstrated some activity of the HDAC inhibitor suberanilohydroxamic acid (SAHA) in patients with metastatic BCA (21). Noteworthy, therapy response may be estimated by analysis of H3Ac (22) or H4Ac (23) quantitation of peripheral blood mononuclear cells, while H3Ac levels, in gastric cancer patients, before chemotherapy with vorinostat predicted treatment success (5).
Thus, we assume that determination of histone acetylation levels may help identify how patients will respond to treatment with HDAC inhibitors; further evaluation of these epigenetic drugs may be reasonable in BCA patients. Furthermore, as histone acetylation levels decrease during progression to MIBC, we hypothesize they could be helpful to identify patients with understaged pT1 tumors after TURB and, thus, identify those who need cystectomy. However, these hypotheses should be clarified in future studies.
- Received June 6, 2016.
- Revision received June 28, 2016.
- Accepted June 29, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved