• Open Access

Global histone acetylation levels: Prognostic relevance in patients with renal cell carcinoma


To whom correspondence should be addressed.
E-mail: joerg.ellinger@ukb.uni-bonn.de


Epigenetic alterations play an important role in carcinogenesis. Recent studies have suggested that global histone modifications are predictors of cancer recurrence in various tumor entities. Global histone acetylation levels (histone H3 lysine 9 acetylation [H3K9Ac], histone H3 lysine 18 acetylation [H3K18Ac], total histone H3 acetylation [H3Ac] and total histone H4 acetylation [H4Ac]) were determined in patients with renal cell carcinoma (RCC) using immunohistochemistry in a tissue micro array with 193 RCC and 10 oncocytoma specimens. The histone acetylation pattern was not different among the diverse histological subtypes of RCC or oncocytoma samples. The H3Ac levels were inversely correlated with pT-stage (= 0.005), distant metastasis (= 0.036), Fuhrman grading (= 0.001) and RCC progression (= 0.029, hazard ratio 0.87). H4Ac deacetylation was correlated with pT-stage (= 0.011) and grading (= 0.029). H3K18Ac levels were an independent predictor of cancer-progression following surgery for localized RCC in the univariate (= 0.001, hazard ratio 0.78) and multivariate (= 0.005, hazard ratio 0.82) analysis. In conclusion, our study supports the concept of global histone modification levels as a universal cancer prognosis marker, and provides evidence for the use of histone deacetylases inhibitors as future drugs in the therapy of RCC. (Cancer Sci 2010; 101: 2664–2669)

In 2009, 57 760 new cases and 12 290 deaths of renal cancer are estimated in the United States.(1) Thus, renal cancer is among the 10 most frequent cancer entities. A recent population-based study in the Netherlands failed to show an improvement of survival in renal cancer patients during the past decade.(2) Renal cell carcinoma (RCC) is a heterogeneous malignancy that includes different histological subtypes (∼85% clear cell RCC [ccRCC]; ∼10% papillary RCC [pRCC]; ∼3% chromophobe RCC [chRCC]; ∼1% sarcomatoid RCC [sRCC]).(3) The most important predictors of a patient’s outcome are TNM stage, Fuhrman grade and the Eastern Cooperative Oncology Group performance status.(3) However, RCC patients with similar tumor characteristics still show heterogeneity in the course and outcome of the disease. So far, evidence for adjuvant treatment of high-risk RCC patients is not available. However, clinical trials with novel agents are ongoing, and future therapy concepts may change. Thus, a subclassification of patients with RCC is desirable to identify patients who would benefit from adjuvant therapy concepts.

Epigenetic alterations including DNA methylation and histone modifications are common in cancer and lead to silencing of tumor suppressor genes. DNA methylation has been widely studied, whereas the knowledge of histone modifications in RCC remains limited. Histone modifications (e.g. acetylation, methylation) are important regulators of transcriptional activity, and occur on the N-terminal tail of the histone.(4) Histone lysine acetylation functions in two manners: (i) the chromatin structure is directly altered; and (ii) it also acts as a molecular tag for the recruitment of chromatin-modifying complexes.(5) Importantly, histone modifications are dynamic and multiple enzymes that add or remove acetylation marks have been identified,(6) and thus, histone acetylation could represent a target for future drug therapy. Indeed, histone deacetylases (HDAC) are overexpressed in RCC,(7) and the HDAC inhibitor valproic acid inhibited RCC growth in vitro and in vivo.(8)

In 2005, Seligson et al.(9) reported that decreased global histone modification levels (H3K4me2 or H3K18Ac) were predictive for prostate cancer recurrence. Later on, other researchers demonstrated a role for global histone deacetylation as a prognostic marker in various tumor entities including breast cancer (H3K9Ac, H3K16Ac, H3K18Ac)(10) and lung cancer (H3K9Ac).(11) Histone modifications are also playing a role in RCC: two smaller studies reported lower levels of H3Ac(12) and H3K9Ac(13) in high-grade RCC. A poor 1-year survival was observed in patients with localized RCC and low levels of H3K4me2 or H3K18Ac.(14) Furthermore, H3K4 methylation was a predictor of recurrence-free survival in RCC patients.(15) Altogether, histone acetylation seems to be a potential prognostic marker in RCC, but further systematic analysis of histone acetylation marks is necessary to rate the value of each modification. Furthermore, the relevance of histone acetylation in non-ccRCC remains uncertain as earlier studies focused on the most common ccRCC entity. We therefore determined total histone H3 lysine acetylation (H3Ac), histone H3 lysine 9 acetylation (H3K9Ac), histone H3 lysine 18 acetylation (H3K18Ac) and total H4 lysine acetylation (H4Ac) in 193 RCC specimens including non-ccRCC entities, and correlated histone acetylation with clinical–pathological parameters and patient outcome.

Materials and Methods

Patients and tissue microarray.  A tissue microarray with 193 RCC samples (ccRCC, n = 142; pRCC, n = 31 [pRCC type I, n = 18; pRCC type II, n = 13]; chRCC, n = 10; sRCC, n = 10) and 10 oncocytoma samples was prepared from formalin-fixed, paraffin-embedded tissue specimens (see Table 1 for the clinical–pathological parameters). Three representative tissue cores per patient were arrayed using a manual device (Beecher Instruments, Sun Prairie, WI, USA) to obtain a representative image of the tumor. Samples were selected from the archival files of the Department of Pathology at the Universitätsklinikum Bonn according to tissue availability and were not stratified for any known preoperative or prognostic factor. The specimens were derived from patients undergoing nephrectomy or nephron-sparing surgery at the departments of urology at the Universitätsklinikum Bonn and the Waldkrankenhaus Bad Godesberg (both in Bonn, Germany) between 1994 and 2008. All cases were re-evaluated by a panel of pathologists for histopathological staging and Fuhrman grading. Follow-up information was available for 127 patients of whom 38 suffered from disease progression; 16 died from RCC. The mean follow-up period was 62 months (median 48 months; range 1–169 months). The study was approved by the local ethics committee.

Table 1.   Clinical–pathological parameters in patients with renal cell carcinoma and oncocytoma
 RCC (n = 193) (%)ccRCC (n = 142) (%)pRCC (n = 31) (%)chRCC (n = 10) (%)sRCC (n = 10) (%)Onco-cytoma (n = 10) (%)
  1. ccRCC, clear cell renal cell carcinoma; chRCC, chromophobe renal cell carcinoma; NA, not applicable; pRCC, papillary renal cell carcinoma; RCC, renal cell carcinoma; sRCC, sarcomatoid renal cell carcinoma.

 Range26–85  26–85  35–82  27–85  51–75  26–73
 Male130 (67.4)89 (62.7)28 (90.3)6 (60.0)7 (70.0)     0 (0.0)
 Female63 (32.6)53 (37.3)3 (9.7)4 (40.0)3 (30.0)    10 (100)
TNM staging
 pT1a52 (26.934 (23.9)16 (51.6)2 (20.0)0 (0)NA
 pT1b33 (17.1)25 (17.6)4 (12.9)4 (40.0)0 (0)NA
 pT243 (22.332 (22.5)6 (19.4)4 (40.0)1 (10.0)NA
 pT362 (32.1)49 (34.5)5 (16.1)0 (0)8 (80.0)NA
 pT43 (1.6)2 (1.4)0 (0)0 (0)1 (10.0)NA
 LN metastasis14 (7.3)8 (5.6)1 (3.2)0 (0)5 (50.0)NA
 Distant metastasis27 (14.0)18 (12.7)3 (9.7)0 (0)6 (60.0)NA
Fuhrman grading
 Grade 159 (30.6)44 (31.0)12 (38.7)3 (30.0)0 (0)NA
 Grade 2119 (61.7)94 (66.2)17 (54.8)7 (70.0)1 (10.0)NA
 Grade 312 (6.2)3 (2.1)2 (6.5)0 (0)7 (70.0)NA
 Grade 42 (1.0)0 (0)0 (0)0 (0)2 (20.0)NA

Immunohistochemistry.  Immunohistochemical staining of H3Ac (#06-599; rabbit polyclonal antibody; Millipore, Lake Placid, NY, USA) and H4Ac (#06-866; rabbit polyclonal antibody; Millipore) was described by our group previously,(16) and the staining procedure was adopted for H3K9Ac (#pAb-ACHAHS-044; rabbit polyclonal antibody; Diagenode, Liège, Belgium) and H3K18Ac (#07-354; rabbit polyclonal antibody; Millipore). In brief, paraffin sections (5 μm) were freshly cut from the tissue microarray block. Deparaffinization was done using xylene and the sections were rehydrated with graded ethanol. Slides were placed in target retrieval solution (citrate buffer, pH 6.0) and heated for 10 min at boiling temperature using a microwave. After cooling for 15 min, endogenous peroxidase activity was blocked by treatment with 3% hydrogen peroxide for 10 min. The sections were washed with Tris-buffered saline. After a 15 min protein block with normal swine serum, the primary antibodies H3Ac (dilution: 1:1000), H4Ac (1:2500), H3K9Ac (1:200) and H3K18Ac (1:2500) were applied overnight at 4°C. Immunohistochemical staining was performed using the streptavidin–biotin–peroxidase complex technique (LSAB+; DAKO Cytomation, Glostrup, Denmark). The biotin-conjugated secondary antibody was incubated for 30 min at room temperature and the avidin biotin enzyme reagent the same. The peroxidase was developed with the AEC system (DAKO). The sections were counterstained with hematoxylin and mounted. Negative control sections were identical array sections without the primary antibody.

Immunostaining results were evaluated using the Remmele scoring system.(17) The number of epithelial cells showing nuclear staining was estimated per core and scaled: 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. These scores were multiplied with an intensity scale (0, negative; 1, weak; 2, moderate; and 3, intensive staining), and the mean staining for a patient was calculated; odd values were rounded up/down. Evaluation of immunohistochemistry was performed without knowledge of the specimen identity.

Statistical analyses.  Clinical–pathological parameters were correlated with the global histone acetylation levels using the Spearman-Rho test, the Mann–Whitney test and the Kruskal–Wallis test as appropriate. The Cox proportional hazard regression analysis was used to correlate the period of progression-free survival and cancer-specific survival following surgery with global histone acetylation levels/clinical–pathological parameters. Statistical analyses were performed using PASW Statistics 17.0 (SPSS Inc., Chicago, IL, USA); significance was concluded at P < 0.05.


Histone acetylation levels are similar in different histological subtypes of RCC.  Antibodies against H3Ac, H3K9Ac, H3K18Ac and H4Ac showed nuclear staining (Fig. 1). The staining of the different histone marks was different: H4Ac, H3Ac and H3K18Ac levels were predominantly high, whereas H3K9Ac levels were lower. Histone acetylation staining scores were significantly correlated with each other (all P < 0.001; r2 = 0.259–0.709). Global levels of all evaluated histone acetylation marks were similar in the different histological RCC and oncocytoma samples (Fig. 2; P > 0.05). We detected lower H4Ac levels in type II pRCC than in type I pRCC (P = 0.025). Global acetylation levels were not correlated with age or sex (P > 0.05). We also compared histone acetylation with histone H3 lysine 4 methylation (H3K4me), which was reported in an earlier study:(15) all acetylation marks were significantly correlated with H3K4me1, H3K4me2 and H3K4me3 (all P < 0.001). Interestingly, the correlation coefficient was higher for H3K4me versus H3Ac/H4Ac (r2 = 0.698–0.788) than for H3K4me versus H3K9Ac/H3K18Ac (r2 = 0.268–0.409) (Table 2).

Figure 1.

 Immunohistochemical staining of (A) his-tone H3 lysine 9 acetylation (H3K9Ac) (Remmele Score 12), (B) histone H3 lysine 18 acetylation (H3K18Ac) (Remmele Score 12), (C) total histone H3 acetylation (H3Ac) (Remmele Score 8) and (D) total histone H4 acetylation (H4Ac) (Remmele Score 12). Original magnification, ×400.

Figure 2.

 Global histone acetylation levels are similar within the different histological subtypes of renal cell carcinoma and oncocytoma: (A) total histone H3 acetylation (H3Ac); (B) total histone H4 acetylation (H4Ac); (C) histone H3 lysine 9 acetylation (H3K9Ac); (D) histone H3 lysine 18 acetylation (H3K18Ac). The boxplots indicate the distribution of histone acetylation levels in each histological entity. The median is indicated by the bold line. The box represents the middle 50% of the data sample, and the remaining 50% is indicated by the whiskers. Outliers and extremes are indicated by circles/asterisks. ccRCC, clear cell renal cell carcinoma; chRCC, chromophobe renal cell carcinoma; pRCC, papillary renal cell carcinoma; sRCC, sarcomatoid renal cell carcinoma.

Table 2.   Spearman’s rank correlation coefficients for histone acetylation and histone H3K4 methylation
  1. All P-values were <0.0001. H3Ac, total histone H3 acetylation; H4Ac, total histone H4 acetylation; H3K9Ac, histone H3 lysine 9 acetylation; H3K18Ac, histone H3 lysine 18 acetylation.


Histone deacetylation is associated with poor prognosis.  Global histone H3 deacetylation was correlated with pT-stage (P = 0.005), Fuhrman grading (P = 0.001) and the presence of distant metastasis (P = 0.036). H4Ac was inversely correlated with pT-stage (P = 0.011) and grading (P = 0.029). H3K18Ac and H3K9Ac were not correlated with any clinical–pathological parameter (Table 3).

Table 3.   Correlation of histone acetylation with clinical–pathological parameters
  1. †Kruskal–Wallis test. ‡Mann–Whitney test. H3Ac, total histone H3 acetylation; H4Ac, total histone H4 acetylation; H3K9Ac, histone H3 lysine 9 acetylation; H3K18Ac, histone H3 lysine 18 acetylation.

Lymph node metastasis‡0.1440.3480.2290.878
Distant metastasis‡0.3460.3860.0360.554
Fuhrman grading†0.0510.4020.0010.029

We next determined whether global histone acetylation was correlated with progression-free survival in patients with localized RCC (n = 161). Low levels of H3K18Ac were a significant predictor of RCC progression (P = 0.001, hazard ratio 0.78). The multivariate model, which also included pT-stage and Fuhrman grading, demonstrated that low H3K18Ac levels were independently predictive of RCC progression (P = 0.003, hazard ratio 0.81). However, none of the acetylation markers were significantly correlated with cancer-specific survival; it should be noted that the established parameters pT-stage and grading also failed to reach significance. Interestingly, the H3K4me-Score, which was predictive for cancer recurrence in the earlier study on H3K4 methylation,(15) also failed to reach significance in the multivariate model (Fig. 3 and Table 4).

Figure 3.

 Kaplan–Meier estimates for global histone H3 lysine 18 acetylation (H3K18Ac) levels. H3K18 deacetylation is correlated with a shorter period of progression-free survival in patients with localized renal cell carcinoma. H3K18Ac scores were categorized into low (Remmele Score ≤3), moderate (4–7) and high (≥8) for purposes of clarity (log rank P = 0.0078).

Table 4.   Prediction of progression-free survival and cancer-specific survival in patients with localized renal cell carcinoma
 Progression-free survivalCancer-specific survival
Univariate analysisMultivariate model IUnivariate analysis
PHR (95% CI)PHR (95% CI)PHR (95% CI)
  1. CI, confidence interval; H3Ac, total histone H3 acetylation; H4Ac, total histone H4 acetylation; H3K9Ac, histone H3 lysine 9 acetylation; H3K18Ac, histone H3 lysine 18 acetylation; HR, hazard ratio.

pT-stage0.0131.65 (1.11–2.46)0.0111.70 (1.13–2.56)0.8441.07 (0.51–2.24)
Fuhrman grading0.0143.24 (1.27–8.26)0.0862.39 (0.88–6.48)0.0806.76 (0.79–57.40)
H3K9Ac0.6041.04 (0.88–1.23)  0.2311.18 (0.89–1.56)
H3K18Ac0.0010.78 (0.68–0.90)0.0050.82 (0.72–0.94)0.7950.96 (0.71–1.28)
H3Ac0.0540.86 (0.74–1.00)  0.2140.83 (0.62–1.11)
H4Ac0.0830.86 (0.74–1.01)  0.6870.93 (0.68–1.28)
H3K4me-Score0.0211.57 (1.07–2.30)0.3541.21 (0.81–1.80)0.4671.34 (0.61–2.97)

We also determined the prognostic relevance for the whole cancer cohort including those patients with lymph node and/or distant metastasis. Low H3Ac (P = 0.029) and H3K18Ac (P = 0.002) levels were significantly correlated with RCC progression in the univariate analysis. However, both markers (P > 0.28) failed to reach significance in the multivariate Cox model, which also included pT-stage, lymph node/distant metastasis and grading. However, the H3K4me-Score was correlated with cancer progression. H4Ac, H3Ac, H3K9Ac and H3K18Ac were not correlated (P > 0.05) with cancer-specific survival (Table 5).

Table 5.   Prediction of progression-free survival and cancer-specific survival in patients with renal cell carcinoma
 Progression-free survivalCancer-specific survival
Univariate analysisMultivariate modelUnivariate analysisMultivariate model
PHR (95% CI)PHR (95% CI)PHR (95% CI)PHR (95% CI)
  1. CI, confidence interval; H3Ac, total histone H3 acetylation; H4Ac, total histone H4 acetylation; H3K9Ac, histone H3 lysine 9 acetylation; H3K18Ac, histone H3 lysine 18 acetylation; HR, hazard ratio.

pT-stage0.0011.80 (1.27–2.55)0.0551.45 (0.99–2.13)0.0072.45 (1.28–4.68)0.6121.20 (0.59–2.43)
Lymph node metastasis<0.0016.27 (2.35–16.72)0.4951.49 (0.47–4.72)<0.00123.99 (7.85–73.28)0.0145.80 (1.43–23.50)
Distant metastasis<0.0017.26 (2.48–15.15)0.0133.26 (1.28–8.31)<0.00124.12 (7.45–78.04)0.0087.86 (1.72–35.99)
Fuhrman grading<0.0013.27 (1.69–6.32)0.0403.26 (1.28–8.31)<0.00111.20 (4.10–30.59)0.0253.20 (1.15–8.83)
H3Ac0.0290.87 (0.77–0.98)0.3181.13 (0.84–1.06)0.0510.83 (0.69–1.00)  
H4Ac0.2080.91 (0.80–1.04)  0.5550.94 (0.77–1.14)  
H3K9Ac0.1731.09 (0.96–1.24)  0.3861.09 (0.89–1.33)  
H3K18Ac0.0020.84 (0.76–0.94)0.3160.94 (0.84–1.06)0.0770.86 (0.73–1.01)  
H3K4me-Score0.0041.58 (1.15–2.15)0.0491.59 (1.01–2.53)0.1261.46 (0.90–2.36)  


Our study shows that global H3Ac and H4Ac levels were correlated with adverse clinical–pathological parameters (pT-stage, grade, presence of metastasis). Furthermore, global H3K18Ac levels were a significant and independent predictor of RCC recurrence following nephron-sparing surgery or radical nephrectomy for localized RCC. This finding is in good agreement with the results of former studies on histone acetylation;(12,13,14) the largest study (n = 359) on histone modifications in RCC patients was published by Seligson et al.(14) They reported a poorer probability of 1-year survival in patients with localized RCC (n = 189) if global levels of H3K4me2 (<60%) or H3K18Ac (<35%) were low. However, histone modifications were not predictive for survival in patients with metastatic disease.(14) The smaller series published by Minardi et al. (H3K9Ac; n = 54) and Kanao et al. (H3Ac; n = 54) also demonstrated significantly decreased acetylation levels in high-grade RCC.(12,13) We therefore assume that histone acetylation levels are potential markers for the risk stratification of patients with localized RCC, and they help to identify those who could benefit from adjuvant therapy concepts.

Earlier, we also studied histone H3K4 methylation in patients with RCC.(15) The staining scores for histone methylation and acetylation were significantly correlated, but the degree of correlation was different: H3K9Ac and H3K18Ac had lower correlation coefficients with H3K4 methylation than H3Ac/H4Ac. However, the correlation of histone acetylation and methylation suggests epigenetic confusion in RCC cells may be due to aberrant activity and/or expression of histone modifying enzymes. Indeed, several histone demethylases are direct HIF-1 targets (i.e. JARID1B, JMJD1A, JMJD2B, JMJD2C)(18) and are transactivated by this key player of RCC carcinogenesis. Despite the correlation of histone methylation and acetylation, these modifications may have different relevance depending on the stage of RCC carcinogenesis. We herein demonstrate that the H3K18Ac mark is the strongest predictor of recurrence for patients with localized RCC, but not for patients with metastatic disease. In comparison, the H3K4me-Score (a combination of all mono-, di- or tri-methylation of H3K4) is an independent predictor of RCC progression also in patients with metastatic RCC.(15) Thus, H3K18Ac is an early step in RCC carcinogenesis, whereas H3K4 methylation occurs later. The analysis of H3K18Ac may help to identify patients who profit from adjuvant therapeutic strategies.

Our study also provides valuable information about histone acetylation among non-ccRCC subtypes. Previous studies were either limited to a few patients with non-RCC subtypes (ccRCC, n = 32; pRCC, n = 7; chRCC, n = 5),(12) included only the ccRCC subtype(13) or did not report details on histological RCC subtypes.(14) Our study shows that global histone acetylation levels are similar among different RCC subtypes; it should be kept in mind that the number of pRCC (n = 31), chRCC and sRCC (both n = 10) was low compared to ccRCC (n = 142), thereby inhibiting a robust statistical analysis. Currently used targeted therapeutics (i.e. sunitinib, sorafenib and temsirolimus) seem to be less effective in patients with non-ccRCC.(19) Thus, histone acetylation may represent a novel therapeutic target for future medical therapy of all RCC subtypes; HDAC are highly expressed in RCC.(7) Several studies have highlighted the antineoplastic effects of HDAC inhibitors in RCC: depsipeptide induced apoptosis and increased global H3Ac levels in RCC cell lines.(12) Valproic acid inhibited RCC growth in vitro and in vivo.(8) Vorinostat (suberoylanilide hydroxamic acid), already FDA-approved for the treatment of cutaneous T-cell leukemia, increased the anticancer activity of temsirolimus in vitro and in vivo.(20)

Recent studies have also investigated whether RCC and normal tissue display a different histone modification pattern. Minardi et al.(13) observed a decreased H3K9Ac staining intensity as well as a decreased percentage of positive nuclei in ccRCC tissue compared with normal tissue. Kanao et al.(12) also reported increased H3Ac levels in healthy tissue, and furthermore specified that H3Ac expression was mainly detectable in the proximal tubule and collecting duct of healthy tissue. These findings indicate that histone modification levels already start to decrease at an early phase of carcinogenesis, and that further reduction of histone acetylation contributes to RCC progression.

Earlier studies demonstrated that reduced global levels of histone modifications (i.e. H3K9me2 in prostate cancer cell lines,(14) or H4K20me3 and H4K16Ac in leukemia cells(21)) were mainly attributable to loss at repetitive DNA sequences (e.g. subtelomeric repetitive element D4Z4; acrocentric chromosomal region NBL2; juxtacentromeric satellite element Sat2), and not to loss at specific tumor suppressor genes or oncogenes; DNA hypomethylation in repetitive regions contributes to chromosomal instability.(22) With the knowledge of the closed interplay of DNA methylation and histone modifications, we assume that loss of histone acetylation may also cause genomic instability. Genomic instability is common in RCC,(23) and was also related to aggressive RCC.(24)

In conclusion, global histone acetylation levels are decreased in patients with RCC. In particular, H3K18Ac levels indicate a poor prognosis in patients with localized RCC. Inhibition of histone deacetylases is a potential therapeutic target.


The authors thank Doris Schmidt for technical assistance. Davit Mosashvilli was supported by a research grant from the German Academic Exchange Services.

Disclosure Statement

The authors have no conflict of interest.


clear cell renal cell carcinoma


confidence interval


total histone H3 acetylation


histone H3 lysine 18 acetylation


histone H3 lysine 9 acetylation


total histone H4 acetylation


hazard ratio


not applicable


papillary renal cell carcinoma


renal cell carcinoma


sarcomatoid renal cell carcinoma