HDAC 1 and 6 modulate cell invasion and migration in clear cell renal cell carcinoma

Abstract

Background: Class I histone deacetylases (HDACs) have been reported to be overexpressed in clear cell renal cell carcinoma (ccRCC), whereas the expression of class II HDACs is unknown.

Methods: Four isogenic cell lines C2/C2VHL and 786-O/786-OVHL with differential VHL expression are used in our studies. Cobalt chloride is used to mimic hypoxia in vitro. HIF-2α knockdowns in C2 and 786-O cells is used to evaluate the effect on HDAC 1 expression and activity. Invasion and migration assays are used to investigate the role of HDAC 1 and HDAC 6 expression in ccRCC cells. Comparisons are made between experimental groups using the paired T-test, the two-sample Student's T-test or one-way ANOVA, as appropriate. ccRCC and the TCGA dataset are used to observe the clinical correlation between HDAC 1 and HDAC 6 overexpression and overall and progression free survival.

Results: Our analysis of tumor and matched non-tumor tissues from radical nephrectomies showed overexpression of class I and II HDACs (HDAC6 only in a subset of patients). In vitro, both HDAC1 and HDAC6 over-expression increased cell invasion and motility, respectively, in ccRCC cells. HDAC1 regulated invasiveness by increasing matrix metalloproteinase (MMP) expression. Furthermore, hypoxia stimulation in VHL-reconstituted cell lines increased HIF isoforms and HDAC1 expression. Presence of hypoxia response elements in the HDAC1 promoter along with chromatin immunoprecipitation data suggests that HIF-2α is a transcriptional regulator of HDAC1 gene. Conversely, HDAC6 and estrogen receptor alpha (ERα) were co-localized in cytoplasm of ccRCC cells and HDAC6 enhanced cell motility by decreasing acetylated α-tubulin expression, and this biological effect was attenuated by either biochemical or pharmacological inhibition. Finally, analysis of human ccRCC specimens revealed positive correlation between HIF isoforms and HDAC. HDAC1 mRNA upregulation was associated with worse overall survival in the TCGA dataset.

Conclusions: Taking together, these results suggest that HDAC1 and HDAC6 may play a role in ccRCC biology and could represent rational therapeutic targets.

Fig. 1

Class I and II HDACs are overexpressed in a subset of ccRCC tumors. Tissues from ccRCC tumors and adjacent non tumor tissues were homogenized and 50 μg of total protein was analyzed by Western blot for the expression of class I and II HDACs. a) Semi-quantitative analysis of HDAC bands were performed in Image J. Each tumor was normalized to GAPDH before normalizing it to their respective non-tumor tissue. b) Representative Western blots of two different matched tumor and non-tumor tissues show differential expression of class I HDACs. c) Semi-quantitative analysis of HDAC bands were performed in Image J. Each tumor was normalized to GAPDH before normalizing it to their respective non-tumor tissue. d) Representative Western blots of two different matched tumor and non-tumor tissues show differential expression of class II HDACs. The dotted line at 1 indicates the expression of HDAC 1 in adjacent non-tumor tissue. *p < 0.05 indicates statistically different HDAC 2 protein expression in tumor tissues as compared to the non-tumor tissue

Fig. 2

HDAC 1 and HDAC 6 increase cell invasion and migration in RCC cells, respectively. HDAC 1 was knocked down in 786–0 and C2 cells using retroviral supernatants. a) Three different clones with HDAC 1 knock down were generated and Sh2 was chosen in both cell lines for further experiments. b) BD biocoat matrigel chambers were used for measuring the invasion capacity of parental and HDAC 1 knocked down cells. C2 cells were incubated in the chamber for 24 h, whereas 786–0 cells were incubated for a short time point of 6 h. Cells at the bottom of the wells were visualized by crystal violet staining. c) Cells at the bottom of the well were counted blindly using a bright field microscope. *p < 0.05 indicates statistically different number of cells in shHDAC1 cells as compared to the parental cells. The error bars represent standard errors from biological triplicate experiments with technical replicates within each experiment. d) Representative images of renal tumor cell lines analyzed for HDAC 6 expression by immunofluorescence are shown. Scale bar indicates 50 μM distance and images are taken at 20X magnification. e) Representative images of scratch assays performed on C2, C2 overexpressing HDAC 6 and 786-O cells at time 0 and at the end of 24 h are shown

Fig. 3

HDAC 1 and HDAC 6 increase MMP-2/9 activity and decrease acetylated α-tubulin levels, respectively. a) C2 and 786–0 parental and knocked down HDAC 1 were analyzed for MMP activity by gelatin zymography assay. MMP activity was measured both in cell lysates as well as cell supernatants. b) Renal tumor cell lines were analyzed for HDAC 6 and acetylated α-tubulin expression by immunofluorescence. The staining in red indicates acetylated α-tubulin and the staining in green indicates HDAC 6 expression. Scale bar indicates 50 μM distance and images are taken at 20X magnification. c) Image J analysis measured immunofluorescence by calculating integrated density values of at least three representative fields per cell line. *p < 0.05 indicates statistically significant difference in acetylated α-tubulin levels as compared to C2 cells. The error bars represent standard errors from triplicate experiments and p-value was calculated using students t-test

Fig. 4

HDAC 1 expression is regulated by HIF in RCC cell lines. a) HDAC 1 expression was compared between C2 and C2VHL cells under normoxic and hypoxic conditions (mimicked by the use of 100 μM cobalt chloride) after overnight serum starvation. The numbers below the bands represent densitometry performed by Image J analysis on representative immunoblots relative to C2 bands in normoxic conditions with total histone H3 serving as loading control. b) HDAC 1 promoter region was analyzed for the presence of hypoxia response elements (HREs) by using the genomatix software. HREs represented by RCGTG and are highlighted in yellow and the black arrow represent the location of the forward and reverse primer. c-e) Chromatin immunoprecipitation assays were carried out in VHL null cell line C2 that expressed both HIF isoforms and 786–0 that only expressed HIF-2α. Pull downs with HIF isoforms were analyzed for HDAC 1 promoter/non-promoter region (1 kb upstream of the transcription start site) and the VEGF promoter (a known target of HIF) by qPCR. The error bars represent standard errors from biological duplicate experiments with technical replicates within each experiment. f) HIF-2α was knocked down in both models of renal cell line tumors by mammalian shRNA and analyzed for HDAC 1 expression and acetylated histone H3 by Western blot. The numbers below the bands represent densitometry performed by Image j analysis on representative immunoblots relative to parental cell lines in normoxic conditions with total GAPDH serving as loading control

Fig. 5

Acetylated α-tubulin is regulated by HDAC6/ER-α interaction in RCC cell lines. a) HDAC 6, ER-α and acetylated α-tubulin protein expression were measured by western blot analysis in ccRCC tumors and adjacent non-tumor tissue. b) A representative ccRCC tumor showed HDAC 6 (in red) and ER-α (in green) localization in the cytoplasm. c-d) Representative immunofluorescent images of acetylated α-tubulin (in red) and HDAC 6 (in green) expression in parental and ER-α knockdown cell lines are shown. e) Knockdown of ER-α in MCF-7 and renal cell tumors as measured by Western blot analysis. f) Image J analysis measured immunofluorescence by calculating integrated density values of at least three representative fields per cell line. *p < 0.05 and **p < 0.01 indicates statistically significant difference of acetylated α-tubulin levels in ER-α knockdown cells as compared to the parental cell lines. The error bars represent standard errors from triplicate experiments and p-value was calculated using students t-test. g) C2H6 cells treatment with 10 μM hydroxy tamoxifen and/or 50nM panobinostat for 4 h. Representative immunofluorescence images of acetylated α-tubulin (in red), HDAC 6 (in green) and ER-α (in green) are shown. h) Image J analysis measured immunofluorescence by calculating integrated density values of at least three representative fields per cell line. *p < 0.01 indicates statistically different HDAC 6 levels in treated cell lines as compared to control cells. The error bars represent standard errors from biological triplicate experiments with technical replicates for each experiment. Scale bar indicates 50 μM distance and images are taken at 20X magnification

Fig. 6

HDAC 1 and HIF positively correlate in clinical ccRCC. a) HDAC 1 expression in TMA was examined by immunohistochemistry. HDAC 1 positive nuclei were quantitated using Image J software and percent positive nuclei (indicated by brown staining) were used for further analysis. Black arrows indicate positive nuclei that were identified by the Image J software. The scale bar indicates 200 μM and the image was captured at 4X using the Aperio software. b) The association between HDAC1 and HIF1 & HIF2 statuses was examined using Wilcoxon rank sum and Kriskall-Wallis exact tests. All analyses were conducted in SAS v9.3 (Cary, NC) at a significance level of 0.05. c) Kaplan Meir curves (overall survival) of unaltered and upregulated HDAC 1 mRNA levels in the TCGA data are shown. The blue line indicates unaltered HDAC 1 mRNA and the red line indicates upregulated HDAC 1 mRNA levels. d) The table shows the numbers of patients in each group and survival data in months

Fig. 7

Schema of HDACs, hypoxia inducible factors and ER-α in clear cell renal cell carcinoma. Our studies show HDAC 1 can be upregulated by hypoxia inducible factors that is in turn stabilized by class II HDACs. Class II HDACs, specifically HDAC 6 can interact with ER-α. These interactions lead to increased motility and invasive capacity of ccRCC cell lines