An atlas of histone deacetylase expression in breast cancer: fluorescence methodology for comparative semi-quantitative analysis

Abstract

The histone deacetylase inhibitors, suberoylanilide hydroxamic acid (Vorinostat, Zolinza™) and depsipeptide (Romidepsin, Istodax™) have been approved by the US Food and Drug Administration for the treatment of refractory cutaneous T-cell lymphoma. Numerous histone deacetylase inhibitors are currently undergoing clinical trials, predominantly in combination with other cancer modalities, for the treatment of various haematological and solid malignancies. Most of the traditional compounds are known as broad-spectrum or pan-histone deacetylase inhibitors, possessing activity against a number of the 11 metal-dependent enzymes. One of the main questions in the field is whether class- or isoform-specific compounds would offer a therapeutic benefit compared to broad-spectrum inhibitors. Therefore, analysis of the relative expression of the different histone deacetylase enzymes in cancer cells and tissues is important to determine whether there are specific targets. We used a panel of antibodies directed against the 11 known mammalian histone deacetylases to determine expression levels in MCF7 breast cancer cells and in tissue representative of invasive ductal cell carcinoma and ductal carcinoma in situ. Firstly, we utilized a semi-quantitative method based on immunofluorescence staining to examine expression of the different histone deacetylases in MCF7 cells. Our findings indicate high expression levels of HDAC1, 3 and 6 in accordance with findings from others using RT-PCR and immunoblotting. Following validation of our approach we examined the expression of the different isoforms in representative control and breast cancer tissue. In general, our findings indicate higher expression of class I histone deacetylases compared to class II enzymes in breast cancer tissue. Analysis of individual cancer cells in the same tissue indicated marked heterogeneity in the expression of most class I enzymes indicating potential complications with the use of class- or isoform-specific compounds. Overall, our approach can be utilized to rapidly compare, in an unbiased semi-quantitative manner, the differential levels of expression of histone deacetylase enzymes in cells and tissues using widely available imaging software. It is anticipated that such analysis will become increasingly important as class- or isoform-specific histone deacetylase inhibitors become more readily available.

Keywords: Chromatin; breast cancer; histone acetylation; histone deacetylase inhibitor; immunofluorescence.

Figure 1

Classification of the classical histone deacetylases (HDACs). The classical HDACs are grouped into three classes based on their homology and sequence identity to yeast. They share a highly conserved deacetylase (DAC) domain shown as long orange cylinders. The class I HDACs (HDAC1, 2, 3 and 8) share homology to the yeast reduced potassium dependency-3 (Rpd-3) and are found primarily in the cell nucleus. Class II HDACs share sequence homology to the yeast histone deacetylase-1 (Hda1) and are sub-divided into two further classes; Class IIa (HDAC4, 5, 7 and 9) which shuttle between the cell cytoplasm and nucleus and Class IIb (HDAC6 and 10) which are found primarily in the cytoplasm. Class IV HDACs is comprised of HDAC11 which shares sequence homology to both Rpd-3 and Hda-1. Shown as small blue cylinders are nuclear localization sites and C- and N- terminal tail are shown as thick black lines. Total numbers of amino acid residues are shown at the N-terminal of each enzyme. SE14, Ser-Glu-containing repeats; ZnF, ubiquitin-binding zinc finger; Leucine-rich domain.

Figure 2

Histone deacetylase expression MCF7 and A431 cells. Immunofluorescence staining for relative expression of the HDAC enzymes in a representative field of (Ai) MCF7 and (Bi) A431 cells is shown. Cells were fixed and stained for rabbit polyclonal anti-HDAC1-11 (red) and goat polyclonal anti-β-actin (green) to highlight cell cytoskeleton. Merged image: nucleus stained with DAPI (blue), HDAC1-11 (red), and β-actin (green). Images were acquired using an Olympus BX61 fluorescence microscope automated with FVII Camera. Bar = 100μm; x20 magnification. (Aii, Bii) HDAC expression was analysed both by observation and given a score from 0-5 and semi-quantitatively by measuring fluorescence using Image J. Shown is the sum of the total fluorescence in the nucleus and cytoplasm measured by both methods. Relatively, strong staining was observed for HDAC1, 6 and 8 in MCF7 cells and HDAC4 and 7 in A431 cells. (C) Equal amounts of MCF7 and A431 cells (40 pg/lane) of whole-cell lysates were fractionated on SDS -PAGE gels, transferred to nitrocellulose membranes and immunoblotted for the indicated HDACs. GAPDH protein was blotted as a loading control.

Figure 3

Distribution epidermal growth factor (EGFr) and transferrin receptors. MCF7 and A431 cells were stained with monoclonal rabbit anti-phospho-EGFr (pT693); polyclonal rabbit anti-phosho-EGFr (s695) and monoclonal rabbit anti-CD71 (TFRC) antibodies. (A) Non specific background staining of EGF receptor phosphorylated on threonine 693 (pT693) was observed in MCF7 cell in contrast to (B) strong positive staining observed A431 cells. (C) Medium staining of EGF receptor phosphorylated on serine 695 (s695) was observed in MCF7 cells in contrast to (D) strong staining observed in A431 cells. Transferrin receptor (red) was found to be negative in both (E) MCF7 and (F) A431 cells. Merged image: nucleus stained with DAPI (blue), EGFr and CD71 (red), β-actin (green). Images were acquired using an Olympus BX61 fluorescence microscope with automated FVII camera. Bar = 200μm x20 magnification.

Figure 4

Histological representation of the morphological changes of breast tissue in cancer. Representative photomicrographs of haematoxylin and eosin stained sections of an active breast (a) and invasive ductal carcinoma (b). (i) Lobules of the breast arise from the lactiferous duct and through successive branching diminishing in size to form the terminal duct lobular unit (TDLU) (A). (ii) The TDLU is a continuation of the interlobular ducts (F) lumen (H) and end protruding into blunt or round saccules called ductules (G). Each TDLU is embedded in specialized hormonally responsive dense connective tissue called the intralobular stroma (C). The perimeter of the TDLU is surrounded by loose connective tissue (B) and is separated by dense interlobular connective tissue called stroma (D) which contains fibroblasts (K). Blood vessels (E) circulate throughout the intralobular stroma and interlobular stroma. (iii) During lactation, the ductules differentiate into the secretory units called acini, producing milk (L). Each ductules or acini are lined with cuboidal epithelium (I) and an outer layer of myoepithelium (J). (iv) Adipose tissue is found surrounding TDLUs, (M) highlighting were adipocytes are found. (b) Breast tissue with regions of both ductal carcinoma ‘in situ’ and invasive ductal carcinoma. (i) representative image of ductal carconima ‘in situ’ with small irregular clusters of calcifications of secretory material within the lumens of the duct. (ii) infiltrating ductal carcinoma cells originating from the ducts invade the stroma. (iii) A magnified region showing intraductal inflammation. (iv) Evidence of radial scarring shown bystallate lesions of trapped ductal cells in hyalinised stroma. Images were acquired using an Olympus FSX100 microscope. Bar = 500μm, 100μm, 50μm, 25μm; x4, x20, x40, x80 magnification, respectively.

Figure 5

Histone deacetylases expression in normal breast tissue. Immunofluorescence staining for relative expression of Class I HDACs (A), Class IIa HDACs (B), Class IIb HDACs (C) and Class IV HDACs (D) in two representative normal breast lobules. Paraffin embedded tissue sections were fixed and stained for rabbit polyclonal anti-HDAC1-11 (red) and goat polyclonal anti-β-actin (green) to highlight tissue structure. Weak staining was observed across Class I and Class IIb HDACs and strong staining was observed in class IIa HDACs. Strong non-specific staining of HDAC11 was seen in the connective tissue of the stroma with very weak staining seen in the actual cells. Show is an artefact for DAPI staining in the representative image of HDAC10 and for red channel in HDAC5. Merged image: nucleus stained with DAPI (blue), HDAC1-11 (red), and β-actin (green). Images were acquired using an Olympus BX61 fluorescence microscope automated with FVII Camera atx20 magnification. Bar = 200μm; x10 magnification.

Figure 6

Histone deacetylases HDAC expression in breast carcinoma. Immunofluorescence staining for relative expression of Class I HDACs (A), Class IIa HDACs (B), Class IIb HDACs (C) and Class IV HDACs (D) in two representative fields of breast carcinoma. Paraffin embedded tissue sections were fixed and stained for rabbit polyclonal anti-HDAC1-11 (red) and goat polyclonal anti-β-actin (green) to highlight tissue structure. Relatively, strong staining was observed across Class I HDACs and weak staining was observed in class IIa HDACs. Strong non-specific staining of HDAC10 and HDAC11 was observed in the connective tissue of the stroma with very weak staining seen in the cells. Merged image: nucleus stained with DAPI (blue), HDAC1-11 (red), and β-actin (green). Images were acquired using an Olympus BX61 fluorescence microscope automated with FVII Camera. Bar = 200μm; x10 magnification.

Figure 7

Overall expression scores of histone deacetylases in normal breast and invasive breast carcinoma. (A) Observational analysis of relative HDAC expression and distribution of the class I HDACs (i); class IIa HDACs (ii); and class IIb and IV HDACs (iii). Fluorescence intensity was examined and scored by three individuals on a scale of low to high (0-5) intensity for a minimum of five breast lobules. Shown is the global expression levels calculated as the sum of regional expression values for breast carcinoma compared to normal breast tissue. (i) Class I HDACs are significantly overexpressed in the breast cancer tissue relative to the normal breast in the cell cytoplasm and over expressed in breast cancer tissue in the cell nucleus for HDAC2, 3 and 8. (ii) Both HDAC5 and HADC9 are significantly decreased in the breast cancer tissue relative to the normal breast tissue in cell nucleus and cell cytoplasm. HDAC4 is significantly decreased in the cytoplasm and HDAC7 is decreased in the nucleus (iii) there are significant increases in HDAC 10 expressed in the cell cytoplasm of the cells found in the cancer breast tissue in comparison to the normal breast tissue. N = 7 in control breast and n=5 in breast cancer; *P<0.05, **P<0.01, ***P<0.001. (B) Total HDAC expression scores in control breast tissue in comparison to breast cancer tissue. Class I HDAC are overexpressed in the breast cancer and class IIa HDACs are under expressed. HDAC8 is most expressed in breast cancer tissue and HDAC4 is most expressed in normal breast tissue.

Figure 8

The relative distribution of histone deacetylase expression in ductal carcinoma cells in breast tissue. Breast carcinoma tissue stained with HDAC1-11 as described in the materials and methods. Immunofluorescence analysis was performed using Image J software to measure the mean fluorescence intensity of the expression of each HDAC the whole cell (A) or the nucleus (B) in 100 individual ductal carcinoma cells. Shown are box plots of the fluorescence intensity values for each HDAC. (A) Class I HDACs showed the greatest amount of variability in HDAC expression within cancer cells of breast tissue, with HDAC1 and HDAC3 showing the largest amount of heterogeneity. (B) In comparison greater variability of HDAC expression between cells was seen within the nucleus for the different HDACs.

Figure 9

Comparison of HDAC4 and 8 using two different antibodies. Im-munofluorescence images of control breast (A) and breast cancer (B) stained with poly-clonal rabbit anti-HDAC4 (Biovision); monoclonal rabbit anti-HDAC4 (Epitomics); polyclonal rabbit anti-HDAC8 (Biovision) and monoclonal mouse anti-HDAC8 (Sigma). Merged image: nucleus stained with DAPI (blue), HDAC4, 8 (red), β-actin (green). Images were acquired using an Olympus BX61 fluorescence microscope automated with FVII Camera atx20 magnification. Bar = 200μm; x10 magnification. Overall expression levels of HDAC4 (C) and HDAC8 (D) calculated as the sum of regional expression values for invasive breast carcinoma against control breast tissue.

Figure 10

Distribution of epidermal growth factor and transferrin receptors in control and breast cancer tissue. Representative tissues were stained with (A) monoclonal rabbit anti-phospho-EGFr (pT693); polyclonal rabbit anti-phosho-EGFr (s695) (B) and monoclonal rabbit anti-CD71 (TFRC). (A) Weak staining of both EGFr (pT693) and EGFr (s659) was observed in the control breast tissue in contrast to strong staining observed in the invasive breast carcinoma tissue. (C) Control breast expressing transferrin receptor (red) (i) showed relatively equal amounts of expression in breast cancer (ii). Merged image: nucleus stained with DAPI (blue), EGFr and CD71 (red), β-actin (green). Images were acquired using an Olympus BX61 fluorescence microscope with automated FVII camera. Bar = 200μm, 100μm; x10, x20 magnification respectively.

Figure 11

Invasive breast cancer has a higher predisposition to apoptosis in comparison to control breast cancer cells. Immunofluorescence staining for a membrane apoptotic marker, Annexin V (red) is shown in representative control and breast cancer tissue sections. Goat polyclonal anti-β-actin (green) was used to highlight the cytoskeleton. (A) Weak Annexin V staining was observed in the columnar epithelial cells of ductules in the control breast and (B) very weak staining was seen in epithelium of blood vessels in the stroma of the control breast. (C) In contrast, strong staining was seen in an invasive ductal carcinoma mass where high concentrations of malignant cells are found. (D) Strong staining of Annexin V in the columnar epithelial cells of a ductule in a tissue region of invasive ductal carcinoma. Merged image: nucleus stained with DAPI (blue), Annexin V (red), β-actin (green). Images were acquired using an Olympus BX61 fluorescence microscope with automated FVII camera. Bar = 200μm, 100μm; x10, x20 magnification respectively.