HEXIM1 modulates vascular endothelial growth factor expression and function in breast epithelial cells and mammary gland

Abstract

Recently, we found that mutation of the C-terminus of transcription factor hexamethylene bisacetamide-inducible protein 1 (HEXIM1) in mice leads to abnormalities in cardiovascular development because of aberrant vascular endothelial growth factor (VEGF) expression. HEXIM1 regulation of some genes has also been shown to be positive transcription elongation factor b (P-TEFb) dependent. However, it is not known whether HEXIM1 regulates VEGF in the mammary gland. We demonstrate that HEXIM1 regulates estrogen-induced VEGF transcription through inhibition of estrogen receptor-alpha recruitment to the VEGF promoter in a P-TEFb-independent manner in MCF-7 cells. Under hypoxic conditions, HEXIM1 inhibits estrogen-induced hypoxia-inducible factor-1 alpha (HIF-1alpha) protein expression and recruitment of HIF-1alpha to the hypoxia-response element in the VEGF promoter. In the mouse mammary gland, increased HEXIM1 expression decreased estrogen-driven VEGF and HIF-1alpha expression. Conversely, a mutation in the C-terminus of HEXIM1 (HEXIM1(1-312)) led to increased VEGF and HIF-1alpha expression and vascularization in mammary glands of heterozygous HEXIM1(1-312) mice when compared with their wild-type littermates. In addition, HEXIM1(1-312) mice have a higher incidence of carcinogen-induced mammary tumors with increased vascularization, suggesting an inhibitory role for HEXIM1 during angiogenesis. Taken together, our data provide evidence to suggest a novel role for HEXIM1 in cancer progression.

Figure 1. Increased HEXIM1 expression inhibits E₂-induced transcription of VEGF via ERα in breast cancer cells

A. MCF-7 cells were transfected with pCMV-Tag2B-HEXIM1 (Flag-HEXIM1 expression vector) or control vector and treated with ethanol (vehicle) or 1 or 10 nM 17-beta estradiol (E₂) for 4 hours. Graph shows fold change of VEGF mRNA expression levels measured by reverse transcriptase PCR (RT-PCR). Data represents mean±s.e.m. from 4 independent experiments performed in duplicate; ‘*’ and ‘**’ indicate statistical significance (p < 0.05 and p < 0.005 respectively). B. MDA-MB-231 cells were transfected with Flag-HEXIM1 expression vector or control vector and treated with ethanol (vehicle) or 1 or 10 nM 17-beta estradiol (E₂) for 4 hours. Graph shows fold change of VEGF mRNA expression levels measured by reverse transcriptase PCR (RT-PCR). Data represents mean±s.e.m. from 3 independent experiments carried out in duplicate. C. MCF-7 cells were transfected with Flag-HEXIM1 expression vector or control vector and treated with ethanol or 10 nM E₂ for 12 hours. Secreted VEGF protein levels were measured by ELISA. Data represents mean±s.e.m. from 4 independent experiments assayed in duplicate; ‘*’ and ‘****’ indicate statistical significance (p < 0.05 and p < 0.00005 respectively). D. Primers used in ChIP assays are directed at regions indicated for VEGF promoter. E. MCF-7 cells were transfected with Flag-HEXIM1 expression vector or control vector and treated with ethanol or 100 nM E₂ for 45 minutes. Results show ChIP analyses of lysates immunoprecipitated with ERα, HEXIM1, Cyclin T1, RNA polymerase II (RNAP II) and rabbit immunoglobulin (IgG) antibodies. PCR amplification of the GC-rich/Sp1 proximal fragment in the VEGF promoter (Figure 1D) was performed and graph shows quantification of PCR products as indicated. Data represents mean±s.e.m. from 3 independent experiments.

Figure 2. Increased HEXIM1 expression inhibits E₂-induced VEGF mRNA expression under hypoxia that correlates with a decrease in E₂-induced HIF-1α protein expression

A. MCF-7 cells were transfected with Flag-HEXIM1 expression vector or control vector, treated with ethanol or 10 nM E₂ and grown under high oxygen (21% O₂) or low oxygen (1% O₂) conditions for 12 hours. Results show fold change of VEGF mRNA expression levels measured by RT-PCR. Data represents mean±s.e.m. from 3 independent experiments performed in duplicate; ‘*’ and ‘**’ indicate statistical significance (p < 0.05 and p < 0.005 respectively). B. MCF-7 cells were transfected with Flag-HEXIM1 expression vector or control vector, treated with ethanol or 10 nM E₂ and grown under 21% or 1% O₂ conditions as indicated for at least 12 hours. Western blot analyses show changes in HIF-1α protein expression and protein expression of HEXIM1 and GAPDH (loading control). Data represents mean±s.e.m. from 4 independent experiments performed in duplicate; ‘**’ and ‘***’represents statistical significance (p < 0.005 and 0.0005 respectively). C. MDA-MB-231 cells were transfected with Flag-HEXIM1 expression vector or control vector, treated with ethanol or 10 nM E₂ and grown under 21% or 1% O₂ conditions as indicated for at least 12 hours. Western blot analyses show changes in HIF-1α protein expression and protein expression of HEXIM1 and GAPDH (loading control). Data represents mean±s.e.m. from 3 independent experiments performed in duplicate.

Figure 3. Increased HEXIM1 expression inhibits E₂-induced recruitment of HIF-1α to VEGF Hypoxia Response Element

A. MCF-7 cells were treated with ethanol or 100 nM E₂ and subjected to high (21%) or low (0.5%) oxygen conditions as indicated or treated with 100 μM cobaltous chloride (CoCl₂) for 6 hours. Results show ChIP analyses of lysates immunoprecipitated with HIF-1α and rabbit immunoglobulin (IgG) antibodies and DNA fragments were analyzed by PCR primers specific for the hypoxic response element (HRE) in the VEGF promoter (region indicated in Figure 1D). Data represents mean±s.e.m. from 3 independent experiments; ‘**’ indicates statistical significance (p < 0.005). B. MCF-7 cells were transfected with Flag-HEXIM1 expression vector or control vector, treated with ethanol or 100 nM E₂ and subjected to 21% or 1% O₂ conditions as indicated for 16 hours. Results show ChIP analyses of HIF-1α and rabbit IgG immunoprecipitates with PCR amplification of fragment containing VEGF HRE. Graphs show quantification of HIF-1α immunoprecipitates and data represents mean±s.e.m. from at least 4 independent experiments; ‘*’ and ‘**’ indicates statistical significance (p < 0.05 and p < 0.005 respectively). C. MDA-MB-231 cells were transfected with Flag-HEXIM1 or control vector, treated with ethanol or 100 nM E₂ and subjected to 21% or 1% O₂ conditions as indicated for 16 hours. Results show ChIP analyses of HIF-1α and rabbit IgG immunoprecipitates with PCR amplification of fragment containing VEGF HRE. Graphs show quantification of HIF-1α immunoprecipitates and data represents mean±s.e.m. from at least 3 independent experiments.

Figure 4. HEXIM1 modulates VEGF and HIF-1α expression and vascularization in mouse mammary gland

A. MMTV/HEXIM1 mice were treated as described in Materials and Methods and mammary gland tissue extracts were subjected to Western blot. Antibodies for VEGF, HIF-1α, and HEXIM1 were used for immunoblotting. Anti-cytokeratin 18 was used as an epithelial cell marker and a loading control. Graph panel shows quantification of VEGF and HIF-1α expression from mice treated with or without doxycycline (DOX); Data represents mean±s.e.m. from 4 mice per group (−/+DOX). ‘*’ and ‘**’ indicate statistical significance (p < 0.05 and p < 0.005 respectively). B. Western blot analyses of mammary gland extracts from adult wild-type mice (WT) and mice heterozygous for the HEXIM1 1-312 mutant allele (HEXIM1 het) using VEGF and HIF-1α antibodies. Blots were probed for cytokeratin 18 to normalize for epithelial cell content. Graph shows quantification of VEGF expression from WT and HEXIM1 1-312 heterozygous mice. Data represents mean±s.e.m. from 8 mice per group (VEGF) and at least 3 mice per group (HIF-1α); ‘*’ indicates statistical significance (p < 0.05). C. Immunohistochemical detection of CD31 in mammary glands of lactating adult WT or HEXIM1 het mice. Panel is representative of 4 mice per group and graph shows quantification of %CD31-positive staining with mean±s.e.m.; ‘**’ indicates statistical significance (p < 0.005).

Figure 5. Expression of HEXIM1 C-terminus mutant enhances carcinogen-induced mammary tumorigenesis and correlates with increased vascularization of tumors

A. Mammary gland extracts from adult WT mice and mice homozygous for the HEXIM1 1-312 mutant allele (HEXIM1 1-312) were subjected to Western blot analyses using antibodies for VEGF, cyclin D1, serine 2 phosphorylated, and hypophosphorylated RNAP II. Blots were probed for cytokeratin 18 to normalize for epithelial cell content. Panel is representative of at least 3 mice per group. B. The graph describes DMBA-induced tumor incidence in HEXIM1 1-312 mice and their WT littermates assessed by palpitation and histopathological examination of excised tumors. DMBA was administered at 8 weeks of age by oral gavage. The frequency of palpable mammary tumors in HEXIM1 1-312 mice was statistically significant from that of the WT mice (p < 0.001); (n = 12 mice per group). C. Immunohistochemical detection of CD31 in DMBA-induced mammary tumors excised from adult WT or HEXIM1 1-312 mice. Panel is representative of at least 3 mice per group and graph shows quantification of %CD31-positive staining with mean±s.e.m.; ‘*’ indicates statistical significance (p < 0.05).

Figure 6

Model: HEXIM1 regulates VEGF expression via ERα and HIF-1α to modulate angiogenesis and tumorigenesis.