Abnormal histone methylation is responsible for increased vascular endothelial growth factor 165a secretion from airway smooth muscle cells in asthma

Abstract

Vascular endothelial growth factor (VEGF), a key angiogenic molecule, is aberrantly expressed in several diseases including asthma where it contributes to bronchial vascular remodeling and chronic inflammation. Asthmatic human airway smooth muscle cells hypersecrete VEGF, but the mechanism is unclear. In this study, we defined the mechanism in human airway smooth muscle cells from nonasthmatic and asthmatic patients. We found that asthmatic cells lacked a repression complex at the VEGF promoter, which was present in nonasthmatic cells. Recruitment of G9A, trimethylation of histone H3 at lysine 9 (H3K9me3), and a resultant decrease in RNA polymerase II at the VEGF promoter was critical to repression of VEGF secretion in nonasthmatic cells. At the asthmatic promoter, H3K9me3 was absent because of failed recruitment of G9a; RNA polymerase II binding, in association with TATA-binding protein-associated factor 1, was increased; H3K4me3 was present; and Sp1 binding was exaggerated and sustained. In contrast, DNA methylation and histone acetylation were similar in asthmatic and nonasthmatic cells. This is the first study, to our knowledge, to show that airway cells in asthma have altered epigenetic regulation of remodeling gene(s). Histone methylation at genes such as VEGF may be an important new therapeutic target.

Fig. 1

Increased asthmatic HASM cell VEGF secretion is transcription dependent. (A) Supernatants were collected 24 hours after serum deprivation and VEGF protein secretion assessed by ELISA. N= 3 NA and 3 A HASM cell lines. *, P <0.05 compared to corresponding NA result. (B) Cells were lysed and mRNA levels were assessed by quantitative real time PCR. N= 3 NA and 4 A HASM cell lines. *, P <0.05 compared to corresponding NA result. (C) HASM cells were cultured for 4 hours post serum deprivation in the presence of 5μg/ml Actinomycin D or DMSO vehicle control. Cells were lysed and VEGF mRNA levels assessed by quantitative real time PCR. N= 3 NA and 3 A HASM cell lines. *, P <0.05 compared to DMSO control. PCR data are expressed as fold change relative to the mean control NA value.

Fig. 2

VEGF165a is specifically upregulated in asthmatic HASM cell. (A and B) Schematic of different VEGF isoforms. (A) Conventional pro-angiogenic isoforms; (B) More recently identified anti-angiogenic isoforms. (C) Isoform specific, quantitative real time PCR was performed to assess splice variant expression, 4 hours post serum starvation. N = 3 NA and 3 A HASM cell lines. *, P <0.05 compared to corresponding NA result. PCR data are expressed as fold change relative to the mean control NA value.

Fig. 3

Basal VEGF secretion is Sp1 dependent. (A) VEGF protein levels following incubation with Mithramycin A. N = 4 NA and 3 A HASM cell lines. *, P <0.05 compared to A DMSO control. ++, P <0.01 compared to NA DMSO control. (B) VEGF protein levels after transfection with empty vector, pEBGV or D/N Sp1. N = 3 NA and 3 A HASM cell lines. *, P <0.05 compared to A DMSO control. +++, P <0.005 compared to NA DMSO control. Sp1 binding to the (C) proximal and (D) distal VEGF promoter was assessed by ChIP. TAF1 binding to the (E) proximal and (F) distal VEGF promoter was assessed by ChIP. Sp1/TAF1 binding were measured in confluent, serum deprived NA and A HASM cells 0 and 2 hours after 24 hours serum deprivation. IgG negative controls are shown. Data is expressed as fold change relative to the mean 0hr NA value. N = 3NA and 2A HASM cell lines. (G) VEGF protein levels after incubation with COX inhibitors, 1μM Indomethacin and 1μM NS398. N= 4 NA and 3 A HASM cell lines. (H) Nuclear Sp1 expression levels were assessed by western blotting. Three NA and three A cell nuclear lysates are shown.

Fig. 4

Methylation status of the VEGF gene. (A) Schematic diagram of the VEGF gene showing exons 1-8 (top) and the position of CpG sites and CpG islands (bottom). Exon positions relative to CpG sites are approximated. (B) Preliminary screen of the methylation status of CpG sites within the three VEGF CpG islands. (C) Comparison of NA and A CpG islands 2 and 3 methylation status. (D) Methylation status of NA VEGF CpG island 1 over time. ● = methylated cytosine, ○ = unmethylated cytosine, x = non-sequenced site. For CpG 1.3 only = site methylated in at least one clone. (E) VEGF protein levels after incubation with 5-aza. N = 4 NA and 3 A HASM cell lines.

Fig. 5

Histone Acetylation is not in different between the asthmatic and non-asthmatic VEGF promoter. Acetylated Histone H3 (AcH3) association with the (A) proximal and (B) distal VEGF promoter was assessed by ChIP. Acetylated Histone H4 (AcH4) association with the (A) proximal and (B) distal VEGF promoter was assessed by ChIP. AcH3/4 association were measured in confluent, serum deprived NA and A HASM cells 0 and 2 hours after 24 hours serum deprivation. IgG negative controls are shown. N = 3NA and 3 A HASM cell lines. (E) VEGF protein levels after incubation with the histone deactylase (HDAC) inhibitor SAHA. N= 3NA and 3A HASM cells lines. (F) VEGF protein levels after incubation with the histone deactylase (HDAC) inhibitor MS275. N= 6NA and 6A HASM cells lines.

Fig. 6

Histone methylation is altered at the A VEGF promoter. Tri-methylated lysine 4 on Histone H3 association with the (A) proximal and (B) distal VEGF promoter was assessed by ChIP. H3K4me3 association was measured in confluent, serum deprived NA and A HASM cells immediately following 24 hours serum deprivation. IgG negative controls are shown. Data is expressed as fold change relative to the mean 0hr NA value. Tri-methylated lysine 9 on Histone H3 association with the (C) proximal and (D) distal VEGF promoter was assessed by ChIP. H3K9me3 association was measured in confluent, serum deprived NA and A HASM cells 0 and 2 hours after 24 hours serum deprivation. IgG negative controls are shown. Data is expressed as fold change relative to the mean 0hr NA value. N= 3NA and 3A HASM cell lines.

Fig. 7

H3K9me3 regulates VEGF expression and VEGF promoter Sp1 binding. (A) VEGF protein levels after incubation with BIX-01294. N = 3 NA and 4 A HASM cell lines. (B) VEGF protein levels after transfection of empty vector, pcDNA3.1 and D/N G9A. N = 3 NA and 3 A HASM cell lines. (C) VEGF protein levels after transfection with empty vector CMV and D/N ΔSET SUV39H1. N = 3 NA and 3 A HASM cell lines. Sp1 association with the (D) proximal and (E) distal VEGF promoter, in the presence of BIX-01294 was assessed by ChIP. Sp1 association was measured in confluent, serum deprived NA HASM cells incubated with BIX-01294, at 2 hours post 24 hour serum deprivation. IgG negative controls are shown. Data is expressed as fold change relative to the mean 2hr NA value. N = 4NA HASM cells. +, P <0.05 and +++, P <0.005 compared to NA DMSO control.

Fig. 8

H3K9me3 affects RNA polymerase II binding to the VEGF promoter. RNA pol II association with the (A) proximal and (B) distal VEGF promoter was assessed by ChIP. RNA Pol II association was measured in confluent, serum deprived NA and A HASM cells 0 and 2 hours after 24 hours serum deprivation. IgG negative controls are shown. Data is expressed as fold change relative to the mean 0hr NA value. N= 3NA and 6 A HASM cell lines. RNA pol II association with the (C) proximal and (D) distal VEGF promoter, in the presence of BIX-01294 was assessed by ChIP as per Sp1. N = 3 NA and 3 A HASM cell lines.

Fig. 9

G9A is not recruited to the asthmatic VEGF promoter. G9A association with the (A) proximal and (B) distal VEGF promoter was assessed by ChIP. G9A association was measured in confluent, serum deprived NA and A HASM cells 0 and 2 hours after 24 hours serum deprivation. IgG negative controls are shown. Data is expressed as fold change relative to the mean 0hr NA value. N= 4 NA and 4 A HASM cell lines.

Fig. 10

Differences in complex formation at the NA and A VEGF promoter regulate VEGF transcription. Both A and NA cells have the same DNA methylation and histone acetylation pattern. A repressive complex is associated with the NA promoter composed of histone H3 trimethylated in lysine 9, the histone methyltransferases G9A and SUV39H1 and low levels of Sp1. Repression of VEGF transcription to maintain ‘normal’ levels is dependent on G9A recruitment and H3K9me3. The A promoter fails to recruit G9A and lacks the repressive complex. Instead the asthmatic VEGF promoter is occupied by a transcriptionally active complex containing histone H3 trimethylated at K4, components of the basal transcriptional machinery, namely RNA pol II and TAF1 and heightened levels of Sp1.