Endothelial Bmx tyrosine kinase activity is essential for myocardial hypertrophy and remodeling

Significance

During the last decades, heart failure has developed into a major burden in the western world, increasingly affecting millions. Cardiac hypertrophy is an adaptive response to myocardial infarction or increased blood pressure, and it often leads to heart failure. Understanding the underlying regulatory processes in the development of pathological hypertrophy is needed for the development of effective therapies. Our results show that the kinase activity of the endothelial bone marrow kinase in chromosome X (Bmx) protein is necessary for the development of pathological cardiac hypertrophy. This finding could provide significant therapeutic applications when specific Bmx kinase inhibitors become available in the clinics.

Abstract

Cardiac hypertrophy accompanies many forms of heart disease, including ischemic disease, hypertension, heart failure, and valvular disease, and it is a strong predictor of increased cardiovascular morbidity and mortality. Deletion of bone marrow kinase in chromosome X (Bmx), an arterial nonreceptor tyrosine kinase, has been shown to inhibit cardiac hypertrophy in mice. This finding raised the possibility of therapeutic use of Bmx tyrosine kinase inhibitors, which we have addressed here by analyzing cardiac hypertrophy in gene-targeted mice deficient in Bmx tyrosine kinase activity. We found that angiotensin II (Ang II)-induced cardiac hypertrophy is significantly reduced in mice deficient in Bmx and in mice with inactivated Bmx tyrosine kinase compared with WT mice. Genome-wide transcriptomic profiling showed that Bmx inactivation suppresses myocardial expression of genes related to Ang II-induced inflammatory and extracellular matrix responses whereas expression of RNAs encoding mitochondrial proteins after Ang II administration was maintained in Bmx-inactivated hearts. Very little or no Bmx mRNA was expressed in human cardiomyocytes whereas human cardiac endothelial cells expressed abundant amounts. Ang II stimulation of endothelial cells increased Bmx phosphorylation, and Bmx gene silencing inhibited downstream STAT3 signaling, which has been implicated in cardiac hypertrophy. Furthermore, activation of the mechanistic target of rapamycin complex 1 pathway by Ang II treatment was decreased in the Bmx-deficient hearts. Our results demonstrate that inhibition of the cross-talk between endothelial cells and cardiomyocytes by Bmx inactivation suppresses Ang II-induced signals for cardiac hypertrophy. These results suggest that the endothelial Bmx tyrosine kinase could provide a target to attenuate the development of cardiac hypertrophy.

Heart failure is a continuously increasing global problem in aging populations. Despite some progress in treatment options, the prognosis of heart failure is worse than that for most cancers (1). Cardiac hypertrophy due to exercise training is referred to as physiological hypertrophy, in which the architecture and contractile function of the heart are maintained or enhanced, and it is reversible if the training is discontinued. Numerous diseases, such as myocardial infarction, aortic stenosis, hypertension and metabolic stress, induce pathological hypertrophy, which is related to activation of maladaptive cellular events in the heart and progressively leads to heart failure (2). Understanding the underlying regulatory processes in the development of pathological hypertrophy has provided effective therapeutic targets to attenuate disease progression, but there are currently no treatments to reverse cardiac hypertrophy.

The Bmx (Etk) gene in the chromosome region Xp22.2 encodes a tyrosine kinase that was originally identified and cloned from hematopoietic cells (3). Striking new evidence indicates that Bmx phosphorylates a phosphotyrosine-primed motif mediating the activation of multiple receptor tyrosine kinases (4). Bmx is highly expressed in the endocardium and in the endothelium of large arteries, starting between embryonic days 10.5–12.5 (5), indicating that Bmx is a signal transducer mainly in the arterial endothelium. However, Bmx deletion does not result in any obvious developmental phenotype in mice (5), suggesting that it has a redundant function during embryogenesis. In contrast, Bmx has been shown to be important in a variety of pathological states, including tumor growth (6–8), and its overexpression promotes ischemia-induced arteriogenesis and inflammatory angiogenesis (9).

Previous reports have indicated that Bmx deficiency attenuates pressure overload-induced cardiac hypertrophy in response to thoracic aortic constriction or vascular endothelial growth factor-B (VEGF-B) overexpression (10, 11), but the mechanisms of how this regulation occurs are as yet not known. We have here analyzed the molecular mechanisms of how Bmx regulates cardiac hypertrophy by using angiotensin II (Ang II) treatment that leads to cardiac hypertrophy and remodeling (12). Importantly, we show that pathological cardiac growth is suppressed in a mouse model where the Bmx kinase is inactivated by a missense point mutation. Our findings provide therapeutic proof of principle indicating that the blocking of Bmx tyrosine kinase activity could be used to inhibit pathological cardiac hypertrophy.

Results

Bmx Tyrosine Kinase Activity Is Necessary for Ang II-Induced Cardiac Hypertrophy.

Previous studies have shown that cardiac hypertrophy in response to aortic coarctation or VEGF-B overexpression is greatly attenuated in mice deficient of the Bmx protein (10, 11). Because of the continuing development of increasingly specific Bmx tyrosine kinase inhibitors, it would be important to know whether specific inactivation of the Bmx tyrosine kinase activity would be enough to inhibit cardiac hypertrophy. Thus, our aim was to determine whether attenuation of cardiac hypertrophy detected in Bmx knockout (KO) mice is dependent on the tyrosine kinase activity of the Bmx protein. We analyzed cardiac growth upon Ang II administration to mice in which the Bmx gene had been deleted, or replaced with a K421R mutant allele (equivalent to human mutation K445R) (13), rendering the kinase catalytically inactive (referred to as TK mice) (14).

To compare the effects of Bmx gene deletion or tyrosine kinase inactivation on Ang II-induced cardiac hypertrophy, we first implanted Ang II minipumps into Bmx KO and WT mice for 2 wk. The WT mice developed cardiac hypertrophy, but, as expected, the hypertrophic changes were significantly milder in KO mice than in WT mice (Fig. 1A). Immunostaining for dystrophin demonstrated that the cardiomyocyte size was increased in both groups after Ang II infusion; however, the increase in Bmx KO mice was significantly smaller than in WT mice (Fig. 1 B and C). Importantly, the Ang II-induced cardiac growth and cardiomyocyte size were reduced also in mice expressing the kinase-inactive Bmx allele (Fig. 1 E–G). The mRNA level of skeletal α-actin, a marker for pathological cardiac hypertrophy, was significantly increased in WT mice, but not in Bmx KO mice or in TK mice (Fig. 1 D and H). Accordingly, after a prolonged Ang II infusion for 6 wk, the extent of cardiac hypertrophy was also less in Bmx KO mice than in WT mice (Fig. S1). However, there were no significant differences in the blood pressure between KO and WT mice before or 2 or 6 wk after Ang II infusion measured by the tail cuff method. Similar results were obtained when a catheter was inserted into the carotid artery to monitor blood pressure more accurately. These results confirmed that Bmx deficiency does not affect blood pressure (Fig. S2). Taken together; these data indicate that the development of cardiac hypertrophy is dependent on Bmx tyrosine kinase activity and that the difference in the cardiac Ang II response is not due to differences in blood pressure.

Fig. 1.

Deletion of Bmx protein or inactivation of its tyrosine kinase activity suppresses Ang II-induced cardiac hypertrophy. (A) Heart weight (HW)-to-body weight (BW) ratios in the indicated experimental groups (Ang II/WT, n = 17; Ang II/KO, n = 16; sham/KO, n = 14; sham/WT n = 10) at 2 wk. Data are pooled from two independent experiments. (B) Dystrophin-stained cardiac sections. (C) Quantification of the cardiomyocyte cross-sectional areas. (D) Relative levels of skeletal α-actin mRNA in Bmx KO mice and respective control WT mice. (E) Heart weight-to-body weight ratios in the indicated experimental groups (Ang II/WT, n = 4; Ang II/TK-, n = 5; sham/TK-, n = 7; sham/WT n = 8) at 2 wk. (F) Dystrophin-stained cardiac sections. (G) Quantification of the cardiomyocyte cross-sectional areas. (H) Relative levels of skeletal α-actin mRNA in Bmx TK mice and respective control WT mice. *P < 0.05. ^#P < 0.05 vs. WT sham. Data are presented as mean ± SEM. Statistical analysis was performed with one-way ANOVA. (Scale bar: 20 μm.)

Fig. S1.

Bmx deficiency attenuates Ang II-induced cardiac hypertrophy after 6 wk of treatment. (A) Heart weight-to-body weight ratios (Ang II/WT, n = 5; Ang II/KO, n = 7; sham/KO, n = 5; and sham/WT, n = 5) after 6 wk on Ang II infusion, sham-operated KO and WT mice were used as baseline controls. (B) Quantification of the cardiomyocyte cross-sectional areas. (C) Representative cardiac sections immunostained for dystrophin. *P < 0.05. (Scale bar: 20 μm.)

Fig. S2.

Bmx deficiency does not affect blood pressure. Arterial blood pressure was measured from the carotid artery using Millar solid-state micropressure transducer-tipped catheters. There was no difference in systolic (A) or diastolic (B) blood pressure between Bmx KO and WT mice. Heart rate was also similar in both groups (C).

The Bmx Tyrosine Kinase Is Important for Endothelial Cell–Cardiomyocyte Cross-Talk.

A previous report has suggested that Bmx is expressed also in cardiomyocytes in addition to its expression in the arterial endothelium, in endocardium, and in cells of the hematopoietic myeloid lineage (15). Thus, we compared Bmx mRNA levels in isolated arterial and microvascular endothelial cells as well as in fibroblasts and cardiomyocytes from human heart. As shown in Fig. 2A, Bmx expression was highest in arterial endothelial cells. The microvascular endothelial cells contained about 35% of this amount whereas expression in the cardiomyocytes was about 0.1% of that found in microvascular endothelial cells, and no Bmx RNA could be detected in cardiac fibroblasts. These results strongly suggest that the effects of Bmx on cardiac hypertrophy are mediated via the endothelial cells of the coronary vasculature.

Fig. 2.

Bmx is expressed mainly in cardiac endothelial cells and phosphorylated upon Ang II treatment. (A) Bmx mRNA expression in human cardiac cells. HCMEC, human cardiac microvascular endothelial cells; HCAEC, human cardiac arterial endothelial cells; HCM, human cardiomyocytes; HCF, human cardiac fibroblasts; ND, not detected. (B) Phosphorylation of Bmx (pBmx) immunoprecipitated (IP) from HUVECs and stimulated with 0.1 and 1.0 nM Ang II for 10 min or 30 min. VEGF was used as a positive control. (C) Western blot showing Bmx expression in isolated adult cardiomyocytes (CMC) and HUVECs (EC). An equal amount of total protein was loaded in each lane. Hsc70, heat shock cognate 70.

Next, we analyzed whether Ang II can induce Bmx phosphorylation in endothelial cells. Human umbilical venous endothelial cells (HUVECs) were stimulated with different concentrations of Ang II, and vascular endothelial growth factor (VEGF) was used as a positive control. Similarly to VEGF, Ang II induced strong Bmx phosphorylation in HUVECs after 10 min and 30 min of stimulation (Fig. 2B). In contrast, Bmx was not detected in isolated adult mouse cardiomyocytes even after Ang II stimulation, further indicating that Bmx is specific for cardiac endothelial cells (Fig. 2C).

Reduced Myocardial Capillary Density in Ang II-Treated WT Mice but Not in Bmx-Deleted Mice.

Pathological pressure overload-induced cardiac growth is often associated with a decrease of capillary density, especially in the left ventricular myocardium, because vessel growth does not match the cardiomyocyte growth (16). Staining for the vascular endothelial marker CD31 showed decreased blood vessel density and area in the hearts of Ang II-treated WT mice compared with sham-operated control WT mice whereas the capillary density in the Bmx-deleted hearts was not altered by Ang II treatment (Fig. 3 A and B). Similar results were obtained when the TK mice were used for the analysis (Fig. 3C). This difference was due to the inhibition of cardiomyocyte growth because there was no statistical difference in the capillary-to-cardiomyocyte ratio between the groups (WT sham, 1.68 ± 0.05; WT Ang II, 1.75 ± 0.03; KO sham, 1.60 ± 0.04; KO Ang II, 1.77 ± 0.06; P > 0.05).

Fig. 3.

Attenuated Ang II-induced cardiac responses in Bmx KO and TK mice. (A) Myocardial blood vessels stained for CD31 antigen. (B and C) Blood vessel area/grid. (D and E) Collagen I (Col I) and collagen III (Col III) mRNA expression in Bmx KO (D) and TK (E) mice. *P < 0.05. ^#P < 0.05 vs. WT sham. Data are presented as mean ± SEM. Statistical analysis was performed with one-way ANOVA. (Scale bar: 100 μm.)

To analyze how the loss of Bmx kinase activity regulates the pathogenesis of Ang II-triggered hypertrophy, cardiac samples from Bmx KO, TK, and WT mice with and without preceding systemic Ang II infusion were subjected to genome-wide RNA expression analysis. Without Ang II treatment, only a few genes were differentially expressed between Bmx KO and WT hearts (Table S1). Ang II-treated WT hearts had alterations in several gene clusters attributable to cardiac hypertrophic response, including increased expression of genes encoding extracellular matrix proteins. Expression of this gene cluster was significantly lower in Bmx KO or TK hearts than in WT hearts after the Ang II treatment (Tables S2 and S3). To confirm microarray findings, we analyzed a set of mRNAs with quantitative PCR (qPCR). For example, procollagen I and III mRNAs were up-regulated in Ang II-treated WT mice but not in Bmx KO or TK mice (Fig. 3 D and E). Of note, Ang II treatment did not affect the mRNA expression of Bmx in WT hearts in any of the experiments.

Table S1.

Functional annotation clustering of the genes differentially regulated in WT vs. Bmx KO hearts at baseline

Category	Term	Count	P value	Benjamini
Up-regulated in KO mice
Annotation cluster 1	Enrichment score: 2.827997834828164
SP_PIR_KEYWORDS	Biological rhythms	6	6.21E−08	1.14E−05
GOTERM_BP_FAT	GO:0048511∼rhythmic process	8	1.84E−06	0.0014
GOTERM_BP_FAT	GO:0007623∼circadian rhythm	5	6.66E−05	0.0131
Annotation cluster 2	Enrichment score: 2.6313785048480707
GOTERM_BP_FAT	GO:0031327∼negative regulation of cellular biosynthetic process	12	4.63E−05	0.0181
Annotation cluster 3	Enrichment score: 2.564466543199008
GOTERM_BP_FAT	GO:0045638∼negative regulation of myeloid cell differentiation	4	2.51E−04	0.0164
Down-regulated in KO mice
Annotation cluster 1	Enrichment score: 2.911608673974152
GOTERM_CC_FAT	GO:0005578∼proteinaceous extracellular matrix	7	2.48E−04	0.0083
GOTERM_CC_FAT	GO:0031012∼extracellular matrix	7	3.07E−04	0.0077
Annotation cluster 2	Enrichment score: 2.677806961647269
GOTERM_CC_FAT	GO:0044421∼extracellular region part	12	1.22E−05	0.0012
GOTERM_CC_FAT	GO:0005576∼extracellular region	16	5.82E−05	0.0029
Annotation cluster 3	Enrichment score: 2.4687947339329064
GOTERM_BP_FAT	GO:0048514∼blood vessel morphogenesis	8	2.59E−06	0.0017
GOTERM_BP_FAT	GO:0001568∼blood vessel development	8	1.03E−05	0.0034
GOTERM_BP_FAT	GO:0001944∼vasculature development	8	1.20E−05	0.0027
GOTERM_BP_FAT	GO:0001525∼angiogenesis	6	5.95E−05	0.0099

Gene clusters up-regulated and down-regulated in cardiac samples isolated from Bmx KO mice versus WT mice at baseline.

Table S2.

Differentially activated gene clusters in whole-genome microarray analysis of Ang II-treated Bmx KO and WT hearts

Category	Term	Count	P value	Benjamini
Up-regulated in KO mice
Annotation cluster 1	Enrichment score: 4.730228073692788
GOTERM_CC_FAT	GO:0005743∼mitochondrial inner membrane	17	2.20E−08	4.33E−06
SP_PIR_KEYWORDS	Mitochondrion	28	3.32E−08	8.07E−06
GOTERM_CC_FAT	GO:0019866∼organelle inner membrane	17	4.59E−08	9.04E−06
GOTERM_CC_FAT	GO:0005739∼mitochondrion	34	1.65E−07	3.25E−05
GOTERM_CC_FAT	GO:0031966∼mitochondrial membrane	17	4.37E−07	8.61E−05
GOTERM_CC_FAT	GO:0044429∼mitochondrial part	20	5.45E−07	1.07E−04
GOTERM_CC_FAT	GO:0005740∼mitochondrial envelope	17	9.79E−07	1.93E−04
SP_PIR_KEYWORDS	Mitochondrion inner membrane	12	2.84E−06	6.90E−04
KEGG_PATHWAY	Mmu00190:oxidative phosphorylation	11	3.15E−06	2.33E−04
Annotation cluster 2	Enrichment score: 2.1738561052301186
GOTERM_BP_FAT	GO:0022900∼electron transport chain	9	2.59E−05	0.0215
Down-regulated in KO mice
Annotation cluster 1	Enrichment score: 7.4427666279094735
SP_PIR_KEYWORDS	Secreted	47	1.25E−12	3.12E−10
GOTERM_CC_FAT	GO:0005576∼extracellular region	50	4.74E−11	7.39E−09
SP_PIR_KEYWORDS	Signal	68	4.94E−11	6.15E−09
UP_SEQ_FEATURE	Signal peptide	68	2.90E−09	2.26E−06
Annotation cluster 2	Enrichment score: 4.062149945242736
GOTERM_MF_FAT	GO:0030246∼carbohydrate binding	15	6.87E−06	0.0017
GOTERM_MF_FAT	GO:0030247∼polysaccharide binding	10	8.64E−06	0.0011
GOTERM_MF_FAT	GO:0001871∼pattern binding	10	8.64E−06	0.0011
GOTERM_MF_FAT	GO:0005539∼glycosaminoglycan binding	8	2.09E−04	0.0171
Annotation cluster 3	Enrichment score: 3.777794734126123
SP_PIR_KEYWORDS	Extracellular matrix	13	2.08E−06	8.64E−05
GOTERM_CC_FAT	GO:0044421∼extracellular region part	24	1.35E−05	7.04E−04
GOTERM_CC_FAT	GO:0005578∼proteinaceous extracellular matrix	13	1.17E−04	0.0036
GOTERM_CC_FAT	GO:0031012∼extracellular matrix	13	1.69E−04	0.0044
Annotation cluster 4	Enrichment score: 3.599001910351243
SP_PIR_KEYWORDS	Methylation	15	6.76E−08	5.61E−06
INTERPRO	IPR007125: Histone core	8	1.95E−07	8.43E−05
SP_PIR_KEYWORDS	Nucleosome core	8	2.95E−07	1.83E−05
GOTERM_BP_FAT	GO:0034622∼cellular macromolecular complex assembly	13	2.46E−06	0.0029
Annotation cluster 5	Enrichment score: 2.977165098955894
GOTERM_CC_FAT	GO:0030017∼sarcomere	7	2.59E−04	0.0058
GOTERM_CC_FAT	GO:0044449∼contractile fiber part	7	3.84E−04	0.0075
GOTERM_CC_FAT	GO:0030016∼myofibril	7	5.20E−04	0.0090

Functional annotation clustering performed after Ang II infusion for 2 wk.

Table S3.

Differentially activated gene clusters in whole-genome microarray analysis of Ang II-treated Bmx TK and WT hearts

Category	Term	Count	P value	Benjamini
Up-regulated clusters in WT hearts compared with TK hearts after Ang II treatment
Annotation cluster 1	Enrichment score: 13.903375261257448
GOTERM_CC_FAT	GO:0031012∼extracellular matrix	55	2.38E−18	8.61E−16
SP_PIR_KEYWORDS	Extracellular matrix	43	4.35E−18	9.54E−16
GOTERM_CC_FAT	GO:0005578∼proteinaceous extracellular matrix	52	4.66E−17	8.43E−15
Annotation cluster 2	Enrichment score: 10.550385469611433
SP_PIR_KEYWORDS	Signal	214	1.25E−19	5.50E−17
SP_PIR_KEYWORDS	Secreted	123	2.59E−16	3.24E−14
UP_SEQ_FEATURE	Signal peptide	214	1.24E−14	2.52E−11
GOTERM_CC_FAT	GO:0005576∼extracellular region	133	2.02E−11	1.83E−09
SP_PIR_KEYWORDS	Disulfide bond	155	5.25E−09	5.75E−07
SP_PIR_KEYWORDS	Glycoprotein	207	8.53E−09	7.47E−07
UP_SEQ_FEATURE	Disulfide bond	146	2.00E−05	0.01345
UP_SEQ_FEATURE	Glycosylation site: N−linked (GlcNAc...)	196	5.31E−05	0.02133
Annotation cluster 3	Enrichment score: 6.562173968199845
GOTERM_MF_FAT	GO:0030247∼polysaccharide binding	23	4.86E−09	3.12E−06
GOTERM_MF_FAT	GO:0005539∼glycosaminoglycan binding	20	9.12E−08	2.93E−05
GOTERM_MF_FAT	GO:0008201∼heparin binding	15	3.90E−06	0.00083
SP_PIR_KEYWORDS	Heparin-binding	12	6.27E−06	0.00027
GOTERM_MF_FAT	GO:0030246∼carbohydrate binding	31	7.99E−06	0.00102
Annotation cluster 4	Enrichment score: 6.130005753729857
GOTERM_BP_FAT	GO:0007155∼cell adhesion	51	1.94E−07	0.00046
GOTERM_BP_FAT	GO:0022610∼biological adhesion	51	2.03E−07	0.00024
SP_PIR_KEYWORDS	Cell adhesion	35	1.02E−05	0.00040
Annotation cluster 5	Enrichment score: 4.569864347028035
GOTERM_BP_FAT	GO:0030198∼extracellular matrix organization	17	3.31E−06	0.00157
GOTERM_BP_FAT	GO:0043062∼extracellular structure organization	19	3.94E−05	0.01336
GOTERM_BP_FAT	GO:0030199∼collagen fibril organization	7	0.00014	0.03503
Annotation cluster 6	Enrichment score: 3.5597990346793367
UP_SEQ_FEATURE	Short sequence motif:Prevents secretion from ER	13	8.24E−06	0.00832
INTERPRO	IPR000886:Endoplasmic reticulum, targeting sequence	10	3.32E−05	0.01783
GOTERM_CC_FAT	GO:0005788∼endoplasmic reticulum lumen	11	0.00131	0.03341
Annotation cluster 7	Enrichment score: 3.323175383704086
KEGG_PATHWAY	mmu05410:Hypertrophic cardiomyopathy (HCM)	15	2.16E−05	0.00153
KEGG_PATHWAY	mmu05414:Dilated cardiomyopathy	14	0.00024	0.00691
Up-regulated clusters in TK hearts compared with WT hearts after Ang II treatment
Annotation cluster 1	Enrichment score: 5.380207516944994
SP_PIR_KEYWORDS	Mitochondrion	23	2.31E−09	3.93E−07
GOTERM_CC_FAT	GO:0005739∼mitochondrion	30	2.49E−09	4.60E−07
GOTERM_CC_FAT	GO:0044429∼mitochondrial part	18	4.57E−08	4.23E−06
GOTERM_CC_FAT	GO:0031966∼mitochondrial membrane	15	1.20E−07	7.38E−06
GOTERM_CC_FAT	GO:0019866∼organelle inner membrane	14	1.25E−07	5.77E−06
GOTERM_CC_FAT	GO:0005740∼mitochondrial envelope	15	2.52E−07	9.31E−06
GOTERM_CC_FAT	GO:0005743∼mitochondrial inner membrane	13	5.35E−07	1.65E−05
GOTERM_CC_FAT	GO:0031090∼organelle membrane	20	1.03E−06	2.71E−05
SP_PIR_KEYWORDS	Mitochondrion inner membrane	10	1.96E−06	1.66E−04
GOTERM_CC_FAT	GO:0031967∼organelle envelope	16	2.24E−06	5.18E−05
GOTERM_BP_FAT	GO:0022900∼electron transport chain	8	3.49E−06	0.00131
KEGG_PATHWAY	mmu05010:Alzheimer's disease	9	9.92E−06	7.84E−04
KEGG_PATHWAY	mmu05012:Parkinson's disease	8	1.20E−05	4.74E−04
SP_PIR_KEYWORDS	Transit peptide	13	2.49E−05	0.00141
UP_SEQ_FEATURE	Transit peptide:Mitochondrion	13	5.43E−05	0.02107
KEGG_PATHWAY	mmu00190:Oxidative phosphorylation	7	1.14E−04	0.00299
Down-regulated clusters in WT hearts compared with TK hearts after Ang II treatment
Annotation cluster 1	Enrichment score: 10.074002769569203
GOTERM_MF_FAT	GO:0000166∼nucleotide binding	173	1.97E−14	1.52E−11
GOTERM_MF_FAT	GO:0017076∼purine nucleotide binding	154	4.47E−14	1.73E−11
SP_PIR_KEYWORDS	Nucleotide-binding	137	3.02E−12	3.89E−10
GOTERM_MF_FAT	GO:0001883∼purine nucleoside binding	129	4.93E−12	1.27E−09
GOTERM_MF_FAT	GO:0030554∼adenyl nucleotide binding	128	5.82E−12	1.12E−09
GOTERM_MF_FAT	GO:0001882∼nucleoside binding	129	7.64E−12	1.18E−09
GOTERM_MF_FAT	GO:0032553∼ribonucleotide binding	140	5.99E−11	7.72E−09
Annotation cluster 2	Enrichment score: 9.87022181487833
GOTERM_CC_FAT	GO:0005739∼mitochondrion	155	1.90E−30	6.90E−28
SP_PIR_KEYWORDS	Mitochondrion	109	3.63E−25	1.40E−22
SP_PIR_KEYWORDS	Transit peptide	73	1.70E−20	3.30E−18
UP_SEQ_FEATURE	Transit peptide:Mitochondrion	70	7.44E−17	2.64E−13
GOTERM_CC_FAT	GO:0044429∼mitochondrial part	69	1.62E−15	3.03E−13
GOTERM_CC_FAT	GO:0031980∼mitochondrial lumen	29	8.64E−10	1.05E−07
GOTERM_CC_FAT	GO:0005759∼mitochondrial matrix	29	8.64E−10	1.05E−07
Annotation cluster 3	Enrichment score: 8.316364079657035
GOTERM_CC_FAT	GO:0044429∼mitochondrial part	69	1.62E−15	3.03E−13
GOTERM_CC_FAT	GO:0031966∼mitochondrial membrane	45	2.94E−09	2.68E−07
GOTERM_CC_FAT	GO:0031967∼organelle envelope	57	4.09E−09	2.98E−07
GOTERM_CC_FAT	GO:0031975∼envelope	57	4.65E−09	2.82E−07
GOTERM_CC_FAT	GO:0005740∼mitochondrial envelope	46	6.42E−09	3.34E−07
GOTERM_CC_FAT	GO:0005743∼mitochondrial inner membrane	38	1.70E−08	7.72E−07
Annotation cluster 4	Enrichment score: 5.906581894470631
GOTERM_BP_FAT	GO:0051186∼cofactor metabolic process	29	1.27E−08	2.79E−05
GOTERM_BP_FAT	GO:0006732∼coenzyme metabolic process	24	1.11E−07	1.21E−04
GOTERM_BP_FAT	GO:0051188∼cofactor biosynthetic process	16	1.32E−05	0.00719
GOTERM_BP_FAT	GO:0009108∼coenzyme biosynthetic process	12	1.26E−04	0.03866
Annotation cluster 5	Enrichment score: 5.341426509172577
GOTERM_MF_FAT	GO:0048037∼cofactor binding	34	2.62E−09	2.90E−07
UP_SEQ_FEATURE	Nucleotide phosphate−binding region:FAD	16	1.31E−07	1.55E−04
SP_PIR_KEYWORDS	Flavoprotein	20	1.98E−07	9.58E−06
SP_PIR_KEYWORDS	FAD	20	4.97E−07	2.14E−05
Annotation cluster 6	Enrichment score: 3.3586616451023303
GOTERM_MF_FAT	GO:0030170∼pyridoxal phosphate binding	12	2.51E−05	0.00176
GOTERM_MF_FAT	GO:0070279∼vitamin B6 binding	12	2.51E−05	0.00176
SP_PIR_KEYWORDS	Pyridoxal phosphate	12	7.81E−05	0.00215
GOTERM_MF_FAT	GO:0019842∼vitamin binding	17	1.15E−04	0.00681
Annotation cluster 7	Enrichment score: 3.2142061926138865
GOTERM_BP_FAT	GO:0006631∼fatty acid metabolic process	22	9.18E−05	0.03289
SP_PIR_KEYWORDS	Fatty acid metabolism	11	4.25E−04	0.01021
KEGG_PATHWAY	mmu00071:Fatty acid metabolism	9	9.04E−04	0.04596
Annotation cluster 8	Enrichment score: 2.9810208005966414
INTERPRO	IPR006089:Acyl-CoA dehydrogenase, conserved site	6	3.61E−05	0.02362
INTERPRO	IPR006092:Acyl-CoA dehydrogenase, N-terminal	6	6.38E−05	0.02775
GOTERM_MF_FAT	GO:0050660∼FAD binding	13	7.63E−05	0.00490
INTERPRO	IPR006090:Acyl-CoA oxidase/dehydrogenase, type 1	6	1.65E−04	0.04272
Down-regulated clusters in TK hearts compared with WT hearts after Ang II treatment
No significant clusters

Functional annotation clustering performed after Ang II infusion for 2 wk.

Mitochondrial Content Is Maintained in Bmx-Deficient Hearts After Ang II Treatment.

Whole-genome gene expression analyses showed that Ang II induces markedly fewer changes in KO and TK mice than in WT mice (Fig. 4A and Tables S2 and S3). Interestingly, genes encoding components of the inner mitochondrial membrane and the electron transport chain were more abundantly expressed in Ang II-treated Bmx KO or TK hearts than in Ang II-treated WT hearts. PGC-1-α is a major regulator of mitochondrial biosynthesis, and decreased levels of PGC-1-α and citrate synthase are associated with the development of pathological cardiac hypertrophy (reviewed in ref. 17). Decreased PGC-1-α RNA level and decreased citrate synthase activity were detected in the Ang II-treated WT hearts but not in the Bmx KO and TK hearts (Fig. 4 B–E). Together with the microarray data, these results indicate a better sustained mitochondrial content in the hearts of Bmx KO and TK mice than in WT mice upon Ang II-induced cardiac stress.

Fig. 4.

Comparison of global and mitochondrial gene expression in Bmx TK and WT hearts in response to Ang II treatment. (A) An overview of the total number of significantly changed genes and functional gene clusters from Database for Annotation, Visualization and Integrated Discovery (DAVID) analysis showing markedly fewer changes in Bmx TK mice compared with WT mice after Ang II treatment. Statistical values are presented in Table S3. (B and C) PGC-1α mRNA expression in cardiac samples of Bmx WT, KO and TK mice after 14 d of Ang II infusion. The values shown indicate fold change in comparison with sham WT (set to 1). (D) Citrate synthase activity in the hearts from Ang II- or sham-treated Bmx WT and KO mice after 14 d of infusion (n = 3 in each group). Note that citrate synthase activity is statistically significantly reduced after Ang II infusion in WT mice but not in the Bmx KO mice. (E) Relative change of citrate synthase activity after Ang II treatment. Data are presented as mean± SEM. Statistical analysis was performed with one-way ANOVA. *P < 0.05. ^#P < 0.05 vs. WT sham.

Bmx Regulates Ang II-Induced Inflammatory Cytokine Responses.

Increased expression of inflammatory cytokines has also been implicated in cardiac hypertrophy. Because Bmx has been shown to regulate inflammatory responses (9, 14, 18), we examined whether Bmx affects the expression of inflammatory genes upon Ang II induction. Indeed, mRNA levels of interleukin-6 (IL-6), IL-8, tumor necrosis factor receptor-2 (Tnfr2), and matrix metalloproteinase 2 (Mmp2) were significantly lower in Bmx KO and TK mice than in WT mice after Ang II induction (Fig. 5 A–D). These data indicate that Bmx regulates at least a subset of inflammation-related cytokine responses in Ang II-induced hypertrophy.

Fig. 5.

Bmx inactivation abolishes the increased expression of genes encoding inflammatory and matrix remodeling proteins in Ang II-treated hearts. (A–F) qPCR analysis of the indicated genes. Bmx WT and KO mice (A–C) and Bmx WT and TK mice (D–F) after 14 d of Ang II or sham treatment. Values are shown as fold of change in comparison with sham-operated WT hearts. Data are presented as mean ± SEM. Statistical analysis was performed with one-way ANOVA. *P < 0.05. ^#P < 0.05 vs. WT sham.

Tissue inhibitor of matrix metalloproteinases-1 (Timp1) mRNA was markedly induced upon Ang II treatment in WT, but not in Bmx KO or TK mice (Fig. 5 B and E). Heat shock protein 70 (Hsp70) has been shown to be up-regulated in hypertrophic cardiomyopathy (19). Interestingly, a clear trend of Hsp1a mRNA induction was detected in the hypertrophic WT hearts whereas lack of Bmx abolished this response (Fig. 5 C and F).

Involvement of Bmx Tyrosine Kinase Activity in Ang II-Induced Downstream Signaling.

To elucidate the influence of Bmx inactivation on signaling pathways mediating cardiomyocyte growth, we analyzed activation of the mechanistic target of rapamycin complex 1 (mTORC1) pathway, which is a major regulator of protein synthetic pathways during cell growth (20). Consistent with the smaller cardiomyocyte size, we found that the mTORC1 downstream target ribosomal protein S6 (rpS6) phosphorylation at Ser-240/244 was less activated in Bmx KO hearts than in WT hearts after the 2-wk Ang II infusion (Fig. 6 A and B). To further analyze the in vivo signaling induced by Ang II, we injected Ang II into the tail vein of TK and WT mice and analyzed the phosphorylation of rpS6, Akt, and STAT3 in the heart after 10 min of stimulation. Ang II induced phosphorylation of all these three signaling proteins in WT mice, but the response was blunted in TK mice (Fig. 6 C–F). In addition, microarray data indicated that Ang II treatment increased Stat3 [false discovery rate (FDR) = 0.005] and decreased Stat1 (FDR = 0.00001) mRNA expression in WT hearts, but there was no effect of Ang II on either gene in KO hearts.

Fig. 6.

Lack of Bmx tyrosine kinase activity alters intracellular signaling responses to Ang II treatment. (A) Representative Western blots from cardiac samples of Bmx WT and KO mice after 14 d of Ang II or sham treatment. The blots were probed with antibodies against phosphorylated rpS6 (Ser-240/244), rpS6, and GAPDH. The uppermost and lowermost signals are from the same blot; the middle blot contains the same samples. (B) Quantitative analysis of phosphorylated rpS6 Ser-240/244 normalized to GAPDH from three samples per treatment group. (C) Western blots from WT and TK hearts after 10 min of Ang II stimulation in vivo. (D–F) Quantification of the Western blots for phospho-Akt (Ser-473), phospho-rpS6, and phospho-STAT3. (G) Quantification of STAT3 phosphorylation in HUVECs transduced with shScr or shBmx and stimulated with Ang II. The data are pooled from three experiments using three different Bmx silencing constructs, and shScr is set to a relative expression of 1. (H) Schematic summary of the suggested mechanism by which endothelial Bmx regulates cardiomyocyte growth. EC, endothelial cell; CMC, cardiomyocyte; FB, fibroblast; AT1R, angiotensin type 1 receptor. Data are presented as mean ± SEM. Statistical analysis was performed with one-way ANOVA or two-tailed Student’s t test. *P < 0.05, ^#P < 0.05 vs. sham/WT.

To study the effect of Bmx on Ang II signal transduction in more detail, we silenced Bmx in HUVECs with three different shRNA constructs and analyzed the effect of Ang II stimulation on STAT3 activation. Ang II induced STAT3 phosphorylation in control cells transduced with shScr, but not in Bmx-silenced cells (Fig. 6G), indicating that Bmx is needed for Ang II-induced STAT3 signaling in endothelial cells.

Discussion

Our results show that Bmx tyrosine kinase activity is necessary for the cardiac hypertrophy induced by angiotensin II. This finding contrasts with results from a macrophage-driven inflammatory arthritis model where Bmx protein, but not tyrosine kinase activity, was found to be essential (14). In the Ang II hypertrophy model, both the deletion and tyrosine kinase inactivation of Bmx prevented increased inflammatory cytokine and matrix gene activation, stimulation of the mTORC1 pathway, and the decrease of mitochondrial gene expression and activity that characterize the hypertrophic and dysfunctional myocardium in the Ang II hypertension model. The present results on Bmx expression in human and mouse cardiac cells, together with our previous findings (5), indicate that Bmx is expressed mainly in the arterial and microvascular endothelium in the heart. Very little, if any, Bmx mRNA was detected in primary human cardiomyocytes, and no Bmx protein was detected in adult mouse cardiomyocytes. This observation highlights the importance of the cross-talk between vascular endothelial cells and cardiomyocytes in the development of pathological hypertrophy. Furthermore, these mouse models show that deletion of Bmx does not affect normal cardiac homeostasis, which provides a wide therapeutic window for possible use of Bmx tyrosine kinase inhibitors in heart disease.

Angiotensin II activates the angiotensin type 1 receptor (AT1R), which then transactivates the epidermal growth factor receptor (EGFR) to mediate cellular growth. Knockdown of Bmx was shown to attenuate tyrosine phosphorylation of the EGFR by angiotensin II stimulation in mammary epithelial cells, but Bmx did not have an effect on direct stimulation of the EGFR with EGF, indicating that Bmx functions between the activated AT1R and EGFR (21). Bmx also induces the tyrosine phosphorylation and DNA binding activity of STAT1, STAT3, and STAT5 (22). Of these proteins, STAT3 has been shown to be important for physiological homeostasis and stress-induced remodeling of the heart (23). In our experiments, Bmx silencing completely abrogated STAT3 activation by Ang II in endothelial cells and also significantly attenuated phosphorylation in vivo. Although we cannot fully exclude the possibility that Bmx functions also in cardiomyocytes, our results and those of others strongly suggest that Bmx mainly acts in endothelial cells (24). Thus, the present results highlight the importance of endothelial cell-to-cardiomyocyte signaling in cardiac remodeling (see the schematic summary in Fig. 6H).

Our transcriptomic analyses of the Ang II-treated hearts indicated a better preserved mitochondrial content in Bmx kinase-deficient mice upon pathological cardiac growth. Mitochondrial and NADPH oxidase dysfunction and the resulting oxygen radicals have been shown to induce cellular inflammatory responses leading to increased cytokine expression (17, 25–28). Ang II also activates proinflammatory and profibrotic pathways. In this regard, it is interesting that Bmx kinase activity contributed to the regulation of a range of inflammatory cytokine responses associated with pathological cardiac hypertrophy. Inflammatory markers have not been previously analyzed in the context of cardiac hypertrophy in Bmx-deficient mice. In other models, such as in ischemia-mediated arteriogenesis in the hind limb, Bmx has been shown to interact with the TNFR2 pathway to promote adaptive arteriogenesis and angiogenesis (29). In a model of rheumatoid arthritis, Bmx was shown to act in a kinase-independent manner downstream of, or at the same level as, TGF-β activated kinase 1 (TAK1) to mediate inflammatory cytokine signaling, including IL-8 (14). Recently, it was shown that deletion of IL-6 prevents cardiac inflammation, fibrosis, and dysfunction after Ang II (30). Our results showed that inactivation of Bmx abolished the Ang II-induced increase in IL-6 and IL-8, placing Bmx upstream of the inflammatory cascade.

Our data on the prevention of Ang II-induced cardiac hypertrophy when Bmx tyrosine kinase is rendered inactive provide a proof of principle that inhibitors of Bmx tyrosine kinase activity could do the same. Bmx-deficient, as well as tyrosine kinase-inactivated, mice are healthy and fertile, suggesting that Bmx-specific tyrosine kinase inhibitors could provide a wide therapeutic window. This finding could turn out to be significant for therapeutic applications when Bmx kinase inhibitors become available in clinics (31–33). At the moment, many Bmx inhibitors also inhibit other kinases, such as EGFR, which may cause cardiotoxicity, whereas the more specific inhibitors have not yet been tested in vivo. Additional studies will be necessary to explore the possible benefits of modulating Bmx activity as a therapeutic target.

Materials and Methods

A detailed description of all materials and methods can be found in SI Materials and Methods.

Ang II-Induced Cardiac Hypertrophy.

Bmx KO mice (described in ref. 5), backcrossed to the C57BL/6J background at least eight times, and age- and gender-matched WT C57BL/6J mice were used in the present studies. Ang II infusion at 0.1 mg⋅kg⁻¹⋅h⁻¹ for 14 d was induced by s.c. implantation of osmotic minipumps (Alzet model 1002; Durect Corporation). Sham-operated KO mice (n = 14) and WT mice (n = 10) were used as controls. At 2 or 6 wk, the mice were weighed and killed, and then the hearts were excised, weighed, and processed for further analysis as described in SI Materials and Methods. We also studied mice with the kinase-deficient Bmx K421R mutation (equivalent to the human K445R mutation) (13) in the BALB/c background, referred to here as Bmx TK mice (14).

The National Animal Board for Animal Experiments at the Provincial State Office of Southern Finland approved all animal experiments, which were performed in accordance with Finnish legislation regarding the humane care and use of laboratory animals.

Ang II–Bmx Signal Transduction, RT-qPCR, and Western Blot Analysis.

Ang II–Bmx signal transduction, RT-qPCR, and Western blot analysis are detailed in SI Materials and Methods. Primers used are listed in Table S4.

Table S4.

RT-PCR primer sequences

Name	Sequence
Mouse primers
Skeletal α-actin Fwd	5′-TCCTCCGCCGTTGGCT-3′
Skeletal α-actin Rev	5′-AATCTATGTACACGTCAAAAA-3′
Procollagen I Fwd	5′-TGTTGGCCCATCTGGTAAAGA-3′
Procollagen I Rev	5′-CAGGGAATCCGATGTTGCC-3′
Procollagen III Fwd	5′-TGGTCCTCAGGGTGTAAAGG-3′
Procollagen III Rev	5′-GTCCAGCATCACCTTTTGGT-3′
Interleukin-6 Fwd	5′-AGTTGCCTTCTTGGGACTG-3′
Interleukin-6 Rev	5′-AGGTCTGTTGGGAGTGGTATC-3′
Interleukin-8 Fwd	5′-ATGACTTCCAAGCTGGCCGT-3′
Interleukin-8 Rev	5′-TTACATAATTTCTGTGTTGGC-3′
TNFR2 Fwd	5′-CAGGTTGTCTTGACACCCTAC-3′
TNFR2 Rev	5′-GCACAGCACATCTGAGCCT-3′
TIMP1 Fwd	5′-TCCTCTTGTTGCTATCACTGATAGCTT-3′
TIMP1 Rev	5′-CGCTGGTATAAGGTGGTCTCGTT-3′
MMP2 Fwd	5′-CCGAGGACTATGACCGGGATAA-3′
MMP2 Rev	5′-CTTGTTGCCCAGGAAAGTGAA-3′
TGFβ1 Fwd	5′-CACTGGAGTTGTACGGCAGT-3′
TGFβ1 Rev	5′-AGAGCAGTGAGCGCTGAATC-3′
HSPA1A Fwd	5′-CAAGATCACCATCACCAACG-3′
HSPA1A Rev	5′-GCACCTTCTTCTTGTCAGCC-3
TNFα Fwd	5′-CATCTTCTCAAAATTCGAGTGACAA-3′
TNFα Rev	5′-TGGGAGTAGACAAGGTACAACCC-3′
TNFR1 Fwd	5′-GGGCACCTTTACGGCTTCC-3′
TNFR1 Rev	5′-GGTTCTCCTTACAGCCACACA-3′
Interleukin-1β Fwd	5′-GTGTGACGTTCCCATTAGAC-3′
Interleukin-1β Rev	5′-CATTGAGGTGGAGAGCTTTC-3′
Mcp1 Fwd	5′-TGCATCTGCCCTAAGGTCTTC-3′
Mcp1 Rev	5′-GAAGTGCTTGAGGTGGTTGTG-3′
PGC-1α Fwd	5′-CCCTGCCATTGTTAAGACC-3′
PGC-1α Rev	5′-TGCTGCTGTTCCTGTTTTC-3′
TBP Fwd	5′-GAAGCTGCGGTACAATTCCAG-3′
TBP Rev	5′-CCCCTTGTACCCTTCACCAAT-3′
Human primers
Bmx Fwd	5′-CAGATTGTCTATAAAGATGGGC-3′
Bmx Rev	5′-TGTAATGCTTTCAACCACTG-3′
Hprt1 Fwd	5′-TGAGGATTTGGAAAGGGTGT-3′
Hprt1 Rev	5′-TCCCCTGTTGACTGGTCATT-3′
β-actin Fwd	5′-AGGCCAACCGCGAGAAGATGA-3′
β-actin Rev	5′-GCCGTGGTGGTGAAGCTGTAG-3′
GAPDH Fwd	5′-CCACTAGGCGCTCACTGTTC-3′
GAPDH Rev	5′-CCCCATACGACTGCAAAGAC-3′

Sequences used for RT-PCR analysis. Fwd, forward; Rev, reverse.

Whole-Genome Microarray and Data Analysis.

RNA samples were analyzed with genome-wide Illumina MouseWG-6 v2 microarrays (Illumina Inc.). Detailed microarray data analysis is described in SI Materials and Methods. The data have been deposited in the Gene Expression Omnibus (GEO) data repository with accession number GSE47420.

Immunohistochemical Analyses.

Five- to 7-μm frozen sections were fixed with acetone, immunostained, and analyzed as detailed in SI Materials and Methods.

Statistical Analysis.

Values are indicated as mean ± SEM in the figures. Statistical analysis of multiple groups was performed with one-way ANOVA, followed by Tukey’s post hoc test for groups with equal variances and by Games–Howell’s post hoc test for groups with unequal variances. Statistical analysis of two groups was performed with unpaired t test. All statistical tests were two-tailed. Differences were considered statistically significant at P < 0.05.

SI Materials and Methods

Ang II-Induced Cardiac Hypertrophy.

Bmx KO mice (5), backcrossed to the C57BL/6J background at least eight times, and age- and gender-matched WT C57BL/6J mice were used in these studies. Ten- to 12-wk-old C57BL/6J male Bmx KO mice (n = 17) and WT mice (n = 16) were anesthetized with xylazine (Rompun vet; Bayer Healthcare) and ketamine (Ketalar; Pfizer), and Ang II infusion at 0.1 mg⋅kg⁻¹⋅h⁻¹ for 14 d was induced by s.c. implantation of osmotic minipumps (Alzet model 1002; Durect Corporation). Sham-operated KO mice (n = 14) and WT mice (n = 10) were used as controls. Two weeks later, the mice were weighed and killed, and then the hearts were excised and weighed. The samples were cut into three parts, and the hearts were processed for histology, for RNA extraction using RNAlater (Qiagen), and for protein analysis by snap-freezing in liquid nitrogen. For histological processing, the samples were embedded in TissueTek OCT compound (Sakura Finetek Europe) and frozen in 2-methylbutane with 2% (vol/vol) pentane over liquid nitrogen. For the Ang II induction for 6 wk, osmotic minipumps (Alzet model 2006; Durect Corporation) were implanted into 12- to 16-wk-old male Bmx KO and WT C57BL/6J mice (n = 5 and n = 7, respectively), and sham-operated age- and gender-matched Bmx KO and WT C57BL/6J mice (n = 5 in both groups) were used for baseline controls. The analysis at the end point at 6 wk was performed as outlined above.

In the experiments using the previously described kinase-deficient Bmx K421R (equivalent to the human K445R mutation) (13) knock-in mice (C-BMX^{tm1(K421R)Npa} mice in the BALB/c background, hereafter referred to as Bmx TK mice) (14), the osmotic minipumps were implanted into 10- to 11-wk-old Bmx TK mice (n = 4), and the age-matched WT BALB/c mice (n = 5). Sham-operated Bmx TK– mice (n = 7), and WT BALB/c mice (n = 8) were used as baseline controls. Ang II infusion was continued for 8 d. The cardiac samples were analyzed and processed for histology as detailed above. The experiments described above were performed at least twice with essentially similar results.

The National Animal Board for Animal Experiments at the Provincial State Office of Southern Finland approved all animal experiments, which were performed in accordance with Finnish legislation regarding the humane care and use of laboratory animals.

Blood Pressure Analysis.

Blood pressures were measured before and after Ang II treatment with the tail cuff system in the experiments described above. For more accurate arterial blood pressure analysis, another set of animals (nine WT mice plus seven Bmx KO mice, 20 wk of age) were analyzed using a Millar solid-state micropressure transducer-tipped catheter. Mice were anesthetized with thiobutabarbital (Inactin), and catheters were inserted into the carotid artery. After a calibration period, blood pressure was continuously monitored for 24 min.

Human Cardiac Cells.

Arterial endothelial cells, microvascular endothelial cells, cardiomyocytes, and fibroblasts isolated from human hearts were purchased from PromoCell (C-14022, C-14029, C-14080, C-14036) . RNA was isolated from the cells, and the expression of Bmx was analyzed with qPCR as described above using three different housekeeping genes (Hprt1, beta-actin, and GAPDH).

Ang II–Bmx Signal Transduction in Endothelial Cells.

Confluent human umbilical venous endothelial cells were stimulated with Ang II (0.1 and 1 nM) for 10 min and 30 min. VEGF (100 ng/mL) was used as a positive control. After stimulation, the cells were washed briefly and lysed. The total protein concentrations were measured using the BCA Protein Assay Kit (Thermo Scientific). Samples were immunoprecipitated with mouse anti-human Bmx antibody (610793; BD Transduction Laboratories) and blotted with rabbit anti-human phospho-Bmx Y566 (ab59409; Abcam).

Isolated Adult Cardiomyocytes.

Adult mouse ventricular myocytes were obtained by enzymatic dissociation and cultured as described before (34) (AfCS Procedure Protocol PP00000125). Briefly, mice were injected with heparin, and, 20 min after the injection, mice were killed by cervical dislocation. Hearts were removed and placed in a Langendorff apparatus for perfusion (37 °C, 3 mL/min) with a trypsin (13.8 µg/mL) (Sigma) and liberase (12.5 µg/mL) (Roche Applied Science) solution. Ventricles were cut into small pieces, and cardiomyocytes were gently dissociated with a Pasteur pipette. Ca²⁺ concentration was then increased gradually up to 1 mM. After calcium reintroduction, cardiomyocytes were suspended in plating medium [minimum essential medium (catalog no. 21575-022; Gibco) supplemented with 5% (vol/vol) FBS, 10 mM 2,3-Butanedione monoxime, 100 U/mL penicillin, and 2 mM l-glutamine] and plated on laminin-coated culture plates. Cells were stored in a CO₂ incubator at 37 °C. After 1 h, plating medium was changed to culture medium [minimum essential medium (catalog no. 21575-022; Gibco) supplemented with 0.1 mg/mL BSA, 100 U/mL penicillin, and 2 mM l-glutamine) containing 150 ng/µL Ang II. Cardiomyocytes were exposed to Ang II for 15 min. For protein analysis with Western blot, cells were lysed with a solution containing 0.5% Triton X-100, 0.5% Nonidet P-40, phosphatase inhibitors (50 mM NaF, 5 mM Na₃VO₄), and protease inhibitors (10 µg/mL Aprotinin, 0.5 mM PMSF, 10 µg/mL Leupeptin) in PBS.

In Vivo Signal Transduction by Ang II.

Angiotensin II was injected i.v. into Bmx TK and WT mice, and hearts were harvested 10 min after injection. Protein lysates were analyzed for phosphorylated signaling proteins as described in Western Blot Analysis.

Silencing of Bmx in Human Umbilical Venous Endothelial Cells.

The human shBmx (TRCN0000006360, TRCN0000006361, TRCN0000006362) and scramble controls were from the TRC1 library (Sigma). The lentivirus-containing supernatant of 293FT cells was used to transduce HUVECs. Cells were incubated with virus-containing media together with polybrene (8 μg/mL) overnight, and transduced cells were selected with puromycin (2 μg/mL) for 1–2 d. Cells were starved for 3–4 h and stimulated with VEGF (80 ng/mL) or Ang II (150 ng/mL) for 10 min.

Western Blot Analysis.

Corresponding pieces of the heart from sham and Ang II-treated WT mice and Bmx KO mice after 2 wk of infusion or WT and Bmx TK⁻ mice after 10 min of Ang II stimulation (3 μg per mouse) were immersed into 1 mL of homogenization buffer [20 mmol/L Hepes, pH 7.4, 1 mM EDTA, 5 mM EGTA, 0.2% sodium deoxycholate, 10 mmol/L MgCl₂, 2 mmol/L DTT, 1% (vol/vol) Nonidet P-40, 3% (vol/vol) protease phosphate inhibitor mixture (Cell Signaling Technology), 1 mmol/L Na₃VO₄, and 100 mmol/L β-glycerophosphate] in Lysing Matrix tubes (MP Biomedicals) and homogenized. Cells were rinsed with PBS and lysed in radioimmunoprecipitation assay (RIPA) buffer after stimulation. The total protein concentrations were measured using the BCA Protein Assay Kit (Thermo Scientific). Lysates were boiled in Laemmli sample buffer (LSB). Equal amounts of total protein samples were separated by SDS/PAGE and transferred onto a PVDF membrane, which was then blocked and incubated with antibodies against p-rpS6 S240/244, p-Akt S473, pSTAT3, rpS6, and STAT3 (all from Cell Signaling) and p-VEGFR2 and VEGFR2 (R&D Systems). Probing with anti-GAPDH or anti–Hsc70 antibodies were used as total protein loading controls. Antibody complexes were visualized with a chemiluminescent substrate (Thermo Scientific) and quantified (band intensity × volume) using a ChemiDoc XRS device in combination with Quantity One software (version 4.6.3; Bio-Rad Laboratories) or Image J.

Immunofluorescence Staining.

Five- to 7-μm frozen sections were fixed with acetone. The sections were blocked and further permeabilized with a mixture containing 5% (vol/vol) normal donkey serum, 0.2% BSA, and 0.05% NaN3 in PBS-0.3% Trixon-X100 (Sigma-Aldrich) for 1 h at room temperature. The primary antibodies against dystrophin-1 (Novocastra Leica Biosystems) or a pan-endothelial blood vessel marker PECAM-1/CD31 (BD Biosciences) were applied overnight at 4 °C. The sections were washed repeatedly with PBS-0.3% TX and then incubated with Alexa Fluor-conjugated secondary antibodies (Molecular Probes/Invitrogen) for 1 h. After several PBS-0.3% TX washes, the samples were postfixed with 4% (vol/vol) paraformaldehyde for 5 min and mounted with Vectashield with DAPI (Vector Laboratories).

Histological Image Analysis.

The sections were viewed and imaged with an Axioplan2 fluorescence microscope (Carl Zeiss AG). The analysis was carried out using the ImageJ program (NIH).

Citrate Synthase Activity.

The activity of citrate synthase was analyzed from Ang II-treated WT mice and Bmx KO mice after 2 wk of infusion and from baseline WT mice and Bmx KO mice. Corresponding pieces of the heart were homogenized in ice-cold buffer [20 mM Hepes (pH 7.4), 1 mM EDTA, 5 mM EGTA, 10 mM MgCl₂, 100 mM, β-glycerophosphate, 1 mM Na₃VO₄, 2 mM DTT, 1% (vol/vol) Nonidet P-40, 0.2% sodium deoxycholate, and 3% (vol/vol) protease and phosphatase inhibitor mixture (P78443; Pierce Biotechnology)]. Homogenates were rotated for 30 min at 4 °C and centrifuged at 10,000 × g for 10 min at 4 °C to remove cell debris. Citrate synthase activity was then measured using a kit (CS0720; Sigma-Aldrich), following the instructions provided by the manufacturer, with an automated KoneLab device (Thermo Scientific).

RT-qPCR Analysis.

Comparable midcardiac tissue pieces were homogenized, and total RNA was isolated with the Trisure reagent (Bioline) and purified with a NucleoSpin RNA II kit (Macherey-Nagel GmbH). cDNA was synthetized from RNA using the reverse transcriptase kit (Thermo Scientific) according to the manufacturer’s instructions. Real-time PCR primer sequences are listed in Table S4. Quantitative PCR was performed with a real-time PCR (C1000 Thermal Cycler; Bio-Rad). TATA-binding protein (TBP) was used as a housekeeping gene control. Data analysis was carried out by using relative quantification with -ddCT formula (35).

Illumina Microarray and Microarray Data Analysis.

RNA samples were analyzed with genome-wide Illumina MouseWG-6 v2 Expression BeadChip (Illumina Inc.). The microarray hybridizations were performed according to the manufacturer's instructions. Illumina’s GenomeStudio software was used for initial data analysis and quality control. Detailed data analyses were performed with the Chipster software (chipster.csc.fi/). The data were normalized with quantile normalization. Statistically significant differences in individual genes between groups were tested using Empirical Bayes statistics and the Benjamini–Hochberg algorithm controlling false discovery rate (FDR). Adjusted values of P < 0.05 (FDR) were considered statistically significant.

Gene Functional Classification Analysis.

The clustering of differentially expressed genes into functional groups and their significance of overrepresentation among the groups were estimated with the DAVID functional annotation tool (https://david.ncifcrf.gov/home.jsp) (36, 37). Genes were clustered according to enrichment of Gene Ontology (GO) terms within differentially expressed genes (P < 0.05). The statistical significance of clusters was estimated by modified Fisher exact P value. Gene set enrichment analysis (GSEA, www.broadinstitute.org/gsea) was performed for the normalized dataset (38, 39).

Microarray Data Deposition.

The microarray data have been deposited in the National Center for Biotechnology Information gene expression and hybridization array data repository (GEO, www.ncbi.nlm.nih.gov/geo) with the accession number GSE47420.

Statistical Analysis.

Values are displayed as mean ± SEM in the figures. Statistical analysis of multiple groups was performed with one-way ANOVA, followed by Tukey’s post hoc test for groups with equal variances and by Games–Howell’s post hoc test for groups with unequal variances. Statistical analysis of two groups was performed with unpaired t test. All statistical tests were two-tailed. Differences were considered statistically significant at P < 0.05.