|Home | About | Journals | Submit | Contact Us | Français|
Current therapies for diseases of heart muscle (cardiomyopathy) and aorta (aortopathy) include inhibitors of the renin-angiotensin system, β-adrenergic antagonists, and the statin class of cholesterol-lowering agents. These therapies have limited efficacy, as adverse cardiovascular events continue to occur with some frequency in patients taking these drugs. Although cardiomyopathy and aortopathy can coexist in a number of conditions (for example, Marfan’s syndrome, acromegaly, pregnancy, and aging), pathogenetic molecular links between the two diseases remain poorly understood. We reasoned that identification of common molecular perturbations in these two tissues could point to therapies for both conditions. Here, we show that deficiency of the transcriptional regulator Kruppel-like factor 15 (Klf15) in mice leads to both heart failure and aortic aneurysm formation through a shared molecular mechanism. Klf15 concentrations are markedly reduced in failing human hearts and in human aortic aneurysm tissues. Mice deficient in Klf15 develop heart failure and aortic aneurysms in a p53-dependent and p300 acetyltransferase–dependent fashion. KLF15 activation inhibits p300-mediated acetylation of p53. Conversely, Klf15 deficiency leads to hyperacetylation of p53 in the heart and aorta, a finding that is recapitulated in human tissues. Finally, Klf15-deficient mice are rescued by p53 deletion or p300 inhibition. These findings highlight a molecular perturbation common to the pathobiology of heart failure and aortic aneurysm formation and suggest that manipulation of KLF15 function may be a productive approach to treat these morbid diseases.
Cardiac and vascular smooth myocytes respond to stress through tightly orchestrated gene-regulatory pathways (1, 2). Dysregulation of these signaling pathways can drive pathologic tissue remodeling in the heart (cardiomyopathy) (1, 3, 4) and aorta (aortopathy) (5–8) and ultimately lead to organ failure. Progressive heart failure results in reduced blood flow to vital organs, fluid retention, and lethal cardiac rhythm disturbances. Deterioration of aortic integrity can lead to deadly complications, including aortic aneurysm formation (dilation of the aorta), rupture, and dissection (tearing of the aortic wall). Although cardiomyopathy and aortopathy can coexist in a number of conditions, including Marfan’s syndrome, acromegaly, pregnancy, and aging (7, 9–11), the pathogenetic molecular links between the two diseases are not known. We sought to identify common molecular perturbations occurring in these diseases of the two tissues. Given the residual morbidity and mortality associated with current treatment strategies for heart and vascular disease (7, 12, 13), the elucidation of such potential drug targets would be of clinical value.
Here, we show that deficiency of the transcriptional regulator Kruppel-like factor 15 (Klf15) occurs in human cardiomyopathy (3, 4) and aortic aneurysms (14) and that deficiency of Klf15 in mice (15) causes both cardiomyopathy and aortopathy in a p53-dependent and p300 acetyltransferase–dependent fashion.
We first screened failing human hearts (3, 4) and human aortic aneurysm samples (14) and found both tissues to be markedly deficient in Klf15 messenger RNA (mRNA) (Fig. 1, A and B). Klf15 is a zinc finger transcription factor expressed in cardiomyocytes and cardiac fibroblasts that can repress hypertrophic signaling (15, 16). Although it is also expressed in vascular smooth muscle cells (SMCs) (17), its role in vascular biology is unknown. We therefore hypothesized that Klf15 mRNA deficiency might be common to the pathogenesis of both cardiomyopathy and aortopathy. We first confirmed that Klf15 mRNA concentrations were also reduced in mice by using a well-established angiotensin II (AngII) infusion model to simultaneously stress the heart and vasculature (fig. S1, A and B) (8, 18, 19). Chronic AngII stimulation reduced Klf15 mRNA expression in the heart and aorta of mice in vivo (Fig. 1, C and D) and in cultured cardiomyocytes and vascular SMCs (fig. S1C). Klf15 mRNA concentrations were also reduced in another model of AngII-mediated cardiomyopathy in which the angiotensin type I receptor (AT1R) is overexpressed in a cardiac-specific fashion (20) (Fig. 1E). Together, these data demonstrate that Klf15 mRNA expression is significantly reduced in both human and rodent cardiomyopathy and aortopathy.
Given this expression pattern, we hypothesized that Klf15 deficiency might cause cardiomyopathy and aortopathy. To test this hypothesis, we studied mice with germline deficiency of Klf15 (15) at baseline and after chronic AngII infusion. At baseline, Klf15−/− mice have mild pathologic remodeling of the heart [mild left ventricular (LV) hypertrophy with preserved systolic function; fig. S2, A to D, and table S1] and aorta (mild medial hypertrophy; fig. S2E). However, after AngII infusion, Klf15−/− mice develop a severe cardiomyopathy characterized by cardiomegaly, LV dysfunction, and cavity dilation with attenuated wall thickening (Fig. 2, A to C, and table S2). These pathologic changes were associated with augmented cardiac mass (fig. S3A). This phenotype was not due to differential blood pressure responses between wild-type and Klf15−/− mice (fig. S3B).
The aorta of Klf15−/− mice was also more sensitive to AngII than those of control mice. Thirty-five percent of Klf15−/− mice developed grossly visible aortic aneurysms (72% abdominal aorta, 28% thoracic aorta), whereas no aneurysms were observed in wild-type controls (P = 0.0003, n = 31 per group) (Fig. 2D). A subset of these aneurysmal segments had intramural hematoma formation confirmed by histologic examination (Fig. 2E). No intramural hematoma or dissections were seen in wild-type aortas. The non-aneurysmal portions of Klf15−/− aortas also had diffuse abnormalities, many of which have also been seen in the pathogenesis of AngII-induced aortopathy (19, 21, 22). We observed a severe disruption of medial architecture characterized by prominent elastolysis (Fig. 2F), increased SMC apoptosis (fig. S4, A and B), and reduced medial thickness (Fig. 2G). In the tunica adventitia, Klf15−/− mice showed a blunted adventitial growth response to AngII (fig. S4C). These medial and adventitial abnormalities were associated with generalized aortic dilation (Fig. 2G). In addition, the SMC of Klf15−/− mice elaborated excessive matrix metalloproteinase-3 (MMP-3) (fig. S5, A and B), an enzyme implicated in aneurysm formation (23). Together, these data demonstrate that Klf15−/− aortas are highly sensitive to AngII and develop a diffuse aortopathy involving the tunica media and adventitia, with significantly increased frequency of aneurysm and intramural hematoma formation.
Given the parallel phenotypes of organ decompensation in the heart and aorta in Klf15−/− mice, we hypothesized that Klf15 deficiency may cause common signaling perturbations in both tissues. We focused attention on p53 as it is activated with AngII stimulation (24), plays a critical role in cardiac decompensation (18), and has been associated with aneurysm pathogenesis (6). Analysis of mouse heart and aortic tissue revealed a substantial accumulation of p53 in Klf15−/− mice after AngII infusion (Fig. 3, A and B). In addition, direct targets of p53, such as the proapoptotic protein Bax and the potent antiangiogenic and proapoptotic factor thrombospondin-1 (TSP-1) (25), were also significantly elevated in both the heart and the aorta (Fig. 3, A and C). The increase in TSP-1 was predominantly seen in the tunica adventitia and is consistent with our finding that adventitial growth is blunted in Klf15−/− aortas after AngII infusion. There was no significant difference in p53 mRNA concentrations in these tissues (fig. S6, A and B), suggesting that its accumulation was a result of posttranslational regulation. Consistent with this in vivo data, KLF15 overexpression in cultured cardiomyocytes potently inhibited basal and AngII-mediated induction of TSP-1 (Fig. 3D), suggesting that KLF15 may inhibit expression of important p53 transcriptional targets. Because p53 accumulation in the heart is causally implicated in the transition from compensated hypertrophy to heart failure through its ability to inhibit myocardial angiogenesis (18), we quantified myocardial capillary density. AngII increases capillary density in wild-type hearts by 30%, whereas it has no angiogenic effect in Klf15−/− hearts (Fig. 3E). The diffuse vessel dilation and blunted medial thickening in Klf15−/− aortas after AngII stimulation were also associated with a paucity of adventitial angiogenesis (Fig. 3F).
To establish a causal role for p53 in the cardiomyopathy and aortopathy of Klf15−/− mice, we backcrossed into the p53−/− strain (18, 26) and assessed for rescue. Klf15−/− p53−/− mice were born in Mendelian ratios and viable into adulthood. As hypothesized, p53 deficiency rescued AngII-induced systolic dysfunction, chamber dilation, and wall thinning in Klf15−/− hearts (Fig. 4, A and B). p53 deficiency also ameliorated the aortopathy in Klf15−/− mice as evidenced by improved medial integrity (Fig. 4C). Furthermore, the p53-dependent rescue of these pathologic abnormalities was associated with attenuation of TSP-1 abundance (Fig. 4D). There was also rescue of the angiogenic defect in Klf15−/− hearts, with a significant augmentation of myocardial capillary density in Klf15−/− p53−/− relative to Klf15−/− p53+/+ mice (Fig. 4E). p53 deficiency had no differential effect on cardiac mass, consistent with its promotion of myocardial angiogenesis and compensated cardiac growth (18) (table S3). Consistentwith the antiangiogenic role of p53 in the heart (18), endothelial staining of aortic sections revealed that p53 deficiency normalized capillary density and adventitial growth in Klf15−/− mice (Fig. 4F).
We next explored how KLF15 might regulate p53 function. In Klf15−/− tissues, the increased abundance of p53 protein occurs in the absence of p53 mRNA accumulation (Fig. 3, A and B, and fig. S6). Because p53 function is tightly controlled by an intricate network of posttranslational modifications and protein interactions (27, 28), we reasoned that KLF15 might alter this aspect of p53 regulation. KLF15 overexpression did not induce Mdm2 or Stub1 (two E3 ubiquitin ligases that can inactivate p53) (27, 29) in cultured cells, and neither protein was reduced in Klf15−/− hearts. KLF15 overexpression in cultured cardiomyocytes caused an attenuation of basal and AngII-induced p53 acetylation on Lys379 (equivalent to Lys382 in humans) (Fig. 5A), a residue that is specifically modified by the acetyltransferase p300 (28). KLF15 overexpression did not alter phosphorylation of p53 at Ser15, a modification also important for p53 activation (27) (Fig. 5A). Consistent with our overexpression data, Klf15−/− hearts (Fig. 5B) and aortic SMCs (Fig. 5C) had significantly augmented abundance of acetylated p53 (Lys379). Finally, we found that substantial p53 hyperacetylation also occurs in human disease, as failing human hearts showed an increased abundance of acetylated p53 (Lys382) (Fig. 5D).
The acetyltransferase p300 is an important regulator of p53 function via its ability to acetylate p53 at multiple lysine residues, a modification critical for p53 transcriptional activity and protein stability (27, 28). We reasoned that the KLF15-mediated inhibition of p53 acetylation might reflect its ability to disrupt p300 function. However, KLF15 overexpression did not reduce p300 abundance or acetyltransferase activity in cultured cardiomyocytes. Furthermore, there was only a modest effect on nuclear p300 accumulation in Klf15−/− hearts (fig. S7A) or in failing human hearts (Fig. 5D).
Previous studies demonstrate a causal role for excessive p300 activity in heart failure (30, 31), whereas its role in aortopathy is unknown. We hypothesized that when Klf15 is deficient, p53 hyperacetylation reflects an enhanced state of p300 function that leads to cardiovascular decompensation. As such, targeting p300 function might rescue the cardiomyopathy and aortopathy observed in Klf15-deficient mice. Because germline deficiency of p300 results in early embryonic lethality (32), we tested this hypothesis by using the drug curcumin, a potent p300 acetyltransferase inhibitor (31, 33). Mice were pretreated with curcumin for 7 days, followed by a 14-day AngII infusion with continued curcumin therapy (fig. S7B). Curcumin significantly ameliorated decompensated heart failure in Klf15−/− mice, with a reduction in cardiac mass (Fig. 6, A and B), improvement in LV systolic function (Fig. 6, C and D), attenuation of eccentric remodeling (Fig. 6, C and D, and table S4), and decrease in acetylated p53 abundance in tissues (Fig. 6E). Finally, we examined the aorta to assess whether curcumin could also rescue the aortopathy in Klf15−/− mice and found that curcumin caused a near normalization of elastolysis (Fig. 6F) and restoration of medial thickness (fig. S7C) without any significant effect on mean vessel diameter (fig. S7C).
Our central finding is that deficiency of the transcriptional regulator Klf15 leads to cardiomyopathy and aortopathy via a shared mechanism in which there is derepression of p53 and p300 acetyltransferase. We further demonstrate that, in the setting of Klf15 deficiency in mice, inhibition of this pathway can normalize these abnormalities in vivo. These conclusions are based on several observations. We first show that Klf15 expression is reduced in both human and murine cardiomyopathy and aortopathy. Klf15 deficiency can result in these disease phenotypes, as Klf15−/− mice develop severe heart failure and aortopathy in response to stress. KLF15 inhibits p53 function by reducing abundance of acetylated p53. Conversely, Klf15 deficiency leads to p53 hyperacetylation and enhanced p53 effector functions, findings that are recapitulated in human disease. A causal role for this pathway is established by the observation that both genetic ablation of p53 and pharmacologic inhibition of p300 acetyltransferase with curcumin can both ameliorate the cardiovascular abnormalities seen in Klf15−/− mice. Thus, deficiency of a single transcription factor can exert detrimental effects in the heart and vasculature through a unifying molecular mechanism.
On the basis of these observations, we propose the model shown in Fig. 6G. Pathologic stressors such as AngII down-regulate KLF15 and promote p53 accumulation (24). Excessive p53 accumulation leads to diminished angiogenesis, apoptotic cell loss, decompensated heart failure (18, 24), and aneurysmal aortopathy. These diseases are accelerated in Klf15−/− mice in a p53-dependent manner. KLF15 can inhibit p53 function via interfering with p300-mediated acetylation. As such, we propose that KLF15 is a molecular brake on p53 and that genetic deficiency of Klf15 leads to unbridled p53 activity in response to stress. The partial reduction of KLF15 concentrations in the heart and aorta in response to AngII might represent a molecular program that initially facilitates compensated forms of tissue remodeling. However, sustained reduction of KLF15 or its complete absence may sensitize both tissues to grossly decompensate in response stress. The fact that excessive p300 acetyltransferase activity also contributes to this phenotype implicates hyperacetylation of p53 as a critical component of this maladaptive signaling pathway. We recognize, however, that p300 can acetylate other factors that have important roles in pathologic cardiovascular remodeling, such as GATA4, MEF2, histones, and Smad (30, 31, 34). Indeed, KLF15 might also disrupt the ability of p300 to acetylate and coactivate these factors, suggesting how it might exert its potent repressive effect on multiple transcriptional pathways (2, 15). In addition to its ability to disrupt p300-mediated acetylation of p53, KLF15 might simultaneously facilitate the ability of p53 to interact with deacetylases (27) or other inhibitory proteins (27, 35). Such possibilities are consistent with our model and may contribute to selective regulation of p53 effector functions (27).
Recent studies have elegantly elucidated an intricate network of AngII–TGF-β (transforming growth factor–β) cross talk that is central to the pathogenesis of heart failure (36) and aortopathy (5, 37, 38). Our model has numerous points of intersection with the TGF-β signaling pathway. We show that Klf15−/− mice have exaggerated expression of TSP-1, a well-characterized transcriptional target of p53 (25, 39) that is an endogenous activator of latent TGF-β (40) and one of the most potent antiangiogenic proteins in mammalian biology (41). The observation that p53 deficiency attenuates TSP-1 concentrations in tissues of Klf15−/− mice confirms that its maximal induction is p53-dependent in vivo. Because p53-mediated inhibition of angiogenesis causes cardiac decompensation (18), TSP-1 might be an important downstream mediator of this effect. Although the role of angiogenesis in aortic disease is less well established, we find that the aortopathy in Klf15−/− mice is associated with a p53-dependent attenuation of adventitial angiogenesis. These observations raise the possibility that localized angiogenic insufficiency in the tunica adventitia might contribute to aortopathy and are consistent with a recent study implicating pathologic changes in the adventitia as contributors to the progression of aortic disease (22). In addition to its antiangiogenic capabilities, TSP-1 can markedly potentiate TGF-β signaling at the cell surface by its ability to activate latent TGF-β (40). The final point of intersection between the two pathways is in the nucleus, where p300 and p53 can both cooperate with Smads and potentiate Smad-mediated transcription (42). Definition of the role of Klf15 as a regulator of the AngII–TGF-β axis in vivo may provide insight into the pathogenesis of Marfan’s disease and related disorders (5, 7, 38).
The current study extends our previous observations (15, 16) and implicates Klf15 as a central regulator of stress responses in both the heart and the vasculature. In addition, we identify the Klf15−/− mouse as a model for aortic aneurysm formation in which lesions develop at modest AngII doses without requiring an ApoE−/− background (19). On the basis of our observations in humans and rodents, we propose that a subset of human heart and vascular disease may be due to KLF15 deficiency and that the Klf15-p53-p300 pathway is a potential therapeutic target for cardiovascular disease.
All protocols concerning animal use were approved by the Institutional Animal Care and Use Committee at Case Western Reserve University and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Klf15−/− mice (15) (10 weeks old, male) and littermate controls in a pure C57Bl/6 background were used for all studies. p53−/− mice (18, 26) were from The Jackson Laboratory. For infusion studies, mice were anesthetized with ketamine (1.5 mg per kilogram of body weight)–xylazine (0.3 mg/kg). Mini-osmotic pumps (Alzet 2002, Durect) were filled with AngII (Sigma) or vehicle (normal saline) and implanted subcutaneously. AngII was infused at a dose of 1.4 mg/kg per day for 14 days, after which mice were killed for further analysis. Curcumin (Sigma) was freshly suspended in 0.5% carboxymethylcellulose (Sigma) and administered to mice via oral gavage at a dose of 100 mg/kg per day divided into two daily doses. For perfusion fixation, the thorax was opened, a 21-gauge needle was placed into the left ventricle, and the inferior vena cava was severed. The animals were perfused with normal saline until the perfusate cleared and then with 4% paraformaldehyde at 100 mmHg for 7 min. RNA samples from AT1R-transgenic and nontransgenic mice were gifts from J. Sadoshima (20).
For transthoracic echocardiography, mice were anesthetized with 1% inhalational isoflurane and imaged using the Vevo 770 High Resolution Imaging System (Visual Sonics) and the RMV-707B 30-MHz probe. Measurements were obtained from M-mode sampling and from integrated EKV images that were taken in the LV short axis at the midpapillary level (15). Conscious systolic blood pressure was measured using the BP2000 Blood Pressure Analysis System (Visitech Systems) as recommended by the manufacturer. To allow mice to adapt to the apparatus, we performed daily blood pressure measurements for 1 week before beginning experiments. Blood pressure was recorded at days 0 and 14 of the infusions.
LV samples from nonfailing control (n = 30) and nonischemic cardiomyopathy (NICM) (n = 36) subjects were obtained as described (3, 4) in accordance with the Investigation Review Committee at the Hospital of the University of Pennsylvania (Philadelphia, PA). Aneurysmal tissue was collected from patients undergoing abdominal aortic aneurysm repair operations as described (14). Control abdominal aortic specimens were matched to cases using sex, age, and ethnicity as described (14). All aortic specimens were obtained by protocols approved by the Institutional Review Board of Wayne State University (Detroit, MI). RNA was isolated from each sample as described (3, 4, 14). Taqman quantitative polymerase chain reaction (QPCR) was used to quantify expression of Klf15 with Applied Biosystems Assays-on-Demand Hs00362736_m1 using the ΔΔCt method as described (3, 4).
Neonatal rat ventricular myocytes (NRVMs) were isolated from the hearts of rat pups as described (15). Primary mouse aortic SMCs (MASMCs) were cultured from aortas of wild-type and Klf15−/− mice as described (43). Rat aortic SMCs (RASMCs; used at passages 3 to 7), human embryonic kidney 293, and HCT116 cells were obtained from the American Type Culture Collection. NRVMs and RASMCs were placed in serum-free medium before stimulation with AngII (1 µM). Adenovirus (Ad-GFP versus Ad-KLF15-GFP) has been described (15). NRVMs were infected at 50 multiplicities of infection (MOIs) for 24 hours and RASMCs at 100 MOIs for 24 hours.
Immunoprecipitation histone acetyltransferase (HAT) assays were performed using the p300 IP HAT Assay kit (Upstate-Millipore) according to the manufacturer’s instructions.
Total RNA was isolated from cells or tissues using Trizol (Invitrogen). Northern blot analysis for Klf15 was performed as described (15). For QPCR, total RNA was deoxyribonuclease-treated (Invitrogen) and transcribed to complementary DNA with oligo(dT) or random hexamer priming (Invitrogen) and Moloney murine leukemia virus reverse transcriptase (New England Biolabs). QPCR was performed with the SYBR Green method on a Stratagene Mx3005P system. Relative expression was calculated using the ΔΔCt method with normalization to constitutive genes as indicated in the figure legends. Specific primer sequences are available on request.
Total cellular or tissue homogenates were processed in radioimmunoprecipitation assay buffer (Sigma) supplemented with protease or phosphatase inhibitors (Roche). Nuclear protein was isolated using a Nuclear Extraction kit (NE-Per kit, Pierce) according to the manufacturer’s instructions. Ten to 40 µg of whole protein extracts or 20 µg of nuclear extracts was subjected to Western blot analysis. Antibodies against p53 (DO-1), p300 (N-15), β-actin (AC-15), myc (9E10), and Ku70 (E-5) were from Santa Cruz Biotechnology. Antibodies against acetyl p53 (Lys379 and Lys382), phospho-p53 (Ser15), and Bax were from Cell Signaling Technology. Antibody against TSP-1 (Ab4, clone A6.1) was from NeoMarkers–Thermo Fisher Scientific. Antibodies against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and FLAG (M2) were from Sigma. Antibody against Mdm2 (Ab-3, clone 4B11) was from Calbiochem–EMD Biosciences. Stub1 polyclonal antibody serum was a gift from C. Patterson (University of North Carolina, Chapel Hill, NC) (29).
Primary MASMCs were serum starved for 24 hours, followed by stimulation with AngII (1 µM, 6 hours). Supernatants were subjected to enzyme-linked immunosorbent assay (ELISA) for MMP-3 using the SearchLight Proteome Array or multiplex sandwich ELISA (Pierce Biotechnology).
Perfusion-fixed ventricles and aortas were embedded in paraffin and cut into 5-µm-thick sections. Medial thickness and area were quantified using Image-Pro Plus software (Media Cybernetics). Trichrome staining of heart tissue was performed using the One-Step Trichrome kit (Biocare Medical) according to the manufacturer’s instructions. Elastin staining of aortic sections was performed on perfusion-fixed aortic cross sections (5 µm thick) using Elastin Stain kit (Sigma) according to the manufacturer’s instructions. Quantification of elastin degradation was a direct adaptation of the semiquantitative aortic wall architecture score (5). An elastin break was defined as a distinct discontinuity in a singular concentric lamellar ring of elastin. The total number of elastin breaks in a vessel cross section was counted as a single data point. The aortic sections being compared were from the same aortic level for each mouse. Measurements were performed by two independent and blinded observers. Vessel diameter was quantified from elastin-stained circular aortic sections. We defined diameter as the mean distance between internal elastic laminae on diametrically opposite sides of the vessel ring. TSP-1 immunostaining was performed using antibody against TSP-1 (Ab4, clone A6.1). Antibody against MMP-3 for immunostaining was from Abcam. To assess capillary density in aortic cross sections, we performed CD34 antibody immunofluorescence staining (Chemicon) on fresh-frozen aortic sections using Alexa Fluor 488–labeled secondary antibody and 4′,6-diamidino-2-phenylindole (DAPI) counterstaining. All comparisons were made between sections cut from the same aortic level.
Terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining was performed on perfusion-fixed sections (8 µm thick) of heart and aortas using the ApopTag Plus kit (Chemicon) according to the manufacturer’s instructions. MASMCs were stimulated with AngII (1 µM) for 72 hours and assessed for TUNEL using the In Situ Cell Death Detection kit (Roche) according to the manufacturer’s instructions. DAPI counterstaining and mounting was done with Vectashield H-1500 mounting medium (Vector Labs). TUNEL-positive nuclei per section were normalized to total nuclei using Image-Pro Plus software.
Fresh-frozen sections of mouse LV were cut into 8-µm-thick sections, fixed in 100% ethanol, and dual-stained for platelet endothelial cell adhesion molecule–1 (PECAM-1) and wheat germ agglutinin (WGA) as described (18, 44). Antibody against PECAM-1 was from Millipore and used at 1:100 dilution. Secondary antibody (antibody against rat immunoglobulin labeled with Alexa Fluor 488) was from Invitrogen and used at 1:800 dilution. WGA-rhodamine (5 µg/ml; Vector Labs) counterstaining was applied after immunostaining. Random high-power fields (HPFs) (400×) were imaged using a fluorescent microscope. Intramyocardial capillaries and myocytes were counted and quantified as described (18, 44).
All data are presented as the mean ± SEM. For continuous variables, the statistical significance of differences between two groups was analyzed with two-tailed unpaired Student’s t test. The statistical significance for the frequency of occurrence of a binary variable (that is, presence or absence of aortic aneurysm) was analyzed using χ2 test of association and two-tailed Fisher’s exact probability test. Values of P < 0.05 were considered significant.
Fig. S1. Cardiovascular abnormalities in AngII-treated mice and cultured cells.
Fig. S2. Baseline abnormalities in Klf15−/− heart and aorta.
Fig. S3. Cardiac mass and systolic blood pressure after AngII infusion.
Fig. S4. Histologic parameters in aortas.
Fig. S5. MMP-3 abundance in aortic smooth muscle.
Fig. S6. p53 mRNA concentrations in heart and aortic tissue.
Fig. S7. p300 abundance in hearts, curcumin administration protocol, and aortic morphometry after curcumin therapy.
Table S1. Baseline cardiac parameters in Klf15−/− and wild-type mice.
Table S2. Cardiac parameters in Klf15−/− and wild-type mice after AngII infusion.
Table S3. Cardiac parameters in Klf15 or p53 compound mutant mice after AngII infusion.
Table S4. Cardiac parameters in curcumin-treated Klf15−/− and wild-type mice after AngII infusion.