Aortic pathology from protein kinase G activation is prevented by an antioxidant vitamin B12 analog

People heterozygous for an activating mutation in protein kinase G1 (PRKG1, p.Arg177Gln) develop thoracic aortic aneurysms and dissections (TAAD) as young adults. Here we report that mice heterozygous for the mutation have a three-fold increase in basal protein kinase G (PKG) activity, and develop age-dependent aortic dilation. Prkg1R177Q/+ aortas show increased smooth muscle cell apoptosis, elastin fiber breaks, and oxidative stress compared to aortas from wild type littermates. Transverse aortic constriction (TAC)—to increase wall stress in the ascending aorta—induces severe aortic pathology and mortality from aortic rupture in young mutant mice. The free radical-neutralizing vitamin B12-analog cobinamide completely prevents age-related aortic wall degeneration, and the unrelated anti-oxidant N-acetylcysteine ameliorates TAC-induced pathology. Thus, increased basal PKG activity induces oxidative stress in the aorta, raising concern about the widespread clinical use of PKG-activating drugs. Cobinamide could be a treatment for aortic aneurysms where oxidative stress contributes to the disease, including Marfan syndrome.

: Telemetric blood pressure, ultrasound measurements of aortic diameters, and aortic ring myography in heterozygous Prkg1 RQ/+ mice compared to wild type (WT) littermates. (a-c) Diastolic blood pressure and heart rate measured by telemetry over 24 h in 4-month-old WT and Prkg1 RQ/+ male mice; pulse pressure was average during rest (7 am -7 pm) and activity (7pm-7am) periods (n=4 per genotype). ). (d,e) Isometric tension measurements performed in aortic rings of 8 monthold WT and Prkg1 RQ/+ female mice (n=5 per genotype) after pre-contraction to 0.15g with prostaglandin F2α. Dose-response curve for acetylcholine-induced relaxation (d), and % relaxation in response to 100 µM 8-(4-chorophenylthio)cGMP (e). (f) cGMP-induced PKG activity in aortic extracts was calculated as the difference between activity in the presence and absence of 3 µM cGMP (derived from Fig. 1a). (g-i) Diameters of the thoracic aorta measured by ultrasound as described in Fig. 1e,f. For separate analysis of 12 month-old females (g) and males (h), the 7 female and 6 male animals shown in Fig. 1f were combined with untreated animals shown in Fig. 4d (7 WT and 11 RQ/+ females; and 7 WT and 9 RQ/+ males). Panel i shows 4-to 6-month-old WT and Prkg1 RQ/+ mice of both genders (WT: n= 6M + 7F and Prkg1 RQ/+ mice: n= 7M+ 8F). (j) Media thickness measured in 12-month-old mice of both genders, using hematoxylin-and Van Giessen-stained cross-sections of the ascending aorta (n= 7M + 6F per genotype). *p<0.05 and **p<0.01 by two-sided t-test (panel f), by Mann-Whitney test (panels e,g,h), or by two-way ANOVA (panel d). O 2 production measured by Amplex Red fluorescence in human aortic SMCs infected with virus encoding GFP (control, black), wild type PKG1 (blue) or mutant PKG1 RQ (red); cells were re-plated 48 h after infection at a density of 2 x10 4 cells/cm 2 , and Amplex Red was added 16 h later. (c) NOX4 mRNA knockdown by NOX4-specific shRNA, measured by RT-PCR in human SMCs. (d) Effect of NOX2-specific shRNA on H 2 O 2 production in human SMCs infected with control or PKG1 RQ virus. NOX2 mRNA knockdown could not be reliably determined, because the amount of NOX2 mRNA in control SMCs was at the limit of detection. (e) JNK activation assessed in control and PKG1 RQ -expressing SMCs by Western blotting with phospho-specific antibody, as shown in Fig. 3h; some cells were treated with TGF-β for 2 h. (f) Protein aldehyde groups in control and PKG1 RQ -expressing human SMCs were assessed by OxyBlot TM as described in Fig. 2g. (g,h) Murine 10T1/2 mesenchymal cells were co-transfected with expression vector encoding PKG1 RQ , c-Jun, or empty vector, and a luciferase reporter under control of the murine Nox4 promoter. Cells were treated with vehicle (0.1% DMSO), TGF-β (3 ng/ml), the TGF-β receptor-1 inhibitor SB505124 (3 µM), or the JNK inhibitor (25 µM), as indicated. Luciferase activity was measured 24 h later and normalized to protein concentration. The graphs show means ± SEM of at least three independent experiments; *p<0.05, **p<0.01, ***p<0.001 for the indicated comparisons, and # p<0.05, ### p<0.001 for the comparison between SMCs expressing PKG1 RQ (or c-Jun) versus control cells receiving the same treatment. Panel c, two-sided t-test; panels g,h one-way ANOVA; panels a,d,e two-way ANOVA.

Supplementary Fig. 5: Decreased Proliferation and increased apoptosis in human SMCs expressing PKG1 RQ ; effect of NO/sildenafil-mediated activation of endogenous PKG1 on proliferation and JNK activity. (a-c)
Human SMCs were infected with control virus or virus expressing PKG1 RQ . Fourty-eight hours later, cell proliferation was measured in the absence and presence of fetal bovine serum, (FBS, 10%) or platelet-derived growth factor (PDGF, 30 ng/ml) by counting cells 72 h after re-plating (a). Apoptosis was assessed by cleaved caspase-3 immunofluorescence staining of cells cultured in the presence of 10% FBS or 0.1% FBS for 24 h (b), with representative images shown (c). (d,e) SMCs were cultured in media containing 10% FBS and were treated with 8-CPT-cGMP (100 µM), the NO-donor DETA-NONOate (3 µM), the PDE5 inhibitor sildenafil (100 µM) or both agents for 48 h, and BrdU uptake into S-phase cells was determined as described in Fig. 3l. (f) After 12 h in media with 0.5% FBS, cells were treated for 10 min with the agents described in panel d, and JNK and VASP phosphorylation were assessed by Western blotting with phospho-specific antibodies as described in Fig. 3h and 3b, respectively. Blots were quantified using LI-COR Odyssey with ImageStudio, V5. Graphs show means ± SEM of at least three independent experiments; *p<0.05, **p<0.01, ***p<0.001 for the indicated comparisons, and # p<0.05 and ### p<0.001 for the comparison between SMCs expressing PKG1 RQ versus control cells receiving the same treatment (two-way ANOVA for panels a,b and one-way ANOVA for panels d,f). Figure 6: Effects of Cbi on oxidative stress markers in SMCs in vitro and on apoptosis and collagen content in the aorta of Prkg1 RQ/+ mice. (a,b) Mice were treated with cobinamide (Cbi) as described in Fig. 4. SMC apoptosis was assessed by TUNEL staining (a, brown nuclei) as in Fig. 4c (80x). Collagen content (b, blue) of the media was quantified by ImagePro on Masson-Trichrome-stained cross-sections of the ascending aorta (40x with bar 50 µm; n=6-7 mice per group). (c) Human aortic SMCs were treated for 1h with H 2 O 2 (100 or 250 µM); some cells were pre-treated with Cbi (100 µM) for 1 h. JNK activation was assessed on Western blots using a phospho-specific antibody. In the bar graph, JNK phosphorylation was quantified by ImageJ and normalized to β-actin, with untreated control cells assigned a value of one (means ±SEM of three independent experiments). (d) Protein oxidation was assessed in SMCs treated with 100 µM H 2 O 2 for 16 h in the absence or presence of Cbi. Protein carbonyl groups were detected by OxyBlot TM after derivatization with 2,4-dinitrophenylhydrazine (DNPH); non-derivatized extracts served as control, blots were reprobed with a β-actin antibody. (e) SMCs were infected with control virus or virus encoding PKG1 RQ as described in Fig. 3j, some cells were treated with 10 µM Cbi for 48h. Protein oxidation was measured as in panel d. **p<0.01, ***p<0.001 for the indicated comparisons; # p<0.05 for the comparison between absence versus presence of Cbi, by 2-way ANOVA.

Supplementary Figure 7: Effect of Cobinamide and H 2 O 2 on PKG activity and PKG oxidation. (a)
Lack of effect of increasing Cbi concentrations on the basal activity of purified PKG1α (R177Q), measured as in Suppl. Fig. 1f, but in the absence of cGMP. (b) Eleven-month-old wild type and PKG1 RQ/+ mice were treated for one month with Cbi (1 mM) in the drinking water or were left untreated, and PKG activity was measured in aortic extracts in the absence and presence of cGMP, as described in Fig. 1a (n= 5 female mice per group; #p<0.05 for the comparison to basal PKG activity in untreated wild type mice, by 2-way ANOVA). (c) Body weight of 12-month-old mice described in Fig. 4, with males and females shown separately; some mice were treated with Cbi for 6 months, as described in Fig. 4. (d) Oxidation of Cys 43 in PKG1α assessed by gel shift assay under non-reducing conditions, with the cross-linked PKG1α dimer representing Cys 43 -oxidized enzyme. 4 The ratio of oxidized (dimeric)/reduced (monomeric) PKG1 present in aortic extracts from wild type and PKG1 RQ/+ mice was measured using LI-COR Odyssey with ImageStudio, V5. Some mice were treated with Cbi for one month as described in panel b (n= 4 female mice per group). (e) Human SMCs were treated for 1 h with 100-200 µM H 2 O 2 or were left untreated as indicated. Cell extracts were analysed by SDS-PAGE either under non-reducing conditions to determine Cys 43 oxidation in PKG1α as described in panel d, or under regular, reducing conditions to assess VASP phosphorylation as a measure of PKG activity. Note that PKG activity is not increased by H 2 O 2 despite an increase in the amount of oxidized PKG dimer. GAPDH served as a loading control, and the asterix indicates a contaminating band below the reduced PKG monomer.

Supplementary Figure 8: Aortic pressure gradients, aortic dilation, and hemothorax after TAC; effect of NAC on DNA oxidation and media collagen content. (a)
Aortic arch shown by ultrasound in Prkg1 RQ/+ mice before and 2 weeks after TAC surgery, using the same settings as shown for wild type mice in Fig. 5a (A=aortic root and B=ascending aorta; RPA, right pulmonary artery; LSA, left subclavian artery). (b) Invasive aortic blood pressure measurement 2 weeks after TAC in the ascending aorta (proximal) and descending aorta (distal to the constriction); some mice were treated with NAC as described in Fig. 5 (male mice only: n=17 WT untreated, n=8 Prkg1 RQ/+ untreated, n=11 WT treated with NAC, n=9 Prkg1 RQ/+ treated with NAC; ***p<0.001 for the indicated comparisons by 2-way ANOVA). (c) Representative images of PKG1 RQ/+ mice that died from aortic rupture within two weeks after TAC surgery: hemothorax on the left, hemothorax and hemopericardium on the right. (d) Ultrasound measurements of the aortic root diameter in WT and Prkg1 RQ/+ mice before and after TAC surgery (only mice surviving 14 d post TAC were included, as described in Fig. 5c; ***p<0.001 for the indicated comparisons by 2-way ANOVA). (e) Representative 8-OH-deoxyguanosine immunohistochemical stains of ascending aorta sections from the animals described in Fig. 5h (bars 50 µm, arrows show some examples of brown nuclei counted as positive). (f) Representative Masson-Trichrome stains of cross-sections of ascending aortas from the animals described in Fig. 5i, used to measure media collagen content (bars 50 µm).

SUPPL. FIG. 9 a b
Supplementary Figure 9: Inter-and intra-observer variations between aortic ultrasound measurements. Repeated measurements of ascending aortic diameters were performed by two independent observers (a) or by a single, highly experienced operator (b). Bland Altman analyses were performed to determine inter-and intra-observer variability of echo measurements. Measuring the diameter of the ascending aorta, the average of the differences between two observers was -0.061 mm (with 95% limits of agreement between -0.18 and 0.06 mm), and the average of differences between two observations by the same observer were 0.024 mm (with 95% limits of agreement between -0.12 and 0.17 mm).

Supplementary Table 1. Blood Chemistry
Eleven month old wild type and RQ/+ mice received 1 mM cobinamide in their drinking water or were left untreated for one month (n =5 female mice for each group). Mice were euthanized with an overdose of ketamine and xylazine, and blood was obtained by cardiac puncture. The blood urea nitrogen, creatinine, albumin, alanine transaminase, alkaline phosphatase, and total bilirubin were measured in a VetScan2 analyzer. Homocysteine was measured spectrophotometrically using a kit from Crystal Chem, Inc., and methylmalonic acid was measured by ELISA using a kit from Abbexa, Ltd. For all blood components, there was no significant difference by a two-tailed t test between wild type and RQ/+ samples, and between untreated and cobinamide-treated samples.