Glibenclamide reverses cardiovascular abnormalities of Cantu syndrome driven by K_ATP channel overactivity

Abstract

Cantu syndrome (CS) is a complex disorder caused by gain-of-function (GoF) mutations in ABCC9 and KCNJ8, which encode the SUR2 and Kir6.1 subunits, respectively, of vascular smooth muscle (VSM) K_ATP channels. CS includes dilated vasculature, marked cardiac hypertrophy, and other cardiovascular abnormalities. There is currently no targeted therapy, and it is unknown whether cardiovascular features can be reversed once manifest. Using combined transgenic and pharmacological approaches in a knockin mouse model of CS, we have shown that reversal of vascular and cardiac phenotypes can be achieved by genetic downregulation of K_ATP channel activity specifically in VSM, and by chronic administration of the clinically used K_ATP channel inhibitor, glibenclamide. These findings demonstrate that VSM K_ATP channel GoF underlies CS cardiac enlargement and that CS-associated abnormalities are reversible, and provide evidence of in vivo efficacy of glibenclamide as a therapeutic agent in CS.

Introduction

Cantu syndrome (CS) is a complex disorder with multiple cardiovascular abnormalities, including edema, dilated and tortuous blood vessels with decreased systemic vascular resistance, patent ductus arteriosus (PDA), and marked cardiac hypertrophy (1). CS is caused by gain-of-function (GoF) mutations in KCNJ8 and ABCC9, which encode pore-forming Kir6.1 and regulatory SUR2 subunits, respectively, of ATP-sensitive potassium (K_ATP) channels (2–12). These subunits are prominently expressed in smooth muscle (SM) cells, and vascular SM (VSM) K_ATP channel activation underlies the chronically dilated vasculature observed in patients with CS (13–17). Notably, Kir6.1 is not a major component of cardiomyocyte K_ATP channels (wherein the related Kir6.2 [KCNJ11] is the predominant pore-forming isoform [refs. 3, 18]) and so how CS-associated mutations in both KCNJ8 and ABCC9 result in cardiac hypertrophy is therefore unclear. We recently developed murine CS models in which disease-causing ABCC9 or KCNJ8 mutations were knocked-in to the equivalent mouse loci using CRISPR/Cas9. These animals exhibit increased VSM K_ATP channel activity and consequent chronic vasodilation, which we propose triggers systemic feedback mechanisms aimed at maintaining perfusion — including increased cardiac output and cardiomyocyte hypertrophy — in CS (19).

There are currently no targeted therapies for CS and it is not known if, or to what extent, cardiovascular abnormalities can be reversed once manifest. K_ATP channel inhibitors, including the sulfonylurea glibenclamide (glyburide), are used clinically to treat diabetes due to their inhibitory action on pancreatic K_ATP channels (formed of Kir6.2/SUR1). These drugs also inhibit cardiovascular K_ATP channels and thus may potentially be repurposed for the treatment of CS (20). In this study we thus sought to directly test the hypothesis that cardiac hypertrophy occurs secondary to K_ATP GoF in VSM, to investigate whether cardiac remodeling in CS is reversible, and to test the potential for glibenclamide treatment of cardiovascular abnormalities in Cantu mice.

Results and Discussion

Cardiovascular abnormalities in CS result from K_ATP channel GoF in VSM cells.

To directly test whether cardiac remodeling occurs as a secondary response to VSM K_ATP channel GoF, we crossed CS (SUR2^wt/AV) mice with animals expressing smooth muscle myosin heavy chain promoter-driven Cre-recombinase (SM-Cre) and dominant-negative KCNJ8 (Kir6.1-AAA) transgenes, allowing inducible suppression of K_ATP in smooth muscle of WT and CS mice (Figure 1A). Induction of expression at 8 weeks resulted in complete loss of K_ATP function, determined by whole-cell patch clamp recordings from isolated aortic myocytes (Figure 1, B and C). As previously reported (19), SUR2^wt/AV mice exhibit lower mean arterial pressure (MAP) than WT, and dominant-negative suppression of smooth muscle K_ATP on this CS background (in SM-DN^wt/AV mice) resulted in significant MAP elevation (Figure 1, D and E). Most strikingly, cardiac hypertrophy was essentially completely reversed in SM-DN^wt/AV mice 4 weeks after transgene induction (Figure 1F). These findings confirm a principal role for VSM K_ATP overactivity in the generation of cardiac hypertrophy. Importantly, they also show that cardiac hypertrophy can be reversed once manifest, and hence establish VSM K_ATP channels as appropriate molecular targets for pharmacological treatment of CS cardiovascular abnormalities.

Figure 1. Downregulation of VSM K_ATP overactivity abolishes cardiac hypertrophy.

(A) Transgenic approach to generate inducible, tissue-specific, dominant-negative Cantu mice (see text). (B) Representative whole-cell recordings of K_ATP channel activity in aortic SM cells from WT (left) and SM-DN^wt/wt mouse following tamoxifen induction (right). Cells were voltage-clamped at –70 mV and currents recorded in high-Na⁺ or -K⁺ as indicated. Pinacidil (Pin) and glibenclamide (Glib) were administrated as indicated. (C) K_ATP channel current density from experiments as in C. Data for VSM cells isolated from WT (black bar), SM-DN^wt/wt without tamoxifen induction (white bar), and SM-DN^wt/wt with tamoxifen administration (gray bar). (D) BP recordings from anesthetized WT (black), SUR2^wt/AV (orange), and SM-DN^wt/AV (brown) mice. (E) Mean arterial pressure (MAP) in nontransgenic (Non TG), single-transgenic (STG), and double-transgenic (SM-DN) WT and SUR2^wt/AV mice. (F) Left: Representative images of excised hearts from WT (top), SUR2^wt/AV (middle), and SM-DN^wt/AV (bottom) mice. Right: Heart size (heart weight normalized to tibia length; HW/TL) from nontransgenic (Non TG), single-transgenic (STG), and double-transgenic (SM-DN), WT and SUR2^wt/AV mice. For all figures, individual data points are represented as open circles, bars show mean ± SEM. Statistical significance was determined by 1-way ANOVA and post hoc Tukey’s test for pairwise comparison. *P < 0.05; **P < 0.01 from pairwise post hoc Tukey’s test.

Pharmacological reversal of CS-associated cardiovascular abnormalities in Cantu mice.

We next hypothesized that reversal might also be achieved by pharmacological inhibition of overactive VSM K_ATP channels. Mice were implanted with subcutaneous, slow-release pellets formulated to release a moderate or high dose (approximately 1 or approximately 19 mg/kg/day) of glibenclamide for 4 weeks, which resulted in measured plasma concentrations of 30 ± 8 ng/mL (approximately 60 nM) and 147 ± 51 ng/mL (approximately 300 nM), respectively. Cardiac hypertrophy was reversed in a dose-dependent manner (Figure 2A), almost completely at the highest dose, comparable to the effect of genetically induced VSM K_ATP downregulation in SM-DN^wt/AV mice (Figure 1). Consistent with an action on VSM K_ATP channels, high-dose glibenclamide elevated arterial pressure (MAP) and fully restored vascular resistance (SVR) in SUR2^wt/AV mice (Figure 2, B and C). Glibenclamide also induced a partial reversal of the elevated cardiac index observed in SUR2^wt/AV mice (Figure 2D). Hypertrophy in SUR2^wt/AV mice is not associated with significant fibrosis, and fibrosis was not induced by glibenclamide (Figure 2E). Glibenclamide induced no impairment of cardiac function as determined by echocardiographic measurements of ejection fraction (Figure 2F).

Figure 2. Glibenclamide reverses cardiac hypertrophy in SUR2^wt/AV mice.

(A) Left: Representative hearts from placebo-implanted WT (black), placebo-implanted SUR2^wt/AV (orange), and approximately 19 mg/kg/day glibenclamide pellet implanted SUR2^wt/AV (brown) mice. Right: Summary of heart size (weight normalized to tibia length; HW/TL) for WT and SUR2^wt/AV mice implanted with either placebo pellets (Glib = 0), or pellets releasing approximately 1 mg/kg/day and approximately 19 mg/kg/day. (B) Summary of MAP in anesthetized placebo-pellet (Glib = 0) and approximately 19 mg/kg/day glibenclamide pellet–implanted WT and SUR2^wt/AV mice. In all experiments, pellets were implanted at 8 weeks of age, and phenotypes were assessed 4 weeks later. (C) Systemic vascular resistance (SVR) and (D) cardiac index in placebo-implanted WT mice and placebo- or glibenclamide pellet–implanted SUR2^wt/AV mice. (E) Gomori-stained left ventricular free wall sections. Scale bars: 500 μm. (F) Ejection fraction of placebo-implanted WT mice and placebo- or glibenclamide pellet–implanted SUR2^wt/AV mice. Carotid artery compliance measurements from (G) placebo-implanted or approximately 19 mg/kg/day glibenclamide pellet–implanted WT and SUR2^wt/AV mice, or (H) WT, SUR2^wt/AV, and SMDN^wt/AV mice. Individual data points are represented as open circles, bars show mean ± SEM. Statistical significance was determined by 1-way ANOVA (A–F) and 2-way ANOVA (G and H) with subsequent post hoc Tukey’s test for pairwise comparison. *P < 0.05; **P < 0.01 from pairwise post hoc Tukey’s test. For G and H, color-coded statistical significance indicators are shown for comparison with placebo-implanted WT mice (black).

Notably, high-dose glibenclamide did not reverse the marked carotid diameter enlargement observed in SUR2^wt/AV mice (Figure 2G and ref. 19), and a similar resistance to reversal was observed in SMDN^wt/AV mice (Figure 2H). This suggests that vascular structural abnormalities may be relatively refractory to K_ATP inhibition, but that reversal of conduit vessel structural remodeling is not required to reverse cardiac remodeling.

High-dose glibenclamide induces only transient hypoglycemia in mice.

Glibenclamide is used clinically to treat diabetes, due to its inhibitory action on pancreatic Kir6.2/SUR1-dependent K_ATP channels (which exhibit markedly higher sensitivity than cardiovascular Kir6.1/SUR2 channels) (20). High doses, as required to reverse CS cardiovascular remodeling, will therefore also unavoidably inhibit pancreatic K_ATP channels and are thus naively predicted to increase insulin secretion and lower blood glucose (BG), a potentially important side effect that could limit clinical utility. As expected, fed BG was not different between WT and SUR2^wt/AV mice prior to pellet implantation (Figure 3A) and both low- and high-dose glibenclamide indeed significantly lowered BG on day 1 after implantation. However, BG returned to normal by approximately day 2 (Figure 3, B–E). Moreover, fasted BG was normal in mice that had received high-dose glibenclamide for over 30 days — evidence of long-term glycemic stability (Figure 3F). Transient, spontaneously resolving, hypoglycemic effects of chronic glibenclamide have been demonstrated before, and are explained by chronic downregulation of insulin secretion with continued K_ATP-inhibition (21). Consistent with this, a mild glucose intolerance phenotype was observed in high-dose–treated WT and SUR2^wt/AV mice (Figure 3, G and H). Notably, in a single human CS case thus far treated with glibenclamide, transient hypoglycemia only was also observed at initiation of glibenclamide treatment or dose escalation (22), and thus chronic hypoglycemia may not prove to be a significant complication for glibenclamide therapy in patients with CS.

Figure 3. Chronic high-dose glibenclamide induces only transient hypoglycemia.

(A) Summary of blood glucose levels in fed WT and SUR2^wt/AV mice on day 0 prior to pellet implantation. (B) Mean blood glucose measurements from WT mice implanted with placebo pellets (black diamonds, solid line; n = 6), approximately 1 mg/kg/day glibenclamide pellets (light gray circles, dotted line; n = 4), and approximately 19 mg/kg/day glibenclamide pellets (dark gray triangles, dashed line; n = 4). (C) Summary of blood glucose measurements for WT mice implanted with placebo pellets (black bars) or approximately 19 mg/kg/day glibenclamide pellets (gray bars) on day 0, 1, and 18. (D) Mean blood glucose measurements from SUR2^wt/AV mice implanted with placebo pellets (dark orange circles, solid line; n = 4), approximately 1 mg/kg/day glibenclamide pellets (light orange diamonds, dotted line; n = 7), and approximately 19 mg/kg/day glibenclamide pellets (brown squares, dashed line; n = 8). (E) Summary of blood glucose measurements for SUR2^wt/AV mice implanted with placebo pellets (orange bars) or approximately 19 mg/kg/day glibenclamide pellets (brown bars) on day 0, 1, and 18. (F) Fasted BG in mice which had been implanted with either placebo or high-dose glibenclamide more than 30 days prior. Glucose tolerance test data for WT (G) and SUR2^wt/AV (H) mice implanted with placebo or approximately 19 mg/kg/day glibenclamide pellets. For summary figures, individual data points are represented as open circles, bars show mean ± SEM. Statistical significance was determined by 1-way ANOVA and subsequent post hoc Tukey’s test for pairwise comparison. *P < 0.05 from pairwise post hoc Tukey’s test.

Glibenclamide-induced correction of low blood pressure in Kir6.1^wt/VM mutant mice.

Although the vast majority of patients with CS carry mutations in ABCC9 (SUR2), there are patients with mutations in the pore-forming Kir6.1 (KCNJ8) subunit. To examine the potential for glibenclamide therapy in such patients, we also implanted CS model Kir6.1[V65M] knockin mice (Kir6.1^wt/VM) (19) with high-dose glibenclamide pellets. This resulted in a significant although incomplete (approximately 13 mmHg) improvement of the otherwise severe hypotensive phenotype and an incomplete effect on heart size (Figure 4, A and B). The Kir6.1[V65M] mutation results in a drastic GoF of K_ATP channels and causes severe CS features in humans (7, 8, 19). Unlike the SUR2[A478V] mutation, which does not significantly affect glibenclamide sensitivity (23), the Kir6.1[V65M] mutation markedly decreases glibenclamide inhibition in recombinant channels (8), potentially explaining the incomplete reversal of CV abnormalities. Alternatively, incomplete reversal might reflect the more severe phenotype requiring longer administration times for reversal. In either case, the reduced efficacy of glibenclamide in Kir6.1^wt/VM, compared with SUR2^wt/AV, suggests that sulfonylurea treatment efficacy may depend on the severity of the underlying mutation, and underlines the importance of thorough understanding of the molecular consequences for personalized therapy.

Figure 4. Partial reversal of cardiovascular features by glibenclamide in Kir6.1^wt/VM mice.

(A) Summary of mean arterial pressure (MAP) in anesthetized placebo-implanted (Glib = 0) and approximately 19 mg/kg/day glibenclamide pellet–implanted WT and Kir6.1^wt/VM mice. (B) Summary of heart size (weight normalized to tibia length; HW/TL) for WT and Kir6.1^wt/VM mice implanted with either placebo pellets (Glib = 0) or pellets releasing approximately 19 mg/kg/day. For all figures, individual data points are represented as open circles, bars show mean ± SEM. Statistical significance was determined by 1-way ANOVA and post hoc Tukey’s test for pairwise comparison. **P < 0.01 from pairwise post hoc Tukey’s test.

Inhibition of VSM K_ATP channels as a strategy to treat CS.

The above results establish that cardiac remodeling in CS arises secondary to K_ATP channel GoF in VSM, and provide key preclinical evidence for in vivo efficacy of glibenclamide in treatment of CS. The link between VSM K_ATP channel GoF and cardiac hypertrophy is not yet established, but is likely to involve systemic feedback mechanisms that seek to normalize systemic perfusion in response to vasodilation. Directly inhibiting VSM K_ATP overactivity with K_ATP inhibitors may thus reverse both the primary vascular defect and these secondary features.

Excessive hair growth (hypertrichosis) is a defining CS feature (1), and is also mimicked by the K_ATP channel opener, minoxidil, which is used as a topical treatment for alopecia (24, 25). It is possible that K_ATP inhibition might alleviate hypertrichosis in patients with CS, and that topical administration of K_ATP inhibitors may be of cosmetic use for hair removal in the future (26). PDA is observed in most patients with CS and likely arises from the vasodilatory effect of excessive K_ATP activity in the DA after birth. PDA, which can be lethal without correction, is also present in approximately 1:2000 full-term births but in 20%–60% of premature births (27). K_ATP inhibitors may thus also prove useful for correction of PDA of various etiologies, an application that should be the subject of future study. Increased VSM K_ATP channel expression has been reported in septic shock, and previous animal studies suggested that K_ATP inhibition may also prove beneficial in treating endotoxic hypotension (28, 29), although acute glibenclamide treatment failed to reverse hypotensive shock in humans, despite inducing hypoglycemia (30, 31). Such studies illustrate the different sensitivity of pancreatic and cardiovascular K_ATP channels, and raise the question whether longer term and higher dose treatment might be necessary and appropriate for cardiovascular applications.

Potential adverse effects of high-dose glibenclamide, including actions in skeletal and cardiac muscle, as well as the drug sensitivity of specific CS mutations, require further study, and ideal therapy for CS may ultimately require an agent with much improved selectivity or potency for VSM Kir6.1/SUR2B channels. However, there is immediate need for a targeted therapy for CS, and the present findings clearly demonstrate the in vivo potential of glibenclamide for correcting CS cardiovascular abnormalities. Moreover, they suggest that the undesired glucose-lowering effects in nondiabetic animals are temporary, and may not therefore be prohibitive for the use of glibenclamide as a therapy in CS.

Methods

Mouse models.

CRISPR/Cas9 genome-edited SUR2^wt/AV and Kir6.1^wt/VM Cantu mice were previously reported (19) (see also Supplemental Methods; supplemental material available online with this article; https://doi.org/10.1172/JCI130571DS1). Dominant-negative Kir6.1-AAA mice were crossed with Cantu mice as illustrated in Figure 1A and described in detail in Supplemental Methods. Electrophysiological recordings of acutely isolated aortic smooth muscle cells, blood pressure measurements in anesthetized mice, echocardiographic analysis and heart weight measurements, Gomori stain, vascular compliance, and blood glucose measurements were made as described in Supplemental Methods. Plasma glibenclamide concentrations were measured by LC-MS/MS analysis using an ion trap mass spectrometer following the method described in detail in Supplemental Methods.

Study approval.

Mouse studies were performed in compliance with the standards for the care and use of animal subjects defined in the NIH Guide for the Care and Use of Laboratory Animals and were reviewed and approved by the Washington University Institutional Animal Care and Use Committee.

Statistics.

Statistical analysis was carried out with Microsoft Excel (Real Statistics Resource Pack software, www.real-statistics.com). Significance values were calculated using 1-way ANOVA and subsequent post hoc Tukey’s test for pairwise comparison. For carotid compliance measurements, where groups with 2 variables were compared, 2-way ANOVA with post hoc Tukey’s test was performed using GraphPad Prism 8 for OS X. A P value of less than 0.05 was considered significant. All values are expressed as mean ± SEM.

Author contributions

CM and CGN designed research studies. CM, YH, ZY, CMH, RC, TMH, and AK conducted experiments and acquired and analyzed data. CM and CGN wrote the manuscript. GVH and MSR provided advice. All authors critically reviewed the manuscript.

Version 1. 12/10/2019

In-Press Preview

Version 2. 02/17/2020

Electronic publication

Version 3. 03/02/2020

Print issue publication