Cardiac ion channel modulation by the hypoglycaemic agent rosiglitazone

Abstract

The hypoglycaemic thiazolidinedione rosiglitazone is used clinically in the treatment of type 2 diabetes. However, in 2010, information relating to rosiglitazone-associated increased cardiovascular risk led the European Medicines Agency to recommend suspension of marketing authorizations for rosiglitazone-containing anti-diabetes drugs, while the US Food and Drug Administration recommended significant restriction on the agent's use. Two timely studies in this issue of the British Journal of Phrarmacology provide new information regarding modification of cardiac cellular electrophysiology by rosiglitazone. Szentandrássy et al. demonstrate canine ventricular action potential modification and concentration-dependent suppression of L-type Ca current and of transient outward and rapid delayed rectifier K currents. Jeong et al. demonstrate concentration-dependent inhibition of recombinant K_v4.3 channels, providing mechanistic insight into the likely molecular basis of transient outward K current inhibition by the compound. Further studies using diabetic models would be of value to determine whether, in a diabetic setting, rosiglitazone modification of these channels could affect the risk of arrhythmia at clinically relevant drug concentrations.

LINKED ARTICLES

This article is a commentary on Szentandrássy et al., pp. 499–509 of this issue and Jeong et al., pp. 510–520 of this issue. To view these papers visit http://dx.doi.org/10.1111/j.1476-5381.2011.01215.x and http://dx.doi.org/10.1111/j.1476-5381.2011.01210.x

Rosiglitazone and pioglitazone, which are thiazolidinediones and high-affinity ligands for the peroxisome proliferator-activated receptor γ, have been used widely as hypoglycaemic agents in the treatment of type 2 diabetes (McGuire and Inzucchi, 2008; Doshi et al., 2010). Due to a propensity to increase peripheral oedema and the risk of heart failure, the use of these agents in patients with heart failure is to be avoided (McGuire and Inzucchi, 2008; Doshi et al., 2010; European Medicines Agency, 2010), while rosiglitazone has been the subject of specific attention in recent years, due to debated evidence for increased cardiovascular event risk.

In 2007, two meta-analyses concluded that sustained rosglitazone use was associated with a significantly increased risk of myocardial infarction (MI) (Nissen and Wolski, 2007; Singh et al., 2007). The subsequent multi-centre open-label ‘RECORD’ study confirmed an increased risk of heart failure with rosiglitazone, but did not show a conclusive effect regarding MI risk, and suggested that rosiglitazone does not increase cardiovascular mortality/morbidity compared with other glucose-lowering drugs (Home et al., 2009). A recent study of rosiglitazone-receiving patients showed no adverse effects on markers of cardiac function examined using cardiac magnetic resonance imaging (McGuire et al., 2010). However, in 2010, two further risk analyses raised significant questions regarding the continued use of rosiglitazone-containing medications in type 2 diabetes. One, focusing on elderly diabetic patients receiving either rosiglitazone or pioglitazone who had undergone 3 years follow-up after thiazolidinedione initiation, reported increased risk for rosiglitazone compared with pioglitazone of stroke, heart failure, all-cause mortality and of the composite of acute MI, stroke, heart failure or all-cause mortality (Graham et al., 2010). The second was an updated meta-analysis of randomized controlled trials up to February of 2010 of at least 24 weeks duration reporting adverse cardiovascular events; this study confirmed increased risk with the rosiglitazone treatment of MI, though not of increased cardiovascular or all-cause mortality (Nissen and Wolski, 2010). Following a review of rosiglitazone by the Committee for Medicinal Products for Human Use concluding that the benefits of rosiglitazone no longer outweighed its risks, in September 2010 the European Medicines Agency recommended suspension of marking authorization for rosiglitazone-containing medicines (European Medicines Agency, 2010). In the same month, the US Food and Drugs Administration recommended restricting rosiglitazone to those patients who are unable to control their type 2 diabetes with other medicines (FDA, 2010).

Against such a background, it is perhaps timely that two studies published in the present volume of this journal provide new information regarding cardiac actions of rosiglitazone, specifically on the ability of the drug to inhibit ion channels that are important in the generation of normal cardiac activity. Szentandrássy et al. (2011) have focused on establishing the effects of the drug on ventricular action potentials and major ventricular ionic currents from a large (canine) animal model; Jeong et al. (2011) have focused on studying recombinant K_v4.3 channels, as this K_v channel sub-type is involved in mediating transient outward potassium current (I_TO) that contributes significantly to the early phase of repolarization of ventricular action potentials (see Dixon et al., 1996).

In studying dog ventricular myocytes, Szentandrássy et al. (2011) have chosen a species that exhibits closer similarities to human in ventricular action potential configuration and repolarizing potassium channel currents than small rodents do (mice or rats). They report a concentration-dependent reduction in action potential (AP) maximum upstroke velocity (V_max; significant at drug concentrations of 10 µM or greater), which represents indirect, though likely reliable, evidence for an inhibitory effect of rosiglitazone on fast Na current (I_Na). A similar concentration-dependent suppression of phase 1 repolarization was observed (Szentandrássy et al., 2011). By contrast, AP duration (APD) was comparatively little affected by rosiglitazone, except for a small (though significant) shortening of APD at 50% repolarization by 30 µM of the drug, and a lengthening of APD at 90% repolarization with 100 µM rosiglitazone (Szentandrássy et al., 2011). Whole-cell recordings of I_TO, of rapid delayed rectifier K current (I_Kr) and of L-type Ca current (I_Ca,L) revealed concentration-dependent inhibition of each of these currents with half-maximal effective concentrations (EC₅₀s), respectively, of ∼25 µM, ∼72 µM and ∼83 µM (Szentandrássy et al., 2011). The inwardly rectifying K current, I_K1, was not significantly altered by the drug at concentrations below 300 µM. The multi-channel blocking effects of the drug, with differing concentration-dependences for effects on the different channels, are likely to underlie the comparative lack of effect of the drug on APD, a supposition borne out by the results of AP voltage clamp experiments in the study (Szentandrássy et al., 2011).

The second rosiglitazone study in this issue by Jeong et al. (2011) is complementary to that by Szentandrássy et al. (2011), in that it provides increased insight into the likely basis of I_TO inhibition by rosiglitazone. Recombinant K_v4.3 channel current was inhibited with an IC₅₀ (half maximal inhibitory concentration) of ∼25 µM (Jeong et al., 2011), which is in strong agreement with the observed potency of inhibition of native I_TO (Szentandrássy et al., 2011). Voltage-dependent activation of K_v4.3 channels was little-affected by the drug, but voltage-dependent inactivation was negatively voltage-shifted in a concentration-dependent manner, from which an apparent dissociation constant (K_i) for inactivated channels of ∼1.5 µM was estimated (Jeong et al., 2011). Closed state inactivation of K_v4.3 channels was also accelerated by the drug (Jeong et al., 2011).

A key question that arises is whether or not the results of these two studies may provide insight into increased cardiovascular risk with rosiglitazone. In comparing their findings with earlier studies, Szentandrássy et al. (2011) note that rosiglitazone appears to be a weaker inhibitor of cardiac ion channels than troglitazone (e.g. see Ikeda and Watanabe, 1998). Moreover, in healthy individuals, maximal plasma rosiglitazone concentrations following a single 8 mg oral dose of the drug reach approximately ∼370–700 ng·mL⁻¹ (equivalent to ∼1–2 µM; Thompson-Culkin et al., 2002; Park et al., 2004). Thus, the majority of acute ion channel inhibitory effects seen in the two studies reported here (Jeong et al., 2011; Szentandrássy et al., 2011) occur at higher drug concentrations than are likely to occur in vivo, except – as noted by Szentandrássy et al. (2011) – potentially in overdose. An exception to this is inactivation state-dependent inhibition of K_v4.3 /I_TO, as the K_i for this action of the drug occurs at clinically relevant concentrations (Jeong et al., 2011). It is feasible, therefore, that some modulation of I_TO and thereby of early AP repolarization could occur at low µM rosiglitazone concentrations, although it is not possible on the basis of the available information to reach firm conclusions as to whether or not this could be sufficient to affect the risk of arrhythmias.

These observations do not preclude the occurrence of greater cardiac ion channel inhibitory effects of rosiglitazone under different conditions such as ischaemia. Nor do they preclude acute inhibitory effects of the drug on ionic currents other than those investigated in either of the present studies. Indeed, the authors of both reports note earlier work reporting thiazolidinedione inhibition of various ion channels in other tissues (Jeong et al., 2011; Szentandrássy et al., 2011). It is worth noting that thiazolidinedione drugs have been shown to attenuate porcine ventricular monophasic action potential (MAP) shortening induced during ischaemia, with a greater effect of rosiglitazone than of troglitazone (Lu et al., 2008). In the same study, thiazolidinediones attenuated porcine MAP shortening induced by the K_ATP channel opener levcromakalim, which is strongly indicative that this class of drugs can block cardiac K_ATP channels (Lu et al., 2008). Moreover, during coronary occlusion, the median time to ventricular fibrillation was greatly reduced by rosiglitazone (Lu et al., 2008), which is consistent with a reduction by the drug of the protective role of ischaemia-induced K_ATP channel activation. In this regard, data on acute rosiglitazone inhibition of native ventricular K_ATP channel current (I_KATP) would constitute a valuable comparator for the native current data provided here by Szentandrássy et al. (2011).

Whereas diabetic patients receive drug treatment on an ongoing basis, chronic effects of rosiglitazone exposure on ion channel currents have not been investigated in the two studies that are the subject of this commentary (Jeong et al., 2011; Szentandrássy et al., 2011). It is possible that effects of chronic and acute rosiglitazone exposure on cardiac ionic currents could differ and this possibility warrants experimental investigation. Moreover, recent data from a murine diabetic model indicate that rosiglitazone treatment can influence expression of a number of genes, including some for potassium channel/channel interacting proteins (Wilson et al., 2008). Considering this together with the potential for electrophysiological remodelling in diabetic hearts, it may be valuable now to use diabetic animal models in order to conduct additional studies of both acute and chronic effects of rosiglitazone on cardiac ion channels and electrogenic transporters.

Glossary

Abbreviations

AP

action potential

APD

action potential duration

I_Ca,L

L-type calcium channel current

I_K1

inwardly rectifying potassium channel current

I_KATP

ATP-activated potassium channel current

I_Kr

rapid delayed rectifier potassium channel current

I_Na

fast inward Na current

I_TO

transient outward potassium channel current

MAP

monophasic action potential

MI

myocardial infarction

PPAR γ

peroxisome proliferator-activated receptor γ

Conflicts of interest

None.