Is NOS uncoupling the missing link between atrial fibrillation and chronic non-ischaemic cardiomyopathy? Reply

We thank Crijns et al. for their positive comments regarding the importance of our recent study in Cardiovascular Research.1 We also appreciate the detailed attention of Crijns et al. to the methodologies presented in our study. We welcome the opportunity to provide additional information that we hope will clarify our approach and conclusions for the readers of Cardiovascular Research. The two issues raised by Crijns et al. generally address nitric oxide synthase (NOS) biochemistry, related analytical methodology, and acquisition and interpretation of findings related to the electrophysiological data. It is important to note that the presence of NOS monomer is not direct evidence of NOS uncoupling and that changes in the dimer/monomer ratio are not directly related to the functional uncoupling of the enzyme. Moreover, only the dimeric form has been shown to be biochemically active and able to generate NO and superoxide.2–4 Several lines of evidence in our manuscript are consistent with iNOS oxidase activity in the atria of heart failure (HF) dogs, including (i) decreased tetrahydrobiopterin (BH4) levels, (ii) reduced NO production, and (iii) increased superoxide production. These alterations occurred in the face of an approximately five-fold increase in iNOS expression and were attenuated with BH4 supplementation. Although this is not direct evidence of uncoupling, uncoupling provides the most parsimonious explanation of these observations. Collectively, these findings strongly suggest that iNOS uncoupling is involved in the HF-induced atrial electrophysiological perturbations. We drew our conclusion that NOS function was the critical mediator of BH4 effects based on the following reasoning. Although BH4 is known to possess free radical scavenging activity, the levels of BH4 reached in the myocardium following supplementation were several orders of magnitude lower than the levels of endogenous reducing equivalents and would therefore be expected to exert a negligible overall antioxidant effect. As such, we attribute the antiarrhythmic effects of BH4 therapy to improved NOS catalytic activity and reduced superoxide production. The suggestion to include another treatment arm as a further control, such as with tetrahydroneopterin (HN4), is interesting, but given our findings of reduced atrial [BH4], repletion of BH4 was our primary goal. Treatment with HN4 may provide a similar antioxidant effect without an increase in [BH4].5 However, there is insufficient pharmacokinetic data about HN4 to effect an appropriate allometric dosing strategy for translation into our model. Further, in order to serve as a valid control, the metabolism and biodistribution in the heart would need to be established and shown to be similar to that of BH4. An alternative control that we considered was the use of the NOS-inactive 6S-BH4 rather than 6R-BH4; however, this also has issues related to uncertainties about in vivo metabolism and dose delivery as well as cost. We would welcome the opportunity to do such additional controls in the future, should this information and the necessary resources be available. Crijns et al. suggest that multiple techniques be employed to assess superoxide production. We concur and note that we have included both electron paramagnetic resonance (EPR)-based analysis of semiquinone radical and dihydroethidium (DHE) fluorescence techniques. Although each of these techniques has some inherent limitations, we believe that the combination of both fluorescence- and EPR-based methodologies provides more definitive evidence of reactive oxygen species production than prior reports6,7 using DHE and/or lucigenin, which induces concentration-dependent superoxide generation from a broad range of flavoproteins including NOS.8,9 With respect to the robustness of the atrial fibrillation (AF) induction protocols, we opted to utilize the typical stimulation rate of 10 Hz to maintain consistency with existing literature as well as to minimize the duration of anaesthesia and consequent variability in collected tissues or myocytes. We appreciate the comment about the ability to induce AF in our previous report with premature atrial stimuli,10 which was not observed in this study and which may reflect intrinsic variability between groups of dogs. Notably, induction of AF is a surrogate for actual sustained AF, and initiation and sustenance of spontaneous AF are complex, multifactorial processes which are not completely understood at the present time.11 Shortening of atrial ERP in HF without documented AF has not been a consistent finding, but our data are consistent with human HF,12 and we acknowledged that our findings differ from those in shorter term canine HF models.13 The reasons for this discrepancy deserve to be clarified to elucidate the relationship between HF and atrial electrophysiological changes. The question of potential retrograde atrial activation in our model merits a response. Periods of 2:1 conduction with ventricular pacing 250 b.p.m. have been reported.14 Our stimulation rate was 180 b.p.m. for the last 8 weeks of our 4-month HF model, and we did not observe any evidence of 2:1 or 1:1 retrograde conduction on the electrocardiogram. If the atrial ERP changes we observed resulted from rapid retrograde atrial activation, one would have anticipated more inducible and spontaneous AF than we observed.15,16 Considered collectively, it seems more likely that alterations of the atrial ERP resulted from HF rather than atrial tachycardia. In conclusion, we agree that further experimentation is needed to address the contribution(s) of fibrosis, conduction slowing, and other factors to the atrial electrophysiological alterations in chronic HF. As we stated, our study ‘was not designed to explicitly address structural remodelling, and further investigation is warranted to elucidate the mechanism(s) contributing to the reduced fibrosis and improved atrial conduction with BH4 and l-arginine treatment’. Clearly, fibrosis and structural remodelling as mediators of arrhythmias has been an important focus of investigation. Further investigations are needed to further elucidate the mechanisms linking NOS function, redox balance, and atrial fibrosis.