Repurposing polyamines to prevent life-threatening arrhythmias in Short QT Syndrome type 3 Authors: Ana I Moreno-Manuel, Francisco M Cruz, Álvaro Macías, Eva Cabrera-Borrego, José Jalife Published: 2026 ABSTRACT AbstractShort QT Syndrome type 3 (SQTS3) is an inherited disorder marked by an abnormally short QT interval on ECG and a high risk of life-threatening arrhythmias, for which effective targeted therapies are lacking. SQTS3 is caused by Kir2.1 mutations that enhance the outward component of the strong inward rectifier potassium current (IK1), disrupting its normal voltage-dependent regulation and severely abbreviating the cardiac action potential. Using a mouse model of SQTS3 carrying the Kir2.1M301K mutation, we examine whether administration of polyamines, the fundamental blockers of Kir2.1 channels, could restore the normal QT interval and prevent arrhythmia. Male mutant mice display reduced IK1 rectification, reduced sodium channel function, QT interval shortening, and a high susceptibility to induced ventricular tachycardia. Here, we show that systemic polyamine administration restores strong IK1 rectification, rescues sodium current density, prolongs action potential duration and QT interval, and significantly reduces arrhythmia inducibility, identifying a potential therapeutic approach for SQTS, a life-threatening disease.Subject terms: Arrhythmias, Arrhythmias, Potassium channels, Patch clamp, Electrocardiography - EKG Introduction Short QT Syndrome (SQTS) is a rare, but highly lethal, inheritable channelopathy characterized by an abnormally short QT interval on the electrocardiogram (ECG) and an increased risk for atrial and life-threatening ventricular arrhythmia. SQTS is frequently associated with sudden cardiac death (SCD)1, which is often the first manifestation of the disease with devastating consequences to the patient and their family. Approximately 250 cases have been diagnosed in nearly 150 families worldwide, all during the last two decades2–4. Despite their heterogeneous phenotype, the penetrance of the SQTS causative mutations is remarkable, and patients experience a poor quality of life5. Therefore, there is an urgent need to improve the diagnosis and prevent SCD of SQTS patients. The molecular mechanisms leading to channel dysfunction, cardiac rhythm disturbances, and related disorders associated with SQTS remain incompletely understood. SQTS-linked variants in KCNH2 (SQTS1), KCNQ1 (SQTS2), KCNJ2 (SQTS3), and SLC4A3 (SQTS8) are the only validated genotype-positive SQTS subtypes to date6,7. The SQTS1-causative mutations hERGT618I and hERGN588K are the most common (25.9% of genotyped probands) and second most common (18.5%), respectively, of the clinically occurring SQTS variants8,9. Focusing on SQTS3, the Kir2.1M301K mutation was originally described in an 8-year-old girl with a QTc interval of 194 ms, paroxysmal atrial fibrillation (AF) and ventricular fibrillation (VF) inducibility10,11. The patient also presented multiple other disorders, such as severe mental retardation, abnormal proliferation of esophageal blood vessels, epilepsy, and Kawasaki disease11. She suddenly died ten years after diagnosis. The methionine 301 in Kir2.1 forms a pore-facing loop region12, and functional analysis in heterologous expression systems revealed that its replacement with a lysine abolished IK1 in homozygosis and increased outward IK1 in heterozygosis11. SQTS3 is caused by mutations in the KCNJ2 gene that encodes Kir2.1, a strong inward-rectifier potassium channel. Kir2.1 is responsible for IK1, which is essential in maintaining the resting membrane potential (RMP) and the final phase of action potential (AP) repolarization13–15. Kir2.1 conductance changes (rectifies), and the slope of its current/voltage (IV) relationship is modified at voltages positive to the K+ equilibrium potential, being the current abolished at voltages close to 0 mV. Inward rectification of Kir2.x channels results from blockade by intracellular magnesium and polyamines (spermine, spermidine and putrescine)16–18, which upon membrane depolarization penetrate the cytoplasmic pore of the channel, binding to specific negatively charged residues and obstructing the outward flow of potassium19–21. Spermine is the main channel blocker responsible for inward rectification, followed by spermidine, putrescine, and magnesium16,18,22–24. Kir2.1 channels possess multiple polyamine binding sites25. They are negatively charged regions, one in the transmembrane domain, involving D172, and the other in the cytoplasmic region, involving E224, D255, D259 and E29912. The initial step in the rectification process involves the interaction of polyamines with E224 and E299 located in the COOH-terminus of the Kir2.1 channel18,22,26. The second step is more voltage-dependent and implies the binding of polyamines at the D172 residue22,26, which is deeper inside the Kir2.1 pore. Mutations at any of the above residues result in a defective inward rectification of IK1. Polyamines are non-toxic and have been shown to be important in aging, cancer and other diseases, but inward rectification at depolarized potentials is likely their most important function27. However, to our knowledge, polyamines have never been considered as a therapeutic option in SQTS3. The identification of innovative, mechanistically targeted therapies for individuals with SQTS3 would be an important development. Currently, treatment for SQTS patients is poorly defined, and there are no complex, standardized, or comprehensive clinical protocols guiding patient management and arrhythmia prevention. The rarity of the disease and its recent description have prevented the design of clinical trials in appropriate patient groups4. In symptomatic individuals in whom an implanted cardioverter defibrillator (ICD) is not recommended, pharmacological treatment is mandatory. However, the currently available drugs have off-target effects because they interfere with the behavior of various ion channels and their regulators. Recommended drugs are antiarrhythmics like ibutilide, flecainide, sotalol, amiodarone, propafenone, and β-blockers, like metoprolol and carvedilol4. β-blockers reduce adrenergic input to the heart, a mechanism that helps control the heartbeat and prevent symptoms28. The latest European Society of Cardiology (ESC) Guidelines for the management of patients with ventricular arrhythmias and prevention of SCD recommend an ICD and treatment with quinidine29. It is known that quinidine blocks INa, decreasing phase 0 of the action potential (AP); it also reduces IKr, IKs, IK1, Ito, ICa-L and INa-Late, leading to AP duration (APD) prolongation30. Both quinidine and hydroxyquinidine are tested as pharmacological treatments of SQTS patients because they prolong the QTc interval, reducing the incidence of life-threatening arrhythmias31–33. Nevertheless, in some countries, quinidine has been removed from the market because of important side-effects. On this basis, the design of new, safer and more effective drugs is urgent to improve life-quality and prevent arrhythmias and SCD in SQTS patients. This original article documents our experimental results on the underlying mechanisms of SQTS3 caused by Kir2.1M301K, a highly pathogenic SQTS variant11. We report on data derived from in-vivo adeno-associated (AAV)-based mouse models of cardiac-specific SQTS3 mimicking the electrophysiological phenotype of patients expressing Kir2.1M301K. The availability of cardiac-specific SQTS3 mouse models helped identify new molecular targets for the design of novel antiarrhythmic drugs in an inheritable cardiac disease that currently has no defined therapy. Our results demonstrate that polyamines effectively correct the APD and QT interval and reduce arrhythmogenesis in SQTS3 mice, suggesting that repurposing these naturally occurring molecules might be a new therapeutic option to prevent life-threatening arrhythmias in SQTS3 patients. Results Design and generation of the mouse SQTS3 model We generated a mouse model of SQTS3 using specific constructions containing the mutated versions of KCNJ2 (Kir2.1M301K) packaged in AAV serotype 9 (AAV9) (Supplementary Fig. 1ai). AAV9-generated Kir2.1WT animals were used as controls. Ex-vivo hearts analysis confirmed the cardiac expression of Kir2.1WT and Kir2.1M301K through fluorescence emitted by the tdTomato reporter present in the AAV constructs, which was totally absent in untransfected hearts (Supplementary Fig. 1aii). Haematoxylin-Eosin staining revealed that WT and mutant mouse hearts were structurally normal. Immunohistochemistry and immunohistofluorescence with tdTomato in cardiac slices confirmed a mosaic and wide AAV9 distribution throughout myocardium (Supplementary Fig. 1aiii-iv), as reported previously34. Total Kir2.1 protein levels in the hearts of mice expressing Kir2.1WT and Kir2.1M301K were very similar to untransfected hearts (Supplementary Fig. 1b, left). We also confirmed that there was no increase in Kir2.1 expression at the plasma membrane of Kir2.1M301K compared to Kir2.1WT cardiomyocytes (Supplementary Fig. 1b, right). Echocardiographic analysis confirmed normal structure and contractile function of both Kir2.1WT and Kir2.1M301K hearts (Supplementary Fig. 2). We previously reported that the echocardiographic values of untransfected animals were similar to AAV9-Kir2.1WT control mice35, and that genetic haploinsufficiency did not operate after expression in trans of Kir2.1WT and Kir2.1SQTS336. We also checked that ECG, arrhythmia inducibility and cellular electrophysiological parameters (e.g., APD, IK1 and INa) of Kir2.1WT were similar to untransfected mice (see Supplementary Fig. 3). Therefore, since AAV9-Kir2.1WT did not alter total Kir2.1 expression and represented a transfected control that did not disturb the murine electrical phenotype, all subsequent experiments were conducted in Kir2.1M301K animals and compared with Kir2.1WT controls. The QT interval is significantly abbreviated in Kir2.1M301K mice We recorded ECGs from Kir2.1WT and Kir2.1M301K mice. Whether non-corrected or corrected by the Bazett or the Framingham formula37,38, mean QT intervals in SQTS3 were all shorter than Kir2.1WT mice (Kir2.1WT, 30.1 ± 1.2 ms > Kir2.1M301K, 16.4 ± 0.7 ms) (Fig. 1a–c). Therefore, the Kir2.1M301K mouse recapitulated the SQTS3 phenotype in that both patient and mouse ECGs demonstrated extreme QT interval shortening11. The Kir2.1M301K mutation significantly abbreviates the QT interval and prolongs the QRS complex in mice. a Representative ECG traces showing QT interval abbreviation in Kir2.1M301K models (purple) compared to Kir2.1WT mice (blue). b Distribution of QT interval values in all models analyzed (Kir2.1WT: 30.1 ± 1.2 ms, median 29.18, min 21.67, max 43.5, 25% percentile 25.42, 75% percentile 33.79, lower 95% CI 27.75, upper 95% CI 32.53, N = 28; Kir2.1M301K: 16.4 ± 0.7 ms, median 15.9, min 12.6, max 24.1, 25% percentile 14, 75% percentile 18.6, lower 95% CI 15.08, upper 95% CI 17.83, N = 23; p < 0.0001). c Left, QTc intervals of Kir2.1WT (80.3 ± 3.5 ms, median 78.22, min 55.6, max 125.6, 25% percentile 63.42, 75% percentile 93.75, lower 95% CI 73.35, upper 95% CI 87.9, N = 28) and Kir2.1M301K (40.2 ± 1.7 ms, median 37.61, min 30.38, max 61.14, 25% percentile 33.8, 75% percentile 44.13, lower 95% CI 36.69, upper 95% CI 43.75, N = 23) mice using the Bazett’s equation (p < 0.0001). Right, QTc intervals in Kir2.1WT (162.3 ± 1.4 ms, median 161, min 150, max 180, 25% percentile 155.5, 75% percentile 167.8, lower 95% CI 159.3, upper 95% CI 165.2, N = 28) and Kir2.1M301K (144.7 ± 0.8 ms, median 145, min 140, max 154, 25% percentile 141, 75% percentile 147, lower 95% CI 143, upper 95% CI 146.5, N = 23) animals applying the Framingham’s correction (p < 0.0001). d from left to right, PR interval (Kir2.1WT: 43.7 ± 0.4 ms, median 43.65, min 41.8, max 46, 25% percentile 42.3, 75% percentile 45.05, lower 95% CI 42.73, upper 95% CI 44.75; Kir2.1M301K: 44.5 ± 0.6 ms, median 44.7, min 40.7, max 47.6, 25% percentile 43.5, 75% percentile 45.55, lower 95% CI 43.15, upper 95% CI 45.83), QRS complex (Kir2.1WT: 8.8 ± 0.1 ms, median 8.7, min 8.3, max 9.5, 25% percentile 8.5, 75% percentile 9.275, lower 95% CI 8.5, upper 95% CI 9.12; Kir2.1M301K: 9.9 ± 0.3 ms, median 9.9, min 8.5, max 11.3, 25% percentile 9.15, 75% percentile 10.9, lower 95% CI 9.269, upper 95% CI 10.63) (p = 0.0045), and P wave duration (Kir2.1WT: 17.4 ± 1.2 ms, median 18.4, min 12.17, max 22.83, 25% percentile 13.5, 75% percentile 20.87, lower 95% CI 14.76, upper 95% CI 20.14; Kir2.1M301K: 17.78 ± 1.0 ms, median 18.1, min 12.1, max 21.7, 25% percentile 15, 75% percentile 20.9, lower 95% CI 15.52, upper 95% CI 20,04) are shown (N = 10 animals per condition). Data are presented as the mean ± SEM. Two-sided Mann–Whitney test (b–d) and unpaired 2-tailed Students’t test (d, P waves) were used (****p < 0.0001, **p < 0.01). As illustrated in Fig. 1d, the PR interval and P wave duration were unaffected by the mutation. However, the QRS was prolonged in the ECG of Kir2.1M301K mice, indicating that intraventricular conduction was slower than Kir2.1WT mice. The significance of such a defect will become apparent below. Long-lasting and high-frequency ventricular tachycardia in Kir2.1M301K mice We used an intracavitary octopolar catheter and intracardiac programmed electrical stimulation protocols to characterize the in-vivo cardiac electrophysiological properties and arrhythmogenicity of these mice. As expected, the ventricular refractory period was reduced in Kir2.1M301K compared to Kir2.1WT mice (Supplementary Fig. 4), which was consistent with their different QT intervals. To determine atrial and ventricular arrhythmia inducibility in each mouse group, we applied S1-S2 trains of 10 and 25 Hz on either the right ventricle (RV) or right atrium (RA) through the octopolar catheter. RV stimulation induced VT/VF in Kir2.1M301K mice (Fig. 2ai). Seven out of 8 Kir2.1M301K mice (87.5%) demonstrated the highest percentage of longer-lasting (>5 s) and high-frequency VT/VF episodes (Fig. 2b). Ventricular arrhythmia inducibility is significantly increased in SQTS3 Kir2.1M301K mice. a, c ECG recordings during and after the programmed electrical stimulation (PES) are shown for Kir2.1WT (blue) and Kir2.1M301K (purple) animals. Sinus rhythm (SR), PES, and sustained ventricular tachycardia or non-sustained atrial fibrillation (sVT or nsAF, respectively) after ventricular (a) or atrial (c) stimulation. b, d, i Number of mice inducible for >500 ms ventricular (b, p = 0.0003) and atrial (d) arrhythmias. ii Violin plots showing the distribution for ventricular (b, p = 0.0003) and atrial (d) inducibility presented as the percentage of stimuli that triggered arrhythmia in each group. iii Duration (s) of all ventricular (b, p < 0.0001) and atrial (d, p = 0.0702) arrhythmias recorded in each group while stimulation. b, iv Frequency of the VT episodes lasting more than 500 ms recorded in Kir2.1M301K mice measured as peaks per second. N = 10 Kir2.1WT and N = 8 Kir2.1M301K mice were used. Data are presented as the mean ± SEM. Two-sided Fisher’s exact test (b i and d i, presence/absence of arrhythmias) and 2-sided Mann–Whitney test (b ii and d ii, iii) or 2-sided Wilcoxon Signed Rank Test (b iii) were applied (****p < 0.0001, ***p < 0.001). On the other hand, focusing on the atria, some Kir2.1M301K animals were AF inducible although the duration of the AF episodes was brief (Fig. 2aii, c). Altogether, these data demonstrated that the mutation Kir2.1M301K causes a pathogenic ventricular arrhythmia phenotype mediated by mechanisms that were consequently studied. APD is markedly short in Kir2.1M301K cardiomyocytes To characterize the electrophysiological properties of ventricular isolated cardiomyocytes, we carried out patch-clamp experiments in the current-clamp mode and recorded APs in both Kir2.1WT and Kir2.1M301K conditions. We demonstrated that the Kir2.1M301K mutation consistently and significantly abbreviated the ventricular APD at almost all repolarization percentages and frequencies studied (Fig. 3a, b and Supplementary Fig. 5). Moreover, APD shortening was not rate dependent. However, RMP, dV/dt and AP amplitude (APA) were not affected by the Kir2.1M301K mutation (Fig. 3c–e). The SQTS3 Kir2.1M301K mutation significantly shortens the mouse ventricular APD. a Representative APs from isolated Kir2.1WT (blue) and Kir2.1M301K (purple) ventricular cardiomyocytes recorded at 1 Hz (left), and APDs at 20, 50, 70 and 90 % repolarization (right; p < 0.0001). b APD90 at 1-to-10 Hz are shown for each group following the same color pattern (N = 3 Kir2.1WT mice, from which n = 14 cells 1 Hz, 16 cells 2 Hz, 14 cells 4 Hz, 15 cells 5 Hz, 9 cells 10 Hz were recorded; N = 3 Kir2.1M301K animals, from which n = 18 cardiomyocytes 1–4 Hz, n = 17 cells 5 Hz, 15 myocytes 10 Hz were studied; p < 0.0001). c RMP (in mV; Kir2.1WT N = 3 n = 17, Kir2.1M301K N = 3 n = 18), d, maximum upstroke velocity (dV/dt, V/s; N = 3 Kir2.1WT mice with n = 16 cells 1–2 Hz, n = 14 cells 4 Hz, n = 15 cells 5 Hz, n = 10 cells 10 Hz recorded; N = 3 Kir2.1M301K with n = 17 cells 1–4 Hz, n = 15 cells 5 Hz, n = 14 cells 10 Hz recorded), and e AP amplitude (mV; N = 3 Kir2.1WT mice with n = 16 cells 1–2 Hz, n = 14 cells 4 Hz, n = 15 cells 5 Hz, n = 10 cells 10 Hz recorded; N = 3 Kir2.1M301K with n = 18 cells 1–4 Hz, n = 17 cells 5 Hz, n = 15 cells 10 Hz recorded) for 1-to-10 Hz stimulation frequencies recorded from Kir2.1WT and Kir2.1M301K ventricular cardiomyocytes. Data are presented as the mean ± SEM. Unpaired 2-tailed Student’s t test (RMP), two-way ANOVA (panel a, right), and 2-tailed Mann–Whitney test (for the remaining comparisons) were applied (****p < 0.0001). IK1 is increased and lacks inward rectification, while INa density and kinetics are altered in Kir2.1M301K cardiomyocytes We performed whole-cell voltage-clamp experiments to analyze the barium-sensitive IK1 in Kir2.1WT and Kir2.1M301K cardiomyocytes (Supplementary Fig. 6). In Kir2.1M301K cardiomyocytes the inward component of IK1 was reduced at voltages negative to −80 mV, which decreased the channel slope conductance. But there was also a gain-of-function in the outward component of IK1 due to a lack of rectification (Fig. 4a). Here the channel remained opened at voltages at which it should have been closed (i.e., between −35 and 10 mV). Kir2.1M301K impairs IK1 inward-going rectification and reduces INa in ventricular cardiomyocytes. a IK1 IV relationships for Kir2.1WT (blue; N = 3, n = 16) and Kir2.1M301K (purple; N = 3; n = 14) ventricular myocytes. Note the lack of inward-going rectification with increased outward IK1 at voltages positive to −55mV (3.83 ± 0.47pA/pF in M301K vs −0.12 ± 0.17pA/pF in WT, at −14mV) (p = 0.0006 −54 mV; p < 0.0001 −44 to −14 mV; p = 0.0169 −4 mV; p = 0.0208 6 mV) and loss of inward current at voltages negative to −120 mV (p = 0.0001 −154 mV; p = 0.0110 −144 mV). b Superimposed INa IV relationships for Kir2.1WT (N = 3, n = 20) and Kir2.1M301K (N = 3, n = 13) isolated cardiomyocytes (p = 0.0199 −50 mV; p = 0.0251 −45 mV). The protocols and representative IK1 and INa IV traces for Kir2.1WT (blue) and Kir2.1M301K (purple) are presented. c NaV1.5 channel steady-state activation and inactivation curves for Kir2.1WT (N = 3 mice, n = 14 and n = 18 cardiomyocytes from which activation and inactivation curves were recorded, respectively) and Kir2.1M301K (N = 3 animals, n = 16 and n = 13 cardiomyocytes from which activation and inactivation curves were recorded, respectively). Note the higher open probability at negative voltages in the INa activation curves of Kir2.1M301K cardiomyocyte (p = 0.0871 −65 mV). d NaV1.5 open probability quantified at window current voltages (from −85 to −55 mV). Source data are provided as a Source Data file. Data are presented as the mean ± SEM. Two-way ANOVA tests were applied for comparisons (****p < 0.0001, ***p < 0.001, and *p < 0.05). We wondered whether the arrhythmogenic consequences and the QRS complex prolongation in mice carrying the mutation Kir2.1M301K were due to alterations in Kir2.1 alone or if other channel disturbances were involved. It has been previously described that Kir2.1 form channelosomes with the cardiac sodium channel NaV1.5 at the plasma membrane, where both channels are reciprocally regulated35,39–43. Therefore, we performed whole-cell voltage-clamp experiments and, unexpectedly, saw that Kir2.1M301K cardiomyocytes had a reduced INa density (Fig. 4b). In addition, while INa inactivation was unaffected by the mutation, the slope of the INa activation curve was modified at hyperpolarized potentials causing an increase in the window current (Fig. 4c and Supplementary Fig. 7). Both peak INa density reduction and window current increase are potentially arrhythmogenic44, predisposing the heart to higher VT/VF inducibility in the Kir2.1M301K mice and spontaneous arrhythmias in the SQTS3 patient. NaV1.5 expression levels are reduced in Kir2.1M301K cardiomyocytes We aimed to decipher whether the decrease in INa density was due to a reduced NaV1.5 expression. Therefore, we quantified NaV1.5 protein in total lysates of Kir2.1WT and Kir2.1M301K ventricles. We observed that the total levels of NaV1.5 in hearts expressing Kir2.1M301K were lower than in Kir2.1WT hearts (Fig. 5a). Moreover, in expansion microscopy experiments, the NaV1.5 fluorescence intensity in Kir2.1M301K was lower than Kir2.1WT ventricular cardiomyocytes. Still, its tubular distribution pattern was maintained (Fig. 5b), although at lower levels than Kir2.1WT relative to the membrane Na/K ATPase (Fig. 5c). NaV1.5 protein expression is reduced in ventricular cardiomyocytes of Kir2.1M301K mice. a Representative Western blot comparing total NaV1.5 protein expression in ventricles from Kir2.1WT (blue; N = 4) and Kir2.1M301K (purple; N = 4) mice (p = 0.0286). The experiment was performed in triplicate, using 4 biological samples per technical replicate. Graph shows western blot quantification with NaV1.5 normalized to α-Actinin. b Representative confocal images showing NaV1.5 channel distribution and fluorescence intensity in Kir2.1WT (N = 3; n = 24) and Kir2.1M301K (N = 3; n = 22) expanded ventricular cardiomyocytes. Scale bars, 50 μm. Graph shows quantification of NaV1.5 fluorescence intensity in both groups. c Colocalization profiles of NaV1.5 and Na+/K+ ATPase, as plasma membrane marker, in isolated ventricular cardiomyocytes from Kir2.1WT and Kir2.1M301K animals are shown. Quantification of the colocalization between NaV1.5 and Na+/K+ ATPase (N = 3 Kir2.1WT mice, n = 18 cells; N = 3 Kir2.1M301K animals, n = 15 cells; p < 0.0001) and between NaV1.5 and Wheat Germ Agglutinin (WGA) in N = 3 Kir2.1WT, n = 14, and N = 3 Kir2.1M301K, n = 13 (p < 0.0001), was performed using objective thresholding (Otsu method) with identical parameters applied to all images; signal intensity was consistent across samples. Data are presented as the mean ± SEM. Two-tailed Mann–Whitney tests were used for statistical analysis (****p < 0.0001, *p < 0.05). Apart from NaV1.5, we studied the distribution of the already known Kir2.1-NaV1.5 channelosome interactors SAP97 and α1-syntrophin, and revealed they were similar in Kir2.1WT and Kir2.1M301K cardiomyocytes (Supplementary Fig. 8). Exogenous spermine prolongs the QT interval in Kir2.1M301K mice After studying the mechanism underlying the IK1 increase caused by Kir2.1M301K, we hypothesized that polyamines administered systemically could therapeutically restore the QT interval by inducing inward-going rectification of Kir2.1M301K mutant channels. To prove this concept, we treated Kir2.1M301K mice with spermine, the most potent voltage dependent Kir2.1 blocker16,18,22–24. Fifteen- to 17-week-old Kir2.1WT and Kir2.1M301K mice were administered 10 mg/Kg/day spermine (i.p.) for 10 and 20 days. As previously reported45, after 10 days of treatment, all animals, independently of the genotype, manifested a loss of about 7.5% of their body weight and then stabilized throughout the 20-day duration of the treatment (Supplementary Fig. 9). Most importantly, the QT/QTc interval was significantly prolonged and the QRS complex shortened compared to baseline values in mutant mice (Fig. 6a and Supplementary Fig. 10a). Neither the Kir2.1M301K mice nor the Kir2.1WT mouse were inducible for long-lasting (>1 s) atrial or ventricular arrhythmias after 10-day spermine treatment (Fig. 6b–d), suggesting a potential therapeutic action of the drug for the prevention of Kir2.1M301K-associated arrhythmogenic events. Surprisingly, only 1 out of 8 Kir2.1M301K mice presented a couple of long-lasting atrial episodes after 20-day treatment (Fig. 6c). Administration of spermine reduces arrhythmia inducibility and IK1 hyperfunctionality in Kir2.1M301K mice. a Left and middle, QT and QTc (Bazett’s correction) intervals in Kir2.1WT (blue) and Kir2.1M301K (purple) at baseline and after 10- and 20-day treatment with 10 mg/Kg/day spermine (N = 16 mice per condition at time 0 and 10-day, and N = 8 at 20-day treatment) (****p < 0.0001, ***p = 0.0004, *p = 0.0366, and p = 0.0553 QTc 20-day SPM). Right, representative ECG traces after 10-day and 20-day treatment with spermine. b Representative traces of the ventricular arrhythmia recorded after 20 days of treatment. c Duration (ms) of arrhythmias at 0, 10-day and 20-day treatment with spermine in Kir2.1WT (N = 10 non-treated and N = 8 treated mice) and Kir2.1M301K (N = 8 control, N = 8 10-day treated, and N = 7 20-day treated mice). Note the reduction in number and duration of the arrhythmogenic events in Kir2.1M301K treated mice (****p < 0.0001, *p = 0.010). d Arrhythmia inducibility after ventricular electrical stimulation at 0, 10-day and 20-day treatment with spermine in Kir2.1WT (N = 10 non-treated and N = 8 treated mice) and Kir2.1M301K (N = 8 control, N = 8 10-day treated, and N = 7 20-day treated mice) (**p = 0.0037, *p = 0.0124). e Representative traces of APs recorded from WT control (dark blue), M301K control (dark purple), M301K 10-day spermine (light purple), and M301K 20-day spermine (dark pink). APD90 (***p = 0.0002, **p = 0.0025) (f), dV/dt (g), and AP amplitude (h) measured from the AP recorded from WT control (N = 3 animals; n = 10 cardiomyocytes), M301K control (N = 3; n = 9), M301K 10-day spermine (N = 3; n = 14), and M301K 20-day spermine (N = 3, n = 11) ventricular cardiomyocytes, respectively. i Outward IK1 recorded from WT control (N = 3 mice, n = 14 cardiomyocytes), M301K control (N = 3, n = 14), M301K 10-day spermine (N = 3, n = 9), and M301K 20-day spermine (N = 3, n = 9) ventricular cardiomyocytes. When comparing Kir2.1M301K control with 10-day spermine, ****p < 0.0001, ***p = 0.0002, **p = 0.0059, *p = 0.0159 at −4 mV, and *p = 0.0419 at 6 mV. When comparing Kir2.1M301K control with 20-day spermine, ****p < 0.0001, ***p = 0.0006, **p = 0.002, and *p = 0.042. j INa IVs recorded from WT control (N = 3 mice, n = 20 cardiomyocytes), M301K control (N = 3, n = 13), M301K 10-day spermine (N = 3, n = 10), and M301K 20-day spermine (N = 3, n = 10) myocytes. When comparing Kir2.1M301K control with 20-day spermine, *p = 0.0127 at −45 mV, *p = 0.0215 at −40 mV and *p = 0.0401 at −35 mV. Spermine is indicated as “SPM” through the panels. Source data are provided as a Source Data file. Data are presented as the mean ± SEM. Two-sided Mann–Whitney test (panels d, f–h) and two-way ANOVA (corrected by Šídák’s multiple comparisons; panels a, i, j) were applied (****p < 0.0001, ***p < 0.001, **p < 0.01, and *p < 0.05). At the cellular level, current-clamp recordings in ventricular cardiomyocytes isolated from spermine-treated Kir2.1M301K mice revealed a progressive APD prolongation compared to untreated Kir2.1M301K (Fig. 6e–h and Supplementary Fig. 10b), with no change in the RMP (Supplementary Fig. 10c). As shown by the superimposed IV relations in Fig. 6i, chronic i.p. administration of spermine significantly reduced the outward component of IK1 in the Kir2.1M301K cardiomyocytes. Over time, spermine also restored the INa peak current and substantially reduced the Na+ window current (Fig. 6j and Supplementary Fig. 10d, e). Spermine treatment did not significantly change the RMP, dV/dt, APA (Supplementary Fig. 11a, b), and only slightly prolonged the APD of Kir2.1WT cardiomyocytes (Supplementary Fig. 11c, d), attributed to a mild reduction in the outward amplitude of IK1 (Supplementary Fig. 11e). Spermine treatment did not functionally alter the peak INa or its activation/inactivation properties in Kir2.1WT cardiomyocytes (Supplementary Fig. 11f–h). Altogether, the results thus far indicate that the systemic administration of 10 mg/Kg/day spermine for up to 20 days has beneficial effects in SQTS3 mice expressing Kir2.1M301K. It increases IK1 rectification prolonging the APD. Consequently, it increases the QT interval and reduces arrhythmia inducibility. Chronic oral spermidine administration prolongs the QT interval and reduces arrhythmia inducibility in Kir2.1M301K animals Oral administration is the most convenient, cost-effective, and commonly used route of drug therapy. Oral administration of spermidine, the second most potent voltage dependent Kir2.1 blocker, is available at the pharmacies as a food supplement16,18,22–24. It has been suggested to be cardioprotective and some authors have reported promising results, although in different contexts44–49. Therefore, we decided to add 3 mM spermidine to the drinking water of 15- to 17-week-old Kir2.1WT and Kir2.1M301K mice for 20 days (to compare with spermine) and 6 months (for a long-term treatment study). The treatment significantly prolonged the QT/QTc interval over time without significantly altering other ECG parameters in ether Kir2.1WT or Kir2.1M301K mice (Fig. 7a and Supplementary Fig. 12a). PR interval was transiently prolonged in Kir2.1WT mice after 20 days of spermidine treatment, but the effect was not sustained over time; and the QRS was shortly reduced in Kir2.1M301K animals (Supplementary Fig. 12a). Most importantly orally administered spermidine reduced the arrhythmia duration and inducibility in Kir2.1M301K animals (Fig. 7b–d). Underlying the QT prolongation, the APD was longer, along with an almost complete restoration of IK1 function (Fig. 7e-i and Supplementary Fig. 12b, c), and a full reversal of INa defects in Kir2.1M301K ventricular cardiomyocytes following long-term oral spermidine administration (Fig. 7j and Supplementary Fig. 12d, e). The electrophysiological effects of spermidine after 20 days of treatment were comparable to those observed after 6 months, meaning that effects plateau after 3 weeks. Chronic oral administration of spermidine prolongs the QT interval, reduces arrhythmia inducibility, and ameliorates IK1 and INa dysfunction in Kir2.1M301K mice. a Left and middle, QT (****p < 0.0001, **p = 0.0069, *p = 0.0381) and QTc (Bazett’s correction; ****p < 0.0001, *p = 0.0127) intervals in Kir2.1WT (blue) and Kir2.1M301K (purple) at baseline and over the time after starting the treatment with 3 mM spermidine. Sample size is indicated in every time point. Right, representative ECG traces of the same Kir2.1M301K animal untreated (control) and after 10-day, 20-day and 6-month treatment with spermidine via drinking water. b Representative traces of the ventricular arrhythmia recorded after 6 months of treatment. c Duration (ms) of the arrhythmias induced (****p < 0.0001, *p = 0.0253), and d arrhythmia inducibility (****p < 0.0001, *p = 0.0396) after ventricular pacing in Kir2.1WT (N = 10 control, N = 14 20-day treatment, and N = 5 6-month treated mice) and Kir2.1M301K (N = 8 control, N = 14 20-day treated, and N = 5 6-month treated mice). Note the reduction in number and duration of the arrhythmogenic episodes in Kir2.1M301K mice. e Representative traces of APs recorded from WT control (dark blue), M301K control (dark purple), M301K 20-day spermidine (dark pink), and M301K 6-month spermidine (light pink). AP parameters: APD90 (***p = 0.0002, ****p < 0.0001, ***p = 0.0003) (f), dV/dt (g), and AP amplitude (h) measured from WT control (N = 3 mice; n = 10 cardiomyocytes), M301K control (N = 3; n = 9), M301K 20-day (N = 3; n = 8-9), and M301K 6-month spermidine (N = 3, n = 9-11) ventricular cardiomyocytes. i Outward IK1 recorded from WT control (N = 3 mice, n = 14 cardiomyocytes), M301K control (N = 3, n = 14), M301K 20-day spermidine (N = 3, n = 11), and M301K 6-month spermidine (N = 3, n = 11) ventricular cardiomyocytes. When comparing Kir2.1M301K control with 20-day spermidine, ****p < 0.0001, **p = 0.0052 at −54 mV, **p = 0.0060 at −4 mV, and **p = 0.0064 at 6 mV. When comparing Kir2.1M301K control with 6-month spermidine, ****p < 0.0001, **p = 0.0024 at −44 mV, *p = 0.0140 at −4 mV, and *p = 0.0154 at 6 mV. j INa IVs recorded from WT control (N = 3 animals, n = 20 cells), M301K control (N = 3, n = 13), M301K 20-day spermidine (N = 3, n = 8), and M301K 6-month spermidine (N = 3, n = 9) myocytes. When comparing Kir2.1M301K control with 20-day spermidine, *p = 0.0235 at −45 mV, *p = 0.0147 at −40 mV and *p = 0.0357 at −35 mV. When comparing Kir2.1M301K control with 6-month spermidine, *p = 0.0132 at −45 mV, **p = 0.0038 at −40 mV, **p = 0.0073 at −35 mV, **p = 0.0083 at −30 mV, and *p = 0.0263 at −25 mV. Spermidine is indicated as “SPD” through the panels. Source data are provided as a Source Data file. Data are presented as the mean ± SEM. Two-sided Mann–Whitney test (panels d, f–h) and two-way ANOVA corrected by Šídák’s multiple comparisons test (panels a, i, j) were applied (****p < 0.0001, ***p < 0.001, **p < 0.01, and *p < 0.05). Additionally, spermidine did not modify the RMP, dV/dt or APA (Supplementary Fig. 13a, b), but it did prolong significantly the APD90 in Kir2.1WT cardiomyocytes (Supplementary Fig. 13c, d). Underlying this prolongation, was a reduction in the outward amplitude of IK1 both 20-day and 6-month after starting spermidine administration (Supplementary Fig. 13e). The peak INa of Kir2.1WT cardiomyocytes was unchanged by spermidine, although it slightly shifted toward depolarized voltages (Supplementary Fig. 13f–h). Oral spermidine exhibits a favorable safety profile while maintaining outward IK1 reduction in long-term treated mice Young (16-week-old) and old (55-week-old) Kir2.1WT and Kir2.1M301K animals under long-term (6-month) oral treatment with 3 mM spermidine underwent functional, structural, and serological analyses to determine drug toxicity. Similar to young adult mice, aged Kir2.1M301K animals exhibited prolonged QT/QTc intervals over time following the initiation of long-term spermidine treatment (Supplementary Fig. 14). After 3 weeks, these values reached levels comparable to those of Kir2.1WT mice, without significant disturbances in any other ECG parameter. In contrast, untreated Kir2.1WT and Kir2.1M301K old mice—littermates of those receiving 3 mM spermidine for 6 months—displayed shortened QT/QTc intervals, consistent with the pattern seen in young adults (Supplementary Fig. 14a, b). Like with spermine, both groups showed a body weight reduction at 6 months spermidine treatment (Supplementary Fig. 15a, b), but the body weight/heart weight ratio did not change compared to untreated untransfected control animals (Supplementary Fig. 15c). Echocardiography showed no differences in ejection fraction or fractional shortening between untreated and long-term treated mice, either young or old adult (Fig. 8a, b). Histological analysis following long-term treatment in young animals revealed no structural abnormalities or evidence of fibro-adipose replacement in the liver, kidney, lung or spleen (Supplementary Fig. 15d). Additionally, the hearts from treated mice exhibited normal structure, comparable to those from untreated untransfected control animals (Fig. 8c). Serological analyses and biochemical parameters measured in the same young adult mice were within normal ranges (Supplementary Fig. 16)50,51. Together, these results demonstrate that chronic oral spermidine administration is safe and effective, means to restore the QT interval and prevents arrhythmia inducibility in our SQTS3 mouse model. Long-term oral spermidine administration is safe and nontoxic in both Kir2.1WT or Kir2.1M301K mice. a, b Echocardiography analysis showing ejection fraction and fractional shortening of untreated and 6-months spermidine treated Kir2.1WT and Kir2.1M301K young adult (42-week-old) (a) and old adult (81-week-old) (b) mice. Kir2.1WT untreated (N = 10 Kir2.1WT control young, N = 3 Kir2.1WT control old) are shown in dark blue, Kir2.1M301K untreated (N = 10 Kir2.1M301K control young, N = 4 Kir2.1M301K control old) in dark purple, Kir2.1WT treated (N = 6 Kir2.1WT young 6-month SPD, N = 5 Kir2.1WT old 6-month SPD) in dark green, and Kir2.1M301K treated in light pink (N = 5 Kir2.1M301K young and 5 old 6-month SPD). c Haematoxylin and Eosin (H&E), and Masson’s trichrome staining of untreated untransfected Kir2.1WT (N = 3) and 6-months treated Kir2.1WT (N = 3) and Kir2.1M301K (N = 3) hearts. Scale bar, 1 mm. Top, The right ventricle (RV) and left ventricle (LV) areas are presented. Bottom, Graphs show quantification of the RV, LV, and septum wall thickness. Spermidine is indicated as “SPD” through the panels. Data are presented as the mean ± SEM. Two-way ANOVA tests were applied for comparisons. Discussion Currently, the absence of mechanistically targeted therapies for SQTS3 is in part due to the unknown mechanisms underlying some of the identified mutations. We aimed to decipher the arrhythmogenic and molecular mechanisms that make the human Kir2.1M301K mutation causing a truly abbreviated QT interval11 with the final goal of enhancing the effective therapeutic approaches in SQTS3. Unlike cardiomyocytes, HEK cells and other heterologous expression systems do not express all the interactors necessary for appropriate function of the ion channels; hence we generated an in-vivo SQTS3 mouse model carrying the same variant, using Kir2.1WT as a control. Introduction of exogenous human Kir2.1M301K in mice expressing the endogenous murine Kir2.1 reproduces a heterozygous condition where the dominant negative effect of the mutation mimics the patient’s genetic environment in the heart. Importantly, we confirmed that transduction using the human WT or mutant Kir2.1 sequences did not cause neither Kir2.1 overexpression nor structural alterations in the cardiac tissue (Supplementary Fig. 1). Our generated AAV9-mediated mouse model presented a significantly shortened QT interval and long-lasting ventricular arrhythmias induced by intracardiac stimulation, mirroring the patient’s electrophysiological phenotype. Kir2.1M301K cardiomyocytes had markedly abbreviated APD, linked to impaired inward rectification and increased IK1 at voltages above -80 mV. Amino acid M301 is close to E299 in the Kir2.1 channel structure, and polyamines bind to this region to block the channel at depolarized voltages52,53. The replacement of methionine (M) with lysine (K)—a polar, positively charged amino acid—at position 301 electrostatically impairs polyamine binding. This results, on the one hand, in a gain of Kir2.1 function due to reduced inward-going rectification when expressed in heterozygosis. Heterotetrameric Kir2.1WT-Kir2.1M301K channels remain open at voltages where they would normally be closed, leading to a gain-of-function. On the other hand, the addition of positively charged lysine residues at position 301 of the four monomers comprising the Kir2.1 channel interferes with its potassium conductance, and as noted by Hattori et al.11, may also induce conformational changes in the near-pore region, effectively closing it. Consequently, when Kir2.1M301K is expressed in homozygosis, the resulting mutant homotetrameric conformation of Kir2.1—with 4 lysines in the cytoplasmic pore—electrostatically repels the potassium flow through the channel and leads to its loss-of-function. Therefore, the electrophysiological consequences of Kir2.1M301K differ depending on whether it is expressed in homo- or heterozygous conditions. While in the homozygous condition the mutation reduces IK1, it dramatically increases the current when co-expressed with Kir2.1WT 11. While analyzing the ECG from Kir2.1M301K mice, we realized that the QRS complex was significantly prolonged compared to Kir2.1WT animals. Unfortunately, the only published ECG of the patient carrying the Kir2.1M301K mutation was recorded during childhood. ECGs were not available close to the time of death. Therefore, we could not assess whether there were ECG changes related to disease progression, treatment, or developmental stage, including possible QRS prolongation. To study the possible reason for the QRS prolongation in Kir2.1M301K relative to Kir2.1WT mice, we isolated ventricular cardiomyocytes. We demonstrated that the genetic variant disrupts NaV1.5 function by reducing INa density and shifting the INa activation curve to more negative voltages, even though dV/dtmax was not altered. This is not surprising, given ample evidence of a nonlinear relationship between dV/dtmax and peak INa in cardiac ventricles54,55. In other words, experiments and simulations have shown that dV/dtmax does not always scale linearly with the availability of sodium channels and INa, meaning that small changes in INa can cause disproportionate changes in dV/dtmax, or vice versa56, due to factors like the interaction with other ionic currents and the complex kinetics of the sodium channels. In our experiments, Kir2.1M301K mice showed a peak INa that was 20% lower than Kir2.1WT, consistent with the 10% prolongation of the QRS complex. However, the voltage-gated activation of NaV1.5 channels is shifted toward hyperpolarized potentials, increasing the availability of sodium channels during phase 0 of the AP. This shift mitigates the impact of reduced peak sodium current on the rate of depolarization (dV/dtmax). Still, it increases the window current, with the corresponding arrhythmogenic consequences, as described for the I151V mutation in NaV1.5, among others57. These results call for caution, should the above effects apply to patients with the M301K variant, as sodium channel blocking agents like quinidine might be contraindicated. Kir2.1 and NaV1.5 work synergistically, and loss-of-function mutations in Kir2.1 reduce the NaV1.5 current density and vice versa35,39–43. Recently, we demonstrated that the SQTS3-related mutation Kir2.1E299V modified INa in a cardiac chamber-specific manner36. Moreover, in-silico simulations predicted that both channels have a physical interaction interface with a favorable binding energy: the voltage domains DII-S1 of NaV1.5 interact with the transmembrane helix M1 of Kir2.158. Kir2.1 has been shown to interact with as many as 218 high-confidence interactors encompassing various molecular mechanisms of Kir2.1 function, ranging from intracellular trafficking to lysosomal degradation59. In a different study, membrane-enriched cardiac lysates of male mice, identified 61 protein interactors for KCNJ2 and 59 for SCN5A60. Kir2.1 was not listed as an interactor of SCN5A, and NaV1.5 was not among the detected interactors for KCNJ2. However, both channels share some common proteins, such as apolipoproteins, polypeptides and E3 ubiquitin-protein ligases60, hence a third partner could mediate the functional relationship between them. Proteins like SAP97 and α1-syntrophin act as scaffolds that keep Kir2.1 and NaV1.5 together at the membrane through their PDZ binding domains. In-vitro experiments have demonstrated that inhibition or absence of these interacting proteins causes alterations in both IK1 and INa40,61–66. However, we have not seen any changes in the distribution of these specific interactors in ventricular cardiomyocytes expressing the Kir2.1M301K mutation. Unlike SAP97 and α1-syntrophin, we did see a reduction in the total and sarcolemmal protein levels of NaV1.5 in Kir2.1M301K ventricular cardiomyocytes, so other unknown common interactors might mediate, promote, or even enhance these alterations. Hattori et al.11 demonstrated that the homozygous expression of Kir2.1M301K in HEK cells resulted in an absence of IK1, so the INa reduction seen in heterozygous cardiomyocytes could be the result of a destabilization of NaV1.5 when it is in contact with Kir2.1M301K homotetrameres. While the molecular basis of how Kir2.1M301K reduced INa is beyond the scope of this study, one possible explanation is that the mutation may affect the interaction of both channels with 14-3-3 proteins and indirectly alter NaV1.5 phosphorylation43,67–69. Alternatively, NaV-β subunits could mediate the Kir2.1M301K effects, as they form multiprotein complexes with ion channels and are essential for NaV1.5 function in cardiomyocytes70. Yet another explanation is that Kir2.1M301K may alter NaV1.5 degradation, as both channels share some E3 ubiquitin-protein ligases60. However, further research is needed to decipher how the mutation Kir2.1M301K reduces the expression and current density of NaV1.5 channels. On the other hand, in this work we also aimed to increase the mechanistically targeted treatment options for SQTS3 patients. Since the pathogenic mechanism of the Kir2.1M301K mutation involves a loss of inward rectification, and polyamines can bind to the Kir2.1 cytoplasmic pore to block the channel in a voltage-dependent manner23,71, we sought to evaluate the effects of their chronic administration. Polyamines are well-established physiological mediators of inward rectification in Kir2.1 channels through voltage- and charge-dependent pore blockade. Our findings demonstrate that the Kir2.1M301K mutation attenuates this process by introducing electrostatic repulsion that limits polyamine binding to critical negatively charged residues within the pore. Notably, exogenously increasing intracellular polyamine concentration partially restored voltage-dependent blockade in Kir2.1M301K mice by enhancing vestibular occupancy, thereby promoting interaction with accessible pore residues and reinforcing inward rectification at depolarized potentials. In contrast to conventional potassium channel blockers, polyamines represent endogenous modulators that preserve physiological channel gating while selectively promoting IK1 rectification when this process is defective. This distinction is indeed particularly relevant in the context of SQTS3, where controlled enhancement of a native regulatory mechanism prolongs repolarization without introducing the proarrhythmic liabilities commonly associated with pharmacological agents. Supporting the above concept, polyamine supplementation has not been associated with QT prolongation in healthy individuals, and clinical and epidemiological data suggest that they have cardioprotective and anti-aging roles mediated by promoting cardiac autophagy45,46,72, further underscoring their potential therapeutic relevance45–49,73. A phase I study reported that administration of the spermine analog PG-11047 to cancer patients did not affect the ECGs of 88.9% of participants74. The remaining 11.1% showed abnormal ECG findings after treatment; unfortunately, the authors did not present the ECGs or provide further details. Moreover, it is not possible to determine how the patients’ ECGs would have appeared without the spermine analog. Other studies have reported that polyamine levels negatively correlate with cardiac diseases, such as heart failure and myocardial infarction, indicating that lower polyamine concentrations are associated with these conditions75,76. Most notably, a recent prospective cohort study that included 184,732 participants without cardiovascular disease (CVD) at baseline from the UK Biobank, with a median follow-up of 11.5 years, found that dietary intake of 27.1–33.5 mg/day polyamines (spermidine, spermine and putrescine) were associated with a lower risk of all-cause mortality or incident CVD77. Polyamines are also effective in neurodegenerative diseases by inducing autophagy, oxidative stress, and amyloid-β accumulation78. In hypertension, heart failure, and myocardial infarction, polyamines have been shown to promote cardiac autophagy and longevity, and to reduce inflammation and oxidative damage77,79. In metabolic disorders, such as type 2 diabetes, obesity, and fatty liver disease, they improve insulin sensitivity, lipid metabolism, and reduce hepatic steatosis80. In different cancers, they enhance anti-tumor immune responses81; and in inflammatory and autoimmune diseases like rheumatoid arthritis and inflammatory bowel disease, polyamines reduce pro-inflammatory cytokines and promote intestinal barrier integrity82. We recently demonstrated that increasing the cytoplasmic concentration of spermine interfered with the IK1 gain-of-function of some mutant Kir2.1 channels that lacked inward-going rectification36. Therefore, we hypothesized that repurposing polyamines by administering them to control hyperfunctional Kir2.1M301K channels would have therapeutic benefit. Spermine is the polyamine with the most powerful capacity for blocking Kir2.1 channels16,18,24, but spermidine, the second most potent blocker of Kir2.1, is already available as a food supplement, having demonstrated beneficial effects at different levels (see above). The polyamine doses employed in our study are comparable to those commonly used in humans, supporting the translational validity of our findings. In mice, spermine was administered i.p. at 10 mg/Kg per day (approximately 0.3 mg for a 30 g mouse), which is close to the lower range of the doses typically administered to humans (0.6–40 mg/day). Similarly, spermidine was provided at 3 mM in drinking water (436 mg/L), a concentration consistent with non-toxic regimens reported in previous mouse studies, and corresponding to ~0.275 mM/day in humans. Despite the substantial differences in body size and physiological complexity between mice and humans, Kir2.1—the main molecular target of our investigation—is highly conserved across species. Therefore, the close correspondence in dosing, combined with the evolutionary conservation of Kir2.1, reinforces the translational relevance of our results obtained in mice to human physiology. As a checkpoint after polyamine administration in mice to ensure that both treatments were having an effect, we monitored the mice´s weight and noticed that all animals lost about 7.5% of their weight after 10 days of treatment. It has been described that polyamines improve glucose metabolism in mice45. Most importantly, we demonstrated that chronic administrations of spermine or spermidine prolonged the APD and the QT interval, and reduced the arrhythmia incidence in Kir2.1M301K mice. Using untransfected, untreated mice of the same age as controls in some experiments, we also showed that neither AAVs, Kir2.1M301K expression, nor long-term SPD treatment altered cardiac function, structure, or serological profiles. Moreover, polyamines reversed QRS prolongation by restoring the INa defects caused by the Kir2.1M301K mutation. Further research is needed to understand the mechanisms underlying these effects, but it seems that the Kir2.1M301K-derived INa alterations depend on the IK1 gain-of-function at depolarized potentials. When this gain-of-function is alleviated, the INa disturbances provoked by Kir2.1M301K are similarly reduced. Interestingly, polyamines were shown to prolong the QT interval in mutant mice regardless of whether treatment began at early or late stages of aging. Therefore, based on our studies in mice identifying polyamines as a potential therapeutic approach for SQTS, and given the reported beneficial effects of spermidine in humans, the question arises whether patients with various types of SQTS can benefit from a dietary supplement of this polyamine. Unfortunately, while highly lethal, SQTS is a rare condition, with only about 250 cases diagnosed worldwide2–4, which renders them of little commercial interest to the pharmaceutical industry and challenging for conventional randomized clinical trials. As such, the long and difficult path to translate our academic proof-of-concept study into approved medicines means that very few therapies ever reach patients. Repurposing polyamines for the treatment of SQTS, represents a “disruptive” innovation83,84. Therefore, even if the prospect of saving SQTS patients does not provide sufficient motivation to take immediate action, investing resources in this overlooked area will likely lead to significant breakthroughs that can impact the development of treatments for both rare and common diseases. The absence of arrhythmic events in most polyamine-treated mutant mice is highly promising; polyamine treatment reduces the arrhythmogenic potential of a highly pathogenic mutation in mice, which opens a new pathway for investigation at the preclinical and clinical levels. Therefore, based on our findings, we conclude that chronic treatment with polyamines could be considered a potential therapeutic strategy for patients with SQTS3. Targeting the underlying electrophysiological abnormalities of Kir2.1M301K, using polyamines or novel derivatives, may help prevent life-threatening arrhythmias and reduce the risk of SCD in these patients. Limitations Data derived from our AAV9-mediated cardiac-specific mouse model of SQTS3 should not be directly extrapolated to the human disease and should be cautiously interpreted. First, heart rate and repolarization parameters differ substantially between mice and humans, as some ionic currents that govern the human AP are absent in mice. There are also remarkable structural differences between murine and human hearts, particularly in heart size. Therefore, additional studies using appropriate preclinical models would be needed to ensure rigorous translation of our results to the patient. Second, although cardiac-specific AAV9 vectors delivered intravenously show a 95% transduction rate, and over 50% of cardiomyocytes take up 2–4 copies of the transgene, this system always generates a mosaic cellular distribution of mutant and wildtype expression throughout the heart. However, this mosaicism is also present in AAV9-Kir2.1WT, the control condition. Third, our study does not include human data directly assessing the effects of polyamines on cardiac electrophysiology. The rarity and severity of SQTS3 make such investigations profoundly challenging, as only a few patients worldwide have been identified. Moreover, polyamines such as spermidine are widely consumed as dietary supplements without reports of QT interval prolongation in healthy individuals, suggesting minimal effects in non-pathological conditions. Nevertheless, the Kir2.1 channel is highly conserved across species, with its structure and function preserved from yeast to humans. Thus, despite the absence of direct human data, the strong evolutionary conservation of Kir2.1 supports the translational relevance of our findings and their potential applicability to human SQTS3 patients. Therefore, we stand by our results despite the limitations outlined above. They provide novel insights into mechanisms underlying this arrhythmogenic and potentially lethal channelopathy and into the innovative role that polyamines could take in the treatment of SQTS3 patients. Methods Study approval All experimental and other scientific procedures using animals conformed to EU Directive 2010/63, EU and Recommendation 2007/526/EC, enforced in Spanish law under Real Decreto 53/2013. The local ethics committees and the Animal Protection Area of the Comunidad Autónoma de Madrid (PROEX 226.5/23) approved all animal protocols. PROEX modification, including polyamines administration ref. 10/874770.9/25. Mice Four-week-old C57BL/6J male mice were obtained from Charles River Laboratories. Mice were reared and housed in accordance with institutional guidelines and regulations. The CNIC animal facility ensures 12 h light/12 h dark (12 h LD) cycles, providing normal circadian rhythms, with temperatures at 20–24 °C and 40–60% humidity, and uses individually ventilated cages as a standard to maintain high sanitary quality and reduce cage-to-cage contamination. The mice had access to food and water ad libitum. Fifteen- to 35-week-old mice were used in experiments when we needed untreated or short-term-treated animals. For long-term spermidine treatment, we used 16-week-old and 55-week-old Kir2.1WT and Kir2.1M301K mice and studied their electrophysiological phenotypes for up to 6 months after treatment initiation (i.e., at 42 and 81 weeks of age, respectively). Untransfected untreated animals were used as controls when indicated. Experimental procedures were conducted in anesthetized animals, which is known to pause the circadian clock85. Nevertheless, experiments in which the circadian rhythm could be relevant (ECGs, echocardiography, cardiomyocyte isolation, and intracardiac stimulation) were done during daylight hours, specifically, from 9 am to 2 pm. AAV production and purification Briefly, this was achieved by first generating enhanced tdTomato-reporter AAV vectors driven from the cardiomyocyte-specific cardiac Troponin T proximal promoter (cTnT), strengthened by a cis-regulatory motif number 4 enhancer (CRM4)43, and encoding wildtype Kir2.1 (Kir2.1WT), or the SQTS3 Kir2.1 mutant (Kir2.1M301K). These AAV plasmids were packaged into AAV serotype 9 (AVV9) capsids with the use of pAdDF6 helper plasmids (providing the three adenoviral helper genes) and pAAV2/9 (providing rep and cap viral genes), obtained from PennVector by the CNIC Viral Vectors Unit (https://www.cnic.es/en/investigacion/viral-vectors)34. AAV vectors were produced by triple transfection of HEK-293A cells, as described previously86,87. AAV vector titer (viral genomes, vg, per ml) was carried out by quantitative real-time PCR (qRT-PCR) as described88. Known copy numbers (105–108) of the respective plasmid (pAAV-empty vector, pAAV-Kir2.1WT, and pAAV-Kir2.1M301K) were used to construct standard curves. AAV injection Mice were anesthetized with 60 mg/kg Ketamine and 20 mg/kg Xylazine administered intraperitoneally (i.p.). Once asleep, animals were placed on a heated pad at 37 ± 0.5 °C to prevent hypothermia. Thereafter, we administered 3.5 × 1010 vg to 4- to 5-week-old mice intravenously (i.v.) via the femoral vein in a final volume of 50 μL, taking care to prevent the introduction of air bubbles. Animals were then maintained on the heating pad until recovery. AAV-mediated gene distribution We assessed the corporal distribution of AAV-mediated protein expression eight weeks after AAV9 administration34. The ex vivo fluorescent signal was detected in hearts from transfected mice, confirming the cardiac expression of protein tdTomato used as the reporter gene in the viral construct. The infection rate in mouse hearts (measured as vg/cell) was quantified blindly by immunohistochemistry using an anti-tdTomato antibody and a home-made Fiji plugin, as previously described34. Echocardiography Mice were lightly anesthetized with 0.5–2% isoflurane in oxygen, and placed on a 37 °C heating platform in the supine position. A base-to-apex electrocardiogram (ECG) continuously monitored the animal, adjusting isoflurane delivery to maintain the heart rate at 450 ± 50 beats per minute (bpm). Warmed ultrasound gel was used to maintain normothermia. Transthoracic echocardiography was performed blindly using a high-frequency ultrasound system (Vevo 2100, Visualsonics Inc., Canada) with a 40-MHz linear probe. Two-dimensional (2D) guided M-mode echocardiograms were obtained at a frame rate above 230 frames/s, and pulse wave Doppler was acquired at a frequency of 40 kHz. Images were transferred to a computer and analyzed blindly offline using the Vevo 2100 Workstation software. The M-mode sample gate was placed at the level of the papillary muscles. Parasternal standard long and short axis views were acquired to assess left ventricular (LV) systolic function. LV ejection fraction (EF) and chamber dimensions were calculated from these views in systole and diastole. Surface ECG recordings Mice were anesthetized adjusting isoflurane (1.5­2% volume in oxygen) delivery to maintain the heart rate at 350 ± 50 bpm and the efficacy of the anesthesia was controlled by monitoring the breathing rate and the absence of response to painful stimuli. A subcutaneous 23-gauge needle electrode connected to an MP36R amplifier (BIOPAC Systems) was attached to each limb, and six-lead surface ECGs were recorded. We used AcqKnowledge 4.1 software and LabChart 8 Reader to analyze, blindly and, sometimes in pairs, the ECG complexes, segments and interval durations. For each animal, we measured the duration of these parameters in 5 to 10 beats per time of interest, and the mean ± standard error of the mean (SEM) was presented. We applied Bazett’s and Framingham’s formulas to correct the QT interval considering the heart rate. Polyamine administration First, 15- to 17-week-old male mice were used to test the effects of polyamine administration. Following previously published indications, 10 mg/Kg spermine (Sigma-Aldrich, 71-44-3), was administered i.p. daily for 10 or 20 days, changing the administration area to avoid inflammation45. In addition, 3 mM spermidine (Sigma-Aldrich, S2626) was administered via drinking water for 10 or 20 days, or a maximum of 6 months, changing the water and preparing fresh spermidine every 3 days89. Fifty-two-week-old mice also underwent long-term treatment with spermidine. In-vivo intracardial electrophysiological studies After anesthesia (60 mg/kg Ketamine and 20 mg/kg Xylazine i.p.), an octopolar catheter (Science) was inserted through the jugular vein and advanced into the right atria and right ventricle90,91. The animal was placed on a 37 °C heating platform in the supine position. To measure the duration of the refractory period, we used an S1-S2 protocol consisting of a train of 10 stimuli (S1) at a cycle length of 100 ms (10 Hz) followed by an extra-stimulus (S2) reducing the S1-S2 interval in 2-ms steps starting at 100 ms. Atrial and ventricular arrhythmia inducibility was assessed by applying 10 stimuli at 10 Hz (S1), followed by 10 stimuli at 25 Hz (S2), all of them 2 ms (duration) and 7 V (amplitude)-stimuli. Atrial fibrillation was defined as the occurrence of rapid and fragmented atrial electrograms (lack of P waves) with irregularly irregular atrio-ventricular (AV)-nodal conduction and ventricular rhythm. Ventricular tachycardia/ventricular fibrillation (VT/VF) was defined as an exceedingly fast monomorphic/polymorphic ventricular rhythm independent of atrial excitation. The frequency of VT episodes lasting more than 500 ms was measured as the number of peaks per second. Cardiomyocyte isolation To isolate ventricular cardiomyocytes we used the procedure of Macías et al.35. Briefly, after cervical dislocation, the mouse heart was isolated, mounted on a modified Langendorff-perfusion apparatus and the aorta was retrogradely perfused with Ca2+-free perfusion buffer (PB) containing (in mmol/L): 113 NaCl, 4.7 KCl, 0.6 KH2PO4, 0.6 Na2HPO4, 1.2 MgSO4·7H2O, 12 NaHCO3, 10 KHCO3, 0.032 Phenol Red, 9.8 HEPES, 30 taurine, 5.5 glucose, 10 2,3-butanedione-monoxime (pH 7.4 with NaOH). Perfusion rate was 1 mL/min for 7 min at 37 °C. Enzymatic digestion was then started for 20 min with digestion buffer (DB) consisting of PB supplemented with Liberase™ (0.2 mg/mL), Trypsin 2.5% (5.5 mmol/L), and CaCl2 (12.5 µM) at 37 °C. Thereafter, ventricular cardiomyocytes were gently disaggregated and isolated in 3 mL DB. The resulting cell suspension was filtered through a 200 µm sterile mesh (SEFAR-Nitex) and transferred for enzymatic inactivation to a tube containing 12 ml stopping-buffer-1 (SB-1), consisting of PB supplemented with fetal bovine serum (FBS, 10% v/v) and CaCl2 (12.5µmol/L). After gravity sedimentation for 20 min, the supernatant was removed and cardiomyocytes were resuspended in stopping-buffer-2 (SB-2) containing the same concentration of CaCl2, but lower FBS (5% v/v) for another 20 min. Cardiomyocyte Ca2+-reintroduction was performed in SB-2 with two progressively increased CaCl2 concentrations (0.112 and 1 mmol/L). Cells were resuspended and allowed to decant for 15 min in each step, contributing to the purification of the cardiomyocyte suspension. Patch-clamping of isolated cardiomyocytes For the patch-clamp experiments in the whole-cell configuration, we used the external (bath) and internal (pipette) solutions that are listed in Supplementary Table 135,36,39. Cardiomyocytes were placed in a superfusion chamber (RC-26, Warner Instruments) mounted on the stage of an inverted microscope (DMi8, Leica). Cells were allowed to settle on the bottom of the superfusion chamber before being perfused with the corresponding solution. After sealing the membrane, whole-cell voltage-clamp or current-clamp recordings were made using an Axopatch 200B Amplifier (Axon Instruments) and a MultiClamp 700B Microelectrode Amplifier (Molecular Devices). Pipettes made from borosilicate glass (GD-1, Narishige, OD: 1 mm; ID: 0.6 mm) had resistances of 2–5 MΩ after using the internal solutions. Series resistance compensation of 80–90% was achieved. All voltage-clamp currents were low-pass filtered at 2 kHz with an analog filter and digitized at 4–10 Hz. Data were recorded using pClamp 10.6 software with the Clampex 10.6 program and analyzed with the Clampfit 10.6 program (Axon Instruments, Foster City, USA). Current amplitudes were normalized to the cell capacitance to account for differences in cell size and expressed as current densities (pA/pF). Whole-cell current-clamp recordings We patch-clamped the cardiomyocytes in the current clamp configuration using a conventional patch-clamp amplifier to record changes in the voltage through the membrane (Vm). Threshold current was determined using 1 ms pulses at increasing amplitudes (0.2 nA/pulse) and frequency of 1 Hz. Thereafter, action potentials (APs) were evoked by the injection of 1 ms pulses of constant amplitude. We recorded AP stimulated at 1, 2, 4, 5 and 10 Hz. APD was measured at 20, 50, 70 and 90% of repolarization, and, as part of the AP characterization, we also analyzed the resting membrane potential (RMP), the AP amplitude (APA), and the AP upstroke velocity measured as dV/dt. In these experimental conditions, the liquid junction potential correction applied was 13.4 mV. Whole-cell voltage-clamp recordings For these experiments, we changed the patch-clamp mode from current-clamp to voltage-clamp, where we used a feedback amplifier to apply voltage steps and measure the current needed to maintain the membrane potential at any given level. This enabled us to construct IV relationships for IK1 and INa, as follows: Potassium currents. IV relationship for IK1 densities were constructed by applying 500-ms voltage-clamp steps in 10 mV increments from −140 to 0 mV from a holding potential of -80mV at a frequency of 0.1 Hz at room temperature (RT). IK1 was calculated by subtracting currents recorded in the absence and presence of 500 µM BaCl2. In these experiments, the liquid junction potential was 13.4 mV; it was applied when calculating and presenting the IVs. Potassium currents. IV relationship for IK1 densities were constructed by applying 500-ms voltage-clamp steps in 10 mV increments from −140 to 0 mV from a holding potential of -80mV at a frequency of 0.1 Hz at room temperature (RT). IK1 was calculated by subtracting currents recorded in the absence and presence of 500 µM BaCl2. In these experiments, the liquid junction potential was 13.4 mV; it was applied when calculating and presenting the IVs. Sodium currents. To record INa, cells were held at −120 mV and stepped for 100 ms from −90 up to +40 mV in 5 mV increments at 0.33 Hz. INa was measured at the peak of the current trace. To analyze INa inactivation, cells were held at −120 mV and stepped for 100 ms from −140 mV up to −20 mV in 10 mV increments at 0.2 Hz. Leak currents were subtracted using the P/4 protocol. Steady-state activation and inactivation curves were fitted by using a Boltzmann equation, in which V50 is the half-maximum (in)activation potential and k is the Slope factor. Sodium currents. To record INa, cells were held at −120 mV and stepped for 100 ms from −90 up to +40 mV in 5 mV increments at 0.33 Hz. INa was measured at the peak of the current trace. To analyze INa inactivation, cells were held at −120 mV and stepped for 100 ms from −140 mV up to −20 mV in 10 mV increments at 0.2 Hz. Leak currents were subtracted using the P/4 protocol. Steady-state activation and inactivation curves were fitted by using a Boltzmann equation, in which V50 is the half-maximum (in)activation potential and k is the Slope factor. Immunohistochemical procedure We analyzed the amount of AAV9 infection in mouse hearts (see the “AAV-Mediated Gene Distribution” section). After perfusing the heart with PBS, it was fixed in 4% paraformaldehyde (PFA) in PBS for 48 h and conserved in 70% ethanol. Samples were mounted in OCT embedding compound and frozen at −20 to −80 °C. We cut 5–7 µm thick tissue sections using a cryostat and thaw-mounted the sections onto gelatin-coated histological slides. After permeating (0.2% Triton X-100 in PBS) and blocking (10% normal goat serum (NGS) in PBS), we used the goat polyclonal anti-tdTomato antibody (Sicgen, AB8181-200) overnight (O/N) and the anti-goat HRP or Alexa-488 secondary antibodies for 1 h at RT. We also incubated cardiac samples, in addition to liver, kidney, lung, and spleen tissues slices, with Hematoxylin-Eosin to study possible structural changes in infected compared with untransfected hearts, and in non-treated vs treated with 3 mM spermidine for 6 months. Moreover, we used Masson’s trichrome staining to measure the percentage of fibrosis in the same samples and conditions. Conventional confocal microscopy experiments Isolated cardiomyocytes were washed using PBS and fixed for 10 min with 4% PFA in PBS. Cardiomyocytes were permeabilized using Triton X-100 0.2% in PBS for 10 min and then blocked in suspension with 10% NGS in a solution containing Triton X-100 0.1% in PBS for 1 h at RT. We incubated O/N the cells with primary antibodies at 4 °C and secondary antibodies (1:500 in all cases) for 1 h at RT (the antibodies and their dilutions are specified in Supplementary Tables 2, 3). We washed the samples three times after each antibody incubation with PBS. Samples were mounted in Fluoroshield™-DAPI imaging medium (F6057, Merck). Images of individual cardiomyocytes were acquired with a Leica SP8 confocal mounted on an inverted DMI6000 microscope using the LAS X 3.5.6.21594 Software. The objective HC PL APO CS2 100×/1.40 OIL, as well as PMT and hybrid detectors (HyD) were used. Point-scanning confocal with spectral detection and Acousto-Optical Beam Splitter (AOBS) system were employed. Images of 2048 × 2048 with 8-bit resolution were acquired. LUT was linear and covered the full range of data; we did not use any nonlinear adjustment. All the properties of the shown images (fluorochromes with their excitation and emission ranges, pixel dimensions, LUTs, detectors, gains…) are provided as source data. Uniform brightness and contrast adjustments were applied to all images to optimize their visualization in the displayed panels. The analysis of the colocalization percentage between NaV1.5 and the membrane stainers wheat germ agglutinin (WGA; Anti-WGA AF488, 1:200, Invitrogen # W11261) and Na+/K+ ATPase in Kir2.1WT and Kir2.1M301K ventricular cardiomyocytes was done using Fiji/ImageJ. Masks for NaV1.5 and membrane markers were generated using a consistent threshold across all immunofluorescence experiments, and their areas were measured. For membrane protein quantification, thresholding was performed using objective algorithms (Otsu method), applying the same formula to all images; signal intensity was consistent across all samples. Using these masks and the “Process–Image Calculator” tool, overlapped images were generated, a manual threshold of 1 was applied, and their areas were measured. The overlap percentage was calculated by dividing the area of the overlap image by the area occupied by the plasma membrane marker. Cardiomyocytes plating and fixation Laminin-coated coverslips (rounded 13 mm-diameter) were prepared by diluting the original stock of laminin (1 mg/ml, Corning) 1:10 in PBS. After 1–2 h of covering, we removed the excess diluted laminin on the top of the coverslips and the cardiomyocytes were left to adhere for 1–2 h at RT. After that time, they were washed with PBS 1× and fixed using 4% PFA for 10 min at RT. We washed the already fixed cardiomyocytes twice with PBS 1× and we followed the staining procedure for expansion microscopy as will be explained in the coming section. Expansion microscopy After the fixation, cells were permeabilized with 0.5% Triton-X100 in PBS for 15 min at RT; followed by blocking in 5% BSA diluted in washing buffer (PBS + 0.02% Tween 20) for 1 h at RT. For expansion microscopy, we used the primary antibodies more concentrated than for conventional confocal microscopy (Supplementary Table 4); they were diluted in the blocking solution and incubated O/N at 4 °C. The next day, after three washes in the washing solution, the secondary antibodies (Supplementary Table 5) were diluted in blocking buffer and incubated for 2 h at RT, followed again by two washes with washing solution and two more with PBS. We used the crosslinker 10 mg/ml 6-((Acryloyl)amino)hexanoic acid succinimidyl ester (or AcX) in PBS O/N at RT. The next step was the gelation procedure. We washed the coverslips twice with PBS and prepared in ice the following gelation solution: 4 M Sodium Acrylate, 40% Acrylamide, 2% Bis-Acrylamide, PBS 10x, 10% TEMED and 10% APS. A drop of this solution was added in a gelation chamber per sample, we flipped each coverslip on them, and we left them in the incubator (37 °C) for 1 h. After the gelation, we prepared a digestion buffer containing TAE buffer, 10% Triton-X100 in PBS, 5 M NaCl, Proteinase K (26 μl per every 2 ml) and 1:400 DAPI. We moved the coverslips with the thin gel on their top from the gelation chambers to 12-multiwell plates with the digestion solution, and we let them digest for 4 h at 37 °C. With the digestion the gels separated from the coverslips and increased their size three times. We moved them from the 12-multiwell plate to a 150-mm diameter culture plate per condition to add MilliQ water and let them expand O/N at RT. The next day they were imaged using a Leica STELLARIS 8 confocal mounted on an upright DM6 microscope and the LAS X 4.6.1.27508 Software. The objective HC APO L U-V-I 40×/0.80 WATER, as well as four hybrid detectors (HyD) were used. Point-scanning confocal with spectral detection and AOBS system were employed. Images of 2048 × 2048, 0.135 μm × 0.135 μm pixel size, with 8-bit resolution were acquired. LUT was linear and covered the full range of data; we did not use any nonlinear adjustment. All the properties of the shown images (fluorochromes with their excitation and emission ranges, LUTs, detectors, gains…) are provided as source data. Western blot and membrane fractionation procedures To analyze total expression levels, we excised ventricles from untransfected, Kir2.1WT, Kir2.1D172N, Kir2.1E299V, and Kir2.1M301K mice, and lysed them in ice-cold RIPA buffer (150 mM NaCl, 10 mM Tris-HCl, 1 mM EDTA, 1% Triton X-100, 0.1% SDS and 0.1% Sodium deoxycholate). We sonicated the samples for 15 s, 30% power, and conserved them in ice for 30 min, applying 10 s of vigorous shaking every 10 min. We centrifuged the samples for 20 min at 21,000 × g and then recovered the supernatant. We quantified the resulting samples using Pierce™ BCA Protein Assay Kit (Thermo Fisher) and 25–80 µg of protein resolved in 5–10% SDS-PAGE gels. We carried out the transference with the Trans-Blot Turbo Transfer System (Bio-Rad) (9 min, 1.3 A and 25 V for mid-weight proteins; 13 min, 1.3 A and 25 V for high-weight proteins). We blocked the membranes with 5% non-fat dry milk (diluted in PBS-0.05% Tween) when studying the NaV1.5 expression, and with 5% BSA (diluted in PBS-0.05% Tween) for the remaining experiments, for 1 hour at RT. The employed antibodies and their concentrations are listed in Supplementary Tables 6, 7. To determine the Kir2.1 and/or NaV1.5 protein levels at the plasma membrane, we used isolated mouse cardiomyocytes and followed the manufacturer’s instructions for the Plasma Membrane Protein Extraction Kit (Abcam, ab65400). Briefly, we separated the cytoplasmic fraction and, after different centrifugations, the fraction containing plasma membrane proteins was pelleted and resuspended using the homogenization buffer. The enriched plasma membrane fraction was sonicated and kept in ice for 30 min. The protein concentration in each fraction was quantified using Pierce™ BCA Protein Assay Kit (Thermo Fisher). We resolved total, cytosolic, and plasma membrane fractions of three animals per condition in 4–15% (gradient) Mini-PROTEAN TGX Stain-Free Gels (Bio-Rad). The transference, blocking, and blotting were performed as specified above. Note that there was slight contamination between fractions, and consequently, residual expression of membrane proteins in the cytosolic fraction, and vice versa. Uncropped and unprocessed scans of the blots shown in this paper are included in the Supplementary Material and Source Data files. Serological analysis of cardiac, hepatic and renal parameters Following a maxillofacial puncture performed at 9am, ~200 μl of blood was collected into serum separator tubes from 43-week-old untreated, untransfected animals. This was done to compare cardiac, hepatic, and renal serum parameters with those of 43-week-old male mice treated with spermidine and expressing either Kir2.1WT or Kir2.1M301K. To obtain serum, blood samples were left to stand in an upright position for 30 min to allow proper clot formation, and then centrifuged for 15 min at 1200 × g. The resulting serum was collected from the supernatant into clean Eppendorf tubes for subsequent serological analysis using the analyzer Dimension RxL Max (Siemens). Statistical analysis We used GraphPad Prism software version 7.0 and 8.0. We applied the statistical tests after a normal (Gaussian) distribution analysis (Shapiro–Wilk normality test and equality of variances with F test). With 2 groups and normal distribution, comparisons were made using unpaired 2-tailed Student’s t test. When the data do not follow a Gaussian distribution, we apply the nonparametric Mann–Whitney test. Unless otherwise stated, we used one- or two-way ANOVA followed by the post hoc Tukey test for comparison among more than two groups. If data did not follow a normal distribution, the nonparametric equivalent Kruskal–Wallis test was performed. To compare the trend of the sodium activation and inactivation curves across different membrane voltages between groups, an Ordinary one-way ANOVA was applied. To exclude outliers from some data sets we performed Grubbs’ test, also known as the ESD (extreme studentized deviate) method, to determine whether a given value was a statistically significant outlier from the rest. For ECG recordings, we used 10–30 mice per condition and the analysis were carried out by two blinded researchers; on intracardiac stimulation procedures, around 10 animals of each genotype were used; and about 10–15 cells from 3 mice were studied in patch-clamp experiments. Unless other was specified, “N” means number of mice and “n” means the samples or cells analyzed in each case. Data are expressed as mean ± SEM, and differences are considered significant at p < 0.05 (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Reporting summary Transparent Peer Review file Source data Source Data Acknowledgements We thank the Cardiac Arrhythmia Laboratory members, with a special mention to Dr. Patricia Sánchez-Pérez and Dr. Paula G. Socuéllamos, in addition to Dr. Carlos Morillo, for providing discussions and revisions. We gratefully acknowledge Dr. Mario Delmar and Dr. Giorgia Bertoli for generously sharing their expertise and providing valuable guidance on expansion microscopy. We thank the CNIC Viral Vectors Unit for producing the AAV9 used in this article. Author contributions A.I.M.M and J.J. co-designed the experiments; A.I.M.M., A.M. and F.M.C. performed all the experiments contained in that paper; A.I.M.M characterized ex vivo the AAV9 infection level of WT and SQTS3 mouse hearts, studied their electrical phenotype in vivo, isolated cardiomyocytes, performed most of the electrophysiological experiments, and carried out the molecular and cellular biology techniques, along with their analysis; F.M.C. and A.M. contributed equally: F.M.C. infected the C57BL/6 mice using the appropriate AAV9 and contributed with the mouse models in-vivo characterization after treating the mice with polyamines; and A.M. conducted some of the cellular electrophysiology experiments, both before and after starting the treatment with spermine, and provided some funding; E.C.B. analyzed echocardiography data; A.I.M.M. and J.J. co-wrote the manuscript and conceived the study; J.J. provided supervision, funding and revisions; all authors discussed the results and approved the manuscript. Peer review Peer review information Nature Communications thanks Godfrey Smith, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Funding J.J. discloses support for the research of this work from La Caixa Banking Foundation [project code LCF/PR/HR19/52160013]; grant PI23/01039 of the public call “Proyectos de Investigación en Salud 2023” funded by Instituto de Salud Carlos III (ISCIII); MCIU grant BFU2016-75144-R and PID2020-116935RB-I00, and co-funded by Fondo Europeo de Desarrollo Regional (FEDER); and Fundación La Marató de TV3 [736/C/2020]. A.I.M.M. holds a FPU contract [FPU20/01569] from Ministerio de Universidades. We also receive support from the European Union’s Horizon 2020 Research and Innovation programme [grant agreement GA-965286]; the Dynamic Microscopy and Imaging Unit - ICTS-ReDib Grant ICTS-2018-04-CNIC-16 funded by MCIN/AEI /10.13039/501100011033 and ERDF; project EQC2018-005070-P funded by MCIN/AEI /10.13039/501100011033 and FEDER. The CNIC is supported by the Instituto de Salud Carlos III (ISCIII), the Ministerio de Ciencia, Innovación y Universidades (MICIU) and the Pro CNIC Foundation, and is a Severo Ochoa Center of Excellence (grant CEX2020-001041-S funded by MICIU/AEI/10.13039/501100011033). The confocal experiments were carried out in the CNIC Microscopy and Dynamic Imaging Unit – ICTS-ReDib with funding from MCIN/AEI /10.13039/501100011033 and FEDER “Una manera de hacer Europa” (#ICTS-2018-04-CNIC-16). Data availability The data underlying this article are available in the article and in its online supplementary material. Source data are provided with this paper. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Francisco M. Cruz, Álvaro Macías. Supplementary information The online version contains supplementary material available at 10.1038/s41467-026-74259-7. References Associated Data Supplementary Materials Supplementary Information Reporting summary Transparent Peer Review file Source Data Data Availability Statement The data underlying this article are available in the article and in its online supplementary material. Source data are provided with this paper.