2.1 Ventricular Cardiomyocyte Isolation

Ventricular cardiomyocytes were enzymatically isolated from hearts mounted in Langendorff systems [22]. Hearts were cannulated through the aorta and perfused for 5 min with Ca²⁺-free Tyrode's solution containing (in mM): 140 NaCl, 5.4 KCl, 1 MgCl₂, 0.33 Na₂HPO₄, 10 HEPES, and 10 glucose; adjusted to pH 7.4. The hearts were then perfused for 15 min at 37°C with Tyrode's solution containing 0.05 mM Ca²⁺ and 1 mg/mL collagenase type 2 (approximately 300 U/mg; Worthington Biochemical, USA). Then, the ventricle was mechanically separated with forceps, and ventricular cells were dissociated by gentle trituration with a Pasteur pipette. Finally, the cells were filtered through a mesh and gradually returned to normal Tyrode's solution through stepwise washes to 0.1, 0.3, 0.6, 1, 1.6, and 2 mM Ca²⁺. Only rod-shaped cells with distinct striations were used for subsequent functional experiments.

2.2 Intracellular pH Measurements

We measured pH_i from atrial strips and isolated ventricular cardiomyocytes using the pH-sensitive fluorophore 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein (BCECF). Cells and tissues were incubated with 10 μM acetoxymethyl ester of BCECF (Thermo Fisher Scientific, Denmark) for 10 (cardiomyocytes) or 20–25 (atrial strips) minutes. During alternating excitation at 485 and 440 nm, we collected epifluorescence at 510 nm using an Olympus IX70 inverted microscope connected to an EasyRatioPro fluorescence imaging system (Photon Technology International, USA) or a Nikon Diaphot 200 inverted microscope with a Q-Imaging fluorescence camera and VisiView software (Visitron, Germany). To calibrate the BCECF fluorescence signal to pH, we added combinations of NH₄Cl and Na-acetate to atrial strips under CO₂/HCO₃⁻-free conditions and determined steady-state pH_i by the null-point technique [23-25]. We established the dynamic relationship between changes in BCECF ratio and pH from high-K⁺/nigericin calibrations [26].

After loading with BCECF and following a 10-min stabilization period, we induced intracellular acidification by the NH₄⁺-prepulse technique [27]. We added NH₄Cl to the bath solution for 7 min and observed intracellular acidification after NH₄Cl washout. Based on pilot experiments, the NH₄⁺ concentration (40 mM for atrial strips, 10 mM for isolated ventricular cardiomyocytes) was chosen to reach the desired level of intracellular acidification (ΔpH_i around 0.4). We fitted polynomial functions to the subsequent recovery phase where pH_i in the presence of bath Na⁺ returned toward the resting level. From the fitted curves, we calculated the corresponding net acid extrusion activity—as dpH_i/dt × buffering capacity—at pH_i 6.9 (ventricular cardiomyocytes) or 6.8 (atrial strips). The intrinsic buffering capacity was calculated from the change of cardiomyocyte pH_i in response to the addition and washout of NH₄Cl in the absence of CO₂/HCO₃⁻ [11]. Intrinsic buffering capacity values corresponding to pH_i 6.9 and 6.8 were calculated using a one-phase exponential decay function fitted to the buffering capacities plotted as a function of pH_i. The contribution of CO₂/HCO₃⁻ to intracellular buffering was calculated as 2.3 × [HCO₃⁻]_i. Intracellular concentrations of NH₄⁺ and HCO₃⁻ were calculated from the Henderson-Hasselbalch equation. Resting steady-state pH_i was analyzed as the value before the addition of NH₄Cl.

The CO₂/HCO₃⁻-containing solution used for pH_i recordings contained (in mM): [28] 115.88 NaCl, 2.82 KCl, 1.60 CaCl₂, 1.20 MgSO₄, 22 NaHCO₃, 1.18 KH₂PO₄, 10 HEPES, 5.50 glucose, and 0.03 EDTA. For ion substitution, Na⁺ and HCO₃⁻ were replaced with equimolar amounts of N-methyl-D-glucammonium (except for NaHCO₃, which was replaced with choline-HCO₃) and Cl⁻, respectively. The CO₂/HCO₃⁻-containing solutions were aerated with 5% CO₂/balance air and CO₂/HCO₃⁻-free solutions with nominally CO₂-free air before pH was adjusted to 7.4 at 37°C.

2.3 Electrophysiology

Membrane potentials were recorded from isolated ventricular cardiomyocytes using the Amphotericin B-perforated whole-cell current clamp technique [29]. We acquired data with an Axopatch 200B amplifier, a Digidata 1440A analog-to-digital converter, and Clampex 10 Software (Molecular Devices, UK). Borosilicate patch pipettes were pulled with a P-97 puller (Sutter Instruments, USA) to a final resistance of 2–3.5 MΩ. Pipettes were filled with solution containing (in mM): 125 K-gluconate, 20 KCl, 5 NaCl, 0.22 Amphotericin B, and 10 HEPES, adjusted to pH 7.2 with KOH. The extracellular solution consisted of (in mM): 120 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 27 NaHCO₃, 5 glucose, and 13 choline-chloride, adjusted to pH 7.4 after bubbling with 5% CO₂/balance O₂ at room temperature. Action potentials were triggered at 1 Hz by 2-ms square depolarizing pulses. Electrophysiology experiments were performed at room temperature, and the HCO₃⁻ concentration of the physiological saline solution was adjusted to approximately 27 mM to account for the greater CO₂ solubility in water (Henry's law constant of 0.0436 vs. 0.0307 mM/mmHg), the lower saturated water vapor pressure (21 vs. 47 mmHg), and the raised apparent pKa value (6.17 vs. 6.12) at 23°C compared to 37°C [30]. Absence of spontaneous rupture of the sealed membrane patch—which would convert the experiments from permeabilized to conventional whole cell configuration—is supported by the lack of sudden changes in membrane potential or injection pulse amplitudes [31].

2.4 Immunoblotting

We isolated protein from atrial and ventricular tissue stored at −80°C and lysed in a buffer containing 20 mM Tris·HCl, 150 mM NaCl, 5 mM EGTA, 10 mM NaF, 20 mM β-glycerophosphate sodium salt, 1% Triton-X, 0.1% Tween-20, and Halt Protease and Phosphatase Inhibitor Cocktail (#78444; ThermoFisher Scientific, Denmark) at pH 7.5. Protein concentrations were determined by a bicinchoninic acid protein assay (ThermoFisher Scientific). We loaded 10 μg of total protein diluted in sample buffer in each lane of a sodium dodecyl sulfate-polyacrylamide gel for electrophoresis and subsequent transfer to polyvinylidene difluoride membranes blocked with 3% skimmed milk in Tris-buffered saline with 0.1% Tween-20 for 2 h at room temperature. The membranes were probed with rabbit anti-NH₂-terminal NBCn1 antibody (kind gift from Dr. Jeppe Praetorius, Aarhus University) [3] that we affinity-purified using the corresponding immunizing peptide (21^st Century Biochemicals, USA), with mouse anti-NHE1 antibody (#sc-136 239; Santa Cruz Biotechnology, USA; diluted 1:1000), or with rabbit anti-NBCe1 antibody (AB3212-I, Sigma-Aldrich; diluted 1:500). The applied antibodies have previously been used for immunoblotting on murine tissue [12, 32, 33]. The specificity of the antibodies against NBCn1 and NHE1 was previously confirmed using tissue or cells with knockout or knockdown of the transporters [12, 16, 34]. After thorough washing, membranes were incubated for 1–2 h at room temperature with matched anti-rabbit or anti-mouse horseradish peroxidase-conjugated secondary antibodies from Cell Signaling Technology (USA). We visualized total protein amounts loaded in each lane on an Azure Biosystem c600 imager using stain-free gels (Bio-Rad Laboratories, Denmark) or using pan-actin as loading control (primary antibody diluted 1:1000: #4968, secondary antibody diluted 1:2000: #7074; Cell Signaling Technology) [35].

2.5 Histology

Atria and ventricles were fixed overnight in neutral-buffered 10% formalin, paraffin-embedded, and cut to 3-μm sections, and stained with picrosirius red to study collagen in the extracellular matrix. Detection was done using fluorescence microscopy comparing collagen content (excitation: 575 nm, emission: 600 nm) to overall wall mass interpreted from auto-fluorescence (excitation: 475 nm, emission: 525 nm). The relative collagen content was analyzed semi-automatically using Fiji software [36].

2.6 Single-Cell RNA Sequencing

We obtained transcriptomic data at single-cell resolution from a previous study evaluating adult human hearts [37]. Data from cardiomyocyte populations in the left atrium and ventricle were extracted for SLC4A3, SLC4A4, SLC4A7, and SLC9A1.

2.7 In Vivo Experiments

Systemic blood pressure, heart rate, and electrocardiograms were recorded from NBCn1 knockout and wild type mice using radiotelemetry (Data Sciences International, USA) [12]. To implant transmitters (HD-X11, Data Sciences International), mice were initially injected intraperitoneally with 0.1 mg/kg buprenorphine (Temgesic, Indivior Europe Limited, Ireland), and 30 min later, anesthesia was induced with 4% isoflurane and maintained with 1.4% isoflurane in pure O₂ on a mask. Surgical procedures were performed on a homeothermic blanket system (50-7222F, Harvard Apparatus, MA, USA) to maintain core temperature at 37°C. After the operation, mice were allowed to recover for at least 2 weeks before the measurements. Telemetry recordings were analyzed separately for the light (6 AM to 6 PM) and dark (6 PM to 6 AM) periods. Duration of the QRS complex, and RR, PR, and QT intervals were calculated using Ponemah software (v. 6.42; Data Sciences International). From the telemetry recordings, we also derived mouse physical activity data that are a function of both distance and speed of animal movement.

We measured cardiac function and structure by transthoracic echocardiography using a Vevo 3100 ultrasound system (FUJIFILM VisualSonics, Canada). The mice were lightly sedated using 1% isoflurane in air and immobilized in a supine position on a heating pad with integrated ECG electrodes. Body core temperature was measured with a rectal thermal probe and kept constant at 37°C. After a stabilization period of 10 min, parasternal long axis, short axis, and 4-chamber views were acquired and assessed using the Vevo LAB 5.8.2 software (FUJIFILM VisualSonics) to estimate systolic and diastolic heart function.

4.1 NBCn1 Is Differentially Expressed Between Atria and Ventricles

We first evaluate expression by immunoblotting on tissue lysates and find that NBCn1 shows higher protein expression in atria than ventricles (Figure 1A,B). We also note that the estimated molecular weight of NBCn1 from wild type mice is approximately 3 kDa higher in atria than in ventricles (Figure 1C). As expected, the protein expression of NBCn1 is absent in cardiac samples from NBCn1 knockout mice (Figure 1A,B).

Heterogeneity between cell types (e.g., endothelial cells vs. cardiomyocytes) and between species complicates analyses based on tissue lysates. To address this point and provide information with human translational value and potential clinical implications, we next extend our observations using publicly available single-cell RNA sequencing data from adult human heart (Figure 1D–G). Whereas the mRNA levels for SLC9A1 (Figure 1D), SLC4A3 (Figure 1E), and SLC4A4 (Figure 1F) are very uniform between cardiomyocytes from the left atrium and ventricle, SLC4A7 mRNA is more abundantly expressed—with 92% ± 25% higher transcript levels—in cardiomyocytes from atria compared to ventricles (Figure 1G). The preferred expression of NBCn1 in atria is consistent with our previous observation that Slc4a7 transcriptional activity is substantial in murine atrial tissue but undetectable in ventricular tissue [5].

4.2 NBCn1 Facilitates Net Acid Extrusion in the Cardiac Atria

We next assess whether the preferential expression of NBCn1 in cardiac atria is reflected in the pH_i regulatory function within atrial strips. In the presence of CO₂/HCO₃⁻, steady-state pH_i is very similar in atrial strips from wild type (7.15 ± 0.03) and NBCn1 knockout (7.14 ± 0.04) mice (Figure 2A,B). Omission of CO₂/HCO₃⁻ acidifies the atrial tissue, which indicates that atrial cardiomyocytes perform net HCO₃⁻ uptake at steady-state (Figure 2A,B). Yet, as illustrated in Figure 2B, the acidification upon removal of CO₂/HCO₃⁻ in atrial tissue from NBCn1 knockout mice (ΔpH_i = −0.15 ± 0.04) is not significantly different from that in wild type mice (ΔpH_i = −0.23 ± 0.04).

To study the mechanisms of net acid extrusion, we induce intracellular acidification by the NH₄⁺-prepulse technique [27]. We add and subsequently wash out NH₄Cl to acid load the atrial strips and evaluate their ability to recover pH_i in the absence and presence of extracellular Na⁺ (Figure 2A). The atrial tissue from NBCn1 knockout mice shows slower pH_i recovery and attenuated net acid extrusion relative to wild type mice when investigated in the presence of CO₂/HCO₃⁻ (Figure 2A,C). The NBCn1-dependent recovery of pH_i is Na⁺-dependent and abolished in the absence of CO₂/HCO₃⁻ (Figure 2A,C), which is consistent with the function of NBCn1 as a plasmalemmal Na⁺,HCO₃⁻-cotransporter.

4.3 NBCn1 Does Not Noticeably Contribute to Net Acid Extrusion From Ventricular Cardiomyocytes

We next turn to ventricular cardiomyocytes and explore their pH_i regulatory function. Consistent with the lower expression of NBCn1 in ventricular compared to atrial cardiomyocytes (Figure 1B,G), we observe no effect of NBCn1 knockout on steady-state pH_i (Figure 3A) or net acid extrusion activity after NH₄⁺-prepulse-induced intracellular acidification (Figure 3B). The relative contribution of Na⁺,HCO₃⁻-cotransport to net acid extrusion appears overall smaller in ventricles compared to atria (Figures 2C and 3B); and in accordance, omission of CO₂/HCO₃⁻ does not significantly influence steady-state pH_i in the ventricular cardiomyocytes (Figure 3A).

To probe the possibility that knockout of NBCn1 does not reach functional consequence due to compensation by other acid–base transporters, we investigate the expression levels for the two other major net acid extruders, NBCe1 and NHE1. We find no difference in the ventricular protein expression levels of NHE1 or NBCe1 between NBCn1 knockout and wild type mice (Figure 3C).

4.4 Electrophysiology of Ventricular Cardiomyocytes Is Not Markedly Affected by NBCn1 Knockout

Changes to pH_i or the mechanisms of net acid extrusion—particularly shifts from electroneutral transport via NBCn1 and NHE1 to electrogenic transport via NBCe1—have potential electrophysiological consequences [6]. Therefore, we next study telemetrically recorded electrocardiograms from wild type and NBCn1 knockout mice (Figure 4A) and find no differences in RR, PR, or QT interval lengths, or in the duration of the QRS complex (Figure 4B). We report the unadjusted interval durations, because we observe no differences in heart rate between wild type and NBCn1 knockout mice, and previous studies find that heart rate correction using approaches optimized for humans can lead to errors in data interpretation from mice [39]. We speculate that subtle electrophysiological effects of NBCn1 might be revealed when pH_i regulation of atrial cardiomyocytes is challenged by elevated metabolism at higher heart rates. However, even when we plot the PR interval duration as a function of heart rate, we observe no differences between the wild type and NBCn1 knockout mice (Figure 4C). When we evaluate action potentials recorded from ventricular cardiomyocytes by perforated patch clamp technique (Figure 4D), we observe no effect of NBCn1 knockout on the resting membrane potential (Figure 4E), the action potential amplitude (Figure 4F) or the rate of action potential decay (Figure 4G) between isolated ventricular cardiomyocytes from wild type and NBCn1 knockout mice. The normal electrocardiogram and electrophysiology of ventricular cardiomyocytes from NBCn1 knockout mice are consistent with the unperturbed pH_i regulation in the ventricles (Figure 3A,B) and our observation that (electrogenic) NBCe1 does not compensate for the lack of (electroneutral) NBCn1 (Figure 3C).

4.5 NBCn1 Knockout Mice Are Hypertensive With Mild Cardiac Hypertrophy and Delayed Ventricular Relaxation

Using radiotelemetry, we next explore the blood pressure of freely moving NBCn1 knockout and wild type mice. We previously showed [12], and confirm here, that mice lacking NBCn1 are hypertensive with a 15–20 mmHg elevation in diastolic, systolic, and hence mean systemic blood pressure (Figure 5A,B) but no difference in heart rate (Figure 5C). As shown in Figure 5B, the blood pressure difference between NBCn1 knockout and wild type mice is observed both when the mice are awake (night) and during sleep (day). The physical activity of the mice shows the expected circadian pattern—with greater activity at night-time compared to daytime—but it does not vary between NBCn1 knockout and wild type mice (Figure 5D) and therefore cannot explain the genotype-dependent differences in blood pressure.

The hearts from NBCn1 knockout mice show hypertrophy evaluated by postmortem weighing (~8%; Figure 5E) and echocardiography (~20%; Figure 5F–H). Whereas the left ventricular chamber volumes (Figure 5I,J) are unaffected in NBCn1 knockout mice, the wall thickness is increased (Figure 5H). The cardiac hypertrophy appears to compensate for the elevated afterload as we observe maintained stroke volume (Figure 5K), left ventricular ejection fraction (Figure 5L), and cardiac index (Figure 5M). We also see no difference in the left ventricular myocardial global longitudinal strain (Figure 5N). The cardiac hypertrophy is not related to cardiomyocyte enlargement since the projected areas of isolated cardiomyocytes from the atria and ventricles do not reveal genotype-related differences in cell size (Figure 6A). Our finding that mouse atrial cardiomyocytes are smaller than ventricular cardiomyocytes is in line with previous reports [40].

By pulsed wave Doppler, we measure a prolonged left ventricular isovolumic relaxation time (Figure 5O,P). Impaired ventricular relaxation is consistent with the greater ventricular mass (Figure 5E,F) and thicker ventricular wall (Figure 5H). Other echocardiographic variables do not differ significantly between wild type and NBCn1 knockout mice (Table S1). The prolonged isovolumic relaxation time, but normal left ventricular early diastolic filling velocity, late diastolic filling velocity, and early diastolic transmitral flow deceleration time (Table S1) are consistent with findings from humans with hypertension and left ventricular hypertrophy [41].

Increased extracellular matrix deposition represents an alternative explanation for the greater cardiac mass. To evaluate if extracellular collagen accumulation plays a role in the structurally modified heart of the NBCn1 knockout mice, we perform picrosirius red staining, which shows no differences in collagen content between hearts from wild type and NBCn1 knockout mice (Figure 6B).