All procedures involving animals were performed in accordance with the guidelines established by the Canadian Council on Animal Care and the University of Alberta Animal Care and Use Committee.

Experiments were performed using ae3 null mice in a C57BL/6 background. The ae3 null strain has been previously described and characterized [42]. Age-matched WT mice from separate breedings were used as controls.

Hematoxylin and eosin (HE) staining was performed on longitudinal and transverse sections of wildtype and ae3 null adult mouse heart, using previously described protocols [45, 46]. Briefly, paraffin-embedded hearts were sectioned into 3 μm slices, which were trimmed and floated onto a water bath at 42°C, containing 50 mg/l of gelatin while gently stretching the cut sections to avoid wrinkles. Poly-L-lysine coated microscope slides were dipped under the meniscus of the water bath and a tissue slice was carefully mounted onto it. Sections were then air-dried for 16 h at 20°C, after which the slides were placed on edge in an oven and baked for 15 min at 60°C. Sections were deparaffinized by successively immersing them for 5 min with agitation in xylene, 100% ethanol and 70% ethanol, and rehydrated in Tris-EDTA buffer (1 mM EDTA, 0.05% Tween 20, 10 mM Tris, pH 9.0) for 1 min. Slides were rinsed in distilled water for 1 min with agitation. Slides were agitated for 30 s in Mayer’s hematoxylin solution (1.0 g/l hematoxylin (Sigma), sodium iodate (0.2 g/l), aluminum ammonium sulfate · 12 H₂O (50 g/l), chloral hydrate (50 g/l) and citric acid (1 g/l) and rinsed in water for 1 min. Slides were stained in 1% eosin Y solution (1% eosin Y aqueous solution, Fisher) for 30 s with agitation. Sections were dehydrated by successively immersing it twice in 95% ethanol and twice in 100% ethanol for 30 s each. Ethanol was extracted twice in xylene, followed by addition of two drops of mounting medium (Canada Balsam, Sigma), after which the sections were covered with a coverslip. Images of transverse and longitudinal sections were collected, using a Nikon digital camera (DXM 200) mounted on top of a Nikon Eclipse E600 microscope.

Echocardiographic assessment of cardiac performance in male WT and ae3 ^−/− mice was performed by the Cardiovascular Research Centre Core Facility (University of Alberta). Mice were subjected to mild anesthesia by isoflurane inhalation and echocardiography parameters were measured using a Vevo 770 High-Resolution Imaging System with a 30-MHz transducer (RMV-707B; Visual Sonics, Toronto). M-mode images and a four chamber view allowed for the calculation of wall measurements, ejection fraction, fractional shortening, and mitral velocities (E and A). Mitral valve tissue motion (E’) was measured, by tissue Doppler echocardiography of the mitral septal annulus.

Non-invasive blood pressure measurements were performed by the Cardiovascular Research Centre Core Facility (University of Alberta). Mice were comfortably restrained in a 26°C warming chamber (IITC Life Science) for ~15 min prior to taking blood-pressure (BP) measurements. Tail-cuff sensors were secured on the tail to occlude the blood flow, and connected to the recording device. Systolic and diastolic pressure, heart rate, blood volume and flow were obtained, using CODA6 software (Kent Scientific Corporation, Connecticut, USA).

Cardiomyocytes from adult male mouse hearts were isolated and cultured with modifications of published protocols [21, 47]. Briefly, adult mice were euthanized with sodium pentobarbital (50 mg/kg body mass) by intraperitoneal injection. Upon reaching surgical plane, midsection thoracotomy was performed and hearts were quickly excised and placed in 4°C perfusion solution, containing in mM: 120 NaCl, 5.4 KCl, 1.2 MgSO₄, 5.6 glucose, 10 2,3-butanedione monoxime (BDM) (Sigma), 5 taurine (Sigma), 1.2 NaH₂PO₄, 10 HEPES, pH 7.4. Extra-cardiac tissues were removed and hearts were subjected to retrograde perfusion with perfusion solution to remove excess blood. Perfusion was switched for 15 min to perfusion solution at 37°C, supplemented with 0.5 mg/ml collagenase type B (Roche), 0.5 mg/ml collagenase type D (Roche), 0.02 mg/ml protease XIV (Roche) and 50 μM CaCl₂. Ventricles, partially digested at this stage, were removed, cut into several pieces, and digested further in the same enzyme digestion solution by gentle trituration with a transfer pipette. Once the ventricles were completely digested, enzymatic digestion was terminated by addition of Digestion stop buffer I (perfusion solution, containing 10% (v/v) fetal bovine serum (FBS) (GIBCO) and 50 μM CaCl₂). Lysates settled under gravity for 10 min at room temperature and pellets were resuspended in myocyte stopping buffer II (perfusion solution, containing 5% (v/v) FBS and 50 μM CaCl₂). Samples were transferred to 60 mm tissue culture dish and calcium levels were increased by addition of CaCl₂ to obtain final concentrations of 62, 112, 212, 500 and 1000 μM, sequentially at 4 min intervals at 20°C. Cells were transferred to a 14 ml culture tube and allowed to sediment for 10 min under gravity. Cells were resuspended in myocyte culture medium (Dulbecco’s Modified Eagle’s Medium/Nutrient Mixture F12-Ham (Sigma), supplemented with 10 mM BDM, 5% (v/v) FBS, 1% penicillin (GIBCO), 10 mM BDM, and 2 mM L-glutamine (GIBCO)). Myocytes were plated at a density of (0.5-1) x 10⁴ cells/cm² onto 35 mm culture dishes, pre-coated for 2 h with 10 μg/ml mouse laminin (Invitrogen) in PBS. Cells were incubated at 37°C in a 5% CO₂ incubator for 1 h at which point medium was replaced with Dulbecco’s Modified Eagle’s Medium/Nutrient Mixture F12-Ham, containing 10 mM BDM, 1% penicillin, 2 mM L-glutamine, 0.1 mg/ml bovine serum albumin, and 1x ITS Liquid Media Supplement (Sigma). The entire culture procedure was carried out in a sterile laminar flow hood.

Cardiomyocytes were isolated and cultured from adult mouse hearts as described above. Following 18 h of culture, myocytes were treated with solvent carrier (control), 10 μM phenylephrine (Sigma) or 1 μM angiotensin II (Sigma) for another 24 h. Hypertrophy was assessed by analysis of cell surface area of cardiomyocytes pre- and post-treatment with the hypertrophic agonists. Images of characteristic rod-shaped cardiomyocytes were collected with a QICAM fast-cooled 12-bit colour camera (QImaging Corporation). Cell surface areas were measured, using Image-Pro Plus software (Media Cybernetics). Each treatment group contained 100–200 cells of ~ ten different experiments. Cell surface area (% relative to control) = Surface area (post-treatment)/Surface area (pre-treatment) X 100.

Cardiomyocytes were prepared and treated as above. At the end of the culture period, medium was aspirated and cardiomyocytes were harvested in 350 μl Buffer RLT (Qiagen). RNA was extracted from the lysates with an RNeasy Plus Mini Kit, as per manufacturer’s instructions (Qiagen). RNA samples (100 ng) were reverse transcribed, following the manufacturer’s instructions for SuperScript II™ reverse transcriptase (Invitrogen). qRT-PCR was performed in a Rotorgene 3000 real time thermal cycler (Corbett Research), using a reaction mix containing: 5 μl template cDNA, 12 μl 2x Rotor-Gene SYBR Green PCR Master Mix (Rotor-Gene SYBR Green PCR Kit, Qiagen), and 1 μM of each primer. Data were obtained and analyzed, using Rotor Gene 6.0.14 software. Cycle threshold (Ct) values were obtained for carbonic anhydrase II (CAII), NHE1, atrial natriuretic peptide (ANP), β-MHC and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Primers (Table 1) were designed, using Primer3 (http://​Frodo.​wi.​mit.​edu/​primer3/​).

Cardiomyocytes, isolated and cultured from adult mouse hearts, were subjected to drug intervention as described above. Twenty-four h following treatment, medium was aspirated and myocytes were washed with 4°C PBS. Cells were lysed with SDS-PAGE sample buffer (10% (v/v) glycerol, 2% (w/v) SDS, 2% 2-mercaptoethanol, 0.001% (w/v) bromophenol blue, 65 mM Tris, protease inhibitors (1 μg/ml) pH 6.8) and lysates were heated 5 min at 65°C. Protein concentrations were determined by the bicinchoninic acid (Pierce Biotechnology) assay [48], and 20 μg of protein was resolved by SDS-PAGE on 10% acrylamide gels. Proteins were transferred onto PVDF membranes by electrophoresis for 1 h at 100 V in transfer buffer (10% (v/v) methanol, 25 mM Tris, and 192 mM glycine). PVDF membranes were blocked for 30 min with 5% (w/v) nonfat dry milk/ 0.1% (v/v) Tween 20 in TBS (137 mM NaCl, 20 mM Tris, pH 7.5). Immunoblots were incubated with rabbit polyclonal anti-CAII antibody (Santa Cruz Biotechnology; 1:1000), rabbit anti-human SLC26a6 (1:1000) [49], or rabbit polyclonal anti-NHE1 antibody (1:1000) [32] in TBST-M for 16 h at 4°C. Immunoblots were washed with TBST (TBS, containing 0.1% (v/v) Tween 20) and incubated with donkey anti-rabbit IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology; 1:2000) or mouse anti-goat IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology; 1:2000) for 1 h at room temperature. Immunoblots were washed in TBST and visualized, using ECL reagents (Perkin Elmer) and a Kodak Imaging Station 440CF (Kodak, Rochester, NY). Proteins were quantified by densitometry, using Kodak Molecular Imaging software (version 4.0.3; Kodak). Immunoblots were stripped by incubating in 10 ml of stripping buffer (2% (w/v) SDS, 10 mM 2-mercaptoethanol, 62.5 mM Tris, pH 6.8) at 50°C for 10 min with occasional shaking, followed by three washes with TBST. Membranes were incubated with mouse monoclonal anti-β-actin antibody (Santa Cruz Biotechnology; 1:2000) for 1 h at 20°C, washed with TBST, and incubated with sheep anti-mouse IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology; 1:3000) for 1 h. Immunoblots were washed and visualized again, as described above.

Cardiomyocytes were prepared and subjected to hypertrophic stimulation as above. Radiolabeled phenylalanine ([³H]-Phe, 1 μCi/ml, (Perkin Elmer)) was added immediately after drug intervention and cells were incubated for another 24 h. Proteins were precipitated, using trichloroacetic acid (TCA) as described previously with some modifications [50]. Medium was carefully aspirated and 500 μl of 0.5% (v/v) Triton X-100, containing protease inhibitor cocktail (Roche) were added. Lysates were transferred into 1.5 ml microcentrifuge tubes and TCA (100%: 500 g in 350 ml H₂O) was added to each tube to a final concentration of 40% (v/v). Samples were incubated at 4°C for 30 min, after which proteins were sedimented by centrifugation at 12 682 x g for 15 min at 4°C. Pellets were resuspended by adding 200 μl acetone (−20°C) and sedimented again by centrifugation, as above, for 10 min. The acetone wash was repeated one more time, and the pellets air dried for 20 min at room temperature. Pellets were resuspended in 200 μl of 0.2 M NaOH, 1% (w/v) SDS. Scintillation fluid (Perkin Elmer, 3.5 ml) was added to each sample and the radioactivity of [³H]-Phe was counted in a Beckman LS6500 liquid scintillation counter.

The protocol was as described previously with minor modifications [51, 52]. Briefly, cardiomyocytes were isolated as described above and cultured on laminin-coated glass coverslips. Approximately 2 h later, cells were loaded with 2 μM BCECF-AM (Sigma-Aldrich, Canada) for 30 min at 37°C. Coverslips were placed in an Attofluor cell chamber (Invitrogen, Canada), then transferred onto the stage of a Leica DMIRB microscope. Perfusion with HCO₃ ⁻ Ringer’s buffer solution (in mM: 128.3 NaCl, 4.7 KCl, 1.35 CaCl₂, 20.23 NaHCO₃, 1.05 MgSO₄ and 11 glucose, pH 7.4) was initiated at 3.5 ml/min. Solutions were bubbled with 5% CO₂-balanced air. Intracellular alkalosis was induced by switching to a HCO₃ ⁻ Ringer’s solution, containing, 20 mM trimethylamine (TMA) (Sigma) and perfusion was continued for 3 min. Perfusion was switched back to the HCO₃ ⁻ Ringer’s buffer solution. pH_i of individual cardiomyocytes was measured by photometry at excitation wavelengths of 502.5 nm and 440 nm with a Photon Technologies International (PTI, Lawrenceville, NJ, USA) Deltascan monochromator. Emission wavelength, 528.7 nm, was selected, using a dichroic mirror and narrow range filter (Chroma Technology Corp., Rockingham, VT, USA) and was measured with a PTI D104 photometer. At the end of experiments, pH_i was clamped by the high K⁺/Nigericin technique [53] in calibration solutions containing, 140 mM KCl, 1 mM MgCl₂, 2 mM EGTA, 11 mM glucose, 20 mM BDM, 10 mM HEPES. Three pH standards spanned a range of 6.5-7.5. Steady-state pH_i was measured from the pH_i value prior to induction of alkalosis. Rate of pH_i recovery was measured by linear regression from first min of recovery from imposed alkalosis.

To examine the role of AE3 in heart development, we characterized age-matched wildtype (WT) and ae3 ^−/− mice (~three months old). Breeding yielded Mendelian ratios of litters of both WT and ae3 ^−/− mice, indicating that loss of the AE3 does not affect prenatal survival. While the body weights of age-matched WT and ae3 ^−/− mice were not significantly different (Figure 1A), the difference in heart weights between the groups was statistically significant (Figure 1B). As a result, the HW/BW ratio, an index of cardiac hypertrophy, was reduced in the ae3 ^−/− mice relative to their WT counterparts (Figure 1C and D), suggesting that AE3 has a role in heart growth. Despite the observed apparent smaller size of the ae3 null hearts, gross morphology of hearts stained with hematoxylin/eosin transverse and longitudinal sections revealed no differences in the wall size and the chamber diameter of WT (left panel) and ae3 null (right panel) mouse hearts (Figure 2B-C).

Cardiovascular performance of age-matched WT and ae3 ^−/− mice was assessed by echocardiography. No major variations in cardiovascular functional parameters between WT and ae3 ^−/− mice (~three months old males) were observed, except for a significant decrease in the mitral value E/A ratio in ae3 ^−/− mice (Table 2). While this would suggest that more blood is entering the ventricle during the atrial systolic phase than during ventricular relaxation, other parameters of diastolic cardiac function (E/E’ ratio, IVRT) were unaffected. ae3 ^−/− mice therefore likely do not exhibit diastolic dysfunction. Systemic blood pressure measurements were also performed using the non-evasive tail cuffing approach. No significant difference in the systemic blood pressure of WT and ae3 ^−/− mice was observed (Table 2). Overall, these observations suggest that loss of AE3 does not affect cardiovascular performance under basal conditions, consistent with previous findings [42, 44].

Cardiomyocyte hypertrophy is characterized by an increase in cardiomyocyte surface area, resulting in an overall increase in heart size. Cardiomyocytes were isolated from adult WT and ae3 ^−/− mice and the cell surface assessed by morphometry. The cell surface area of ae3 null cardiomyocytes was 20% ± 4% (n = 6) lower than WT (Figure 3). To determine the response of cardiomyocytes to pro-hypertrophic stimulation, adult cardiomyocytes were cultured and treated with PE and ANGII 18 h later. Cell surface area was measured 24 h following treatment with hypertrophic agonists. PE and ANGII induced a 20–25 ± 4% (n = 10) increase in the cell surface area of WT cardiomyocytes, but cardiomyocytes from ae3 ^−/− hearts were not susceptible to pro-hypertrophic stimulation by these agents (Figure 4). This suggests that AE3 has a role in the hypertrophic signaling pathway downstream of PE and ANGII.

Cardiac hypertrophic development is associated with increased expression of marker genes, including ANP [54], β-myosin heavy chain (β-MHC) [55] and α-skeletal actin [56]. mRNA and protein levels of these markers are elevated in hypertrophic hearts [57]. Expression levels of ANP and β-MHC, were assessed by qRT-PCR in cardiomyocytes subjected to pro-hypertrophic stimulation. Transcript abundance of ANP and β-MHC were not significantly different between untreated cardiomyocytes from WT and ae3 ^−/− mice (Figure 5A and B). Stimulation with PE and ANGII, however, led to 8–10 fold increase of ANP and β-MHC expression levels in cardiomyocytes from WT mice. In contrast, ANP and β-MHC increased only 1–2 fold in cardiomyocytes from ae3 ^−/− mice. These data suggest that pro-hypertrophic stimulation induces a much greater upregulation of hypertrophic marker genes in cardiomyocytes from WT mice than ae3 ^−/− cardiomyocytes.

AE3 is implicated as part of a functional complex with NHE1 and CAII, referred to as the hypertrophic transport metabolon (HTM), whose activation has been proposed to induce cardiac hypertrophy [32, 33]. To assess the possibility of functional compensation for loss of AE3 by altered expression of these partner proteins, we examined mRNA expression of NHE1 and CAII by qRT-PCR. Baseline expression level of NHE1 in ae3 ^−/− mice was significantly elevated compared to the WT mice (Figure 6A). Pro-hypertrophic stimulation with PE or ANGII did not influence the NHE1 transcript abundance in the WT or ae3 ^−/− cardiomyocytes (Figure 6A). CAII transcript abundance, however, was higher in cardiomyocytes from ae3 ^−/− mice than from WT mice (Figure 6B). ae3 ablation thus induces a compensatory increase in CAII expression. Interestingly, pro-hypertrophic stimulation markedly increased CAII transcript levels in both WT and ae3 ^−/− cardiomyocytes (Figure 6B).

Since the baseline expression level of CAII was elevated in ae3 ^−/− cardiomyocytes and was further enhanced by hypertrophic stimulants in WT cardiomyocytes, we evaluated CAII protein expression. Cardiomyocytes isolated from ae3 ^−/− and WT male adult mice hearts were probed for CAII on immunoblots (Additional file 1: Figure S1A). CAII migrated at ~27 kDa, consistent with the expected molecular weight of CAII. Both the steady-state level of CAII protein (Additional file 1: Figure S1B) and mRNA level for CAII (Figure 6B) were higher in ae3 ^−/− cardiomyocytes than WT cardiomyocytes. Importantly, however, there was a large quantitative difference in the response, where CAII protein level rose by about 50%, whereas CAII message rose was about eight-fold higher in the ae3 ^−/− mice than WT. PE and ANGII increased CAII levels in the WT cardiomyocytes. Contrastingly, CAII levels in ae3 ^−/− cardiomyocytes were not significantly affected by pro-hypertrophic stimulation (Additional file 1: Figure S1B).

Cardiac hypertrophy is also characterized by an increase in protein synthesis to accommodate cardiomyocyte enlargement [58]. Protein synthesis was measured by determining the amount of radioactive [³H]-Phe incorporated into proteins in cardiomyocytes, isolated and cultured as described above and treated with PE and ANGII. [³H]-Phe incorporation into proteins in the presence of pro-hypertrophic stimuli was increased in cardiomyocytes from WT, but not ae3 ^−/− mice. We conclude that ae3 ^−/− mice do not respond to pro-hypertrophic agonists with increased protein synthesis (Figure 7).

AE3 contributes to cardiomyocyte pH_i regulation [59‐61], but there are no AE3-specific inhibitors that would enable delineation of the role of AE3 in pH_i regulation. We thus examined the rate of recovery of pH_i from an imposed intracellular alkalinization in freshly isolated cardiomyocytes. Cardiomyocytes were cultured on laminin-coated glass coverslips for 2 h, then incubated with the pH-sensitive fluorescence dye, BCECF-AM. Cells were perfused with a HCO₃ ⁻-containing Ringer’s buffer until steady-state pH_i was reached and the perfusion was switched to the HCO₃ ⁻-containing Ringer’s buffer, containing TMA. The presence of TMA induced a rapid intracellular alkalinization (Figure 8A) until a new steady-state pH_i was reached. pH_i regulatory transporters spontaneously restored cell pH_i. Steady-state pH_i was measured as the pH_i 60 s prior to switching to the TMA-supplemented HCO₃ ⁻-containing Ringer’s buffer. The rate of pH_i recovery from imposed intracellular alkalinization was measured by linear regression of the initial minute of pH_i recovery. Baseline pH_i was identical in cardiomyocytes from WT and ae3 ^−/− mice (Figure 8B). The mean peak pH following alkalization for WT group was 7.65 ± 0.05 and 7.60 ± 0.06 for ae3 ^−/− mice. Thus, under basal physiological conditions, loss of AE3 does not significantly affect the steady-state pH_i in cardiomyocytes. The rate of recovery of pH_i from alkalosis was, however, significantly slower in ae3 ^−/− cardiomyocytes than WT (Figure 8C).

We next assessed the expression level of the other pH_i regulatory transporters at the protein level. On immunoblots the proteins migrated at the expected sizes, NHE1 at ~100 kDa, and SLC26a6 at ~80 kDa (Additional file 2: Figure S2 and Additional file 3: Figure S3). Consistent with a previous study [44], expression of NHE1 was elevated in ae3 ^−/− cardiomyocytes compared to WT, but remained unchanged in the heterozygotes (Additional file 2: Figure S2A-B). Expression of SLC26a6 protein was not affected by deletion of the ae3 gene (Additional file 3: Figure S3A-B).

The role of AE3 in cardiac physiology is incompletely characterized, but several lines of evidence suggest that AE3 Cl⁻/HCO₃ ⁻ exchange is required to maintain pH_i homeostasis [10, 61]. Consistent with this, we found that the rate of recovery from an alkaline load was reduced in ae3 ^−/− cardiomyocytes, relative to WT.

The hypertrophic transport metabolon is a proposed pathological pathway in which AE3, NHE1 and CAII are coordinately activated and promote hypertrophic growth [32, 33]. Specifically, pro-hypertrophic agonists, including PE, ANGII and endothelin are coupled to PKC activation. NHE1 and AE3 are both activated by agonists coupled to PKC activation [59, 62, 63]. Co-activation of these respective alkalinizing and acidifying transporters has the net effect of loading the cell with NaCl, with no change of cytosolic pH. Elevated cytosolic Na⁺ in turn reduces the efficacy of Ca⁺⁺-efflux by the plasma membrane Na⁺/Ca⁺⁺ exchanger, resulting in a rise in cytosolic Ca⁺⁺. Ca⁺⁺ is a pro-hypertrophic second messenger [64, 65] and also may promote hypertrophy in a feed-forward cascade by stimulation of PKC [6]. Accumulating evidence suggests that both NHE1 and AE3 activities are influenced by physical and functional interactions with CAII [32, 41], which provides substrates for these transporters. That said, very recent data suggest that intracellular carbonic anhydrase does not activate cardiomyocyte Na⁺/H⁺ exchange [66]. Prevention of PE-induced cardiomyocyte hypertrophy upon CAII blockade [32, 33] may thus arise from reduced AE3-mediated cytosolic acidification, thereby decreasing driving force for NHE-mediated alkalinization. We propose that CAII, NHE1 and AE3 form a hypertrophic transport metabolon, where hypertrophy is promoted by the pathological activation of AE3 and NHE1, stimulated by interactions with CAII.

The functional relationship between AE3, CAII and NHE1 was further supported by our analysis of protein expression in ae3 ^−/− mice. Increased CAII transcript abundance and protein expression in ae3 ^−/− mice compared to WT mice suggest that there is compensation for a loss of AE3. This finding parallels results in retinal tissue from ae3 ^−/− mice, where there was increased CAII expression [67]. Functional complementarity of AE3 and CAII is further supported by the significant increase in AE3 transcript abundance in Car2 (caii ^−/− ) mice, compared to WT mice [33]. The upregulation of NHE1 transcript abundance and increased protein expression provide further support for the HTM. Taken together, these data support a functional interaction between AE3 and CAII, where there is compensation of one for the loss of the other.

This study lends support to the idea that AE3 is the Cl⁻/HCO₃ ⁻ exchanger isoform working in conjunction with NHE1 to promote cardiomyocyte hypertrophy. Non-specific inhibition of Cl⁻/HCO₃ ⁻ exchangers, using stilbene derivatives, prevented hypoxia-induced acidification in rat ventricular myocytes, as well as increases in intracellular Cl⁻ and Ca⁺⁺ concentrations [68, 69], suggesting a role of Cl⁻/HCO₃ ⁻ exchangers in cardiac pathology. When subjected to ischemia and reperfusion, hearts isolated from ae3 ^−/− mice revealed no effect on cardiac performance demonstrated as contractility, ventricular developed pressure or end diastolic pressure relative to wildtype [42]. Double knock-out ae3/nkcc1 (Na⁺-K⁺-2Cl⁻ co-transporter) mice, however, had elevated ischemia/reperfusion injury, which resulted in impaired cardiac contractility and overall cardiac performance [42]. These findings were attributed to impaired Ca⁺⁺ handling in the double knock-out cardiomyocytes [42] compared to the single mutants. In a hypertrophic cardiomyopathy mouse model carrying a Glu180Gly mutation in α-tropomyosin (TM180), disruption of ae3 did not prevent or reverse the hypertrophic phenotype [43]. The TM180/ae3 double knockout mice had reduced cardiac function and compromised Ca⁺⁺ regulation, which accounted for the rapid decline to heart failure [43].

Taken together, these two studies suggest that AE3 loss is not cardioprotective, which contrasts with findings of the present study, which found that loss of AE3 renders cardiomyocytes less susceptible to pro-hypertrophic stimulation. Our data showed that the marked rise in cell surface area, protein synthesis, and fetal gene reactivation observed in response to hypertrophic stimulation in cardiomyocytes from WT mice was not present in ae3 ^−/− mice. In the context of cardiomyocyte hypertrophy mediated by hormonal stimuli, our data demonstrate that ablation of AE3 affords protection against cardiomyocyte hypertrophy.

The discrepancy could arise from the model of hypertrophy and the cardiac pathology being investigated. Hypertrophic cardiomyopathy is a genetic disorder, which occurs as a result of mutations in the genes that encode cardiac contractile proteins [70]. This anomaly manifests as sudden cardiac death, arrhythmias, hypertrophy and heart failure [70]. Overall, hypertrophic cardiomyopathy results in impairment of Ca⁺⁺ sensitivity by the myofibrils. In the present study however, we employed a model of hypertrophy induced by PE or ANGII, which involves interaction of these ligands with their cell surface receptors, GPCR [71]. The resultant intracellular response leads to increased cytosolic Ca⁺⁺ overload, which mediates a cascade of signaling pathway involving activation of PKC, which ultimately induces cardiomyocyte hypertrophy [72].

One other possible explanation for the discrepancy between the present findings and earlier studies of ae3 ^−/− mice is the assay of cardiac hypertrophy. Earlier investigations probed whole animal physiology, whereas we studied isolated cardiomyocytes in culture. Isolated cardiomyocytes are an established model, which may be more sensitive in detecting alterations of heart cell growth than assessments of functional alterations of heart function sued in earlier studies.

Interestingly, AE3 also forms a complex with CAXIV in the myocardium [73]. In the hypertrophic myocardium of SHR rats, exacerbated AE3/CAXIV activation was proposed to elicit hypertrophic growth of the heart [73]. Thus, the etiology, signaling pathway and pathophysiology of hypertrophic cardiomyopathy are distinct from that mediated by hormonal factors. These differences could account for the disparity between the influence of AE3 on hypertrophy in the present report and that shown in previous models of cardiovascular disease [42, 43]. Since hypertrophic interventions by PE and ANGII failed to induce hypertrophy in the ae3 null cardiomyocytes in our study, the PKC-coupled hypertrophic cascade appears to require AE3. Recently, it was demonstrated that when paced, AE3 null hearts had an impaired force-dependent inotropy characterized by an elevation of protein kinase B (PKB, Akt) phosphorylation and a downregulation of AMPK activity [44]. This observation may provide additional support of the present findings that revealed that ae3 null cardiomyocytes are less susceptible to develop hypertrophy in upon pro-hypertrophic stimulation. Akt phosphorylation is central to the signaling pathway mediated by growth factors that induce physiological hypertrophy [57]. Thus, its activation under stressful conditions may trigger a physiological growth response to counter the pathological signaling cascade stimulated the source of stress. The lack of response by the AE3 knock-out cardiomyocytes to hypertrophic stimulants seen here could arise in part from effects on Akt phosphorylation.

ae3 ^−/− mice have been reported to have no apparent defects, and the results of our analysis of the cardiac function of ae3 ^−/− mice is comparable to these previous studies [42, 43, 74]. Combined analysis of echocardiographic measurements of ventricular wall dimensions, chamber diameter and cardiac function between the two genotypes further suggests that loss of AE3 does not affect hypertrophy or cardiovascular performance. A significant decrease in the HW/BW ratio in ae3 ^−/− mice is the result of a reduction in heart size, arising from a decrease in cardiomyocyte size. This suggests a critical role for AE3 in heart development.

Since cardiomyocyte steady-state pH was the same in ae3 ^−/− and WT mice, loss of Cl⁻/HCO₃ ⁻ exchange activity by ae3 deletion is likely compensated for by another protein. The principal Cl⁻/HCO₃ ⁻ exchanger of cardiomyocytes is SLC26a6 [23], making it the most likely acid loading transporter to compensate for loss of AE3. Since we did not see an increase in SLC26a6 expression, SLC26a6 activation may occur through post-translational mechanisms. We did note an increase of CAII expression in ae3 ^−/− mice. Since CAII increases the rate of CO₂/HCO₃ ⁻ inter-conversion, it increases the rate of observed Cl⁻/HCO₃ ⁻ for SLC26a6 [75]. Thus increased CAII expression might in part compensate for loss of AE3. We also noted a much larger increase in CAII message than we saw for CAII protein. Since proteins carry out cellular functions, not mRNA, we consider the protein change to be a more reliable measure of cell response than the mRNA increase. The reason for a muted increase in CAII protein level in comparison to the mRNA rise remains unclear.

The importance of AE3 in intracellular pH regulation was, however, evident in the reduced rate of pH_i recovery from imposed intracellular alkalosis in cardiomyocytes from ae3 ^−/− mice compared to WT. Endothelin 1 stimulation of myocardial Cl⁻/HCO₃ ⁻ exchange activity in isolated rat papillary muscle is almost totally attributable to the AE3 isoform, on the basis of inhibition by an anti-AE3 antibody [61]. This provides a possible explanation for the resistance of ae3 ^−/− mice to pro-hypertrophic stimuli; ae3 ^−/− cardiomyocytes have reduced acidifying activity to counter enhanced NHE1 activity associated with pro-hypertrophic stimulation.