OUP user menu

The heart requires glycerol as an energy substrate through aquaporin 7, a glycerol facilitator

Toshiyuki Hibuse , Norikazu Maeda , Hideaki Nakatsuji , Yoshihiro Tochino , Koichi Fujita , Shinji Kihara , Tohru Funahashi , Iichiro Shimomura
DOI: http://dx.doi.org/10.1093/cvr/cvp095 34-41 First published online: 18 March 2009


Aims Cardiomyocytes require fatty acids and glucose for energy production. However, other nutrients and substrates that may serve as possible candidates for a cardiac energy source have not been fully studied. Several reports showed that a moderate expression of aquaporin 7 (AQP7), a member of the aquaglyceroporin family that is permeated by glycerol and water, is observed in heart tissue. However, the functional role of cardiac AQP7 is not clear. The aim of this study was to investigate the significance of glycerol as a cardiac energy substrate and to clarify the role of cardiac AQP7.

Methods and results Heart function and morphology were examined in AQP7-knockout (KO) mice under basal conditions and during pressure overload [isoproterenol infusion and transverse aortic constriction (TAC)]. Glycerol uptake and glycerol-dependent ATP production were measured in AQP7-knockdown cardiac cells. Cardiac glycerol consumption was analysed in ex vivo beating hearts. Cardiac morphology and function in KO mice were similar to those of wild-type (WT) mice under basal conditions, although low glycerol and ATP content were noted in hearts of KO mice. In H9c2 cardiomyotubes, knockdown of AQP7 was associated with a significant reduction of glycerol uptake. The ex vivo heart study demonstrated that cardiac glycerol consumption levels in KO mice were significantly lower than those of WT mice. Furthermore, isoproterenol challenge induced severe left ventricular hypertrophy in KO mice, and TAC resulted in a higher mortality rate in KO mice than in WT mice.

Conclusion The results indicate that AQP7 acts as a glycerol facilitator in cardiomyocytes and that glycerol is a substrate for cardiac energy production.

  • Aquaporin
  • Glycerol
  • Heart
  • Energy metabolism

1. Introduction

Accumulating evidence suggests that aquaporins (AQPs) play a crucial role in water homeostasis.13 Some members of AQP superfamily are permeated by a variety of small neutral solutes, such as glycerol, and are subcategorized as aquaglyceroporins.4,5 To date, AQP3, 7, 9, and 10 have been recognized to belong to aquaglyceroporins, permeabilizing glycerol as well as water. We and other groups independently identified AQP7 in human adipose tissue and rat testis, respectively, and showed that AQP7 is a member of the aquaglyceroporins.6,7 A series of studies on AQP7-knockout (KO) mice concluded that AQP7 serves as an adipose glycerol channel in vivo and that AQP7-mediated glycerol release from adipocytes determines, in part, plasma glucose levels.810 AQP7 is abundantly expressed in adipose tissue and testis, but a moderate expression of AQP7 is also observed in cardiac tissue.5 Several studies have reported that AQP7 is the sole glycerol facilitator in heart tissue of mice,5,11,12 but its functional role has not been elucidated in the heart.

Cardiomyocytes utilize fatty acids and glucose as their energy source.13,14 Interestingly, the cardiac muscle requires mainly fatty acids under steady state, but shifts its metabolic reliance to glucose under pressure and/or volume overload. In fact, the significant roles of fatty acids and glucose in myocardial metabolic adaptations have been reported by analysing mice lacking their transporters.1518 However, other candidate cardiac nutrients and substrates of energy production remain to be discovered. In addition, the cardiac role of AQP7 as a glycerol facilitator has not been investigated and remains uncertain at present. Thus, the aim of the present study was to clarify the significance of glycerol as a cardiac energy substrate through AQP7 and to investigate the myocardial metabolic adaptations in mice lacking AQP7.

2. Methods

2.1 Animals and physiological experiments

AQP7-deficient mice were generated and maintained as described previously.8 We analysed mice backcrossed to C57BL/6N over five generations. Mice were fed regular chow and were kept in rooms set at 22°C with a 12/12 h dark/light cycle (light cycle: 8 AM–8 PM). Blood pressure was measured with an automatic sphygmomanometer (BP98A, Softron) from the tail artery while mice were passive under the same condition (from 9 AM–11 AM). Five readings were taken for each measurement, and the mean value was assigned to each individual animal. For isoproterenol administration, 10-week-old mice were implanted with osmotic minipumps (Alzet mini-osmotic pump model 2002, Durect Corp.) and infused with isoproterenol at 15 mg/kg/day for 1 or 2 weeks. Transverse aortic constriction (TAC) was performed as described previously19 at 10 weeks of age. Transthoracic echocardiography was performed in each mouse using LOGIQe ultrasound system with a 4.0–10.0 MHz linear probe (i12L-RS) (GE Healthcare). Mice were anaesthetized with 0.05 mg/body weight of pentobarbital sodium salt and fixed on heating pads to maintain body temperature at 35–37°C. Echocardiography was performed in these mice under similar heart rate of 360–390/min. After obtaining a long-axis two-dimensional image of the left ventricle (LV), a two-dimensional guided M-mode trace crossing the septal and posterior walls was recorded. The following parameters were measured on the M-mode tracings: interventricular septal thickness, LV posterior wall thickness, LV end-diastolic diameter (LVDd), LV end-systolic diameter (LVDs), LV fractional shortening [LVFS = (LVDd − LVDs)/LVDd × 100]. The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of Osaka University School of Medicine. This study also conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996).

2.2 Quantification of mRNA levels

Total RNA was isolated from mice tissues and cells by using RNA STAT-60 (Tel-Test Inc.) according to the protocol supplied by the manufacturer. The quality and quantity of total RNA were determined by using ND-1000 spectrophotometer (Nano Drop Technologies). For northern blot analysis, 10 µg of total RNA was electrophoresed on 1% agarose/formaldehyde gel and transferred onto nylon membranes (Hybond-N+, GE Healthcare). The membrane was hybridized with the indicated cDNA probe labelled with [α-32P]dCTP. The hybridized membrane was exposed to X-ray film. For real-time quantitative PCR analysis, first-strand cDNA was synthesized from 180 ng of total RNA by using Transcriptor First Strand cDNA Synthesis Kit (Roche Diagnostics, Mannheim, Germany). Real-time quantitative PCR amplification was conducted with the LightCycler 1.5 (Roche Diagnostics) using LightCycler-First Strand DNA Master SYBR Green I (Roche Diagnostics) according to the protocol recommended by the manufacturer. Primer sets were: mouse GAPDH, 5′-AGTCCATGCCATCACTGCCACCCA-3′ and 5′-TCCACCACCCTGTTGCTGTAGCCG-3′; mouse glycerol kinase (GK), 5′-GCTACAAGCAGACATTCTGTA-3′ and 5′-CTTCATCACAGCTTTCTTCCA-3′; mouse glycerol phosphate dehydrogenase 2 (GPD2), 5′-CTCGCCATCGCCCTCACTG-3′ and 5′-ACCGCTCACTCGCTCTTTGC-3′; rat AQP7, 5′-ATCCTTGTTTGCGTTCTTGG-3′ and 5′-GCGTGAATTAAGCCCAGGTA-3′; rat AQP3, 5′-AGCAGATCTGAGTGGGCAGT-3′ and 5′-CTTGGGCTTAAGAGGGGAAC-3′; rat AQP9, 5′-CTCAGTCCCAGGCTCTTCAC-3′ and 5′-ATGGCTCTGCCTTCATGTCT-3′.

2.3 Histological analysis

The heart was excised from each mouse after euthanasia and the ventricles were 10% formalin-fixed, paraffin-embedded, and subsequently cut at 6 μm sections and mounted on glass slides, using standard procedures. Sections of the ventricle were stained with haematoxylin and eosin (H&E). To determine the cell surface area, at least 200 cells were measured per mouse using Win ROOF 5.5 software (Mitani Co.). To investigate the extent of collagen fibre accumulation, sections of the ventricle were stained with van Gieson's modified method by using Maeda's Resorsin Fuchsin solution, Weigert's iron haematoxylin solution, and van Gieson's stain solution (Muto Pure Chemicals Co.). We randomly selected 15 fields in four individual sections, calculated the ratio of the area of van Gieson-stained interstitial fibrosis to total LV area by using Win ROOF 5.5 software, and used the average of the ratio of the fibrosis area for individuals.

2.4 Cardiac glycerol and ATP content

Mice were sacrificed under 24 h fasting, and hearts were perfused with phosphate-buffered saline (PBS) and immediately excised. For glycerol content assay, the weights of the ventricular tissues were measured exactly with microbalance, and the tissue was then homogenized in PBS buffer (50 mg tissue/500 µL PBS). Chloroform (500 µL) was added to the homogenized tissues, mixed by vortex, and the mixture subsequently centrifuged at 10 000g for 10 min. The supernatant was subjected to glycerol assay by using Free Glycerol Determination Kit (Sigma-Aldrich Inc., St Louis, MO, USA). Cardiac ATP content was measured using ENLITEN ATP assay system (Promega, Madison, WI, USA) according to the protocol supplied by the manufacturer.

2.5 Glycerol uptake and ATP assay in H9c2 cells

The embryonic rat heart-derived H9c2 cells were purchased from ATCC. H9c2 cells were grown and maintained in Dulbecco's modified Eagle's media (DMEM) supplemented with 10% foetal calf serum. To induce differentiation of H9c2 myoblasts into myotubes, the medium was replaced with the differentiation medium (DMEM containing 2% horse serum) when H9c2 cells reached 90% confluence. At day 10 after differentiation, control- or AQP7-small interfering RNA (siRNA) (Qiagen, Hilden, Germany) was introduced into H9c2 myotubes by using Lipofectamin 2000 (Invitrogen, Carlsbad, CA, USA). Twenty-four hours after transfection, cells were subjected to glycerol uptake and ATP assay. For glycerol uptake assay, H9c2 myotubes in 12-well plates were incubated for 3 h in Krebs–Ringer HEPES buffer (128 mM NaCl, 5.2 mM KCl, 1.4 mM MgSO4, 1.4 mM CaCl2, 5 mM KH2PO4, 10 mM HEPES, pH 7.4). The assay was initiated by the addition of 300 µM glycerol containing [2-3H]glycerol (3 µCi per well), 5.5 mM D(+)-glucose, 200 µM palmitic acid, and 100 µM fatty acid-free BSA. After 10 min, uptake was terminated by 200 µM phloretin, and cells were washed vigorously three times in ice-cold PBS. The cells were collected and homogenized in 0.5 N NaOH, and the aliquot was used to determine radioactivity by liquid scintillation counting. For ATP assay, H9c2 cells with control- or AQP7-siRNA were incubated in six-well plates for 24 h in Kreb's Ringer HEPES buffer. After transfection of siRNA, 5.5 mM glucose or 300 µM glycerol was added for 6 h prior to harvest. To exclude any hyperosmolar effect, we added identical concentrations of 5.5 mM or 300 µM mannitol to the medium, respectively. The cells were collected in 10% trichloroacetic acid, and intracellular ATP levels were measured using ENLITEN ATP assay system (Promega).

2.6 Langendorff-perfused heart and energy consumption

Langendorff-perfused heart study was conducted as described previously.20 Briefly, mice were anaesthetized with 0.05 mg/body weight of pentobarbital sodium salt and heparinized (100 U i.v.). To remove blood components, the heart was perfused through the post-cava by injection of about 5 mL Krebs–Henseleit bicarbonate (KHB) buffer (118.5 mM NaCl, 25 mM NaHCO3, 4.7 mM KCl, 1.2 mM MgSO4, 1.2 mM KH2PO4, 2.5 mM CaCl2, and 0.5 mM EDTA, pH 7.4). Hearts were excised and placed in KHB buffer. Extraneous tissues were removed, the aorta was cannulated with a polyethylene tube (PE50, Becton, Dickinson and Company, Franklin Lakes, NJ, USA), and the heart underwent a retrograde Langendorff perfusion with KHB solution. The KHB buffer was bubbled with 95% O2 and 5% CO2 gas mixture and passed at a flow rate of 2–2.5 mL/min using a peristaltic pump, and the buffer was maintained at exactly 37°C. The perfused heart showed spontaneous beating. To measure the consumption of each energy substrate, hearts were perfused with KHB solution containing either (i) 5.5 mM D(+)-glucose, 0.4 mM palmitic acid, and 0.3 mM glycerol, (ii) 5.5 mM D(+)-glucose and 0.3 mM glycerol, or (iii) 0.6 mM glycerol. The perfusate was collected at 5 or 65 min after the appearance of heartbeats. The collected samples were analysed for glucose, palmitic acid, and glycerol by Glucose CII-Test (Wako Pure Chemicals, Osaka, Japan), NEFA C-Test (Wako), and Free Glycerol Determination Kit (Sigma-Aldrich Inc.), respectively. Consumption of each substrate was calculated from its disappearance in the re-circulated KHB solution.

2.7 Statistical analysis

Results were expressed as mean ± SEM of n separate experiments. Differences between groups were examined for statistical significance using Student's t-test or ANOVA with Fisher's protected least significant difference test, and P < 0.05 was considered significant.

3. Results

3.1 Expression of glycerol facilitators in mouse heart

We measured cardiac mRNA expression levels of aquaglyceroporins AQP7, 3, and 9 in wild-type (WT) and AQP7-KO mice. In heart tissue, northern blotting showed AQP7 expression in WT mice and lack of such expression in KO mice (Figure 1A). AQP9 and AQP3 were not expressed in heart tissue of both WT and KO mice, although they were expressed in other organs (Figure 1B). Next, we analysed AQP7 mRNA expression levels in neonatal rat cardiac fibroblasts and myocytes. AQP7 expression was greater in cardiac myocytes than in cardiac fibroblasts (data not shown). These results suggest that AQP7 is the sole glycerol facilitator in the mouse heart and there is no compensatory expression of other glycerol facilitators, AQP9 and AQP3, in KO mice.

Figure 1

Cardiac expression of aquaglyceroporins, blood pressure, and cardiac morphology in AQP7-deficient mice. (A) Cardiac mRNA expression of AQP7 in wild-type (WT) and knockout (KO) mice by northern blotting. (B) Cardiac mRNA expression levels of AQP9 and AQP3 in the heart, liver, and kidney by northern blotting. (C) Blood pressure levels under basal conditions in WT and KO mice (WT, n = 8; KO, n = 8). Blood pressure was measured from the tail artery. sBP, systolic blood pressure; dBP, diastolic blood pressure. (D) Heart weights of WT and KO mice at 8–10 weeks of age (WT, n = 6; KO, n = 6). (E) Representative haematoxylin and eosin-stained heart tissue and quantification of cell surface area (WT, n = 6; KO, n = 6). The area of cardiomyocytes was measured in 200 or more cells per mouse in the respective groups. Scale bar: 50 µm. (F) Relative mRNA levels related to glycerol metabolism in the left ventricle (WT, n = 6; KO, n = 6). 36B4, ribosomal protein, large, P0; GK, glycerol kinase; GPD2, glycerol phosphate dehydrogenase 2. In CF, data are mean ± SEM. **P < 0.01. NS, not significant.

3.2 Knockout mice have normal cardiac histology and morphology under basal conditions

First, we examined blood pressure, cardiac morphology, and mRNA levels in the LV. Systolic and diastolic blood pressure levels increased gradually with age, but there were no significant differences in their levels between WT and KO mice at each week of age (Figure 1C). No differences in heart weight were evident between WT and KO mice at 8–10 weeks of age when these mice had similar body weight (Figure 1D). Lung weight of KO mice (5.50 ± 0.09 mg/g BW) was also similar to that of WT mice (5.42 ± 0.06 mg/g BW; P = 0.702). Examination of H&E-stained sections of heart tissues showed no apparent differences between WT and KO mice (Figure 1E, left panel). Quantification of cell surface areas of cardiomyocytes also showed no significant differences between these mice (Figure 1E, right panel). Cardiac mRNA levels relating to glycerol metabolism, such as GK and GPD2, were slightly but significantly lower in KO than in WT mice, suggesting possible disturbance of cardiac glycerol metabolism in KO mice (Figure 1F).

3.3 AQP7 deficiency reduces glycerol and ATP content in heart tissue and cardiomyotubes

Cardiac glycerol content was significantly lower in KO mice than in WT mice under basal conditions (Figure 2A). Furthermore, ATP levels in heart tissue were significantly lower in KO than in WT mice (Figure 2B).

Figure 2

Decreased glycerol and ATP levels in the AQP7-deficient heart and AQP7-knockdown cardiomyotubes. (A) Glycerol content in heart tissue [wild-type (WT), n = 10; knockout (KO), n = 9]. (B) ATP content in heart tissue (WT, n = 6; KO, n = 6). (C) Relative mRNA levels of AQP7 in H9c2 cardiomyotubes after differentiation. (D) Knockdown of AQP7 in H9c2 cardiomyotubes by introducing small interfering RNA (siRNA) (n = 6 per group). At day 10 after differentiation, control- and AQP7-siRNA was introduced in H9c2 cardiomyotubes and cells were harvested 24 h after transfection. (E) Glycerol uptake in H9c2 cardiomyotubes (n = 6 per group). (F) Intracellular ATP content in H9c2 cardiomyotubes (n = 6 per group). After transfection of siRNA, glucose or glycerol was added for 6 h prior to harvest. Data are mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001.

Next, we examined AQP7-dependent glycerol uptake by using rat cardiomyocytes derived from the cell line H9c2. AQP7 mRNA expression levels increased in parallel with differentiation of cardiomyotubes (Figure 2C), whereas the mRNA levels of another aquaglyceroporins, AQP3 and 9, were not detected in H9c2 cells (data not shown). Transfection of siRNA was performed at day 10 after cell differentiation; AQP7-siRNA markedly suppressed AQP7 mRNA level (Figure 2D). Knockdown of AQP7 was associated with a significant reduction in glycerol uptake into cells (Figure 2E), suggesting that glycerol is taken up into cardiomyotubes partly through AQP7.

Glucose- and glycerol-dependent ATP production was also examined in H9c2 cardiomyotubes. Intracellular ATP levels were significantly higher in cardiomyotubes incubated with glucose for 6 h (Figure 2F, lane 1 vs. 3; P < 0.001), but glucose-dependent ATP production was not influenced by knockdown of AQP7 (Figure 2F, lane 3 vs. 4). The addition of glycerol resulted in a significant rise in intracellular ATP content (Figure 2F, lane 5 vs. 7; P < 0.001). However, the glycerol-dependent increase in intracellular ATP content was significantly lower in AQP7-knockdown cardiomyotubes compared with control cells (Figure 2F, lane 7 vs. 8). These data suggest that glycerol is taken up into cardiomyotubes partly through AQP7 and is utilized for cardiac energy production.

3.4 Deterioration of cardiac glycerol consumption in knockout mice

To understand the utilization of glycerol in heart tissue, cardiac consumption of fatty acid, glucose, and glycerol was analysed by using Langendorff-perfused heart. First, we analysed the substrate consumption in the beating heart when perfusate contained palmitic acid, glucose, and glycerol (Figure 3AC). No marked differences were noted in palmitic acid and glucose consumption levels between WT and KO mice (Figure 3A and B), whereas glycerol consumption was significantly reduced in hearts of KO mice compared with WT mice (Figure 3C). Secondly, the perfusate was changed to a buffer containing glucose and glycerol, and glucose and glycerol consumptions levels in the beating heart were measured (Figure 3D and E). There were no differences in glucose consumption levels between WT and KO mice (Figure 3D), whereas cardiac glycerol consumption levels were significantly lower in KO mice than in WT mice (Figure 3E).

Figure 3

Reduced glycerol consumption in the AQP7-deficient isolated heart. Hearts were excised and underwent retrograde Langendorff perfusion. Consumption of palmitic acid (A), glucose (B), and glycerol (C). The perfusate contained 0.4 mM palmitic acid, 5.5 mM glucose, and 0.3 mM glycerol [wild-type (WT), n = 7; knockout (KO), n = 6]. Consumption of glucose (D) and glycerol (E). The perfusate contained 5.5 mM glucose and 0.3 mM glycerol (WT, n = 4; KO, n = 4). (F) Number of animals with beating hearts. The perfusate contained 0.6 mM glycerol alone (WT, n = 7; KO, n = 8). In AE, data are mean ± SEM. *P < 0.05. NS, not statistically significant.

Finally, the perfusate was changed to a buffer containing glycerol alone (Figure 3F). The isolated heart of WT mice continued to beat even under perfusate containing glycerol alone for 60 min (beating hearts/total hearts: 6/7; 85.7%). However, the majority of KO mice hearts could not continue beating for 60 min (beating hearts/total hearts: 2/8; 25.0%, P = 0.0342; log-rank test). Considered together, these results provide strong evidence that glycerol is an energy source in the beating heart and that AQP7 seems to contribute to glycerol uptake in cardiac cells.

3.5 Impaired myocardial adaptation to pressure overload in knockout mice

As demonstrated in Figures 13, lack of AQP7 did not affect cardiac morphology under basal conditions, although it resulted in low mRNA levels of GK and GPD2, low ATP content, and deterioration of glycerol consumption in the heart. Next, these mice were examined under cardiac overload to determine the role of cardiac AQP7 in pathological state. WT and KO mice were continuously infused with isoproterenol for 2 weeks. Figure 4A (left panel) shows representative M-mode echocardiographic images of the LV recording at pre-infusion, after 1 and 2 weeks of isoproterenol infusion. KO mice exhibited significant increases in LV wall thickness at 1 week after isoproterenol treatment, and increased LV diameter and reduced LVFS at 2 weeks after infusion, compared with WT mice (Table 1 and Figure 4A). Cardiac hypertrophy was noted in KO mice (Figure 4B), and heart weight was greater in KO mice than in WT mice after 2 weeks of isoproterenol infusion (Figure 4C). Further comparisons showed significantly heavier lung of KO mice than WT mice after 2 weeks of isoproterenol infusion (6.62 ± 0.14 vs. 5.65 ± 0.09 mg/g BW; P < 0.0001). Measurement of cell surface area on H&E-stained cardiac sections indicated significant hypertrophy of cardiomyocytes in KO mice (Figure 4D). Furthermore, measurement of areas of fibrosis in modified van Gieson-stained cardiac sections of mice treated with isoproterenol showed significantly larger fibrosis area in KO mice than in WT mice (Figure 4E).

Figure 4

Impaired myocardial adaptation to pressure overload in AQP7-deficient mice. Wild-type (WT) and knockout (KO) mice were challenged with isoproterenol for 2 weeks (AE) or operated on for transverse aortic constriction (TAC) (F). (A) Representative M-mode echocardiography (left) and fractional shortening (right). Pre., pre-infusion; Iso. 1W, isoproterenol infusion for 1 week; Iso. 2W, isoproterenol infusion for 2 weeks. (B) Appearance of heart tissue of WT and KO mice at the end of 2 weeks of isoproterenol infusion. (C) Heart weights at the end of 2 weeks of isoproterenol infusion. (D) Representative haematoxylin and eosin-stained cardiac sections (left) and cell surface area (right). Mice were infused with isoproterenol for 1 week. Scale bar: 50 µm. (E) Representative micrographs of heart sections stained by van Gieson's modified method (left) and quantification of fibrosis area (right). Mice were treated with isoproterenol for 2 weeks. Scale bar: 50 µm. (F) Survival rates of WT and KO mice after TAC operation. Mice were operated on for TAC at 10 weeks of age (WT, n = 13; KO, n = 19). In A, CE, data are mean ± SEM. *P < 0.05, **P < 0.01.

View this table:
Table 1

Results of echocardiography in isoproterenol-infused AQP7-knockout mice

ParameterPre-infusionIsoproterenol 1 weekIsoproterenol 2 weeks
IVSd, mm0.73 ± 0.030.78 ± 0.040.79 ± 0.010.83 ± 0.01**0.83 ± 0.010.81 ± 0.02
LVDd, mm3.06 ± 0.073.24 ± 0.082.91 ± 0.053.09 ± 0.082.96 ± 0.123.22 ± 0.05*
LVPWd, mm0.75 ± 0.040.78 ± 0.020.77 ± 0.020.84 ± 0.01**0.82 ± 0.030.80 ± 0.01
Fractional shortening, %44.2 ± 1.943.2 ± 1.949.3 ± 1.252.9 ± 2.245.3 ± 2.636.8 ± 2.0*
  • Values are mean ± SEM.

  • IVSd, interventricular septal thickness in diastole; LVDd, LV end-diastolic diameter; LVPWd, LV posterior wall thickness in diastole.

  • *P < 0.05; **P < 0.01 compared with WT mice under the same conditions.

Finally, we performed TAC in WT and KO mice to test the effect of direct and severe pressure overload on heart mortality. The mortality rate was higher in KO than in WT mice after TAC (P = 0.0393; log-rank test). The heart and lung weights of surviving mice were measured at day 10 after TAC operation. Heart weights were significantly heavier in KO mice (11.19 ± 0.22 mg/g BW) than in WT mice (8.64 ± 0.20 mg/g BW; P < 0.001). Furthermore, the weight of the lungs of KO mice (21.25 ± 1.75 mg/g BW) was also higher than that of WT mice (13.48 ± 1.77 mg/g BW; P < 0.05). These results suggest that AQP7-deficient hearts do not adapt well to pressure overload.

4. Discussion

The major findings of the present study were: (i) AQP7-deficient hearts exhibited significant reduction of glycerol consumption and ATP content; (ii) AQP7-deficient hearts were prone to cardiac hypertrophy and heart failure when KO mice were challenged with isoproterenol and following TAC; (iii) in H9c2 cardiomyotubes, knockdown of AQP7 resulted in significant reductions of glycerol uptake into cells and intracellular ATP level.

The functional role of cardiac AQPs is not fully understood, although several AQPs are expressed in heart tissue. AQP1 is expressed in human cardiac muscle and co-localizes with t-tubular and caveolar proteins,21 but the significance of AQP1 in the heart tissue remains uncertain. A recent study demonstrated that AQP4 mRNA levels correlate significantly with the size of myocardial infarction area in mice, suggesting that AQP4 might contribute to myocardial oedema after infarction.22 Our results are consistent with the previous studies that reported AQP7 expression in mouse heart and its role as the sole glycerol facilitator.11,12 Immunohistochemical analysis showed the existence of mouse AQP7 in capillary endothelia in the heart,23 although previous11 and present data suggested that AQP7 is also expressed in cardiomyocytes. Differentiated H9c2 cells are more skeletal muscle-like than cardiac muscle-like.24 In fact, a small amount of AQP7 is also expressed in skeletal muscles,23 suggesting that skeletal muscles, as well as cardiomyocytes, require glycerol through AQP7. However, the role of AQP7 in skeletal muscles has not been clarified. The present study suggests that AQP7 may act as a glycerol facilitator and associate with energy production in skeletal muscles.

It is generally accepted that normal hearts maintain cardiac work by utilizing fatty acids and glucose as energy substrates to produce ATP via the TCA cycle pathway.25,26 Oxidation of fatty acids supplies 60–90% of myocardial ATP in the healthy adult mammalian heart, whereas the balance (10–40%) comes from glucose and lactate.27,28 The present study demonstrates that mouse heart uptakes and utilizes glycerol as a substrate for energy production. Other groups also demonstrated increased cardiac glycerol uptake with the elevation of heart rate.29 The hearts of KO mice had low ATP and glycerol content, although no significant changes in cardiac morphology and function were observed at basal state. These results indicate that glycerol might be utilized as an energy substrate even under basal conditions. AQP7-knockdown H9c2 cardiomyotubes and AQP7-null hearts exhibited the ability of glycerol uptake even under AQP7-deficient conditions, suggesting possible existence of other yet undiscovered glycerol facilitators. In addition, before the discovery of AQPs, glycerol was considered to cross the cell membrane by simple diffusion; thus glycerol may be taken into cells, at least in part, by simple diffusion. However, the significance of glycerol uptake remains uncertain under basal conditions, because hearts mainly require fatty acids for energy production under control conditions.

KO mice, but not WT mice, developed cardiac hypertrophy and defective LV contraction following isoproterenol challenge. Furthermore, the mortality rate of KO mice subjected to TAC was higher than that of WT mice. In pressure- or/and volume overload-induced hypertrophy, mitochondrial oxidative capacity is reduced and the heart shifts to reliance on glucose metabolism.13 Glucose is converted to acetyl-CoA via pyruvate, meets the TCA cycle, and contributes to energy production in cardiomyocytes.13,26 In fact, heart-specific Glut4-KO mice develop cardiac hypertrophy,15,16 indicating that myocardial metabolic adaptation is impaired when glucose utilization deteriorates in the heart. Glycerol is taken into cardiomyocytes and is finally converted to pyruvate by GK and GPD2 enzymes.30 Collectively, glycerol, as well as glucose, seems to contribute to energy production through pyruvate and it is possible that cardiac overload may require glycerol as one of energy sources. In this regard, another group demonstrated increased cardiac glycerol uptake in parallel with increased concentrations of glycerol.29 The authors also showed that the increased glycerol uptake suppressed fatty acid oxidation in the heart. Interestingly, as shown in Figure 3D and E, fatty acid withdrawal from the heart perfusate increased glycerol consumption more than glucose consumption (Figure 3B and C), indicating that glycerol may compensate for the lack of fatty acids as an energy substrate. It remains elusive how far glycerol is physiologically utilized for energy substrate compared with fatty acids and glucose in the working heart, and this issue should be clarified in the future.

The present study has several limitations. Although the ex vivo working heart experiments demonstrated the significance of AQP7 as a glycerol facilitator, whole-body glycerol metabolism should be considered especially on the basis of the results of isoproterenol infusion and TAC studies because conventional KO mice were used in the current study. In addition, cardiac phenotype of AQP7 over-expression may be worthy of analysis. Cardiac-specific KO and/or transgenic model mice of AQP7 should be generated and analysed in the future. Echocardiography was performed in anaesthetized WT and KO mice while recordings were made at similar heart rates. Haemodynamic study using cardiac catheter will be needed in the future to examine the cardiac function more accurately in KO mice.

In human, AQP7 is expressed in heart tissue, whereas the existence of other glycerol facilitators, AQP3 and AQP9, is still equivocal.12 Therefore, it is necessary to clarify the physiological and pathological significance of cardiac AQPs, including AQP7, in human subjects in the future. The present study indicates that AQP7 acts as a glycerol facilitator in mouse heart tissue. We reported previously a patient with loss of functional mutation in the AQP7 gene (Gly-264 to Val).31 Another group reported the correlation between single-nucleotide polymorphisms (SNPs) in AQP7 gene and obesity.32 In view of present results, the impact of genetic mutation and/or SNPs in AQP7 gene on cardiac diseases needs to be assessed in human subjects in the future.

In summary, AQP7 acts as a glycerol facilitator in cardiomyocytes, and glycerol is a substrate for cardiac energy production. AQP7-deficient hearts exhibit deterioration of myocardial metabolic adaptation upon overloading. The present results suggest that molecular interventions based on modulation of cardiac AQP7 could be a valid optional therapeutic strategy for heart failure.


This work was supported in part by a Grant-in-Aid for Scientific Research (B) (19390249 to T. F.), Grants-in-Aid for Scientific Research on Priority Areas (13137206 to T. F.) and 15081208 (to S. K.), Takeda Science Foundation (to T. F.), Takeda Medical Research Foundation (to T. F.), The Cell Science Research Foundation (to N. M.), Yamanouchi Foundation for Research on Metabolic Disorders (to N. M.), Japan Diabetes Foundation (to N. M.), Japan Heart Foundation/Novartis Grant for Research Award on Molecular and Cellular Cardiology (to N. M.), The Salt Science Research Foundation (to N. M.), Kanae Foundation for The Promotion of Medical Science (to N. M.), Japan Heart Foundation Research Grant (to N. M.), and Japan Foundation of Cardiovascular Research (to N. M.).


We thank Yulin Liao and Masafumi Kitakaze for providing the TAC technique, and also Mina Sonoda, Fumie Katsube, and Yoko Motomura for the technical assistance. We acknowledge all members of the Adiposcience Laboratory, Osaka University, for the helpful discussions on the project.

Conflict of interest: none declared.


View Abstract