Creatine kinase double knockout mice (i.e. deficient in both M-CK and mitochondrial-CK isoenzymes) were imported from a colony at the University of Würzburg, which in turn originated from the laboratory of Prof Wieringa (University of Nijmegen, The Netherlands) [²⁵, ²⁷]. These represent the standard strain reported in the literature, which are on a non-fixed mixed C57BL/6 and 129Sv genetic background and are therefore denoted M/Mt-CK^{−/−(129/Bl6)} in this manuscript. These mice were used for breeding followed by initial echocardiography and haemodynamic studies reported in Fig. 1, with stock C57BL/6 mice used as control.

Female heterozygotes were backcrossed in our laboratory for ten generations with male C57BL/6J mice obtained from Harlan UK to produce M-CK^−/− and M/Mt-CK^−/− mice on a pure genetic background. These mice were produced by heterozygous mating, so that littermates could be used as the appropriate wild-type controls (WT). All investigations were performed when mice were approximately 1-year-old. Mice were kept under specific pathogen-free conditions in individually ventilated cages with 12-h light–dark cycle, controlled temperature (20–22 °C), and fed standard chow and water ad libitum. This investigation conforms with UK Home Office Guidance on the Operation of the Animals (Scientific Procedures) Act, 1986.

Mice were anaesthetised with 4 % isoflurane in medical oxygen and maintained on a nose-cone at 1–1.5 %. They were placed supine on a homoeothermic mat, chest fur shaved and parasternal short and long-axis images obtained. M/Mt-CK^{−/−(129/Bl6)} mice were imaged with an Agilent Sonos 5500 equipped with a 7/15 MHz linear-array transducer, while all subsequent imaging was obtained under identical conditions using a Visualsonics Vevo 2100 with 22–55 MHz transducer.

Approximately 1 week after echocardiography, mice were anaesthetised with 4 % isoflurane in medical oxygen and maintained on a nose-cone at 1–1.5 % on a homoeothermic mat. The left ventricle was cannulated via the right carotid artery using a 1.4-F Millar Mikro-Tip cannula (SPR-839, Millar Instruments, Houston, Texas). The right jugular vein was cannulated with stretched polyethylene tubing for infusion of dobutamine (16 ng/g body weight/min) to test contractile reserve. Measurements were obtained after 15 min of equilibration via a Powerlab 4SP data acquisition system (ADInstruments, UK). At the end of the experiment, the heart and other organs were excised, washed in heparinised saline, blotted and weighed. Left ventricular samples were snap frozen in liquid nitrogen and stored at −80 °C.

Body composition analysis was carried out by non-invasive magnetic resonance relaxometry in conscious restrained mice using an EchoMRI-100 system (Active Field Resources, Houston, Texas). Accumulation factor was for extra-high precision (3×) resulting in a scan time of approximately 2.5 min.

Total creatine and total CK, CK isoenzyme and citrate synthase activities were measured from LV homogenates as previously described [²¹]. The activity of total adenylate kinase (AK) was quantified spectrophotometrically as described in [¹]. Total RNA was extracted from LV tissue and gene expression of LVH markers quantified using real-time RT-PCR as previously described [¹⁵].

All data were analysed blind to genotype by a single experienced operator. Data are expressed as mean ± standard deviation throughout. Comparison between two groups was by student’s t test, and between three groups by one-way analysis of variance (ANOVA) using Bonferroni’s correction for multiple comparisons. Differences were considered significant when P < 0.05.

All previously published studies have reported on M/Mt-CK^{−/−(129/Bl6)} mice with a mixed C57BL/6 and 129Sv genetic background. When these are maintained by KO × KO breeding, there is no appropriate littermate control, and standard C57BL/6 mice have historically been used for this purpose. However, there are reported differences between C57BL/6 and 129Sv strains for blood pressure and LV mass [⁸], so we sought to recapitulate the finding of LV hypertrophy in M/Mt-CK^{−/−(129/Bl6)} mice under our own standard laboratory conditions and to determine whether choice of control had any influence on the results obtained. Furthermore, we performed in vivo LV haemodynamic measurements, which have never previously been reported for M/Mt-CK^{−/−(129/Bl6)} mice.

At 1+ years of age (mean 58 weeks), male M/Mt-CK^{−/−(129/Bl6)} mice had significantly higher LV weight despite lower body weight and regardless of whether C57BL/6 (+27 %; P < 0.001) or 129Sv (+21 %; P < 0.01) were used as controls (Fig. 1a). This difference was still evident after desiccation, suggesting it was not due to oedema (LV dry weight: 31 ± 2 in C57BL/6 versus 38 ± 6 mg in M/Mt-CK^{−/−(129/Bl6)}; P = 0.008). LV hypertrophy was less pronounced in females, but remained significant (+17 % versus C57BL/6; P < 0.05). Echocardiography showed that male M/Mt-CK^{−/−(129/Bl6)} mice had a +16 % larger cavity cross-sectional area during diastole indicating significant LV dilatation (Fig. 1b), exhibited contractile dysfunction (dP/dt_max −31 %), end-diastolic pressures elevated 4.6-fold and significant pulmonary congestion compared with control mice (Fig. 1c–e), all indicative of congestive heart failure (see supplemental data table 1 for detailed analysis). In female mice, there were clear trends towards LV dysfunction for all these parameters, but differences were not statistically significant (Table 1).

All subsequent experiments were obtained in M-CK^−/− and M/Mt-CK^−/− mice that were backcrossed in our laboratory with C57BL/6 stock mice for >10 generations. Comparisons were made with a single wild-type littermate control group. Measurement of protein activity in LV homogenates confirmed loss of target genes, resulting in residual CK activity of 33 % in M-CK^−/− and 0.02 % in M/Mt-CK^−/− mice. There was an increase in BB-CK isoenzyme activity of 33 and 100 % for M-CK^−/− and M/Mt-CK^−/− hearts, respectively (Table 2), but as this represents only 2 % of normal total CK activity; however, the significance is unclear. There were no compensatory changes in either adenylate kinase or citrate synthase activities. Total LV creatine levels were indistinguishable from wild type (P = 0.46).

There were no significant differences between male and female mice for haemodynamic parameters when compared within genotypes, and therefore both sexes were analysed together. However, a detailed breakdown of these parameters for males and females is provided in supplemental data tables 2 and 3, respectively. Total animal numbers and sex ratios were: WT n = 26 (12 M/14 F), M-CK^−/−n = 31 (14 M/17 F) and M/Mt-CK^−/−n = 30 (12 M/18 F). Mean age was 55 weeks in all groups (range 52–59 weeks).

Haemodynamic parameters were indistinguishable between WT and M-CK^−/− mice (Fig. 2). Compared to wild-type littermates, M/Mt-CK^−/− mice had normal LV end-systolic and end-diastolic pressures (Fig. 2a, b), but had significantly impaired indices of relaxation (dP/dt_min 31 % lower and tau 31 % longer than WT; Fig. 2c, d) and contraction (dP/dt_max 23 % lower than WT; Fig. 2g). Resting heart rate was also significantly lower than WT (10 %; Fig. 2e). Under conditions of maximal β-adrenergic stimulation using dobutamine infusion, both maximum heart rate and dP/dt_max increased in M/Mt-CK^−/− mice, but remained significantly impaired compared to WT (11 and 24 % lower, respectively; Fig. 2f, h). Thus, the relative differences in absolute values were maintained, but all groups showed similar contractile reserve as a percentage of resting baseline values, i.e. increasing dP/dt_max from baseline by 27 % (Fig. 2i).

The lower heart rate in M/Mt-CK^−/− mice may account for some (or all) of the difference observed in contractility, i.e. by occupying different positions on the force–frequency response curve. To investigate this further, dP/dt_max was correlated with heart rate (Fig. 2j) for both genotypes: WT Pearson r = 0.61, P = 0.001; M/Mt-CK^−/− Pearson r = 0.84, P < 0.0001. A runs test indicated that neither relationship deviated significantly from linearity; therefore, linear regression analysis was used to compare WT with M/Mt-CK^−/−. Equation of line for WT was y = 21.3x – 1,562 and for M/Mt-CK^−/−y = 18.3x − 1,273. There was no difference in the slopes (P = 0.58), but the difference in elevation was highly significant (P = 0.0004), indicating that WT and M/Mt-CK^−/− hearts do not share the same relationship for dP/dt_max versus heart rate, but instead run in parallel (i.e. for any given heart rate, dP/dt_max is lower in M/Mt-CK^−/− hearts).

Echocardiographic examination of M/Mt-CK^−/− mice on the pure genetic background did not reveal any difference in cardiac volumes or in ejection fraction and cardiac output (Table 3). Pulmonary artery acceleration and ejection times were not different suggesting normal RV function and absence of pulmonary congestion. Transmitral Doppler was also indistinguishable from wild-type littermates suggesting that LV compliance was unaltered (Table 3).

Postmortem LV weight was not elevated in either M-CK^−/− or M/Mt-CK^−/− mice, but body weight was 37 % lower in M/Mt-CK^−/− mice, resulting in a high LV/body weight ratio (Fig. 3a–c). However, LV weight was in proportion to tibial length and to other major organs such as liver, kidneys, lungs and total heart weight (Fig. 3d–h). The same pattern was observed in both males and females when analysed separately (see supplemental data tables 2 and 3). An absence of LVH was confirmed on a molecular level by quantitative real-time RT-PCR to measure mRNA expression of common hypertrophy markers, which were not up-regulated (Fig. 4). Surprisingly, α-skeletal actin was elevated in M-CK^−/− compared to WT littermate controls; however, such a change in isolation does not suggest hypertrophy.

Differences in body weight were investigated further by magnetic resonance relaxometry in WT and M/Mt-CK^−/− mice. As suggested by organ weight analysis, differences in body weights were driven by altered body composition and in particular by drastically reduced fat in M/Mt-CK^−/− mice. This was accompanied to a lesser extent by lower total body water and lean weight (Fig. 5a–e). These highly significant differences were observed in both male and female mice (supplemental data table 4). There was a linear relationship between percentage body fat and body water for both WT and M/Mt-CK^−/− mice that was indistinguishable, suggesting that they are at different ends of the same continuum (Fig. 5e).

M/Mt-CK^{−/−(129/Bl6)} mice (i.e. on a standard mixed background) have previously been studied up to 41 weeks of age, at which point they had extensive LVH, mild dilatation and normal ejection fraction by MRI [¹⁶, ¹⁷]. We have now demonstrated that at 1 year of age this compensated hypertrophy eventually progresses to heart failure, but only in males, with females much less severely affected. Since previous studies have used stock C57BL/6 as a control for Mt-CK^{−/−(129/Bl6)} mice, we also demonstrated that the choice of control, whether C57BL/6 or 129Sv, does not affect the extent of LVH observed. Previous studies measured cardiac function using MRI, and in our experience it is not unusual to find minimal differences in global imaging parameters, yet still observe large differences in LV pressures. For example, we have previously shown that creatine-deficient mice have normal ejection fraction, but impaired pressure-generating capacity [²⁶]. This should not be surprising, as these parameters report very different phases in the cardiac cycle.

Creatine kinase-deficient mice survive for a long time with only mild dysfunction, so obviously CK activity is not obligatory for cardiac function, but the finding that male M/Mt-CK^{−/−(129/Bl6)} mice eventually develop heart failure suggests that loss of CK places a chronic strain on the heart and that gender modulates the effects of CK depletion. Multiple changes have been described in the ageing heart that impact on its ability to tolerate stress [⁵], and this may result in the gradual unmasking of the functional phenotype over time. A wide range of potentially compensatory adaptations have been described in M/Mt-CK^{−/−(129/Bl6)} mice to explain the mild early phenotype, for example, closer juxtaposition of mitochondria with myofilaments, thereby reducing diffusion distance [¹³], coupling of glycolysis with the SERCA pump [⁴] and increased mitochondrial capacity with flux through alternative phosphotransfer pathways [⁹]. It may be that these changes are inefficient or energy costly and therefore cannot be maintained as a compensatory mechanism in the long term. This also suggests that reduced CK activity, as occurs in the failing heart [¹⁴] and in viral myocarditis [¹⁰], may indeed contribute to disease progression. Although CK dysfunction is not as severe as in the double knockout, the effect is likely to be amplified when overlaid on existing organic disease and in combination with gross changes in cellular processes, inflammation and neuro-endocrine signalling.

We have previously shown that genetic background has a major effect on the phenotype of Mt-CK^−/− mice [¹⁵]. We therefore backcrossed M-CK^−/− and M/Mt-CK^−/− lines until they were congenic with C57BL/6 mice (theoretically 99.9 % identical). This did not alter the phenotype in M-CK^−/− mice, which remained mild to non-existent even when studied at 1-year of age. However, M/Mt-CK^−/− mice had a less severe phenotype compared to M/Mt-CK^{−/−(129/Bl6)}; most notably, they had normal end-diastolic pressure and an absence of LVH and pulmonary congestion. It is likely that this disparity purely reflects genetic differences rather than environmental or experimental influences such as diet or pathogen status, since M/Mt-CK^{−/−(129/Bl6)} obtained from Würzburg but kept in our laboratory developed LV hypertrophy as expected. Clearly, the backcrossed mice are no longer in heart failure, which suggests the existence of powerful genetic modifiers in the C57BL/6 background to account for this. Future identification of these modifiers may suggest therapeutic targets for heart failure.

A lack of LVH is in agreement with our previous findings in Mt-CK^−/− mice, which had up-regulation of M-CK and citrate synthase as potential compensatory pathways [¹⁵]. We predicted that backcrossed M/Mt-CK^−/− mice would therefore have a more pronounced phenotype, since they have no M-CK with which to compensate with. This does appear to be the case since backcrossed M/Mt-CK^−/− had impaired indices of isovolumetric function that were not observed in Mt-CK^−/− mice at the same age. A potential compensatory role for increased mitochondrial capacity (as suggested by elevated citrate synthase) was observed in our backcrossed Mt-CK^−/− mice [¹⁵] and also recently in the M/Mt-CK^{−/−(129/Bl6)} strain [⁹]. However, we did not observe an increase in citrate synthase activity in our backcrossed M/Mt-CK^−/− strain even though they had a milder phenotype, which argues against a compensatory role.

It is increasingly recognised that genetic background can have a major influence on observed phenotypes in terms of both normal cardiac function [⁸, ³¹] and response to pathological stress [³], and we have previously shown this to be true for Mt-CK^−/− mice [¹⁵]. This underlines the importance of using appropriate controls in mouse studies. The optimal control for a double knockout mouse is to use heterozygous mating to produce double WT littermates for comparison, and this is the approach we took for the mice on a pure C57BL/6 background. This would also be optimal for mice on a non-fixed, mixed background, even though random mixing will mean that wild-type mice are not genetically identical. However, we chose to use stock C57BL/6 as controls for the experiments on M/Mt-CK^{−/−(129/Bl6)} mice, because we wanted to emulate the conditions used in previous studies to determine whether we could recapitulate their findings. The limitations of using a poor control in those studies are therefore shared by our experiments on M/Mt-CK^{−/−(129/Bl6)} mice.

An unexpected finding, not previously described in the literature, was lower body weight observed in M/Mt-CK^−/− mice regardless of genetic background (although not in female M/Mt-CK^{−/−(129/Bl6)}). Using non-invasive body composition analysis, we established that this was mainly driven by reduced body fat in knockout mice. Water content was also reduced in M/Mt-CK^−/− mice; however, water is linearly related to fat over a wide range of body compositions in the mouse [⁶], therefore the preservation of the fat-to-water relationship suggests that this change is secondary to alterations in fat content. Although less pronounced, this pattern is very similar to creatine-free (GAMT knockout) mice [²³] and indicates that disruption of the CK system has profound effects on whole body metabolism.

Click here for additional data file (395_2012_276_MOESM1_ESM.docx)

The authors declare that they have no conflict of interest.

This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.

Values are mean ± standard deviation, with P values for unpaired Student’s t test

Mean ± standard deviation

Number of samples is indicated in the headings except where included in parenthesis

* Denotes P < 0.05, ^‡ P < 0.001 and ^† P < 0.0001 compared to wild type. ^# Denotes P < 0.001 for M-CK^−/− versus M/Mt-CK^−/−

Values are mean ± standard deviation, with P values for unpaired Student’s t test