Mechanical loading reveals an intrinsic cardiomyocyte stiffness contribution to diastolic dysfunction in murine cardiometabolic disease

Abstract

Cardiometabolic syndromes including diabetes and obesity are associated with occurrence of heart failure with diastolic dysfunction. There are no specific treatments for diastolic dysfunction, and therapies to manage symptoms have limited efficacy. Understanding of the cardiomyocyte origins of diastolic dysfunction is an important priority to identify new therapeutics. The investigative goal was to experimentally define in vitro stiffness properties of isolated cardiomyocytes derived from rodent hearts exhibiting diastolic dysfunction in vivo in response to dietary induction of cardiometabolic disease. Male mice fed a high fat/sugar diet (HFSD vs. control) exhibited diastolic dysfunction (echo E/e' Doppler ratio). Intact paced cardiomyocytes were functionally investigated in three conditions: non-loaded, loaded and stretched. Mean stiffness of HFSD cardiomyocytes was 70% higher than control. E/e' for the HFSD hearts was elevated by 35%. A significant relationship was identified between in vitro cardiomyocyte stiffness and in vivo dysfunction severity. With conversion from the non-loaded to loaded condition, the decrement in maximal sarcomere lengthening rate was more accentuated in HFSD cardiomyocytes (vs. control). With stretch, the Ca²⁺ transient decay time course was prolonged. With increased pacing, cardiomyocyte stiffness was elevated, yet diastolic Ca²⁺ elevation was attenuated. Our findings show unequivocally that cardiomyocyte mechanical dysfunction cannot be detected by analysis of non-loaded shortening. Collectively, these findings demonstrate that a component of cardiac diastolic dysfunction in cardiometabolic disease is derived from cardiomyocyte stiffness. Differential responses to load, stretch and pacing suggest that a previously undescribed alteration in myofilament-Ca²⁺ interaction contributes to intrinsic cardiomyocyte stiffness in cardiometabolic disease. KEY POINTS: Understanding cardiomyocyte stiffness components is an important priority for identifying new therapeutics for diastolic dysfunction, a key feature of cardiometabolic disease. In this study cardiac function was measured in vivo (echocardiography) for mice fed a high-fat/sugar diet (HFSD, ≥25 weeks). Performance of intact isolated cardiomyocytes derived from the same hearts was measured during pacing under non-loaded, loaded and stretched conditions in vitro. Calibrated cardiomyocyte stretches demonstrated that stiffness (stress/strain) was elevated in HFSD cardiomyocytes in vitro and correlated with diastolic dysfunction (E/e') in vivo. HFSD cardiomyocyte Ca²⁺ transient decay was prolonged in response to stretch. Stiffness was accentuated with pacing increase while the elevation in diastolic Ca²⁺ was attenuated. Data show unequivocally that cardiomyocyte mechanical dysfunction cannot be detected by analysis of non-loaded shortening. These findings suggest that stretch-dependent augmentation of the myofilament-Ca²⁺ response during diastole partially underlies elevated cardiomyocyte stiffness and diastolic dysfunction of hearts of animals with cardiometabolic disease.

Keywords: Ca2+ myofilament interaction; cardiometabolic disease; cardiomyocyte; diastolic dysfunction; relaxation; stiffness.

Figure 1. Systemic indicators of cardiometabolic disease and in vivo diastolic function in high‐fat/sugar diet (HFSD) vs. control (CTRL) mice

A, body weight elevation at 5‐week intervals up to 30 weeks of dietary treatment (last time point full cohort intact; N = 16–17 mice per group). B, glucose tolerance testing at 23–24 weeks’ post dietary intervention; N = 6 mice per group). C, echocardiography exemplar blood flow and tissue Doppler images from which diastolic indices were derived for CTRL and HFSD mice. Panels with orange border: HFSD mitral blood flow and annular wall motion. Panels with blue border: CTRL mitral blood flow and annular wall motion. N = 14–16 mice per group. D, early diastolic mitral annular blood flow (E) was increased in HFSD mice. E, early diastolic mitral annular wall velocity (e′) was decreased in HFSD mice. F, deceleration time of early diastolic mitral annular blood flow was decreased in HFSD mice. G, ratio of early diastolic blood flow to wall velocity (E/e′) was increased in HFSD mice. Time‐course data analysed using two‐way repeated measures ANOVA with Šidák's post hoc test with P‐value for ‘time × diet’ effect presented. Comparisons between two treatment groups performed using Student's t test. Data presented as means ± standard deviation. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 2. Non‐loaded sarcomere and Ca²⁺ dynamics in CTRL and HFSD cardiomyocytes

A, mean sarcomere contraction cycles presented as a percentage of diastolic sarcomere length for CTRL and HFSD cardiomyocytes. B and C, mean diastolic sarcomere length and extent of sarcomere shortening for CTRL and HFSD cardiomyocytes. D, mean sarcomere shortening and lengthening rate records (time derivative of sarcomere contraction–relaxation cycle) for CTRL and HFSD cardiomyocytes. The nadir and zenith represent maximal shortening and lengthening rates respectively. E and F, maximal rate of lengthening and shortening rates for CTRL and HFSD cardiomyocytes. G, mean Ca²⁺ transient records for CTRL and HFSD cardiomyocytes. H–J, mean Ca²⁺ levels (systolic and diastolic) and time constant of transient decay (τ) in HFSD and CTRL cardiomyocytes. Mean data computed from average of 20 cycles per cardiomyocyte to derive group mean (N = 6–8 mice per group, n = 32–49 cells per group). Comparisons between two treatment groups performed using Student's t test (all P > 0.05, ns). Data presented as means ± standard deviation. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 3. Sarcomere shortening and lengthening in non‐loaded and loaded pacing states

A, exemplar non‐loaded and loaded sarcomere contraction cycles for the same cardiomyocyte (CTRL). B and C, extent of sarcomere shortening and maximum rate of lengthening, measured for each cardiomyocyte in non‐loaded and loaded state for CTRL and HFSD groups combined (N = 9 mice, n = 25 cells). Mean data for CTRL and HFSD are presented as grey points to the left (non‐loaded) and right (loaded) of individual cardiomyocyte data. D, correlation of loading associated sarcomere shortening change and maximum rate of lengthening change for CTRL and HFSD combined. E, mean loading effect on maximum deformation rate during lengthening ((µm/s)/µm; maximum rate of lengthening (µm/s) normalized by shortening extent (µm)) for CTRL vs. HFSD (CTRL: N = 5 mice, n = 14 cells; HFSD: N = 4 mice, n = 12 cells). Significant ANOVA interaction effect (diet treatment × load state). F, mean decrement in deformation rate during lengthening ((µm/s)/µm). Comparison of cardiomyocyte responses in different loading conditions analysed by paired t test. Comparison of cardiomyocytes from different treatment groups analysed by unpaired t test. A 2‐way ANOVA with repeated measures used for analysis of two independent variables. Data fitted with linear regression model and calculation of Pearson's correlation coefficient. Data presented as means ± standard deviation. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 4. Cardiomyocyte standardized stretch protocol

A, image of a left ventricular cardiomyocyte longitudinally attached to glass rods. Edge‐tracking used to measure inter‐rod distance (marked as blue bars) for cardiomyocyte segment analyses. The delineated region of interest (marked pink) represents the area from which fast Fourier transform analysis of sarcomere striations was performed to derive sarcomere‐specific data. B, exemplar traces depicting the change in glass rod internal distance in response to successive 0.5 V piezo motor inputs used for calibration of stretch protocol for each individual cardiomyocyte to allow normalized comparisons. C, measurements of stress and strain during progressive cardiomyocyte stretch. Exemplar force, sarcomere length and rod edge‐tracker traces during the first stretch of a stress–length protocol. D, comparison of variability of cell stretches performed using a uniform (pre‐set) stretch constant and the individually derived stretch calibration constant. Strain was measured as the percentage change in segment length relative to the initial segment length. Reproducibility is optimized using the individual stretch calibration constant. Data shown as medians ± interquartile range. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 5. Quantification of diastolic stiffness in CTRL vs. HFSD cardiomyocytes using strain normalization methods

A, exemplar stress recordings of CTRL and HFSD cardiomyocytes undergoing ‘stress–length’ protocol (2 Hz paced). Diastolic baseline stress shifts indicated by dashed line. B, diastolic (pair of lower lines) and systolic (pair of upper lines) sarcomere mean stress–sarcomere length (SL) relations in CTRL cardiomyocytes (N = 5 mice, n = 9 cells). C, diastolic mean stress–sarcomere length relation (linear regression fit) for HFSD and CTRL. D, diastolic stress (mN/mm²) normalized relative to sarcomere stretch (µm) yields sarcomere stress/strain = stiffness ((mN/mm²)/strain (µm)) for HFSD and CTRL. E, diastolic mean cardiomyocyte segment stress–length relation (linear regression fit) for HFSD and CTRL. F, diastolic stress (mN/mm²) normalized relative to segment stretch (% length) yields cardiomyocyte segment stress/strain = stiffness ((mN/mm²)/strain (% length)) for HFSD and CTRL. G, correlation (linear regression fit) of in vitro cardiomyocyte segment diastolic stiffness (stress/strain) with in vivo diastolic dysfunction E/e′ (blood flow/wall motion) in CTRL and HFSD. Cardiomyocyte data averaged per heart. Comparisons between two groups performed using Student's t test. Data fitted with linear regression model and calculation of Pearson's correlation coefficient, presented as means ± standard deviation (D, F) and means ± standard error of the mean (B, C, E). For C and D, CTRL: N = 4 mice, n = 8 cells; HFSD: N = 2 mice, n = 7 cells. For E and F, CTRL: N = 6 mice, n = 16 cells; HFSD: N = 6 mice, n = 18 cells. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 6. Load dependence of cardiomyocyte Ca²⁺ transient decay in CTRL and HFSD cardiomyocytes

A and B, time constant of Ca²⁺ transient decay (τ) and time to peak Ca²⁺ transient in cardiomyocytes with different load conditions: loaded (L) and stretched (S) shown as fold change relative to non‐loaded (N) for CTRL and HFSD. C and D, exemplar Ca²⁺ transient decay records, with a mono‐exponential decay function fit, in non‐loaded and stretched CTRL and HFSD cardiomyocytes. Horizontal black broken lines intersect with mono‐exponential fit at 36.8% (1/e) transient amplitude. Vertical broken lines from intersection points indicate the time constant of Ca²⁺ transient decay (τ). For analysis of two independent variables a 2‐way ANOVA with repeated measures used (CTRL: N = 5 mice, n = 13 cells; HFSD: N = 4 mice, n = 12 cells). Data are presented as means ± standard deviation. [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 7. Pacing induced changes in cardiomyocyte cytosolic Ca²⁺ levels and stiffness in CTRL and HFSD groups

A, exemplar CTRL and HFSD cardiomyocyte Ca²⁺ transients at 2 and 4 Hz pacing. B, diastolic cardiomyocyte Ca²⁺ levels measured before and after transition from 2 to 4 Hz pacing in HFSD and CTRL. C, diastolic cardiomyocyte Ca²⁺ level mean shift with transition from 2 to 4 Hz pacing in HFSD and CTRL. D, diastolic sarcomere length measured before and after transition from 2 to 4 Hz pacing in HFSD and CTRL. E, diastolic mean cardiomyocyte segment stress–length relation (linear regression fit) for HFSD and CTRL at 2 and 4 Hz pacing. F, diastolic mean cardiomyocyte segment stress–length relation slope (mN/mm²) normalized relative to segment stretch (% length) yields cardiomyocyte mean segment stress/strain = stiffness ((mN/mm²)/strain (% length)) for HFSD and CTRL at 2 and 4 Hz pacing. Comparison between two groups performed using unpaired t test. With two independent variables a repeated measures 2‐way ANOVA Šidák's post hoc analysis was performed. Data are presented as means ± standard deviation (C), means ± standard error of the mean (E) or mean with paired individual points (B, D, F). For B and C, CTRL: N = 4 mice, n = 9 cells; HFSD: N = 4 mice, n = 8 cells. For D and E, CTRL: N = 4 mice, n = 7 cells; HFSD: N = 4 mice, n = 9 cells. For F, CTRL: N = 4 mice, n = 8 cells; HFSD: N = 4 mice, n = 9 cells. [Colour figure can be viewed at wileyonlinelibrary.com]