Somatic events modify hypertrophic cardiomyopathy pathology and link hypertrophy to arrhythmia

Methods

Mice. MHC^403/+ mice were produced by homologous recombination as described (19). The MHC^403/+ allele was bred onto the 129SvEv genetic background and maintained in that background for >10 generations. Mice were housed as prescribed by the Association for Assessment and Accreditation of Laboratory Animal Care. The Standing Committee on Animals at Harvard Medical School approved all experimental protocols used in this study.

Electrocardiography, Electrophysiology, and Echocardiography Studies. Measurements of electrocardiographic and electrophysiologic parameters and provocative electrophysiologic testing were performed as described (21, 24). Electrophysiological parameters were measured at baseline, and no drugs were administered during electrophysiologic studies. Based on provocative electrophysiologic testing, mice were scored as not inducible, inducible of a minor (lasting <1 s) arrhythmia, or inducible of a major (lasting >1 s) arrhythmia. Induced premature ventricular contractions or couplets were not considered as induced arrhythmia.

Echocardiography was performed with a Hewlett–Packard SONOS 4500 echocardiograph by using a 12-Mz probe as described previously (20, 25, 26) on anesthetized mice with a heart rates >450 beats per minute. One experienced observer (C.S.) analyzed all studies. Reproducibility of echocardiographic data has been described (22, 26).

Histopathology: Quantification of Disarray and Fibrosis. Hearts were quickly excised, washed in PBS, fixed in 4% formaldehyde, and embedded in paraffin as described (20, 25). Embedded hearts were serially sectioned every 50 μm from apex to base in a transverse plane. Adjacent sections were stained with hematoxylin and eosin to assess myocyte disarray and hypertrophy and with Masson trichrome to assess collagen deposition resulting from fibrosis. Fibrosis was independently scored by two experienced observers (C.M.W. and I.P.G.M.) in a subset of five MHC^403/+ and three WT littermates without knowledge of genotype. Inter-observer variability was 7% for fibrosis content. All data presented were obtained by one observer (C.M.W.). The pattern of fibrosis (diffuse, patchy, or bundled), the location of fibrosis within the myocardium (subepicardial, subendocardial, subepi- and subendocardial, or intramural), and the distribution within the left ventricle (LV) (free wall, interventricular septum, or ventricular free wall and interventricular septum) was also scored.

Total myocardial fibrosis was assessed in two ways. Fifteen to 20 serial sections of the LV and right ventricle were taken at 50-μm intervals (Fig. 1) and stained with Masson trichrome, and fibrosis content was assessed from digitized images (×25 magnification) by using a custom-made program generated on matlab 6.1. Three-dimensional heart images were generated by computerized assembly of 15 sections (oriented from the base to the apex) with fibrosis (blue) and myocardial tissue (red) shown. The maximum fibrosis observed on any section was calculated as the area occupied by blue-stained connective tissue divided by the areas occupied by connective tissue plus cardiac myocytes × 100. Total fibrosis within the heart was calculated from the sum of fibrosis on all sections divided by the sum of the areas occupied by connective tissue plus cardiac myocytes × 100. Intramural vessels, perivascular collagen, endocardium, and trabeculae were excluded.

Myofiber disarray was identified by using criteria described by Maron et al. (27). Myocardial areas were considered disorganized when adjacent longitudinally cut muscle cells were aligned perpendicularly or obliquely to each other, forming a “pinwheel” configuration (type I-A disorganization). Semiquantitative assessment of myocyte disarray (scored as present or absent) was made by one observer (C.M.W.), who examined three fields within the left ventricular free wall and three fields within the interventricular septum on each of 10–12 hematoxylin and eosin-stained sections (×400 magnification). The percentage of myocyte disarray was calculated by dividing the number of fields with disarray by the total number of fields examined (≈60 per heart). To assess for inter-observer variability, two independent observers (C.M.W. and I.P.G.M.) scored disarray and ranked five mutant hearts. The rank order of myocyte disarray in these five mutant hearts was the same by each observer, and the percentage of myocyte disarray differed by 13%.

Heart Isolation and High-Resolution Optical Mapping. Mice were treated with heparin (500 units/kg) to prevent intracardiac blood coagulation, anesthetized with CO₂, and euthanized by cervical dislocation; hearts were surgically removed by a thoracotomy. While fully immersed in oxygenated (95% O₂/5% CO₂) Tyrode's solution (114 mM NaCl, 25 mM NaHCO₃, 10 mM dextrose, 4.6 mM KCl, 1.5 mM CaCl₂, 1.2 mM Na₂PO₄, 0.7 mM MgCl₂), the aorta was cannulated and Langendorff-perfused at a constant pressure at 37°C. Hearts were subsequently stained with the voltage-sensitive dye (Di-4-ANEPPS), and high-resolution optical mapping studies were performed as described in ref. 28. Recordings were made in the bin mode, which allows for an array of 64 × 64 pixels to be acquired at 947 frames per s with 12-bit resolution in the absence of any pharmacological or mechanical motion-reduction techniques. Epicardial conduction velocity measurements were obtained by pacing the right ventricle and LV directly with a silver monopolar electrode at a basic cycle length of 100 ms with 4-ms stimuli at two times diastolic threshold. Activity during pacing was recorded for 2 s. Optical movies were signal-averaged, and conduction velocities were calculated as described in refs. 28 and 29.

Results and Discussion

Four young (8–12 weeks old) and 16 adult (>30 weeks old) male MHC^403/+ mice and 5 adult age-matched WT inbred 129SvEv mice that were housed in a single viral antibody-free facility were studied. With increasing age MHC^403/+ mice developed cardiac hypertrophy, and by age 20–30 weeks hypertrophy was uniformly present (Table 1; maximum LVWT). The range of maximum LVWT was 1.00–1.40 mm for inbred adult MHC^403/+ mice and 0.80–0.92 mm for inbred WT mice. The end-diastolic LV dimensions were significantly smaller, and fractional shortening was increased in adult MHC^403/+ mice compared with age-matched WT mice, findings that are consistent with increased cardiac contractility, as has been previously reported in outbred MHC^403/+ mice (refs. 19, 20, and 30 and Table 1). Cardiac histopathology (Fig. 1) of inbred MHC^403/+ mice was similar to that described previously for mice carrying the Arg403Gln missense mutation on an outbred background (19, 22).

Mice underwent standard electrophysiologic testing and programmed electrical pacing to identify inducible arrhythmias (21, 24, 25). Ventricular tachyarrhythmia, defined by more than five ventricular beats, was induced in 25% (1/4) of young and 69% (11/16) of adult inbred MHC^403/+ mice by rapid ventricular pacing at a cycle length of ≥50 ms, or by double and triple extra-stimuli with a coupling interval of ≥50 ms. Induced ventricular tachyarrhythmias lasted 2.1 ± 3.3 s (range = 0.2–22 s) and were polymorphic in 9/16 mice and monomorphic degrading to polymorphic in 2/16 mice. No ventricular fibrillation or atrial arrhythmias were induced. Consistent with previous studies (21, 22, 24, 25), no arrhythmias were induced in WT mice.

Left ventricular conduction velocities were assessed in 16 independent adult MHC^403/+ and 5 WT mice by high-resolution optical mapping. Conduction velocities in both transverse (minimal signal conduction velocity) and longitudinal (maximal signal conduction velocity) directions and the anisotropic ratio (maximal signal conduction velocity/minimal signal conduction velocity), a measurement of conduction pattern, were similar in MHC^403/+ compared with WT hearts, when paced at 100 ms cycle lengths to approximate physiologic heart rates of 600 beats per minute (Table 1). Among mutant mice, neither conduction velocities nor anisotropic ratios correlated with susceptibility to inducible arrhythmias from programmed electrical pacing (Table 1).

A comparison of the histopathology throughout the entire myocardium demonstrated greater fibrosis in inbred MHC^403/+ mice than in WT mice (Table 1 and Fig. 2). Young MHC^403/+ hearts displayed only 2-fold more maximum and total fibrosis than the WT hearts (data not shown), but with increasing age the fibrosis in mutant mice increased. The maximum fibrosis per section was 4-fold greater and total myocardial fibrosis content was 5-fold greater in adult MHC^403/+ than in WT hearts. Despite the identical genetic background of MHC^403/+ mutant mice, the total amount of fibrosis within each heart varied broadly, ranging from 0.04% to 2.52% of heart area (Table 1 and Fig. 2 B–D). Two MHC^403/+ adult hearts (e.g., Fig. 2B) had very little fibrosis, comparable to WT hearts, whereas 14 mutant mice (e.g., Fig. 2 C and D) had significantly more fibrosis than WT mice.

The location and pattern of the fibrotic regions varied considerably among MHC^403/+ hearts. Although the total amounts of fibrosis in two MHC^403/+ hearts was insufficient (<0.1%) to assess distribution, among other mutant mice fibrosis was found in subepicardial, subendocardial, and intramural portions of the left ventricular free wall as well as the interventricular septal wall. Fibrosis was diffuse in four MHC^403/+ hearts (Fig. 1D), focal in six mutant hearts (Fig. 1 E and F), and both diffuse and focal in four mutant hearts. Focal fibrosis was identified in variable locations, including an extended area from the subvalvular region to the apex (Fig. 2C) and a confined area in the subvalvular region (Fig. 2D).

Neither the total myocardial fibrosis nor the maximal fibrosis per section in MHC^403/+ hearts correlated with inducible arrhythmias produced by in vivo electrophysiologic studies (Table 1). Furthermore, the location and pattern of fibrosis in MHC^403/+ LVs were the same in mice with and without inducible arrhythmias; diffuse, focal, or composite fibrosis patterns were seen equally in these study groups (Fig. 2; P = not statistically significant; data not shown). For example, one MHC^403/+ mouse with long runs of inducible ventricular tachycardia had almost no fibrosis (Fig. 2B). Two MHC^403/+ mice with inducible ventricular tachycardia had different amounts and locations of fibrosis (Fig. 2 C and D). Regression analyses of the total myocardial fibrosis or maximum fibrosis per section did not indicate a relationship to the minimal signal conduction velocity (r² = 0.11), the maximal signal conduction velocity (r² = 0.15), or the anisotropic ratio (r² = 0.01).

Myocyte disarray occurred significantly more in microscopic fields from young (29%) and adult (40%) mutant MHC^403/+ than from WT (8%) hearts (Table 1 and Fig. 1 A and B). Disorganized myocyte architecture in mutant specimens usually clustered in the LV free wall or interventricular septum. No association was observed between the extent or the location of myocyte disarray in mutant mice and inducible arrhythmias (Table 1). Myocyte disarray also did not correlate with transverse and longitudinal conduction velocities or the anisotropic ratio (data not shown).

The MHC^403/+ mice with inducible arrhythmias had significantly greater LVWT and significantly greater hypercontractility than did MHC^403/+ mice without inducible arrhythmias (Table 1). Regression analysis showed no correlation between LVWT and the amount of fibrosis or myocyte disarray (data not shown), indicating that myocyte hypertrophy primarily accounted for increased ventricular wall dimensions.

Because genetically identical MHC^403/+ mice that were housed in a uniform environment exhibited variable amounts of hypertrophy, fibrosis, myocyte disarray, and susceptibility to induced arrhythmias, we conclude that the broad range of responses to a sarcomere protein gene reflects in part stochastic, somatic events. Variation in the amount and distribution of cardiac fibrosis was illustrative of this hypothesis. Focal cardiac fibrosis occurs when myocytes die and are replaced by fibroblasts, whereas diffuse fibrosis reflects increased collagen expression by interstitial fibroblasts that are dispersed between myocytes (31). Whereas the diffuse pattern of fibrosis observed in 8/16 adult MHC^403/+ hearts was consistent with an enhanced overall vulnerability of mutant myocytes to premature death, 10 mutant adult MHC^403/+ hearts displayed prominent foci of fibrosis (4 of these also had diffuse fibrosis), a finding that suggests an additional trigger for myocyte death. Given the underlying genetic cause of HCM, a “two-hit” process aptly models these data, in which a primary sarcomere mutation enhances vulnerability to secondary insults that increased focal myocyte death and fibrosis (Fig. 3). Surprisingly, these second hit events are neither genetically programmed nor clearly related to local hemodynamic parameters or mechanical strain, given the variable anatomic extent and location of focal fibrosis found in inbred mutant mice.

Local ischemia is one potential stimulant for focal myocyte death, and, although the hemodynamics of coronary blood flow would predict greater susceptibility of myocytes located within the subendocardial layer of the ventricle, this was not observed. Other stimuli that might account for stochastic development of focal areas of fibrosis in HCM hearts might include local fluctuations in hormones and/or ions that influence myocyte signal transduction; changes in these signals could trigger cell death and produce a nidus of biochemical factors that promote additional death of surrounding myocytes.

Myocyte disarray in inbred MHC^403/+ mice also varied in amount and distribution and did not correlate with either the total amount of fibrosis (r² = 0.12) or the maximum fibrosis present in sections (r² = 0.18; data not shown). Although the causes of local disorganized myocyte architecture are unknown, these data indicate that myocardial disarray and fibrosis, two characteristic pathologic findings of HCM, may arise by independent mechanisms.

Studies of human HCM have hypothesized that cardiac fibrosis and/or myocyte disarray increase susceptibility to ventricular arrhythmias and increase risk for sudden death (9, 10, 32, 33). However, we identified no correlation between the amount and/or pattern of fibrosis or the quantity of myocyte disarray in MHC^403/+ mice and the propensity for arrhythmia as assessed by two independent techniques, ex vivo high-resolution optical mapping and in vivo electrophysiologic study (Fig. 2). Almost one-third of MHC^403/+ mice were resistant to cardiac arrhythmias and had anisotropic ratios comparable to WT mice, even when displaying a significant amount of fibrosis (Table 1 and data not shown). Sites of fibrosis did not disrupt normal electrical currents, initiate novel currents, or trigger ventricular tachycardia. Our data do not exclude the possibility of very small amounts of fibrosis within a specific myocardial locus that might initiate narrow reentrant pathways; highly selective differences in conduction velocity would not be detected by the overall conduction velocities assessed here.

Although many issues limit direct extrapolation of these data to human disease, our findings suggest a more complex relationship between HCM histopathology and arrhythmogenicity than was previously appreciated. The variable risk for sudden death associated with cardiac troponin T gene mutations is illustrative of this complexity; both human HCM patients (11, 17) and mouse models (34, 35) with one of these mutations exhibit little myocardial fibrosis but increased arrhythmogenicity.

The extent of left ventricular hypertrophy and increased contractility in MHC^403/+ mice was informative for susceptibility to induced arrhythmias. Notably, an association between hypertrophy and arrhythmia was identified despite only mild to moderate increases in LVWT in MHC^403/+ mice. The mechanism by which myocyte enlargement increases arrhythmic risk may relate to changes in intrinsic automaticity, in that other experimental models of hypertrophied myocytes exhibit reexpression of pacemaker currents (36) and enhanced prolongation of the action potential by a down-regulation of the main outward current I_to (37, 38). Calcium signaling may also play a role in arrhythmia vulnerability, both because calcium handling is directly perturbed by sarcomere mutations (26) and because calcium cycling increases with enhanced contractility (39, 40). However, given that high-resolution optical mapping defined comparable conduction velocities and anisotropic ratios in the hearts of MHC^403/+ and WT mice, we presume that any electrophysiologic abnormalities in mutant hearts are focal and produce small reentrant circuits or areas of increased susceptibility to triggered automaticity in HCM mice that, if propagated throughout the myocardium, could cause arrhythmic sudden death.

We conclude that the three cardinal manifestations of HCM histopathology (cardiac hypertrophy, myocyte fibrosis, and disarray) reflect independent pathologic processes within myocytes carrying a sarcomere gene mutation (Fig. 3). The severity of two components of this triad, fibrosis and disarray, is substantially influenced by unknown somatic factors. Myocyte hypertrophy, although genetically programmed, is also influenced by other factors that also increased susceptibility to ventricular arrhythmias. Although direct extrapolation from genetically inbred mouse models of HCM to the human condition is difficult, these observations imply that therapeutic approaches that mitigate even moderate myocyte growth in HCM may also attenuate the risk for cardiac arrhythmia. Identification of somatic factors that influence cardiac responses to a sarcomere protein gene mutation may provide other strategies to limit histopathology and attenuate important adverse clinical events in HCM.