Introduction
The family of four and a half LIM (FHL) proteins is composed of FHL1–4, ACT (activator of CREM) and ARA55, and is characterized by an N-terminal half LIM domain followed by four complete LIM domains [
12,
30]. FHL proteins are components of adhesion complexes, can act as transmitters of Rho signaling pathways and are involved in tissue-specific gene regulation [
5,
39,
47,
60]. The second member, FHL2 plays a role in cell cycle regulation, differentiation and apoptosis, assembly of extracellular matrix, bone formation, and wound healing (for review, see [
30]).
FHL2 is highly expressed in the heart throughout embryonic development and in adults [
13,
32]. FHL2 has been first shown to be located in the Z-disk and to a lesser extent in the M-band of the sarcomere in neonatal rat cardiac myocytes (NRCMs) [
50] and further analysis indicated I-band localization in papillary muscle [
33]. It interacts with several components, particularly with titin in the cardiac-specific N2B domain (I-band) and the IS2 region (M-band), where it couples cardiac metabolic enzymes to sites of high energy consumption [
33]. A body of evidence indicates that
FHL2 inhibits cardiac hypertrophic pathways, such as calcineurin–NFAT (nuclear factor of activated T cells)-dependent gene expression via binding of calcineurin [
28], or the MEK1–ERK1/2 signaling cascade by binding to ERK2 [
44]. Additionally,
FHL2 inhibited serum response factor (SRF)-dependent transcription in a Rho-dependent manner in embryonic stem cells and heart [
42].
FHL2 deficient mice exhibited normal response to short-term TAC [
11], but developed exaggerated cardiac hypertrophy under chronic isoprenaline stimulation [
32], suggesting that the implication of
FHL2 in heart failure depends on the trigger inducing heart failure. Whereas it was reported that
FHL2 is up-regulated upon adrenergic stimulation
in vivo in rodents [
28],
FHL2 protein abundance was markedly reduced in angiotensin II (AngII)-induced cardiac hypertrophy in mice [
40] and in human heart failure [
4].
Integrating the predominantly heart-specific expression of
FHL2, its suggested antihypertrophic role and its lower expression in human heart failure, we hypothesized that
FHL2 altered expression or genetic variants could be associated with HCM. HCM is the most prevalent myocardial disease (1:500; [
17]). Its hallmarks are hypertrophy, predominantly in the interventricular septum, diastolic dysfunction, myocardial fibrosis and disarray. The phenotype is very variable, and diastolic dysfunction can be the first feature of the disease. Symptoms include dyspnea, chest pain, palpitations, lightheadedness, fatigue, and syncope. HCM is a leading cause of sudden cardiac death (SCD) in young athletes and is connected with a significant risk of heart failure [
24,
34]. HCM is a genetic disease mainly transmitted as an autosomal dominant trait. It is caused by mutations in at least 14 genes coding for sarcomeric components (for reviews, see [
20,
45,
49,
51]). More recently, mutations in
FHL1 have been shown to be associated with HCM [
21]. We evaluated
FHL2 expression in patients and mouse models of HCM and screened the
FHL2 gene for genetic variants in a cohort of HCM patients devoid of mutations in established disease genes. We identified six
FHL2 genetic variants and analyzed the molecular and/or functional impact of the nonsynonymous substitutions after gene transfer in rat cardiac myocytes and engineered heart tissues (EHTs).
Methods
A detailed description of materials and methods can be found in the Supplemental Material.
Patients and human samples
We enrolled 121 HCM index cases without mutations in
MYH7,
MYBPC3, TNNT2,
TNNI3, or
MYL2 (data not shown). They were selected out of 299 HCM index cases recruited from the Eurogene Heart Failure cohort supported by the Leducq Foundation [
19,
21]. Diagnosis was grounded on medical history, physical examination, electrocardiogram, and echocardiogram (left ventricle (LV) wall thickness ≥15 mm in probands and >13 mm in relatives) [
8‐
10]. Controls consisted of 262 individuals.
Human myocardial samples were obtained from HCM patients who underwent septal myectomy or heart transplantation, from a patient undergoing aortic valve implantation due to aortic stenosis, and from individuals who had no cardiac disease but died of another cause. All materials from patients and controls were taken with informed consent of the donors and with approval of the local ethical boards.
Mouse models
The study complies with the Guide for the Care and Use of Laboratory Animals published by the NIH (Publication No. 85–23, revised 1985).
Mybpc3-targeted knock-out (KO) and knock-in (KI) mice were developed previously and maintained on the Black swiss genetic background [
6,
59].
Pre-embedding immunoelectron microscopy
Human septal myectomy and mouse ventricular tissues used for the study were fixed in 4 % paraformaldehyde, 15 % saturated picric acid in 0.1 M phosphate buffer (PB) pH 7.4 overnight at 4 °C. Sections were cut on a vibratome (Leica VT 1000S) at a thickness of 50 µm, blocked in 20 % NGS in PBS and were incubated with FHL2 (Abcam 12327) primary antibody in phosphate-buffered saline (PBS) containing 5 % normal goat serum (Vector Laboratories, Burlingame, CA, USA) overnight at 4 °C. After washing in PBS, sections were incubated with 1.4-nm gold-coupled secondary antibodies (diluted 1:100 in PBS; Nanoprobes, Stony Brook, NY, USA) overnight at 4 °C. After several washings sections were postfixed in 1 % glutaraldehyde in PBS for 10 min and then incubated with HQ Silver kit (Nanoprobes). After treatment with OsO4, sections were stained with uranyl acetate, dehydrated and embedded in Durcupan resin (Fluka, Switzerland). Ultrathin sections were prepared (Ultracut S; Leica, Germany) and examined with a ZEISS 910 electron microscope.
Screening of FHL2 for genetic variants
FHL2 gene has eight exons, of which five (exons 4–8) are coding. There are four FHL2 variants mainly differing in the 5′UTR, but encoding the same isoform. Therefore, only the five coding exons for FHL2, including neighboring intron boundaries were screened by PCR amplification performed on 30 ng of genomic DNA from peripheral lymphocytes using specific primer pairs (Supplemental Table 2). Sequences were examined using Codon Code Aligner Software®. Reference FHL2 sequence was taken from NCBI (NG_008844.2) with +1 designing the A of the ATG codon.
Rat cardiac myocytes culture, transduction, and hypertrophy stimulation
Isolation of neonatal rat cardiac myocytes (NRCMs) and stimulation with hypertrophy stimuli, followed by subsequent automated cell size determination was performed as defined previously [
29]. In detail, cells were transduced with adeno-associated virus serotype 6 (AAV6)-
FHL2 at a MOI of 100,000, serum was reduced to 0.2 % during 24 h and then NRCMs were stimulated for 48 h with phenylephrine (PE; 50 µM), endothelin-1 (ET1; 100 nM) or DMSO (
n = 3 different experiments, each in triplicates). Afterwards, cells were stained for alpha-actinin and automated cell size and number quantification were executed.
For mRNA analysis, NRCMs were transduced with AAV6-FHL2 at a MOI of 30,000 (n = 5 per condition), serum was reduced to 0.2 % during 24 h and then NRCMs were stimulated for 48 h with PE (50 µM) or without. Subsequently, cells were harvested and RNA was extracted as described in the appropriate section thereafter.
Engineered heart tissue generation, transduction, contraction measurements, and immunofluorescence
Generation of EHTs from neonatal rat heart cells was performed as previously described [
14,
21,
25]. EHTs were transduced with AAV6 encoding FHL2 WT or variants at a MOI of 1,000. Briefly, transduction was performed directly in the reconstitution mix before pipetting it into the agarose slots. Constructs were then cultured at 37 °C in 7 % CO
2 humidified cell culture incubator for 14–21 days. EHT medium for the first 10 days of culture consisted of DMEM (Biochrom), 10 % horse serum inactivated (Gibco), 2 % chick embryo extract, 1 % penicillin/streptomycin (Gibco), insulin (10 lg/mL, Sigma-Aldrich), and aprotinin (33 µg/mL, Sigma-Aldrich). To exclude any hypertrophic influence by the serum we applied a protocol with horse serum in the medium being successively replaced by triiodothyronine (T
3) and hydrocortisone after day ten of culture [
26]. Therefore, horse serum medium content was reduced to 4 % between day 10 and 13. Up to day 13 medium was changed three times per week, afterwards twice daily. From day 13 onwards EHTs were kept in serum-free medium, i.e. the above medium without horse serum plus T3 (0.5 ng/mL, European Commission-Joint Research Centre IRMM-469) and low concentrations of hydrocortisone (50 ng/mL, Sigma-Aldrich). EHTs started to beat coherently one week after casting. For chronic PE stimulation, phenylephrine (PE, 20 µM, powder dissolved in H
2O) was added to the medium every 12 h for seven consecutive days starting day 14. Contraction measurements were performed on day 8, 10, 13, 14, 17, 20 and 21 as previously described [
14,
21,
25,
48]. Subsequently after chronic stimulation, EHTs were PBS washed three times and directly processed or frozen in liquid nitrogen.
For immunofluorescence analysis, the entire EHTs were analyzed using confocal imaging as specified [
21]. Immunofluorescence was performed as described above (primary antibodies against FLAG 1:800 and cMyBP-C, custom 1:200, nuclear staining by Draq5, Axxor, 1:1000).
Statistical analysis
Data are presented as mean ± SEM. Statistical analyses were performed by one-way or two-way ANOVA followed by Dunnett’s or Bonferroni’s post test, and by Student’s t test as indicated in the figure legends, using the GraphPad software (GraphPad Software), version 5.02. A value of P < 0.05 was considered significant. Quantitative PCR data analyses were carried out using the ΔΔCt method.
Discussion
The recent evidence that FHL2 plays an antihypertrophic role and that its expression is reduced in human heart failure suggested that altered FHL2 expression or variants could be associated with HCM. Therefore, we evaluated FHL2 expression in human and mouse HCM cardiac tissue samples and screened FHL2 for genetic variants in a cohort of 121 HCM unrelated index cases, who do not carry any other known mutations in major sarcomeric genes. The key findings are as follows: (1) FHL2 expression and number of FHL2 immunogold particles were lower in ventricular tissue of HCM patients; (2) out of six identified FHL2 variants in unrelated HCM families, two were novel (T171M, V187L); (3) gene transfer of FHL2 WT or nonsynonymous variants in cardiac myocytes down-regulated Acta1 and partially blunted the hypertrophic response induced by PE or ET1; (4) force and velocity of contraction or relaxation of EHTs were higher in the presence of T171M and V187L mutants than WT under basal conditions; (5) chronic PE stimulation reduced contractile force and velocities in all groups, but had no major effect in T171M-transduced EHTs. These findings support the view that FHL2 expression is negatively associated with HCM and that FHL2 WT partially protects against PE- or ET1-induced hypertrophy, whereas T171M and V187L FHL2 mutants mainly induced hypercontractility.
The molecular mechanisms by which changes in
FHL2 expression and/or
FHL2 variants could contribute to HCM and associated cardiac dysfunction remain elusive at this point. One obvious hypothesis is a loss of antihypertrophic function [
15,
28,
40,
42,
44]. The present study shows not only a downregulation of
FHL2 mRNA levels upon PE stimulation in NRCMs, but also lower
FHL2 mRNA and protein levels as well as a lower number of FHL2 immunogold particles in the I-band of the sarcomere in human and mouse models of HCM. This supports previous findings of reduced FHL2 protein levels in AngII-induced cardiac hypertrophy in mice [
40] and in failing human hearts [
4]. Reduced
FHL2 mRNA levels were also found in models of pathological hypertrophy induced by ET1, PE or afterload enhancement in rat EHT ([
26] and Hirt, unpublished data). Besides a lower grade of antihypertrophic action, low
FHL2 expression could also have consequences on cardiac energy metabolism [
33]. Whether
FHL2 downregulation is a dispensable by-product or whether the same mechanisms leading to hypertrophy and heart failure cause this concomitant downregulation is unclear. Altogether, these data suggest that
FHL2 expression is negatively associated with HCM and more generally with cardiac disease.
However, several arguments suggest that the hypothesis of the lack of antihypertrophic function of FHL2 mutants is unlikely. First, FHL2 mutants showed a stable expression on gene and protein level. Second, in the absence of hypertrophic stimuli, cardiac myocyte area and mRNA levels of
Nppa,
Nppb and
Rcan1.4 did not differ between non-transduced and
FHL2-transduced NRCMs, indicating that neither FHL2 WT nor mutants exert growth effects per se. Conversely,
Acta1 gene expression was markedly down-regulated after gene transfer of FHL2 WT or mutants in NRCMs, whereas downregulation of
Fhl2 and up-regulation of
Acta1 coincided with the appearance of hypertrophy in
Mybpc3-targeted KI HCM mice (Friedrich, unpublished data and [
22,
35]). This supports the previously suggested role of FHL2 as a negative regulator of cardiac hypertrophy [
28,
42,
44]. Finally, FHL2 mutants had similar effects as WT on cardiac myocyte area and
Nppa,
Nppb, Rcan1.4 and
Acta1 mRNA levels, which indicates that at least in the 48 h NRCM assay the genetic variants do not interfere with the antihypertrophic activity of FHL2.
Findings in EHTs rather indicate that the genetic variants may induce hypercontractility, which suggests a gain-of-function effect. Out of the six
FHL2 variants identified in HCM families, variants IV–VI are silent variants that were previously recognized as SNPs (rs137869171, rs3087523 and rs11124029), whereas variants I–III are
FHL2 nonsynonymous substitutions (R177Q, T171M, V187L). These nonsynonymous substitutions were found in HCM-affected individuals, they were expressed into stable proteins, correctly incorporated into the sarcomere and thereby may act in a dominant fashion on endogenous FHL2/sarcomere function. However, there are arguments against variant I (R177Q) as disease causing. First, variant I (recently named as rs1131188481) was found in 201 of 12,805 alleles (frequency 1.5 %) in the NHLBI Exome variant server and was classified as a benign polymorphism by Mutation Taster and Polyphen-2. Second, variant I was detected in a family, which also carries a mutation in
FHL1 [
21]. Finally, the EHT contractile parameters did not differ to WT under basal conditions and after chronic PE. In contrast, the novel
FHL2 variants (T171M, V187L) could be associated with or contribute to HCM for the following reasons. First, variants II and III were not detected in the large cohort of the NHLBI Exome variant server and in silico analysis predicts harmful consequences for variant I. Second, the expression of both variants in EHT induced hypercontractility, characterized by either higher amplitude and/or shorter kinetics of force under basal conditions. Additionally, whereas chronic PE stimulation depressed EHT force of contraction in all groups as expected, it did not have any effect in T171M-transduced EHTs. One explanation could be that the T171M substitution alters post-translational modifications such as phosphorylation or oxidation. In fact, in silico analysis proposes a phosphorylation of FHL2 at position T171 (
http://www.phosida.com). Which kinase performs this modification is unknown, but ERK2 and protein kinase D were previously excluded [
44,
54]. Hypercontractility with consecutive higher LV systolic pressure is a common finding in patients with HCM [
23,
41] and is in agreement with previous findings obtained for HCM-associated
FHL1 and
ANKRD1 variants in EHTs [
14,
21]. It is also consistent with the increased myofilament Ca
2+ sensitivity observed in human or mouse models of HCM with different sarcomere gene mutations [
18,
31,
37,
38,
46,
57,
58], and with enhanced contractile function in muscle fibers, cardiac myofibrils, or cardiac myosins containing different
MYH7-HCM mutations and HCM mouse native or engineered tissues [
1,
3,
7,
37,
43,
53,
55]. Hypercontractility and diastolic dysfunction, but not LVH have also been observed in a transgenic
Tnnt2 HCM mouse model [
56]. Similarly, functional changes (such as increased myofilament Ca
2+ sensitivity, diastolic dysfunction) do exist without hypertrophy in heterozygous
Mybpc3-targeted KI mice [
18,
22,
55]. These data support the findings in human HCM patients who exhibited “supranormal” contractile function or diastolic dysfunction without accompanying LVH [
16,
27,
36]. Hence, changes in contractility as induced by
FHL2 variants could precede or may even be independent of LVH development and consistent with the conception that compensatory mechanisms may play a role in the development of hypertrophy.
In conclusion, this study provides evidence for altered FHL2 expression and novel FHL2 genetic variants in HCM. However, whereas we confirmed that FHL2 has an antihypertrophic role, our data suggest that FHL2 genetic variants did not release this antihypertrophic effect of FHL2. Instead, our data support the view that FHL2 genetic variants could increase cardiac function in HCM.