Transcription factor neuromancer/TBX20 is required for cardiac function in Drosophila with implications for human heart disease

Expression of nmr in Adult Cardiomyocytes.

nmr genes are prominently expressed in the embryonic myocardium (12–14). To determine whether cardiac expression persists, we examined nmr expression in the adult heart. Double labeling with nmr^H15-lacZ (13), and α-Actinin (33), a Z-line marker of myofibers, shows that nmr is expressed in the adult heart and co-labels with Dmef2 and Tinman-positive nuclei of the contractile myocardium (Fig. 1A-D) (34). The two smaller Dmef2-positive nuclei correspond to the ostiae, which form openings in the heart tube controlling movement of the hemolymph into and out of the heart (asterisk in Fig. 1B and arrows in Fig. 1 C and D). This observation is consistent with our real time quantitative RT-PCR results in which nmr1&2 mRNA were both detected in the adult fly heart (see Fig. 4), indicating nmr genes are indeed expressed in adult heart.

Pacing-Induced Heart Arrest/Fibrillation in nmr/Tbx20 Mutants.

Since tinman is expressed and required in the adult heart, we wondered whether nmr also contributes to adult cardiac structure and function. nmr1 null mutants (nmr1⁶¹⁴) are viable and show a mild embryonic heart phenotype, whereas mesodermal nmr2 knockdown is almost as severe a (cardiac) phenotype as nmr1&2 double mutants (12–14). To test the roles of these genes in adult flies, we measured cardiac performance in nmr mutants using an electrical pacing protocol, which stresses the heart and induces fibrillation or arrest, especially in aged or compromised hearts (31). Although we have previously used the term “heart failure” to describe this cardiac dysfunction (24, 31), this term does not imply the “pump failure” used to describe mammalian heart function but rather an alteration or cessation of normal cardiac pumping as a result of the pacing induced stress. First, we examined flies heterozygous for nmr1⁶¹⁴ or a larger deficiency covering both nmr1&2, Df(2L)Exel6012. No significant increase in heart arrest/fibrillation following pacing was observed in these flies compared to wild type (Fig. 1E and supporting information (SI) Table S1). However, when both copies of nmr1 are removed (nmr1⁶¹⁴) or when nmr2 function is reduced with heart-specific knockdown (tinCΔ4-Gal4 driving UAS-nmr2-RNAi) (13), there is a dramatic increase in the rate of pacing-induced heart arrest/fibrillation (Fig. 1E and Table S1), which further increases with age (Fig. S1 A and D). The down-regulation of nmr2 in the heart also results in a severely truncated lifespan (Fig. S1B), similar to what is observed for flies with a cardiac-specific loss of tinman expression (30). To distinguish between an embryonic versus an adult requirement for nmr we knocked down nmr2 in the heart only during the adult stage (using the TARGET system (32): tub-Gal80-ts;tinCΔ4Gal4 or tub-Gal80-ts;24BGal4 crossed to UAS-nmr2-RNAi, see Materials and Methods). In these flies, we also observed an increased rate in pacing-induced heart arrest/fibrillation compared to control flies (Fig. 1F, Fig. S1D, and Table S1). Although we cannot categorically rule out that some RNAi off-target effects may contribute to observed phenotype, these data strongly suggest that nmr function is indeed required for normal cardiac performance in adult flies. This conclusion is further substantiated by the effects on cardiac function and morphology as a result of genetic interactions between nmr deficiencies and tinman (see Fig. 3).

Arrhythmias and Abnormal Myofibrillar Structure in nmr/Tbx20 Mutants.

We conducted image-based contraction analysis in nmr mutant flies. High-speed movies were taken from dissected and exposed hearts and contractions were displayed as M-modes (see Materials and Methods) showing the dynamics of heart wall movements over time (26). Heart specific knock-down of nmr2 causes a considerably irregular beating pattern and an overall increase in the heartbeat length compared to the controls (Fig. 2A-C'). These arrhythmias are due to the increased variability of diastolic (but not the systolic) interval lengths of the heart period (Fig. 2 A-F). The severity of arrhythmia is quantified by the standard deviation of the mean heart period for each fly (“arrhythmia index,” Fig. 2F; see ref. 26). Thus, nmr mutant hearts not only beat more slowly but are also arrhythmic. Depleting nmr levels in the adult heart possibly accelerates the cardiac aging process, since both the incidence of arrhythmias and length of the heart period normally increase with age (26).

Given the heart function defects observed for nmr mutants, we were also interested in examining these hearts for cellular defects. During metamorphosis, the fly heart undergoes a remodeling process producing a regular spiral or transverse muscle fiber pattern within the tinman-expressing myocardial cells (see Fig. 1 A-D). In addition, a new set of longitudinal fibers that express Dmef2 but not tinman forms ventral to the heart (Fig. 1 C and G) (34, 35). We examined nmr2 knockdown hearts using α-Actinin antibodies, which localize to Z-lines of Drosophila muscle fibers (33). Using this marker we find that the regularity of the myofibrillar alignment and the spacing between Z-lines is disrupted within the Tinman-positive myocardium, where nmr2 expression is diminished (Fig. 1 I-K and Fig. S1E), but not in the ventral longitudinal muscle fibers that do not express tinman (Fig. 1 G and H and Fig. S1E). No noticeable change in cell death is observed; however, death in 1–2 cells per heart could possibly contribute to the observed myofibrillar disarray. Changes in cardiomyocyte size can also be estimated from measurements of diastolic diameter (DD) and systolic diameter (SD) of the heart tube (Fig. S1 F and G). We found a moderate decrease in DD and a mild increase in SD for nmr2 adult-specific knockdown mutants. These two changes result in a significant decrease in fractional shortening compared to controls, indicating a change in contractility in nmr knockdown hearts (Fig. S1H). The observed abnormalities in myofibrillar structure in response to decreased nmr expression likely contribute to these functional defects.

Genetic Interaction Between nmr/Tbx20 with tinman/Nkx2–5 in Regulating Adult Heart Function.

One of the enduring problems in the study of cardiac disease loci is the relative contribution of the genetic background or polygenic interactions that may ameliorate or aggravate the phenotypic expression of genetic abnormalities. This is the case for known cardiac disease loci, such as Nkx2–5, Tbx5, and Gata4, which cause a phenotype as heterozygotes (15–18, 36). Therefore, we examined whether the fly's T-box and Nkx homologs interact genetically to control adult heart physiology. We combined the heterozygous nmr1&2 deficiency Df(2L)Exel6012 with heterozygotes for either tinman or Df(3L)DocA (which span all three Drosophila Tbx5/6 related genes) (14, 37). The pacing-induced heart arrest/fibrillation rate for Df(2L)Exel6012/+;tinman/+ or Df(2L)Exel6012/+;Df(3L)DocA/+ is dramatically elevated compared to the single heterozygous controls (Fig. 3A and Table S1). This phenotype could be rescued by expressing wild-type nmr2 cDNA specifically in the heart (Fig. 3B and Table S1), indicating that the observed phenotype is indeed due to the genetic interaction of tinman or the doc genes with the nmr locus. An interaction of Df(2L)Exel6012^+/− with tinman^+/− was also observed with respect to the development of arrhythmias (Fig. 3 C–E, Fig. S2 A–E), the level of cardiac expression of potential downstream effector genes (Fig. 4), and the alignment of myofibers within the cardiac myocytes (Fig. 3 F-H). The genes whose expression is most affected as a result of tinman-nmr interactions are the nmr genes themselves, a potassium channel, an ATPase (SERCA), and dystrophin, all of which are likely to be involved in heart morphogenesis or function (Fig. 4). Indeed a dystrophin deficiency has previously been shown to cause myofibriller misalignment in the fly heart (38). Whether the observed changes in gene regulation are direct, indirect, or the consequence of a compensatory mechanism remains to be examined. Nevertheless, our data suggest that cardiogenic transcription factors interact strongly in establishing and maintaining adult heart structure and function.

Cardiac performance exhibits a progressive deterioration with age, which is manifest by an increase in pacing-induced heart arrest/fibrillation rate and in arrhythmias (24, 26). Since tinman and nmr significantly influence adult heart function, we wondered whether reduced cardiac performance in old flies could be rescued by overexpression of these factors. When nmr2 is overexpressed beginning at the 5th week of adult life (Fig. S2F; see Materials and Methods), pacing-induced arrest/fibrillation rate is significantly reduced compared to non-overexpressors (Fig. S2G). Similar effects were observed with tinman cDNA overexpression in aging fly hearts (Fig. S2H). This suggests that maintaining high levels of tinman or nmr expression in older hearts halts or slows the age-dependent cardiac performance decline.

TBX20 Mutations in Humans with Dilated Cardiomyopathy.

Since most, if not all, transcription factors with a demonstrated role in the fly heart also function in vertebrates, we wondered whether TBX20 sequence variants could be associated with cardiomyopathies in human patients. For this purpose, unrelated probands with clinically apparent dilated cardiomyopathy (DCM), who presented with echocardiographic evidence of left ventricular (LV) dilation, systolic dysfunction, and clinical evidence of heart failure and cardiac rhythm disturbance were enrolled for this study. All subjects and controls had blood drawn, DNA was isolated, and lymphoblastoid cell lines were developed. Patient (n = 96) and control (n = 392) DNA samples were screened for non-synonymous variants (mutations) in TBX20 and NKX2–5 genes by direct DNA sequencing (for primers see Table S2). This revealed three missense mutations (L196V, R334Q and W349R) in the TBX20 gene in Caucasians, two of which occurred with incomplete penetrance in familial cases (see Fig. 5 and Fig. S3). None of these mutations were identified in 392 ethnic-matched controls (nearly 800 chromosomes; see Table S1). Two of the variants had alterations in conserved residues in the DNA-binding T-box of TBX20 (Fig. S3). Notably, variants in the familial cases of DCM demonstrated reduced penetrance, which is a common finding in inherited forms of cardiomyopathy. In Family 4, for instance, only half of the individuals with the R334Q variant had clinically apparent DCM at the time of the evaluation. All non-synonymous variants found in patients and ethnic-matched controls are listed in Tables S3 and S4. No potentially disease-causing mutations in NKX2–5 were identified in the examined patient cohort and no affected subject lacked the variant in family 4. These findings imply that candidate gene functions found in the Drosophila model are potentially relevant for human cardiac disease. However, definitive causality cannot be claimed at this point, in part because of the sporadic or incompletely penetrant nature of the cases identified. Although this is a common observation in human cardiovascular diseases such as cardiomyopathies and arrhythmia disorders (40, 41), it limits our ability to conclude a cause-and-effect relationship at this time. To demonstrate a causal relationship between gene mutation and disease, it will be important to conduct linkage analysis in large families with identified and functional studies of the identified variants. Nevertheless, our observations of a number of patients (but not controls) with non-synonymous variations in conserved regions of the coding region of TBX20 make this gene a promising candidate to potentially be involved in the human disease.

Role of TBX20 in the Genetics of Congenital Heart Disease in Humans.

Kirk, et al. previously reported that, in addition to DCM, mutations in TBX20 are also associated with ASD and mitral valve disease (41). Consequently, we screened children with CHD including ASDs and ventricular septal defects (VSDS), as well as more complex CHD associated with these septal defects. In addition, mitral valve disease was evaluated. We screened 96 children with ASD and 96 with VSD associated disease. In four children, sporadic TBX20 variants were identified. A TBX20 variant in exon 4 resulting in a 597C>G alteration, which changes a histidine to an aspartic acid (H186D), was found in two unrelated children. In one child, an African American female, an atrioventicular canal with secundum and primum ASDs, as well as a cleft mitral valve with moderate mitral regurgitation was noted. The second child, a Hispanic male, had pentalogy of Fallot (tetralogy of Fallot with ASD). The other exon 4 mutation (601T>C; L197P) was identified in an African American female with pentalogy of Fallot and in a white female with isolated ASD. None of these variants were identified in 392 ethnic-matched controls (see Tables S3 and S4).

Drosophila Stocks.

tin³⁴⁶/TM3-ftzlacZ (2), tin^EC40/TM3, Df(3R)tin^GC14 (1), Df(3L)DocA (37), nmr1⁶¹⁴, nmr1²¹⁰ (13), Df(2L)Exel6012 (38), also indicated as Df(nmr). UAS-Gal4 system was used as in (45). Gal4 and UAS lines used: twi-Gal4 (twi>; ref. 44), 24B-Gal4 (24B>; ref. 45), twi-Gal4;24B-Gal4 (twi24B>; ref. 46), tinCΔ4-Gal4 (47), Dmef2-Gal4 (48), UAS-tin (49), UAS-nmr2RNAi, UAS-nmr1, and UAS-nmr2 (13).

Electrical Pacing.

The electrical pacing was conducted as previously described (24, 31). Briefly, 50–100 flies were paced with a square wave stimulator at 40 V and 6 Hz for 30 s and scored for heart arrest/fibrillation rate.

Image Analysis and M-mode Traces on Semiintact Preparations.

Semiintact heart preparations for image analysis were as in Ocorr, et al. (26). M-modes from movies were generated using Simple PCI software and MatLab-based image analysis (26).

Quantitative Real-Time PCR.

RNA was extracted from 20 1-week old hearts for each genotype and subjected to qRT-PCR as described in SI Text.

Patient Evaluation.

Patients were evaluated by chest radiography, electrocardiography, echocardiography, and MRI. Left ventricular size and function were evaluated by M-mode and two-dimensional Doppler and color Doppler echocardiographic images. Further details are described in SI Text.