The HAND1 frameshift A126FS mutation does not cause hypoplastic left heart syndrome in mice

Abstract

Aims

To test if a human Hand1 frame shift mutation identified in human samples is causative of hypoplastic left heart syndrome (HLHS).

Methods and results

HLHS is a poorly understood single ventricle congenital heart defect that affects two to three infants in every 10 000 live births. The aetiologies of HLHS are largely unknown. The basic helix–loop–helix transcription factor HAND1 is required for normal heart development. Interrogation of HAND1 sequence from fixed HLHS tissues identified a somatic frame-shift mutation at Alanine 126 (NP_004812.1 p.Ala126Profs13X defined as Hand1^A126fs). Hand1^A126fs creates a truncated HAND1 protein that predictively functions as dominant negative. To determine if this mutation is causative of HLHS, we engineered a conditional Hand1^A126fs mouse allele. Activation of this allele with Nkx2.5^Cre results in E14.5 lethality accompanied by cardiac outflow tract and intraventricular septum abnormalities. Using αMHC-Cre or Mef2CAHF-Cre to activate Hand1^A126fs results in reduced phenotype and limited viability. Left ventricles of Hand1^A126FS mutant mice are not hypoplastic.

Conclusions

Somatically acquired Hand1^A126FS mutation is not causative of HLHS. Hand1^A126FS mutation does exhibit embryonic lethal cardiac defects that reflect a dominant negative function supporting the critical role of Hand1 in cardiogenesis.

1. Introduction

Hypoplastic left heart syndrome (HLHS) is a poorly understood single ventricle congenital heart defect (CHD) that accounts for 2–3% of all CHDs.^1–3 In HLHS, the LV is diminished in size and unable to support the systemic circulation. To account for the insufficiency, the RV is fed with oxygenated blood via an atrial septal defect directly connecting the right and left atriums or ventricular septal defects (VSD) directly connecting the RV and LV. Blood from the RV then exits the ventricle via the pulmonary artery and enters the aorta via a persistent ductus.

Traditional linkage approaches have identified a growing number of mutations that give rise to survivable CHDs.^4–8 However, it has proved more difficult to identify CHD-causing genes as a result of early in utero lethality due to catastrophic failure of cardiogenesis, or alternatively, failure of other organ systems required for embryo viability. Such gene mutations cannot be inherited. It is suggested that the acquisition of somatic mutation(s) in such genes, which appear in only a subset of the cells in the developing embryo, could result in greater in utero viability or even in survivable CHDs; however, such a mechanism would likely be very rare.⁹ HAND1, a bHLH transcription factor, is a prototypical example of a CHD-causing gene whose analysis has been hampered due to early embryonic lethality.^10–13 The function of Hand1 is regulated by homo- or hetero-dimerization with other bHLH proteins through its bHLH domain. Hand1 forms dimers that bind E-box (CANNTG) and D-box (CGNNTG) cis-elements.^14–16 Deletion of Hand1 or hypomorphic Hand1 expression in mice results in cardiac morphogenic defects, and early embryonic lethality due to placental defects.^{10,11,13,17,18} Recently, somatic HAND1 mutations have been implicated in the genesis of Tetralogy of Fallot,¹⁹ VSDs,²⁰ and HLHS.^9,21 These results were obtained from fixed tissues and subsequent studies interrogating unfixed samples were unable to confirm these results.²² Although it is clear that Hand1 plays an important role in cardiac morphogenesis, mutations within the HAND1 coding domains are unlikely to result in viable offspring; therefore, the idea that HAND1 mutations could occur somatically is an intriguing disease model. The controversy regarding somatic mutation as a disease model and the potential caveats of using fixed samples prompted us to address this question directly. We tested a reported Hand1 HLHS mutation (NP_004812.1 p.Ala126Profs13X)^21,23 by generating a conditional activation knock-in allele (defined as Hand1^A126fs) of this putative Hand1 somatic mutation.

Results show that in tissue culture analysis, the p.Ala126ProfsX13 mutant protein localizes to the nucleus and inhibits both DNA binding and transcriptional activation. Phenotypic and molecular analysis of Hand1^SFA126fs mice generated with Nkx2.5^Cre24 shows that expression of this dominant negative Hand1 protein disrupts cardiac development. Hand1^A126fs/+ mice die by E15.5 and display outflow tract (OFT) abnormalities, thin myocardium and VSDs. Although cardiac Hand1 expression and linage commitment is largely left ventricular,^10,25 we observe normal sized LVs in Hand1^A126fs/+ mutant mice. Using the αMHC-Cre driver,²⁶ we encounter less severe heart defects and mutant viability. We also employed the anterior heart field (AHF) Mef2cAHF-Cre driver²⁷ and show similar OFT defects to Nkx2.5^Cre mutants; however, these mice also survive. Together, these data suggest that although the Hand1 p.Ala126Profs13X mutant protein has altered function, when expressed in mouse cardiomyocytes, it is not causative of HLHS.

2. Methods

2.1 Mouse strains, genotyping

SFHand1^A126FSwas generated from ES cells targeted with the constructs described and genotyping by Southern blot¹⁰ or with primers H1 5′-CTG CCA TTG GCT CCG GCT AGA GGT-3′ and PGK 5′-GGC TGC TAA AGC GCA TGC TCC AGA CTG-3′ using PCR conditions of 94 °C 1 min, 60 °C 1 min, 72 °C 1 min for 35 cycles. Nkx2.5^{Cre 24} mice are a Cre-recombinase knock-in allele that express Cre within the cardiac crescent at E7.5 marking myocardium, endocardium.²⁸Hand1 is not expressed in edocardium.²⁵αMHC-Cre transgenic mice express Cre recombinase at E9.5 within the myocardium after which the Cre is dormant until after birth.²⁶Mef2CAHF-Cre is expressed endocardium and myocardium at E7.5.²⁷

No anaesthetic/analgesic agents were used. Euthanasia was performed using CO₂ gas in a closed chamber followed by cervical dislocation. All experiments were performed conforming to the NIH guidelines following the Indiana University IACUC animal protocol 10809.

2.2 Cell transfection, luciferase assays, and electrophoretic mobility shift assays

HEK293 CaPO₄ transfections and Luciferase assays protocol as described were performed as described.^29,30 Ebox reporter, Dbox reporter, Hand1 pIRESNeo, and E12 pIRESNeo expression plasmids are described.^14,31 Hand1A126fspIRESNeo was mutagenized using QuickChange Mutagenesis Kit (Agilent Technologies). Data represent six-independent experiments. Error bars denote SE. Asterisk represent significance of P ≤ 0.05 by ANOVA.

2.2.1.Electrophoretic mobility shift assays

EMSAs were carried out as previously described.¹⁴ EMSA probe Hand E-box (5′-gga ttc cat tgc atc tgg att cca gag-3′) was used to shift and as specific competitor. Non-specific competitor oligo (5′-gga ttc cat tgGGtc AAg att cca gag-3′).

2.3 Histology

Embryos (E9.5–E18.5) were processed as described.¹³ Eight viable embryos per genotype were examined for all analyses. Bifurcated heart preparations were generated by cryo-sectioning.

2.4 In situ hybridization

Section and whole mount in situ hybridization (ISH) were performed as described^32,33 for probes Hand1, Wnt11, Tbx5, Dkk3, Cited1, Bmp10, Nppa (Anf), Cxcl12, Tbx20, Hey2.

2.5 Lysotracker and EdU immunohistochemistry analysis

Lysotracker and Click-IT EdU Imaging (Life Technologies) was incubated with embryonic hearts as described.^26,32

2.6 Quantitative RTPCR (QPCR)

RNA Isolation and quantitative RT–PCR (QRTPCR) was performed as described.³² Error bars denote the maximum and minimum relative level of gene expression set in the QuantStudio 3&5 software. Statistics employed Student’s two-tailed t-test with P ≤ 0.05 significance. N = 6.

3. Results

3.1 Hand1^A126FS acts as a dominant negative transcription factor

To confirm that the p.Ala126Profs13X protein (Hand1^A126FS) functions as a dominant negative, we employed luciferase analysis, EMSA and cellular localization to test functionality (Figure 1). Hand1^A126FS truncates Hand1 within the NH4-terminal portion of the loop domain and is reported to not bind DNA and to repress transcriptional activation of both E- and D-box luciferase reporters in yeast extracts.²¹ To confirm these findings in a mammalian transcription system, we utilized similar E- and D-box luciferase reporters in HEK293 cells (Figure 1A). Results show that alone neither Hand1, Hand1^A126FS (Hand1FS), or E12 transactivate (Figure 1B). When Hand1 is cotransfected with E12 (Hand1 + E12) along with either an E- or D-box reporter significant transactivation (P ≤ 0.05) is observed. Cotransfection of Hand1^A126FS and E12 (Hand1FS + E12) results in a decrease of the Hand1-E12 heterodimer transcriptional activity consistent with previous analysis.²¹ EMSA shows that E12 and (Hand1 + E12) bind to the Ebox probe as homo- (red arrowhead) and heterodimers (black arrowhead), respectively (Figure 1C). Cotranslation of Hand1A126FS with E12 (Hand1^A126fs + E12) results in undetectable shifting of the Ebox probe (Figure 1C, asterisks). We engineered a Hand1^A126fs–eGFP fusion construct and co-transfected this plasmid with an E12–DsRed fusion plasmid (Figure 1D). Results show Hand1^A126fs–eGFP and E12–DsRed nuclear epifluorescence. These data support a model where Hand1^A126FS acts as a dominant negative factor.

Figure 1.

Hand1^A126FS localizes to the nucleus acting as a dominant negative. (A) Schematic of wild-type HAND1 and HAND1 A126FS and Dbox and Ebox luciferase reporters. (B) Luciferase assays from HEK293 lysates with the indicated plasmids (pIRESNeo), open bars; E12-pIRESNeo (E12), grey bars; Hand1-pIRESNeo (Hand1), black bars; Hand1^A126FS-pIRESNeo (Hand1FS), blue bars; Hand1-pIRESNeo + E12-pIRESNeo black diagonal-striped bars; Hand1FS pIRESNeo+ E12-pIRESNeo blue diagonal-striped bars; Hand1-pIRESNeo + Hand1FS-pIRESNeo, waved bars). *^,#P ≤ 0.05 by ANOVA compared with pIRESNeo and Hand1FS + E12, respectively. Error bars show standard error. (C) EMSA of in vitro expressed HAND1 (H1), Hand1 A126FS (H1^A126fs), and E12 in the indicated combinations. Unprogrammed reticulocyte lysates (UP) show non-specific complexes (ns). E12-programmed lysates reveal homodimer binding (red arrowhead) that is competed with specific competitor (SC) but not non-specific competitor (NS). Hand1-E12 heterodimers (black arrowhead). H1^A126fs-E12 heterodimers do not bind DNA (double asterisk). (D) HEK293s co-transfection of Hand1A^126fs-eGFP and E12-DsRed. DAPI (blue) marks nuclei. Both Hand1^A126fs-eGFP and E12-DsRed co-localize to the nucleus. Transfection and EMSA represent five experiments; only one EMSA is shown. Scale bar in (D), 10 μm.

3.2 Generation and analysis of a conditional Hand1^A126FS allele

To test if expression of Hand1^A126FS is causative of HLHS, we engineered a conditional-activation allele by introducing the identified p.Ala126Profs13X mutation into the murine Hand1 locus that includes a stop-flox (SF) cassette that bocks the transcription and translation of the mutant Hand1 protein (Figure 2A). The mutation deletes the G in codon 126 (alanine) producing a 13 amino acid frame shift ending in a termination codon. ES targeted clones were identified at a frequency of 59% (Figure 2B). Mice carrying the conditional-activation Hand1^SFA126FS/+ allele are viable and fertile. We next generated Nkx2.5^Cre; Hand1^A126FS/+ and Nkx2.5^Cre; Hand1^A126FS/fx embryos and show LV myocardium and myocardial cuff expression of Hand1 is detectable from the conditionally activated mutant Hand1 allele (Figure 2C).

Figure 2.

Targeting strategy and expression validation of Hand1A^126FS/+ mice. (A) Schematic of the Hand1 locus and targeting construct. Hand1 targeting arms¹⁰ were utilized to insert a loxP element-flanked Neomyosin cassette (Flox STOP Neo).³² The 3′ targeting arm includes the A126fs mutation. Upon Cre-recombination and removal of the Stop-Flox cassette, activates the Hand1^A126fs allele. (B) Southern blot analysis shows targeting frequency exceeding 50%. (C) E10.5 wholemount Hand1 ISH in control (Hand1^SF-A126fs/+), recombined (Nkx2.5^Cre/+; Hand1^A126fs/+), and compound Hand1 heterozygotes (Nkx2.5^Cre/+; Hand1^A126fs/fx) single-copy mutants. I2, mandibular component of the first pharyngeal arch; oft, cardiac outflow tract; rv, right ventricle; lv, left ventricle; u, umbilicus. Scale bar 150 μm.

To test if expression of the Hand1^A126FS allele models HLHS, we intercrossed Hand1^A126FS/+ females to Nkx2.5^Cre males and looked for live births. After 59 pups were genotyped and no surviving Nkx2.5^Cre; Hand1^A126FS/+ were encountered, we initiated embryonic analysis (see Supplementary material online, Table S1). At E14.5, we encountered Nkx2.5^Cre; Hand1^A126FS/+ embryos at a frequency of 0.08 (n = 20; expected frequency 0.25) and no mutants were recovered beyond E15.5 (n = 46). We then initiated isolating Nkx2.5^Cre; Hand1^A126FS/+ embryos at E12.5, which we obtained at a frequency of 0.30.

Hand1^SFA126FS/+ embryos appear phenotypically normal exhibiting normal OFT structures, forming IVS and trabeculated RV and LV (Figure 3A–E). Endocardial cushions are well developed and positioned correctly above the forming IVS (Figure 3D, black arrow). Two wholemount examples of Nkx2.5^Cre; Hand1^A126FS/+ show oedema (white arrowheads Figure 3F and K) and elongated OFTs (white bars) indicating that morphological alignment of the forming heart is compromised. Abnormal haemorrhaging is apparent in Nkx2.5^Cre; Hand1^A126FS/+ mutants (Figure 3K).

Figure 3.

Nkx2.5^Cre/+; Hand1^A126FS/+ E12.5 mutants present with an elongation of the cardiac OFT. (A) Nkx2.5^+/+; Hand1^SF-A126FS/+ wholemount control compared with two Nkx2.5^Cre/+; Hand1A^126FS/+ mutants (F, K). Measure of cardiac OFT (white bars) shows that mutant OFTs extend farther into the forming RV, which appears smaller and shows some signs oedema (white arrowhead). Section analysis of the control shown in (A). (B–E) reveals expected structural development that includes a septated pulmonary trunk (pt) developing IVS and LV. Magnification in (E) shows a well-formed compact zone (black bar). Matched sections from Nkx2.5^+/+; Hand1A^126FS/+ mutants shown in (F) and (K). (G–O) A poorly formed IVS (asterisk), small RV, and a thinner compact zone (J, O). Size of the LV is indistinguishable between control and Nkx2.5^+/+; Hand1A^126FS/+ mutant hearts. Data represent an example of each genotype. Ten embryos per genotype were examined. ra, right atria; la, left atria; rv, right ventricle; lv, left ventricle; ivs, intraventricular septum; pt, pulmonary trunk. Scale bars 150 and 500 μm.

Section analysis of Nkx2.5^Cre; Hand1^A126FS/+ mutants reveals a poorly formed IVS that is positioned to the right of the endocardial cushions (compare Figure 3C and D with H, I and M and N), a thin compact zone (black lines Figure 3E, J, and O), and hypotrabeculation. Given that Hand1 is expressed within the myocardium of the left ventricle and a region of the SHF-derived myocardial cuff, but not the majority of the IVS or endocardium^25,29 these phenotypes are consistent with both cell-autonomous and non-cell autonomous defects in cardiogenesis.

3.3 Increased cell death is observed in hearts of Hand1^A126FS embryos

To determine the cause of the Nkx2.5^Cre; Hand1^A126FS/wt mutant heart phenotypes, we looked at cell death using Lysotracker staining at E10.5 and E11.5 (Figure 4). Control Hand1^SFA126FS/+ hearts display nearly undetectable levels of cell death as indicated by punctate epifluorescence (Figure 4A and B). Nkx2.5^Cre; Hand1^A126FS/+ mutant hearts reveal a clearly visible increase in the numbers of dying cells within the cardiac ventricles and OFT (Figure 4C and D, white arrows). Cell death is observed in structures that express Hand1 (LV) and structures that do not express Hand1 (RV) indicating autonomous and non-cell autonomous death. Cell proliferation analysis was performed (Figure 4E–G) and results show that proliferation is indistinguishable between control and Nkx2.5^Cre; Hand1^A126FS/wt mutant hearts (Figure 4E–G).

Figure 4.

Nkx2.5^Cre/+; Hand1^A126FS/+ hearts display an increase in both cell autonomous and non-cell autonomous cell death. Control Hand1^SF-A126fs/+ (A, B) and Nkx2.5^Cre/+; Hand1^A126FS/+ mutants (C, D) were assayed for cell death using Lysotracker staining in wholemount at E10.5 and E11.5. Control hearts show little cell death. Increased cell death is observed in OFT, RV, and, LV of Nkx2.5^Cre/+; Hand1^A126FS/+ mutants (white arrows). (E–H) EdU wholemount proliferation analysis at E10.5 and E11.5. Results show no obvious differences in between control and mutant hearts. Data represent an example of each genotype. Six embryos per genotype were examined. ra, right atria; la, left atria; rv, right ventricle; lv, left ventricle. Scale bars 150 μm.

3.4 Altered gene expression is observed in hearts of Hand1^A126FS embryos

We next interrogated gene expression defining cardiogenesis. Wnt11 marks the precardiac mesoderm of the OFT and is essential for SHF development.^34,35 E11.5 wholemount ISH of Wnt11 confirms that Nkx2.5^Cre; Hand1^A126FS/+ mutant OFTs are elongated (Figure 5A and B; white bars). The T-box transcription factor Tbx5 expression is largely restricted to the LV.³⁶ Compared with Hand1^SFA126FScontrols, Nkx2.5^Cre; Hand1^A126FS/+ mutant hearts reveal a broader domain of Tbx5 expression that includes the RV (Figure 5C and D; white arrowhead). Cardiac expression of Dkk3 generally marks the cardiac mesoderm becoming enriched within the IVS.^37,38 Comparison of Dkk3 expression in Hand1^SFA126FScontrols and Nkx2.5^Cre; Hand1^A126FS/+ mutant hearts shows a largely comparable expression pattern; however, the level of Dkk3 appears upregulated in Nkx2.5^Cre; Hand1^A126FS/+ mutants (Figure 5and F, white arrowhead). Cited1 is a trabecular maker and is reported to be downregulated in Hand1 cardiac conditional knockout mice.¹² Comparison of control and Nkx2.5^Cre; Hand1^A126FS/+ mutant hearts reveals no difference in Cited1 expression (Figure 5G and H). LV size of Hand1^A126FS/+ mutants is indistinguishable from controls.

Figure 5.

Wholemount ISH of Nkx2.5^Cre/+; Hand1^A126FS/+ mutant embryos reveals altered cardiac gene expression during heart formation. E11.5 Nkx2.5^+/+; Hand1^SF-A126FS/+ control (A, C, E, G) and Nkx2.5^Cre/+; Hand1^A126FS/+mutant (B, D, F, H) embryos hybridized to the indicated riboprobes. Wnt11 marks OFT tissue (white bars; A and B). Tbx5 is an LV chamber maker. Tbx5 expression appears more homogeneous in Nkx2.5^Cre/+; Hand1^A126FS/+mutants when compared with controls (white arrowhead; C and D). Dkk3 marks the IVS and ventricular chambers. Dkk3 expression appears more robust in Nkx2.5^Cre/+; Hand1^A126FS/+mutants (white arrowhead; E and F). Cited1 appears unchanged between control and Nkx2.5^Cre/+; Hand1^A126FS/+mutants (G and H). Data represent an example of each genotype. Eight embryos per genotype were examined. Scale bars 150 μm.

Next we performed section ISH at E10.5 and E12.5 on Hand1^SFA126FS controls and Nkx2.5^Cre; Hand1^A126FS/+ mutants (Figure 6). We looked at the trabecular markers Bmp10,³⁹Nppa (Anf),⁴⁰ and Cited1 (Figure 6A–L). Results show that expression is largely unaltered. Analysis of Cxcl12 expression, a cardiomyocyte ligand required for coronary formation,^41,42 shows Cxcl12 expression is markedly upregulated (Figure 6M–P; arrowheads). Similarly, Dkk3 expression (Figure 6Q–T) shows a marked expansion of expression consistent with E11.5 analysis (Figure 5). Examination of the compact zone markers Tbx20^43,44 and Hey2^45,46 show that Tbx20 expression is upregulated in Nkx2.5^Cre; H and 1^A126FS/+ mutants (Figure 6U–Xl, black arrowheads) while Hey2 expression is decreased (Figure 6Y–Ab, asterisks).

Figure 6.

Section ISH in E10.5 and E12.5 control and Nkx2.5^Cre/+; Hand1^A126FS/+ mutants. Bmp10 (A–D) and Anf (E–H) expression marks trabeculae. No significant changes in expression are observed. Cited1 expression (I–L) expression is not altered. Cxcl12 cardiac expression is required for formation of intra-ventricular coronary arteries. At E10.5 (M, N) and E12.5 (O, P), Cxcl12 expression is upregulated (black arrowheads) compared with controls (M and O). Dkk3 (Q–T) shows expanded expression (black arrowheads) in Nkx2.5^Cre/+; Hand1^A126FS/+mutants (R, T) compared with control (Q, S). The compact zone markers Tbx20 (U–X) and Hey2 (Y–Ab) show altered expression in Nkx2.5^Cre/+; Hand1^A126FS/+mutants. Tbx20 is upregulated in Nkx2.5^Cre/+; Hand1^A126FS/+mutants (V, X black arrowheads). Hey2 expression is markedly down in Nkx2.5^Cre/+; Hand1^A126FS/+mutants (Z, Ab; asterisks) compared with Nkx2.5^+/+; Hand1^SF-A126FS/+ controls. Data represent an example of each genotype. Eight embryos per genotype were examined. Scale bars 200 μm.

As we observe a thin walled myocardium in the Nkx2.5^Cre; Hand1^A126FS/+ mutants, we also looked at endocardial and epicardial gene expression which can influence myocardial growth^47–49 and production of cardiac myofibroblasts and coronary vasculature.^50–53 Expression analysis of the endocardial markers Hand2, Flt1, Tie2, Nrg1, Nrp1, VegfR2, and NfactC1 and epicardial markers WT1 and Tcf21 by QPCR from E10.5 RNA control and mutant hearts revealed no observable changes in expression (see Supplementary material online, Figure S1).

As some Nkx2.5^Cre; Hand1^A126FS/+ mutants survive to E14.5, we looked at these mutant phenotypes (Figure 7). H&E sections of E14.5 Nkx2.5^Cre; Hand1^A126FS/+ mutant hearts reveal obvious VSDs both membranous and muscular in nature (Figure 7A and B). Although the LV compact zone is thin when compared with controls (black bars), LV chamber size is clearly proportional to RV size and similar to controls. We observe lower levels of Bmp10 expression (Figure 7C–F), which may reflect either real changes in expression or a near death state of these mutants. Dkk3 maintains its enhanced expression when comparing Nkx2.5^Cre; Hand1^A126FS/+ mutants to controls suggesting that death is not causative of decreased Bmp10 expression.

Figure 7.

Histological comparison and assessment of gene expression in E14.5 Nkx2.5^+/+; Hand1^SF-A126FS/+ control and Nkx2.5^Cre/+; Hand1A^126FS/+mutants. (A) H&E transverse Control hearts show a patent IVS and separate RV and LV with a well-established compact zone (black bar), RA, and LA. (B) Nkx2.5^Cre/+; Hand1^A126FS/+mutants consistently display VSDs (asterisk) and thin poorly developed compact zone (black bar). LV chamber size is unaffected. (C–F) E14.5 Bmp10 expression reveals a loss trabecular expression. Black boxes in (C, D) define the higher magnification images of (E) and (F). Lines through compact zone measures thickness. (G–J) Dkk3 expression is upregulated in E14.5 Nkx2.5^Cre/+; Hand1^A126FS/+mutants. Boxes in (G) and (H) define the higher magnification view shown in (I) and (J). Data represent an example of each genotype. Six embryos per genotype were examined. ra, right atria; la, left atria; rv, right ventricle; lv, left ventricle; ivs, intraventricular septum. Scale bars 500 and 100 μm.

3.5 aMHC-cre and Mef2CAHF-cre activation of the Hand1^A126FS/+ allele results in improved phenotype and viability

One caveat is that the Nkx2.5^Cre knock-in allele could have influence on Hand1 LV expression.⁵⁴ To help address the Nkx2.5 contribution, we employed the αMHC-Cre transgenic mice that recombines within E9.5 cardiomyocytes.²⁶αMHC-Cre^tg/+; Hand1^A126FS/+ embryos were obtained at expected ratios at E10.5 and E14.5 (see Supplementary material online, Table S2). To our surprise, we identified surviving αMHC-Cre^tg/+; Hand1^A126FS/+ neonatal mice at a frequency of 0.11. We looked at cardiac morphology and expression of Nppa (Anf) at E14.5 (see Supplementary material online, Figure S2A and B). Results show indistinguishable differences in phenotype and in Anf expression between control and αMHC-Cre^tg/+; Hand1^A126FS/+ mutants. We then examined P60 adult hearts for phenotype. Bifurcated αMHC-Cre^tg/+; Hand1^A126FS/+ hearts appear slightly hypertrophied and lack a well define apex (see Supplementary material online, Figure S2D and E). Mutant adults do not exhibit HLHS.

Next, we looked at Mef2CAHF-Cre; Hand1^A126FS/+ mutants. Mef2CAHF-Cre²⁷ activates Hand1^A126FS within the SHF-derived myocardial cuff. Mef2CAHF-Cre; Hand1^A126FS/+ mutants survive to birth at normal frequency (see Supplementary material online, Table S3) exhibiting a normal LV chamber; however, RV size appears reduced at E14.5 and at p60 (see Supplementary material online, Figure S2C and F). We assayed expression of Wnt11, Tbx5, Dkk3, and Cited1 by wholemount ISH in both αMHC- and Mef2CAHF-Cre E10.5 mutants (see Supplementary material online, Figure S3). αMHC-Cre mutants exhibit a more normal OFT (Wnt11 staining white bar) and expression of Tbx5 maintains its LV specific expression (see Supplementary material online, Figure S3A, B, D, E). Expression of Dkk3 appears expanded similar to what is observed in Nkx2.5^Cre; Hand1^A126FS/+ mutants (white arrowhead in Supplementary material online, Figure S3G and H). No changes in Cited1 patterning are observed in αMHC-cre; Hand1^A126FS/+ embryos (see Supplementary material online, Figure S3J and K).

Mef2CAHF-Cre; Hand1^A126FS/+ mutants exhibit expanded Wnt11 expression similar to Nkx2.5^Cre; Hand1^A126FS/+ (see Supplementary material online, Figure S3C). Expression of Tbx5, Dkk3, and Cited1 are unaffected, consistent with Mef2CAHF-Cre expression (see Supplementary material online, Figure S3F, I, L).

4. Discussion

Left-sided cardiac defects have poor clinical outcomes.^1–3 In contrast to the increasing mechanistic understanding of congenital defects affecting the OFT,^35,55–57 little is understood of the underlying mechanisms causative of HLHS. Hand1 expression and lineage is largely restricted to the LV.^10,25 In light of several studies implicating HAND1 in the genesis of Tetralogy of Fallot,¹⁹ VSDs,²⁰ and HLHS^9,21 in humans along with published evidence demonstrating a high frequency of HAND1 somatic mutations observed from fixed tissues that are not encountered from fresh tissue²²—the follow up on these controversial findings in an in vivo system was imperative. The p.Ala126Profs13X Hand1 mutation was identified in hypoplastic hearts from the Leipzig University fixed tissue collection²³ and reported to act as a dominant negative in yeast.²¹ We confirm this result in mammalian expression analysis (Figure 1). Given we observe a loss of DNA binding, we hypothesize that the mutant Hand1 protein dimerizes with both wild-type Hand1, Hand2, and E-proteins when they are coexpressed. Twist-family bHLH factors are well-established to be promiscuous in dimer partner choice and the disruption of phosphoregulation of evolutionarily conserved bHLH residues regulating dimerization is directly associated with human disease such as in Saethre Chotzen Syndrome.^{14,15,31,58,59} The Hand1 p.Ala126Profs13X localizes to the nucleus and maintains the integrity of the first helix allowing for protein dimerization. All data are consistent with this model.

In this study, we directly tested the HAND1^A126fs mutation in mice as causative of HLHS. Our results suggest that when this mutation is activated within cardiomyocytes at E7.5 via Nkx2.5^Cre is not causative of HLHS but these mutants are embryonic lethal (Figures 3–7) presenting OFT elongation, thin compact zone, hypotrabeculation, and VSDs. Although Nkx2.5^Cre is expressed in the endocardial lineage,²⁸Hand1 is not²⁵ thus the observed phenotypes are driven by Hand1^A126FS cardiomyocyte expression. As the Nkx2.5^Cre is a knock-in allele genetic interactions could be at play. To account for this, we activated the HAND1^A126fs allele using αMHC-Cre²⁶ and Mef2CAHF-Cre²⁷ mice. The αMHC promoter is expressed at E9.5 and is then down regulated until after birth. Activation of the Hand1^A126fs allele results in a less-severe phenotype where some mice survive (see Supplementary material online, Table S2) displaying normal looking hearts at E14.5 that become enlarged by P60 (see Supplementary material online, Figure S2). Given the later activation of Hand1^A126fs by αMHC-Cre, we cannot distinguish the difference in phenotypes as resulting from genetic interactions with Nkx2.5 or differences in temporal activation or both. Nevertheless, neither Cre driver results in HLHS when crossed to the HAND1^SFA126fs allele. Mef2CAHF-Cre generated mutants are born and exhibit elongated OFTs similar to Nkx2.5^Cre generated mutants (see Supplementary material online, Figure S3). Wholemount expression shows consistent Wnt11 elongation with no effect on LV gene expression. From this data, we conclude that p.Ala126Profs13X Hand1, even if somatically acquired, is not causative of HLHS. As we considered the high frequency reported in the Leipzig samples (0.77)^9,21 for the HAND1^A126fs mutation juxtaposed to the finding that no HAND1 mutations were observed from interrogation of non-fixed HLHS patient samples,²² it suggests to us that DNA modifications via tissue fixation is the source of this discrepancy. Given this, we are not confident that HAND1^A126fs mutation phenotypes are biologically relevant to human CHDs, although they do reveal some insight into Hand1 role in cardiogenesis. Data from HLHS patients induced pluripotent stem cells show reduced levels of Hand1 and Hand2 in differentiating cardiomyocytes, suggesting that although protein mutations may not be commonly encountered the alteration of transcriptional regulation maybe more frequent explaining the association between Hand1 and CDHs.⁶⁰ Indeed, changes in HAND1 DNA methylation status is reported in Tetralogy of Fallot patients.⁶¹ Additionally, two HAND1 point mutations were reported (p.G73S and p.K152N) in a study from screens of peripheral blood.⁶² Such simple point mutations are likely far less deleterious than a truncated protein.

The cardiac developmental abnormalities observed in Nkx2.5^Cre; Hand1^A126FS/+ mouse embryos do present a unique phenotype compared with previously tested Hand1 loss-of and gain-of function mutations. Knockout of Hand1^10,11 and hypomorphic Hand1 expression¹³ are embryonic lethal at E9.5 resulting from extraembryonic insufficiencies. Conditional cardiac deletion of Hand1,¹² results in neonatal lethality due to a transient decrease in LV size accompanied with endocardial cushion defects. Additionally, in an Mlc2v-Hand1 knock-in model,⁶³ hearts were growth expanded and lacked an IVS. Collectively, these models suggest that cardiomyocyte growth/cell death is affected but the molecular mechanisms affected are still elusive; however, we can conclude that the expression of either wild-type⁶³ or a truncated Hand1 protein (this study) can be more deleterious than the conditional deletion of Hand1.¹²

In considering the defects observed in Nkx2.5^Cre; Hand1^A126FS/+ mice, VSDs are a common CHD encountered in human births whereas in mice these defects are often associated with embryonic lethality. VSDs have a wide spectrum of aetiologies and encountering them in mouse mutants is common. Cardiomyocyte cell death is also observed at an abnormal level (Figure 4). This phenotype is not normally encountered in models of CHDs but is common in adult heart disease.^64,65 Interestingly, elongated OFT has been reported in Hand1 gain-of-function studies⁶⁶ supporting a possible active role for the dominant function of Hand1^A126FS. Additionally, hypotrabeculation and thin walls are encountered in Nkx2.5^Cre; Hand1^A126FS/+ hearts. These phenotypes are accompanied by altered trabecular and compact zone gene expression where some genes are down regulated such as Hey2 (Figure 6) and Bmp10 at E14.5 (Figure 7); however, most gene expression changes identified show an increased or expanded expression (CxCl12, Dkk3, Tbx5, Tbx20; Figures 6 and 7). This finding could suggest that Hand1 has repressive transcriptional functions although previous findings do not show such changes in loss-of-function analysis.^10–13 Alternatively, expanded expression could result from feedback regulation. Tbx5 for example is restricted to the LV and ChIP-Seq analysis⁶⁷ suggests that Tbx5 may be a Hand1 regulator. Interestingly, Tbx5 expression is expanded within the RV where Hand1 is not expressed, suggesting a non-cell autonomous mechanism. Tbx5 expansion is only observed within Nkx2.5^Cre; Hand1^A126FS/+ hearts and not when either the αMHC-Cre or Mef2CAHF-Cre is employed (Figure 5; see Supplementary material online, Figure S3). The Nkx2.5^Cre haploinsufficiency combined with the earlier activation of Hand1^A126FS/+ expression than αMHC-Cre (E7.5 vs. E9.5) could both influence this phenotype. Mef2CAHF-Cre recombines Hand1 at E7.5 but not within the LV, collectively limiting our ability to distinguish between a temporal and/or genetic mechanism.

Together, these data suggest that although the idea of acquiring a somatic mutation in Hand1 as a mechanism to account for HLHS or any CHD is attractive, the likelihood of this occurring in HAND1 is low. We speculate that it is possible and perhaps likely that mutations within Hand1 enhancer sequences that alter its spatiotemporal expression would be an inheritable HAND1 mutation and identification of human small nucleotide polymorphisms in HAND1 further supports this idea.⁶⁸

Supplementary material

Supplementary material is available at Cardiovascular Research online.