Identification of HOXA1 variants in index cases with BAV.

a Schematic representation of the human HOXA1 gene and corresponding predicted protein. The HOXA1 protein contains the histidine repeat motif (green; amino acids 65–74), the hexapeptide motif involved in the interaction with PBX proteins (blue; amino acids 204–209), encoded in exon 1, and the DNA-binding homeodomain (HD; yellow; amino acids 229-288), encoded in exon 2. Alignment of the HOXA1 histidine repeat sequences from the human, mouse, rat, cat, cow, chicken and shark species is represented. The NCBI GenBank accession numbers that were utilized for the alignment are as follows: Human-P49639, Mouse-P09022, Rat-O08656, Cat-M3W8N7, Cow-E1B918, Chicken-R4GHI8, and Shark-C7B9D2. b Sanger sequences showing the insertion or deletion variants observed in exon 1 of HOXA1. The polymorphism c.218G>A (p.Arg73His) is shown with asterisks (black, c.218G; green, c.218G>A). We identified different variants in this region (c.[213_215dupCCA;218G>A] named +1His; c.[213_215dupCCA;218G] named +1His^Arg; c.[213_215delCCA;218G>A] named −1His and c.[207_215delCCACCACCA;218 G] named −3His^Arg) with low frequency in gnomAD (see Table 3).

HOXA1 variants induce proteasome-dependent protein degradation and reduce transcriptional activity of the protein.

a–d HEK293T cells were transfected with expression vectors coding for FLAG-HOXA1 (WT^A, + 1His, +1His^Arg, −1His, −1His^Arg and −3His^Arg). To determine the half-life of HOXA1, cells were treated with cycloheximide (CHX; 200 μg/ml) for the indicated time. a Cell lysates were collected every 6 h and then subjected to immunoblot analysis with antibodies against FLAG and β-ACTIN. b Intensity of the HOXA1 bands observed upon cycloheximide treatment was quantified and the relative HOXA1/β-ACTIN ratios were plotted (n = 6; n indicates number of biologically independent experiments (cell transfected independently in different weeks). Error bars indicate standard deviation (SD). c Involvement of the proteasome in the HOXA1 degradation was assayed by treating cells with proteasome inhibitor (MG132; 7 μM) for 20 h prior inhibition of protein translation with CHX (n = 3). Cell lysates were collected every 6 or 12 h and then subjected to immunoblot analysis with antibodies against FLAG and β-ACTIN. d The intensity of the HOXA1 bands observed upon cycloheximide treatment was quantified and the relative HOXA1/β-ACTIN ratios were plotted (n = 6). Error bars indicate standard deviation (SD). e Luciferase assay showing transcriptional activity of wild-type (WT^A) and mutated hHOXA1 (+1His, +1His^Arg, −1His, −1His^Arg, and −3His^Arg). HEK293T cells were transfected with the TSEII-luciferase reporter construct, alone (TSEII-Luc) or together with expression plasmids for PREP1, PBX1a, and FLAG-HOXA1 (WT^A, + 1His, +1His^Arg, −1His, −1His^Arg, and −3His^Arg). Same quantity of DNA was transfected for each plasmid (n = 6). **p = 0.035. f, In order to characterize the dominant negative effect of mutated human HOXA1 proteins, cells were transfected with the TSEII-luciferase reporter together with both wild-type and HOXA1 variant coding vectors (n = 6). ***p = 0.00000012. Boxes and whiskers (min to max) show the values of lower and upper quartile. Statistical values were obtained using Pairwise t test with Holm P value adjustment method test. Statistical test used was two-sided. ‘n’ represents number of biologically independent experiments (cell transfected independently in different weeks) in duplicate. Source data are provided as a Source data file.

Expression of HOXA1 variants in vivo disrupts aortic valve development.

Two-photon images of aortic valves in 7 dpf zebrafish larvae labeled with BODIPY. a, d Representative image of aortic valve leaflets in a wild-type, non-injected larvae (ABwt, NI), valves are outlined with a dashed yellow line (a: n = 12; d: n = 5). bhoxa1a morphant (n = 10). choxa1a morphant co-injected with wild-type human HOXA1 mRNA (n = 11). ehoxb1b morphant (n = 17). fhoxb1b morphant co-injected with wild-type human HOXA1 mRNA (n = 11). g Graph showing quantification of aortic valve defects observed in 7 dpf larvae after knockdown of hoxa1a. Approximatively 80% of hoxa1a morphants exhibited aortic valve defects (**p = 0.0015; *p = 0.03). HOXA1 human mRNA rescued the hoxa1a morpholino aortic valve phenotype. h Graph showing quantification of aortic valve defects observed in 7 dpf larvae after knockdown of hoxb1b (**p = 0.0048). human HOXA1 mRNA does not rescue the hoxb1b morpholino aortic valve phenotype. i, j Representative images of aortic valve leaflets after injection of WT human HOXA1 mRNA (WT) at 5 pg (n = 10) or 25 pg (n = 12), respectively. k Injection of the −3His^Arg variant of human HOXA1 at 5 pg (n = 10). l Injection of the −1His variant of human HOXA1 at 5 pg (n = 12). m Injection of the +1His variant of human HOXA1 at 25 pg (n = 14). n Graph showing quantification of aortic valve defects observed in 7 dpf larvae after injection of the variants of human HOXA1 (***p = 0.001; **p = 0.0037; *p = 0.021). Statistical values were obtained using Fisher exact test. Scale bars: 20 μm. Source data are provided as a Source data file.

Hoxa1 knock-out and Hoxa1^−1His knock-in mice display bicuspid aortic valve.

a, b Hematoxylin and eosin (H&E) staining images showing representative transversal section of wild-type (WT) and Hoxa1^−/− outflow tract at E13.5. Representative pictures from 3 independent hearts of each genotype. c, d Three-dimensional (3D) reconstructions of histological images at E13.5. A slight difference between WT (c) and Hoxa1^−/− (d) aortic valve (AoV) structure is observed. Indeed, the non-coronary leaflet (NC) seems smaller in Hoxa1−/− when compared to wild-type embryos. e, f Cross-sectional H&E images through the aortic valve of WT (e) and Hoxa1^−/− (f) embryos at E18.5. Representative pictures from 3 independent hearts of each genotype. Normal valve with three leaflets (asterisks) is observed in WT embryos (f), whereas bicuspid aortic valve is detected in the mutant (f). Asterisks indicate the aortic sinus. g, h 3D reconstruction of histological images at E18.5 showing three aortic valve leaflets in the WT embryo (g), whereas a persistent small non-coronary leaflet is seen in Hoxa1^−/− embryos (h). i, j Cross-sectional H&E images through the aortic valve of WT (i) and Hoxa1^{−1His/−1His} (j) embryos at E18.5. Representative pictures from 4 independent hearts of each genotype. Normal valve with three leaflets (asterisks) is observed in WT embryos (i), whereas bicuspid aortic valve is detected in the knock-in embryos (f). Left coronary (L; pink), right coronary (R; yellow), non-coronary (NC; blue) leaflets; PV, pulmonary valve. Scale bars: 100 μm (a, b); 200 μm (e, f, I, j).

Analysis of cell lineage contribution in Hoxa1-null mice.

a, b Transversal sections of E13.5 hearts immunostained with an endothelial marker (anti-Pecam antibody, green) and DAPI staining for nuclei (blue). Note the smaller non-coronary (NC) leaflet in Hoxa1^−/− (b) compared with WT (a) embryos. c Total nuclei were counted in the three forming aortic valve leaflets from Hoxa1^−/− (n = 20) and WT (n = 18) embryos at E13.5 spanning a 180 μm depth (***p = 0.0006; **p = 0.0009; *p = 0.01). d–l Fate-mapping of the neural crest-, endothelial- and second heart field-lineages in the aortic valves of WT (d, g, j), and Hoxa1-null (e, h, k) embryos at E13.5. Tomato-reporter is visualized in red and nuclei (DAPI) appear in blue. d, e Immunofluorescence staining using Wnt1-Cre;Rosa^tdTomato/+ reporter mice shows a decrease in the number of neural crest derivatives in absence of Hoxa1. f Quantification of neural crest-derived cells demonstrates significant decrease in Hoxa1^−/− (n = 5) compared to WT (n = 6) littermate embryos (****p = 0.008;***p = 0.004; **p = 0.03). Area of interest is indicated by a dotted line. g, h Immunofluorescence staining using Tie2-Cre;Rosa^tdTomato/+ reporter mice shows a decreased contribution of endothelial lineage to the non-coronary leaflet of Hoxa1^−/− (h) compared to WT (g) littermates. i Quantification confirms the decreased number of endothelial-derived cells in the non-coronary leaflet in Hoxa1^−/− (n = 6) compared to WT (n = 6) littermates (***p = 0.002). Area of interest is indicated by a dotted line. j, k Immunofluorescence staining using Tnnt2-Cre;Rosa^tdTomato/+ reporter mice demonstrates a decreased contribution of the second heart field lineage to the non-coronary leaflet in Hoxa1^−/− (k) compared to WT (j) littermates. l Quantification confirms the reduced number of second heart cells-derived cells in the non-coronary leaflet in Hoxa1^−/− (n = 5) compared to WT (n = 4) littermates (**p = 0.01). Area of interest is indicated by a dotted line. Data are shown as mean ± SEM. Boxes and whiskers (min to max) show the values lower than the 25 percentile and greater than the 75 percentile. Statistical values were obtained using the Mann-Whitney test. LC: left coronary leaflet; NC: non-coronary leaflet; RC: right coronary leaflet. Scale bars: 100 μm (a, b, d, e, g, h, j, k). Source data are provided as a Source data file.

Transcriptomic analysis identifies Hoxa1 targets involved in migration of neural crest cells.

a Scheme showing which region was used to perform RNA-seq analysis on Hoxa1^−/− vs WT embryos. b Volcano plot of fold changes and corresponding FDR (False Discovery Rate) value for each RNA after comparison of Hoxa1^−/− and WT embryos. The red and green dots represent RNAs with fold change >2.0 and FDR < 0.05. The black dots represent RNAs with fold change <2.0 and FDR > 0.05. The statistical test used for data analysis is package R EnhancedVolcano. c Gene ontology (GO) analysis performed with ClusterProfiler system showing enrichment of up-regulated (red) and down-regulated (green) genes in the mutant with ranked by −log₁₀ (p-value). The statistical test used for data analysis is package R clusterProfiler. d Boxplot of FPKM (Fragments Per Kilo base of transcript per Million mapped fragments) expression value for nine selected DE genes in RNA-seq. Boxes and whiskers (min to max) show the values of lower and upper quartile. The y-axis represents the FPKM expression level, and the x-axis the genotype. Data are shown as mean ± SEM. Data are representative of three biologically independent experiments.

Neural crest cell defects in Hoxa1 null mice.

a, b RNA-FISH showing the expression of Sox10 (green) in wild-type (WT) (a) and Hoxa1^−/− (b) E9.5 embryos. Lateral view of Sox10 staining revealed a decrease of expression in the migrating neural crest cells (arrowheads). c, d, RNA-FISH showing Robo2 expression (green) in wild-type (c) and Hoxa1^−/− (d) E9.5 embryos. Expression of Robo2 is higher in Hoxa1^−/− than wild-type littermates. e, f Cross-sections of wild-type (e) and Hoxa1^−/− embryos (f) immunostained with anti-Tfap2-alpha antibody (red). Note the reduction of Tfpa2-alpha positive cells in the mutant (arrowhead). g, hHoxa1-lineage visualized by X-gal staining of Hoxa1-enhIII-cre;R26R-lacZ embryos. g, Lateral view showing β-galactosidase activity in migrating neural crest cells (NCC; arrowhead) in transgenic embryo at E9.5. h, At E12.5, β-galactosidase activity is observed in the outflow tract (OFT). i, j, Whole-mount X-gal staining of Wnt1-cre;R26R-lacZ WT (i) and Hoxa1^−/− (j) E9.5 embryos. Migration of neural crest cells (arrowhead) is disrupted in Hoxa1^−/− embryos (j) compared to WT (i) littermate embryos. k, l Whole-mount X-gal staining of −31/−23.5 Krox20/lacZ WT (k) and Hoxa1^−/− (l) E9.5 embryos. Migration of β-galactosidase-positive cells (arrowhead) is disrupted Hoxa1^−/− (l) compared to WT (k) littermate embryos. ba: branchial arch; ht, heart tube; la, left atrium; lv, left ventricle; NCC, neural crest cells; ov, otic vesicle; oft, outflow tract; ra, right atrium; rv, right ventricle. Data are representative of 3 independent embryos of each genotype. Scale bars: 100 μm (c–g), 200 μm (a, b, h, l).