Mutation of LRP1 in cardiac neural crest cells causes congenital heart defects by perturbing outflow lengthening

Abstract

The recent recovery of mutations in vesicular trafficking genes causing congenital heart disease (CHD) revealed an unexpected role for the endocytic pathway. We now show that mice with a C4232R missense mutation in Low density lipoprotein receptor related protein 1 (LRP1) exhibit atrioventricular septal defects with double outlet right ventricle. Lrp1^m/m mice exhibit shortened outflow tracts (OFT) and dysmorphic hypocellular cushions with reduced proliferation and increased apoptosis. Lrp1^m/m embryonic fibroblasts show decreased cell motility and focal adhesion turnover associated with retention of mutant LRP1 in endoplasmic reticulum and reduced LRP1 expression. Conditional deletion of Lrp1 in cardiac neural crest cells (CNC) replicates the full CHD phenotype. Cushion explants showed defective cell migration, with gene expression analysis indicating perturbation of Wnt and other signaling pathways. Thus, LRP1 function in CNCs is required for normal OFT development with other cell lineages along the CNC migratory path playing a supporting role.

Fig. 1. Homozygous Lrp1^m/m mutants exhibit a spectrum of cardiac defects.

Compared with E14.5 normal control (a–g), Lrp1^m/m mutant embryo showed side-by-side great arteries (h, i, l, n). Echocardiogram using Vevo2100 of control embryo indicated intact ventricular septum with no evidence of interventricular shunt (b), which was confirmed by histopathology with ECM (e). Representative fetal ultrasound of the Lrp1^m/m mutant (i) showed a VSD and side-by-side great vessels with aorta positioned to the right of the pulmonary artery. This was also confirmed by ECM (l), which showed double outflow right ventricle with both the aorta (AO) and pulmonary artery (LPA and RPA) arise from RV and a large VSD. Compared to the wild-type control (c, d), the homozygous Lrp1^m/m mutant exhibited ASD primum (ASDI) (j), a part of AVSD, and ventricular part defect of AVSD (k). While there is aorta to mitral valve continuity in controls (f), the Lrp1^m/m mutant showed loss of the continuity between aortic valve and mitral valve, with a gap between the mitral and aortic valves (m, yellow arrowhead). Instead of aorta being to the right and posterior to the pulmonary artery (g), the Lrp1^m/m mutant exhibited a side-by-side and rightward aorta (AO) to pulmonary artery (PA) relationship (n). Scale bars: 0.5 mm.

Fig. 2. LRP1 expression in different cell linages that are important for cardiac development in the developing heart.

Panel a is schematic representation of embryonic heart showing AVC and OFT (conal part and truncal part). Immunostaining in E9.5 and E10.5 embryos demonstrated LRP1 is expressed in mesenchymal cells in the endocardial cushion, with no expression seen in the CD31 positive endocardium (b–d) at E10.5; epicardium (e white arrow) at E10.5 pharyngeal endoderm and pharyngeal arch (b, f) at E10.5, neural crest/mesenchymal cells (g–i) at E9.5; outflow tract (i) at E9.5, outflow tract (j) at E10.5 and truncal artery (k) at E10.5. LRP1 and Islet-1 (ISL1) are expressed in (f) pharyngeal endoderm and pharyngeal mesoderm. LRP1 and AP2α both expressed in the neural crest/mesenchymal cells (g–i) and outflow tract (j–l). i The magnified view from the white box of figure (h). Scale bars: 50 µm.

Fig. 3. Conditional Cre deletion of floxed Lrp1 allele.

Lrp1 conditional knockout in different cell lineages during cardiac development. a Pie charts illustrate the percentage of each cardiac phenotype. Representative echocardiograms and ECM pictures in each conditional knockout are demonstrated in (a). Targeting Lrp1 deletion in cardiac neural crest cells using Wnt1-Cre recapitulated the cardiac phenotype of Lrp1^m/m mutants with AVSD and DORV in 91% of Wnt1^+/Cre: Lrp1^f/f mutants, yellow arrow indicated pericardial hemorrhage. Deletion of Lrp1 in the AHF derivatives using the Mef2c-AHF-Cre yielded septation defects. Nkx2–5-Cre mediated deletion of Lrp1 yielded mostly DORV with or without AVSD. Totally, 43% of mutants generated with the Tie2-Cre drivers had septation defects. Combined deletion of Lrp1 mediated by both Tie2-Cre and Nkx2–5-Cre recapitulated the high penetrance of the DORV/AVSD phenotype observed in the Lrp1^m/m mutant and increase the penetrance of AVSD. Deletion of Lrp1 in the mesenchymal cells of AV cushion with Twist2-Cre drivers showed 82% of mutants with membranous VSD and 18% mutants with DORV. There were no cardiac abnormalities with the Nfatc1-Cre drivers, consistent with absence of Lrp1 expression in the endocardium of atrioventricular cushion. b Schematic representation of LRP1 expression (yellow) in developing heart around E10.5–E11.5. Lrp1 expresses in pharyngeal endoderm (PE) and pharyngeal mesoderm (PM), mesenchymal cells of outflow tract cushion, atrioventricular cushion, neural crest cell migration pathway as well as epicardial cells. Ablation of Lrp1 in neural crest cells using Wnt1-Cre (Wnt1^+/Cre:Lrp1^f/f) is illustrated in dotted green arrows. Ablation of Lrp1 using Mef2c-AHF-Cre (Mef2c-AHF^+/cre: Lrp1^f/f) is illustrated in orange. Ablation of LRP1 expression using Nkx2–5-Cre (Nkx2–5^+/Cre:Lrp1^f/f) is illustrated in pink, specifically in PM, PE, AHF, OFT cushion and AV canal cushion. Ablation of Lrp1 expression using Tie2-Cre (Tie2^+/Cre: Lrp1^f/f) is illustrated in navy-blue hatch (mesenchymal cells of atrioventricular cushion) and navy-blue dots (endothelium of atrioventricular canal and endocardium of ventricle). The ablation of Lrp1 in double knockout of Nkx2–5-Cre and Tie2-Cre (Nkx2–5^+/Cre Tie2^+/Cre: Lrp1^f/f) is illustrated in purple and navy blue. Ablation of Lrp1 in Twist2-Cre is expressed in azure (royal blue). Ablation of Lrp1 expression in Nfatc1-Cre is expressed in red. Scale bars: 0.5 mm.

Fig. 4. Lrp1^m/m mutant hearts exhibit decreased OFT length, decreased AVC and OFT cushion volume, decreased proliferation and increased apoptosis at E10.5.

ECM in the sagittal plane of an E10.5 wildtype (a) and Lrp1^m/m mutant (d) embryos are shown, indicating reduction in length of mutant OFT versus control (yellow arrow in (a, d)). Insert panels in a and d demonstrated the cardiac diameter (yellow line) in the corresponding embryo. g Quantitative measurement using histopathology images showed a significant decrease in the length of the OFT with cardiac diameter normalization in the homozygous Lrp1^m/m mutant hearts as compared to the controls (combined wildtype and Lrp1^+/m hearts). Three-dimensional (3D) reconstructions of AVC (b, e) and OFT (c, f) were generated from ECM of Lrp1^m/m mutant hearts and controls at E10.5. h Quantitative measurement using histopathology images and 3D slicer to process the images showed a significant decrease in the volume (with normalization of cardiac diameter) of the AVC and OFT in the homozygous Lrp1^m/m mutant hearts as compared with the controls (combined wildtype and Lrp1^+/m hearts). i Quantitative results of cell proliferation using the ratio of positive pH3 cells/total cushion cells in AVC and OFT cushion. k Quantitative results of the ratio of positive TUNEL cells/total cushion cells. Statistical comparison was performed using unpaired two-way Student′s t test. Error bars show standard deviation. Scale bars: 0.5 mm in (a, d); 0.1 mm in (b, c, e, f).

Fig. 5. Abnormal atrioventricular and outflow tract cushion development was observed in Lrp1^m/m mutants.

Immunostaining of NFATC-1 (green) and Periostin (red) in wildtype (a, b) and Lrp1^m/m mutant (c, d) embryos at E12.5. b, d Magnified views of the white box from (a, c). Compared with the well-formed two atrioventricular valves in the wildtype control, Lrp1^m/m mutant demonstrated primitive undivided endocardial cushion morphology. Representative images from X-gal and Eosin stained tissue sections from E10.5 control (Lrp1^m/+) (e–g) and Lrp1^m/m mutant (h–j). f, i The high magnified view of the black box e and h (OFT). g, j The high magnified view of the yellow box from (e, h) (AV cushion). X-gal staining e–j demonstrated diminished expression of LacZ expression in the Lrp1^m/m AVC (e, f, h, i) and OFT cushion (e, g, h, j). k Quantitative analysis of the percentage of positive BAT/LacZ cells in the muscle and mesenchyme portion of OFT and AVC. Lrp1^m/m mutant had hypocellular AVC and OFT cushions (h–j). l–o Immunostaining of AP2α (green) and LRP1 (red) in Lrp1^m/m mutant (n, o) and control (l, m). Lrp1^m/m mutant had decreased expression of AP2α and LRP1 in the outflow tract at E10.5–E11.5 (l–o). p Quantitative analysis of the fluorescence intensity of AP2α and LRP1 demonstrated decreased expression of both AP2α and LRP1 in Lrp1^m/m mutant OFT. (q, r) LacZ staining of OFT in Wnt1^+/cre: Lrp1^{f/f/rosa/rosa} and control (Wnt1^+/cre: Lrp1^{f/+/rosa/rosa}), which demonstrated decreased LacZ expression in the Wnt1^+/cre: Lrp1^{f/f/rosa/rosa} mutant. Scale bars: 200 μm. Statistical comparison was performed using unpaired two-way Student's t test. Error bars show standard deviation.

Fig. 6. Quantitative differential expression of EMT-related genes by real-time PCR.

Quantitative comparison of EMT-related gene expression in Lrp1^m/m vs. WT in the endocardial cushion at E10.5, prepared using Morpheus (Broad Institute) (a). a Genes which were repeatably over- (red) or under-expressed (blue) are shown. After discarding measurements with unacceptably low copy number (gray; raw Ct > 40), replicate ddCt values were filtered with an outlier test (ROUT, Q = 10%) and samples with three replicates were tested against a H0 that ddCt = 0 using a one-sample two-tailed t test. Genes that passed (p < 0.05) are shown, along with other strongly responsive genes (p ≥ 0.05, but mean > variance). Red indicates genes upregulated in Lrp1^m/m and blue indicates downregulated. The functional relationships between these genes is shown in (b); bolded names or outlines indicate genes with significant overexpression or underexpression and lighter names or outlines indicates strongly responsive but not significant. Color coding is the same as in (a); black outlines or names represent key genes that are not overexpressed or underexpressed in this analysis. The major cellular pathways represented by this gene set are shown in (c), as determined by Ingenuity Pathway Analysis.

Fig. 7. Lrp1 mutation causes migration defects.

a, a′ Representative image of wound scratch assay from (a) a wild-type control and (a′) a mutant MEF, 18 h after scratch. b Quantitation of wound gap closure showed Lrp1^m/m mutant MEFs have decreased cell motility compared to control with decreased filling of the scratch “gap” compared with controls. % indicates the percentage of closing the gap, 100% means close the gap completely. c Wild-type and (c′) Lrp1^m/m mutant fibroblast cells were transfected with vinculin-GFP and vinculin turnover was measured using live-cell imaging. d Quantitative measurement revealed slower focal adhesion turn over in MEF from Lrp1^m/m mutant, as measured by the half-life of the overlap between segmented focal adhesions at time 0 and time n. e, e′ Representative imagine of AVC explant migration assay from a wild-type control (e) and Lrp1^m/m mutant (e′). Black line indicated the migration distance at 48 h. f Quantitative analysis of the distance of migration of endocardial cushion explants from E10.5 control and Lrp1^m/m embryos showed near complete loss of cell migration from the Lrp1^m/m mutant explants Statistical comparison was performed using unpaired two-way Student′s t test. Error bars show standard deviation.

Fig. 8. LRP1 C4232R mutant protein affects the expression of the 85 kD (LRP1-β) domain and has increased retention in the endoplasmic reticulum.

Schematic diagram of LRP1 protein. a LRP1 consists of 4545 amino acids. The mutant line 1554 (MGI 96828) harbors a missense (C4232R) mutation in the extracellular region encoding the EGF repeat domain of LRP1. The cysteine residue mutated at position 4232 is conserved among species. b Immunoblotting analysis showed strongly reduced expression of the 85 kD LRP1 protein in liver lysates from the Lrp1^m/m homozygous mutant, as compared with Lrp1^+/m and wild-type embryos. c Quantitative measurement of colocalization between LRP1 and the markers of cellular compartments using Pearson′s coefficient demonstrates (c, d) Lrp1^m/m has increased localization to the ER (KDEL) and decreased localization in early endosomes (EEA1). Error bars showed mean standard deviation. d Immunostaining for LRP1 and markers for the Golgi apparatus (GM130), endoplasmic reticulum (ER) marker (KDEL), clathrin marker (CHC), and the early endosome marker (EEA1). Scale bars: 10 µm in (d); 2 µm in magnified box in (d).