Mutations in LRP2, which encodes the multiligand receptor megalin, cause Donnai-Barrow and facio-oculo-acoustico-renal syndromes

Abstract

Donnai-Barrow syndrome is associated with agenesis of the corpus callosum, congenital diaphragmatic hernia, facial dysmorphology, ocular anomalies, sensorineural hearing loss and developmental delay. By studying multiplex families, we mapped this disorder to chromosome 2q23.3–31.1 and identified LRP2 mutations in six families with Donnai-Barrow syndrome and one family with facio-oculo-acoustico-renal syndrome. LRP2 encodes megalin, a multiligand uptake receptor that regulates levels of diverse circulating compounds. This work implicates a pathway with potential pharmacological therapeutic targets.

Congenital diaphragmatic hernia (CDH), a defect of diaphragm formation, has a persistently high morbidity and mortality often related to pulmonary hypoplasia. To identify genetic pathways of CDH, we study monogenic disorders in which CDH is a component, such as Donnai-Barrow syndrome (DBS; OMIM 222448). Individuals with this rare autosomal recessive disorder have major malformations, including those described above^1,2. DBS has clinical similarity to facio-oculo-acoustico-renal (FOAR) syndrome (OMIM 227920), although the latter is typically reported as having proteinuria but lacking agenesis of the corpus callosum (ACC) and CDH^3,4.

To detect regions of identity by descent (IBD), we applied Affymetrix 10K SNP arrays to four individuals with DBS in a large consanguineous family (kindred 1) (Supplementary Fig. 1 and Supplementary Methods online). The largest region of IBD (~21 Mb) was on chromosome 2q23.3–q31.1, flanked by SNPs rs1020088 and rs1362496 (data not shown). Using microsatellite markers in multiplex kindred 1 and three additional multiplex kindreds (2–4), we refined this region to ~18 Mb between D2S2299 and D2S2284 and generated a maximum two-point LOD score of 4.31 and multipoint LOD score of 6.243 (Supplementary Fig. 2 online). The interval contains 51 known genes, including the 79-exon gene LRP2 (low-density lipoprotein receptor–related protein 2), which encodes megalin.

We analyzed LRP2 coding sequences and intron-exon boundaries in affected individuals from a total of seven kindreds; kindreds 1–6 were clinically diagnosed with DBS, and kindred 7 was clinically diagnosed with features of both FOAR and DBS (Supplementary Tables 1 and 2 and Supplementary Note online). All affected individuals showed missense, nonsense, splice junction or frameshift mutations in evolutionarily conserved residues of both LRP2 alleles (Fig. 1 and Supplementary Table 3 online). The similar phenotype observed among affected individuals, including those with homozygous frameshifts, suggests that most mutations are functionally null. These mutations were absent in NCBI and Ensembl SNP databases; additionally, we did not detect the kindred 1–3 mutations by sequencing DNA from 96 controls of similar ancestry. Thus, DBS and FOAR syndromes should be regarded as the same disorder, DBS/FOAR, and may represent the first human phenotype associated with mutations in LRP2.

Figure 1.

Schematic representation of LRP2 mutations and corresponding megalin domains. LRP2 sequencing uncovered missense, nonsense, frameshift or splice junction mutations in individuals with DBS/FOAR from seven kindreds. Mutations are indicated adjacent to the megalin extracellular domains that they affect; mutation ‘hotspots’ or genotype-phenotype correlations are not apparent. Mat, maternally inherited mutation; pat, paternally inherited mutation (Supplementary Table 3).

Megalin, an endocytic transmembrane receptor of 4,655 amino acid residues, containing low-density lipoprotein receptor class A and B motifs, epidermal growth factor-like repeats and an intracellular protein-binding Asn-Pro-X-Tyr sequence, is located primarily on the apical surface of absorptive epithelia⁵. It is critical for reuptake of numerous ligands, including lipoproteins, sterols, vitamin-binding proteins and hormones^6,7. Megalin also has a role in cell signaling by interacting with sonic hedgehog⁸. Studies in mice demonstrate megalin expression in specialized epithelium of numerous organs, including brain, kidney, and lung^5,9. Its expression has not been examined in the diaphragm.

Megalin knockout mice have high perinatal mortality owing to respiratory insufficiency, with their lungs described as ‘emphysematous’ and ‘atelectatic’¹⁰. A small number of surviving megalin-null mice, as well as those with a conditional knockout affecting the kidneys, demonstrate proteinuria with elevated levels of megalin ligands, including retinol-binding (RBP) and vitamin D–binding (DBP) proteins¹¹. Comparably, urine samples from eight individuals with DBS/FOAR show proteinuria, including increased spillage of RBP and DBP (Supplementary Fig. 3 online). This characteristic proteinuria in all affected individuals tested provides strong evidence that the LRP2 mutations have a negative impact on megalin function in the renal proximal tubule and that elevated urinary RBP and DBP can serve as valuable diagnostic and surrogate markers for DBS/FOAR.

The megalin-null mice also demonstrated forebrain and olfactory bulb anomalies, ACC, mild holoprosencephaly, microophthalmia or anophthalmia, and craniofacial dysmorphology¹⁰. Review of magnetic resonance imaging (MRI) scans available from three unrelated individuals with DBS/FOAR (Fig. 2a–e) showed that none had holoprosencephaly, but all had corpus callosum anomalies (agenesis in kindred 5; hypoplasia involving the splenium and rostrum in kindreds 1 and 3) and enlarged globes with small colobomas at the optic nerve heads. Comparison of cortical surface recontructions between one affected individual who had periventricular nodular heterotopia along the lateral ventricles and one age-matched control (Fig. 2f–i) demonstrated disrupted frontal lobe development with an unrecognizable, abnormally placed central sulcus. Based on previously published and recent reports with adequate imaging information, all clinically diagnosed individuals with DBS have either callosal agenesis (58%) or hypoplasia (42%); by contrast, individuals clinically diagnosed with FOAR have macrocephaly but not ACC.

Figure 2.

MRI abnormalities in individuals with DBS/FOAR. (a) Sagittal T1-weighted magnetic resonance image of subject 1, kindred 5, showing ACC with associated absence of cingulate gyrus. (b) Sagittal T1-weighted magnetic resonance image of subject IV-6, kindred 1, showing arrested development of corpus callosum with truncation of posterior body, absence of splenium and rostrum of corpus callosum, partially empty sella (s) and small-appearing pons (P). (c,d) Axial FSE T2-weighted magnetic resonance image of subject IV-6, kindred 1, showing subependymal nodular heterotopia (c) and enlarged globes with small colobomas at the optic nerve heads (d). (e) Axial computed tomography of temporal bones of subject IV-6, kindred 1, showing globular malformation of the left horizontal semicircular canal and vestibule. Abnormalities are indicated by arrows in a–e. (f–i) Brain cortical surface reconstruction from spoiled gradient-echo (SPGR) structural magnetic resonance image. f shows a frontal right view and g shows a top view for a 4-year-old control. Central sulcus, precentral gyrus and postcentral gyrus are labeled by a red curve, a white arrow and a yellow arrow, respectively. h shows a frontal right view and i shows a top view for subject IV-6, kindred 1, with DBS/FOAR. Note abnormally developed frontal lobe and unrecognizable abnormally placed central sulcus.

Given their phenotypic similarity to individuals with DBS/FOAR, we sequenced LRP2 in three families with Chudley-McCullough syndrome (OMIM 604213) and observed that this syndrome is not allelic. Overlapping syndromes such as acrocallosal syndrome (OMIM 200990) may belong to a family of ‘megalinopathies’ due to mutations in the gene encoding megalin or an interacting gene in the same pathway. The mechanism(s) by which LRP2 mutations cause DBS/FOAR is currently unknown. Megalin is not reported to be expressed in fibroblasts or lymphoblasts, and accordingly, we were unable to detect either protein or mRNA by immunohistochemistry or RT-PCR in these tissues from individuals with DBS/FOAR (kindreds 1, 3, and 7) or from controls (data not shown). Immunohistochemistry for megalin in kidney sections from a week 25 fetus with DBS/FOAR (kindred 2, subject 2) demonstrated a staining pattern comparable to that of a control of matched gestational age, suggesting that the functional defect does not require absent protein. The characteristic pattern of DBS/FOAR anomalies may result from impaired megalin endocytosis and subsequent failure to deliver lipophilic compounds during ontogeny¹². Megalin ligands (such as vitamin A (retinol) and cholesterol) are critical for normal embryonic development. Extensive data on teratogenic and genetic animal models, as well as from recently published accounts of affected individuals with STRA6 mutations, confirm a role of vitamin A in diaphragm and lung development^13,14. In addition, disruption of megalin–sonic hedgehog interactions may contribute to abnormal brain and lung development^12,15. Our data raise the possibility that ligand-selective supplementation, such as with vitamin A analogs, may be valuable in preventing or mitigating some of the specific birth defects seen in DBS/FOAR. Finally, examination of a larger cadre of affected individuals with isolated or nonsyndromic malformations, such as CDH or ACC, can help determine whether megalin pathway defects have a role in their pathogenesis and whether this pathway contains putative therapeutic targets.

Individuals with DBS/FOAR were first recruited and clinically characterized by their geneticists, who, with appropriate consent, provided samples to be analyzed in our study (‘Gene Mutations and Rescue in Human Diaphragmatic Hernia’); the protocol for this study is approved annually by the Massachusetts General Hospital Institutional Review Board.