Genotyping of RET in Patients.

Previous reports have shown that BBS patients with HSCR are more frequent carriers of a RET hypomorphic allele than expected from population frequency (9). Therefore, we hypothesized that the BBS+HSCR patients in our families might also have a RET variant. To test this possibility, we sequenced the patients for mutations in RET. The 3 F1 sibs were homozygous for the common intronic hypomorphic T allele; both parents were heterozygous. In F2, only the 2 BBS+HSCR, but not the other 3 BBS-only patients were homozygous for the same RET allele. Last, the sporadic SB25 case was WT for the intron 1 allele, but had a novel, maternally inherited heterozygous 11-bp deletion (del AAGCAACTGCC). This deletion lies 106-bp downstream of the known intron 1 hypomorphic allele, maps within the highly conserved enhancer region, and was not found in over 800 control chromosomes (Fig. 1; Fig. S1).

Functional Assessment of BBS and RET Mutations.

Our genetic data provide strong evidence that the BBS5 and BBS7 alleles in families F1 and F2 are sufficient to cause BBS. By contrast, the pathogenic potential of the BBS4 S457I allele cannot be established unequivocally, despite its absence from ethnically matched controls. Therefore, we investigated the (non)neutrality of this allele using an established zebrafish BBS model. We have reported previously that suppression of Bbs proteins results in Wnt-dependent gastrulation movement defects in zebrafish (15, 16), which can be rescued by coinjection of WT human mRNA. This system has enabled us establish the pathogenic potential of several novel BBS-causing alleles (12).

Therefore, we suppressed endogenous bbs4 and evaluated the ability of a 457I-encoding BBS4 message to rescue the phenotype. Approximately 63% of embryos injected with bbs4 morpholino (MO) had broadened somites, kinked notochords, and shortened body axes (Fig. 2 A and B). The observed phenotypes not only fully recapitulated our previous observations (15), but were also shown to be specific to the MO, because coinjection of mRNA encoding WT human BBS4 efficiently rescued the morphant phenotypes (Fig. 2B). By contrast, embryos injected with 457I were indistinguishable from embryos injected with bbs4 MO alone (n = 96 embryos, scored blind to injection mixture), suggesting that the introduction of the 457I residue results in functionally null protein (Fig. 2 A and B); expression of 457I-encoding mRNA alone yielded modest phenotypes, indicating that this allele does not act in a dominant-negative fashion (Fig. 2B). To validate the pathogenic nature of this mutation further, we investigated the cellular localization of the 457I protein in ciliated mammalian cells. Previous studies have established the reproducible localization of BBS4 near the 2 centrioles by overexpression of tagged protein in transformed cell lines (17, 18). Upon transient transfection of a BBS4-MYC construct, we observed a similar localization pattern (Fig. 2C). However, expression of a BBS4 S457I mutant protein resulted in complete mislocalization of the protein (Fig. 2C), with the mutant protein forming apparent intracellular aggregates away from the pericentriolar region (visualized by γ-tubulin staining), providing further evidence that the mutation is a likely null.

Open in a separate window

Fig. 2.

In vivo assessment of the pathogenic potential of BBS4 S457I variant. (A) Zebrafish embryos coinjected with a MO against bbs4 and mRNA expressing the S457I variant exhibited defects, including short body axes, broad somites, wider and kinked notochords, and detachment of cells along the dorsal axis. (B) Compared to rescue with WT BBS4 mRNA, rescue with the mutated mRNA produced an increase in the proportion of both moderately (Class I) and severely (Class II) affected embryos, but injection of S457I mRNA alone does not produce severe defects, suggesting that the variant is functionally null. (C) HeLa cells transfected with either WT or S457I mutant BBS4 protein (red). WT protein localizes to the basal body (γ-tubulin, green), but mutant protein completely mislocalizes to form cytoplasmic aggregates.

Next, we turned our attention to the RET alleles in our patients. Although the functional significance of the intronic HSCR predisposing allele in RET is known, the pathogenic potential of the 11-bp deletion within intron 1 is highly suggestive, but not conclusive. We have shown previously that the multispecies conserved sequence (MCS) within intron 1 (MCS + 9.7) acts as an enhancer element, and contains the previously described hypomorphic allele (5); the 11-bp deletion is 106 bp downstream of this SNP. To investigate the function of the deletion, we constructed pDSma-RET vectors with luciferase under the control of an SV40 promoter and the MCS + 9.7 enhancer plus a 1215-bp encompassing region either with (WT) or without (Δ11bp) the 11-bp sequence, and assayed relative luciferase activity in SK-N-SH neuroblastoma cells or HeLa cells. The WT RET enhancer sequence significantly decreased reporter gene expression (P = 0.001) compared with the promoter alone, and also as compared with MCS + 9.7, suggesting that it contains a putative transcription repressor-binding site. However, the vector harboring the 11-bp deletion significantly derepresses reporter gene expression compared with WT (P = 0.05) to a level statistically indistinguishable from the promoter alone (NS), suggesting specific disruption of the repressor element binding motif (Fig. 3). Reporter assays conducted in HeLa cells produced similar results. These findings suggest that deletion of 11 bp within the intronic enhancer results in defects in transcriptional regulation in vitro, although they do not abolish the effect of this genomic segment, raising the possibility that Δ11bp represents a mild allele. Together, these results suggest that the individual BBS and RET mutations identified in the sporadic case of BBS and HSCR are likely detrimental to gene function. Among putative consensus binding sites encompassing the 11-bp deletion, only the zinc finger transcription factor SNAIL is expressed in hNCC. However, ChIP experiments in SK-N-SH cells did not detect an interaction with this protein, suggesting that potentially novel components are associated with the observed suppression regulatory element.

Open in a separate window

Fig. 3.

Deletion in the RET regulatory sequence perturbs reporter gene expression in vitro. (A) Schematic representation of vectors used to assay the effect of the 11-bp deletion on reporter gene expression in vitro. WT or mutated (Δ11bp) RET enhancer sequence (green) was placed upstream of an SV40 promoter (gray) driving firefly luciferase expression (yellow). The pDSma-RET MCS + 9.7, described previously (5), was used as a positive control. (B) Relative luciferase activity of pDSma-RET vectors (depicted in A) in SK-N-SH neuroblastoma cells. Cells were lysed 48 h posttransfection, and relative luciferase activity was normalized to values obtained from the promoter alone (pDSma-Promoter). The WT RET enhancer sequence significantly decreases reporter gene expression (P = 0.001) compared with promoter alone, suggesting that it contains a putative transcription repressor-binding site. However, the vector harboring the 11-bp deletion significantly derepresses reporter gene expression compared with WT (P = 0.05) to a level statistically indistinguishable (NS) from the promoter alone, suggesting specific disruption of the repressor element binding motif. Assays were performed in triplicate wells and repeated thrice.

Genetic Interaction of BBS and RET in Vivo.

Our data are consistent with previous studies, in which a higher-than-expected frequency of RET hypomorphic alleles were found in BBS-HSCR patients, and suggest that this property is not restricted to a specific BBS gene (9), not least because BBS4, BBS5, and BBS7 are 3 of the least commonly mutated BBS genes (13). To test this possibility in vivo, we assayed the effects of suppression of each of Ret and Bbs proteins on enteric neuron migration in the developing zebrafish gut. As shown previously, suppression of ret by MO injection results in a reduction of enteric neurons populating the proximal and distal hindgut, as labeled by in situ hybridization with a phox2b riboprobe (Fig. 4 B and I) (19). Likewise, and consistent with recent data for bbs8, suppression of either bbs4, bbs5, or bbs7 also results in a reduction of enteric neuron migration, although not as severe (Fig. 4 C, E, G, and I). However, coinjection of ret MO with any of the 3 bbs genes tested resulted in a significantly greater decrease in the enteric neurons in the hindgut, with phenotypic synthesis particularly prominent in the distal hindgut (Fig. 4 D, F, H, and I). The increased severity is greater than a mere additive effect, suggesting an epistatic interaction between ret and bbs function. These findings indicate that suppression of ret in addition to loss of bbs function results in more severe defects in ENS development than with loss of these genes independently.

Open in a separate window

Fig. 4.

Genetic interaction of Ret and Bbs in zebrafish. (A) Migration of enteric neurons in 4-day-old embryos normally proceeds along the developing hindgut to reach the anus (asterisks). Embryos injected with MO against either (B) ret or (C) bbs4, (E) bbs5, or (G) bbs7 exhibit premature termination of enteric neuron migration (arrowhead). This defect is exacerbated in embryos injected with a combination of ret MO and either (D) bbs4, (F) bbs5, or (H) bbs7 MO. (I) The extent of the defect was quantified by assessment of the presence of phox2b expression in the proximal hindgut (somites 4–9) or the distal hindgut (somites 10 and beyond) as previously described (31).

Luciferase Reporter Assay.

SK-N-SH or HeLa cells were plated in 24-well plates at a density of 10⁵ cells/well, and were transfected at ≈60% confluence with FuGENE6 transfection reagent (Roche) according to manufacturer's instructions, using a 2:1 ratio of pDSma vector and pRLSV40 Renilla luciferase control vector (Promega), respectively. At 48 h posttransfection, cells were lysed and assayed using the Dual-Luciferase Reporter Assay System (Promega), and a Tecan microplate reader equipped with the XFluor4GENiosPro macro. Assays were performed on triplicate wells, and were repeated at least twice. Relative luciferase activity was normalized to values obtained from the promoter alone (pDSma-Promoter).

Zebrafish Embryo Manipulation and MO Injection.

Embryos were injected with a previously described and validated oligonucleotide MO against bbs4 to suppress translation of the endogenous protein, resulting in phenotypes consistent with PCP defects including shortened body axes, broad notochords, and widened somites (15, 16). They were subsequently coinjected with mRNA encoding WT human BBS4 or the S457I variant of BBS4. Embryos were classified as normal, moderately (Class I), and severely (Class II) affected according to phenotype.

In Situ Hybridization.

Embryos were cultured for 4 days in embryo media at 28 °C, fixed in 4% Paraformaldehyde overnight at 4 °C, and stored at −20 °C in 100% MeOH until in situ hybridization, which was carried out according to previously described protocols (15). Enteric neurons were labeled using a phox2b digoxigenin ripoprobe (31).

Human NC Isolation and Culture.

Human embryos were collected from pregnancies legally terminated using the mefiprestone protocol, in concordance with French bioethics law 2004-800 and with the approval of the Necker hospital ethics committee. Human NCCs were isolated from posterior hindbrain and thoracic-level neural tube and cultured as previously described (32) to 80% confluency on 8-well collagen I-coated slides (BD BioCoat).

Immunocytochemistry.

Twenty-four hours after seeding, hNCC were starved in DMEM/F12 with 0.4 ng/mL FGF2 for a further 72 h. Cells were fixed for 20 min in 4% paraformaldehyde, then treated in 50 mM NH₄Cl 10 min, 0.3% Triton X-100 in PBS 15 min, followed by 1 h in a solution of 1% BSA, 0.1% Tween-20, 10% goat serum in PBS. Cells were then incubated with antiacetylated α-tubulin (TUBA4A; Sigma T6793), anti-RET (SC13104, 1:50; Santa Cruz), or anti-SOX9 (ab5535, 1:200; Chemicon) for 2 h at room temperature, and rinsed. SOX9 and RET were revealed using an Alexa Fluor 488 goat anti-rabbit IgG and TUBA4A with an Alexa Fluor 555 donkey anti-mouse IgG (1:200 each; Invitrogen). The slides were mounted with Prolong Gold antifade reagent with DAPI (Invitrogen).

HeLa cells were grown on glass coverslips in DMEM supplemented with 10% FBS at 37 °C in 5% CO₂. Transient transfections of WT or mutant BBS4 constructs were carried out using calcium phosphate (Invitrogen), and cells processed 24 h later with anti-MYC mouse monoclonal and anti-γ-tubulin rabbit polyclonal antibodies (T3559; Sigma).

We thank the families and their referring clinical teams for participation, the Association Françcaise de la Maladie de Hirschsprung for its support, Sophie Audollent for technical assistance, and the Johns Hopkins University Finz Center for husbandry and technical support with the zebrafish experiments. This work was supported by grants from the Association Nationale pour la Recherche, the Fondation pour la Recherche Médicale, and the National Institute of General Medical Sciences (A.S.M.), by National Institute of Child Health and Development Grant R01HD04260, by National Institute of Diabetes, Digestive, and Kidney Disorders Grants R01DK072301 and R01DK075972, by the Macular Vision Research Foundation (N.K.), by a Visual Neuroscience Training Program fellowship (N.A.Z.), by and National Institute of Diabetes, Digestive, and Kidney Disorders National Research Service Award Fellowship F32 DK079541 (to E.E.D.). A.S.M. and D.M. were supported by award GM071648 (to A.S.M.).