Loss of MAFB Function in Humans and Mice Causes Duane Syndrome, Aberrant Extraocular Muscle Innervation, and Inner-Ear Defects

Abstract

Duane retraction syndrome (DRS) is a congenital eye-movement disorder defined by limited outward gaze and retraction of the eye on attempted inward gaze. Here, we report on three heterozygous loss-of-function MAFB mutations causing DRS and a dominant-negative MAFB mutation causing DRS and deafness. Using genotype-phenotype correlations in humans and Mafb-knockout mice, we propose a threshold model for variable loss of MAFB function. Postmortem studies of DRS have reported abducens nerve hypoplasia and aberrant innervation of the lateral rectus muscle by the oculomotor nerve. Our studies in mice now confirm this human DRS pathology. Moreover, we demonstrate that selectively disrupting abducens nerve development is sufficient to cause secondary innervation of the lateral rectus muscle by aberrant oculomotor nerve branches, which form at developmental decision regions close to target extraocular muscles. Thus, we present evidence that the primary cause of DRS is failure of the abducens nerve to fully innervate the lateral rectus muscle in early development.

Figure 1

Mutations in MAFB Cause DRS (A) MAFB mutations segregate with DRS in pedigrees FA, PM, 0819, and N. Three of the four affected FA pedigree members also have unilateral or bilateral congenital hearing loss (indicated by an asterisk). (B) FA IV:1 has bilateral DRS, characterized by bilaterally limited eye abduction and narrowing of the palpebral fissures with globe retraction during attempted eye adduction. (C) Axial CT images of the right temporal bone of a healthy control individual with normal cochlea and vestibule (arrows) and of individual FA IV:1, who has a cystic common-cavity anomaly (arrow).

Figure 2

Mafb-Mutant Mouse Embryos Demonstrate DRS Pathology (A–F) Whole-mount sagittal confocal images at E11.5. (A) Mafb^WT/WT embryos showed normal hindbrain and cranial nerve development. The white line indicates the region of developing rhombomeres 5 and 6. (B) Mafb^WT/KO embryos had abducens nerve hypoplasia but no other major abnormalities of hindbrain and cranial nerve development. (C) Mafb^KO/KO embryos showed loss of rhombomeres 5 and 6, resulting in loss of the hindbrain area (white line), an absent abducens nerve, and fusion of the glossopharyngeal nerve and the vagus nerve (arrow). (D–F) Medial sagittal sections highlight the developing oculomotor and abducens nerves. (D) In Mafb^WT/WT embryos, the abducens nerve was present (short arrow) and reached the developing eye. (E) Mafb^WT/KO embryos had a hypoplastic abducens nerve (short arrow). (F) Mafb^KO/KO embryos were missing an abducens nerve (short arrow). (G–O) Confocal images of the right orbit in mouse embryos from the inferior view. (G) At E12.5 in Mafb^WT/WT embryos, the abducens nerve (arrowhead) innervated the LR muscle, whereas the oculomotor nerve innervated the IR muscle and developing IO muscle. (H) In E12.5 Mafb^WT/KO embryos, the abducens nerve was hypoplastic (arrowhead) and innervated the LR muscle, whereas the oculomotor nerve began to send aberrant branches toward the LR muscle (arrow) and the RB muscle. (I) In E12.5 Mafb^KO/KO embryos, the abducens nerve was absent (arrowhead), and the oculomotor nerve sent aberrant branches toward the LR muscle (arrow) and the RB muscle. (J) At E13.5 in Mafb^WT/WT embryos, the abducens nerve (arrowhead), oculomotor nerve, and extraocular muscles continued to develop normally. (K) In E13.5 Mafb^WT/KO embryos, the hypoplastic abducens nerve (arrowhead) innervated the LR muscle, whereas the oculomotor nerve formed a second, more distal aberrant branch toward the LR muscle (yellow arrow), in addition to the proximal aberrant branch formed earlier (white arrow). (L) In E13.5 Mafb^KO/KO embryos, the abducens nerve was absent (arrowhead), and the oculomotor nerve sent a distinct distal aberrant branch toward the LR muscle (yellow arrow), in addition to the proximal aberrant branch formed earlier (white arrow) and the other aberrant branches contacting the RB muscle. (M) Mafb^WT/WT embryos at E16.5 had the final developmental pattern of the orbit, whereby the abducens nerve (arrowhead) innervated the LR muscle. (N) In E16.5 Mafb^WT/KO embryos, the abducens nerve remained hypoplastic (arrowhead), and a proximal aberrant branch (white arrow) and a distal aberrant branch (yellow arrow) of the oculomotor nerve also innervated the LR muscle. (O) In E16.5 Mafb^KO/KO embryos, the abducens nerve remained absent (arrowhead), and a proximal aberrant branch (white arrow) and a distal aberrant branch (yellow arrow) of the oculomotor nerve innervated the LR muscle instead. (P) At E16.5, the diameter of the abducens nerve was significantly smaller in Mafb^WT/KO embryos than in Mafb^WT/WT embryos and was absent in Mafb^KO/KO embryos. (Q) At E16.5, the diameter of the oculomotor distal aberrant branch was significantly greater than that of the oculomotor proximal aberrant branch in both Mafb^WT/KO and Mafb^KO/KO embryos. The diameter of the distal aberrant branch was significantly greater in Mafb^KO/KO than in Mafb^WT/KO embryos. These aberrant branches were not present in Mafb^WT/WT embryos. Abbreviations are as follows: III, oculomotor nerve; IV, trochlear nerve; V, trigeminal nerve; VI, abducens nerve; VII, facial nerve; IX, glossopharyngeal nerve; X, vagus nerve; IO, inferior oblique; IR, inferior rectus; LR, lateral rectus; and RB, retractor bulbi. Scale bars represent 100 μm. ^∗p < 0.05, ^∗∗∗∗p < 0.0001; differences were measured by Tukey’s multiple-comparison test, and error bars represent the SEM (GraphPad Prism). n > 10 for each genotype.

Figure 3

Less Than 50% MAFB Function Causes DRS and Inner-Ear Defects (A) Pedigree N has a full gene deletion and therefore no mutant MAFB. Pedigrees 0819 and PM are predicted to have truncated MAFB proteins that lack the EHR, BR, and LZ domains followed by 78 and 10 altered amino acids, respectively. Pedigree FA is predicted to have a MAFB that retains the wild-type EHR, BR, and beginning of the LZ followed by 125 altered amino acids. (B) A luciferase assay showed that wild-type MAFB increased transcription by approximately 150-fold. The 0819 or FA mutant protein alone did not have any transcriptional activity. Co-expression of wild-type MAFB and the FA mutant, but not the 0819 mutant, reduced the transcriptional activity of the wild-type protein in comparison to that of the wild-type alone. ^∗∗p < 0.01; differences were measured by Tukey’s multiple-comparison test, and error bars represent the SEM (GraphPad Prism). Each experiment was performed in triplicate. (C) Threshold model for loss of MAFB function. At greater than 50% MAFB function, Mafb^WT/WT and kr/+ mice had no phenotypic alterations. At 50% MAFB function, Mafb^WT/KO mice and N, 0819, and PM family members with heterozygous loss-of-function mutations had isolated DRS. At some level below 50% MAFB function, kr/kr mice, FA members with a dominant-negative mutation, and Mafb^KO/KO mice had both DRS and inner-ear defects.