Lack of KIF21A mutations in congenital fibrosis of the extraocular muscles type I patients from consanguineous Saudi Arabian families

• Introduction
• Methods
• Results
• Discussion

Abstract

Purpose: Congenital fibrosis of the extraocular muscles type I (CFEOM1), the most common CFEOM worldwide, is characterized by bilateral ptotic hypotropia, an inability to supraduct above the horizontal midline, horizontal strabismus (typically exotropia), and ophthalmoplegia with abnormal synkinesis. This distinct non-syndromic phenotype is considered autosomal dominant and is virtually always from heterozygous missense mutations in kinesin family member 21A (KIF21A). However, there are occasional KIF21A-negative cases, opening the possibility for a recessive cause. The objective of this study is to explore this possibility by assessing CFEOM1 patients exclusively from consanguineous families, who are the most likely to have recessive cause for their phenotype if a recessive cause exists.

Methods: Ophthalmic examination and candidate gene direct sequencing (KIF21A, paired-like homeobox 2A [PHOX2A], tubulin beta-3 [TUBB3]) of CFEOM1 patients from consanguineous families referred for counseling from 2005 to 2010.

Results: All 5 probands had classic CFEOM1 as defined above. Three had siblings with CFEOM. None of the probands had mutations in KIF21A, PHOX2A, or TUBB3.

Conclusions: The lack of KIF21A mutations in CFEOM1 patients exclusively from consanguineous families, most of whom had siblings with CFEOM, is strong evidence for a recessive form of CFEOM1. Further studies of such families will hopefully uncover the specific locus(loci).

Congenital fibrosis of the extraocular muscles type 1 (CFEOM1, OMIM 135700) is the most common form of CFEOM reported worldwide [1]. It is a distinct non-syndromic congenital cranial dysinnervation disorder and is characterized by bilateral ptotic hypotropia, an inability to supraduct above the horizontal midline, horizontal strabismus (typically exotropia), and a variable degree of ophthalmoplegia with abnormal synkinesis [2-4]. The phenotype was first mapped as an autosomal dominant fully penetrant trait to the centromere on chromosome 12 [5]. Screening of transcripts in this region in several affected families lead to the discovery of heterozygous missense mutations in kinesin family member 21A (KIF21A) as the cause [2]. Since then further studies provide strong evidence that the classic CFEOM1 phenotype results from mutations in KIF21A and that sporadic cases are due to de novo mutations in the same gene [1-3,6-14]. A heterozygous KIF21A missense mutation has been found to underlie CFEOM1 patients across populations worldwide with most patients having mutations in exon 21 that affect arginine at position 954 of the protein (p.R954W, p.R954Q, or p.R954L) [1-3]. It has been suggested that methylation of CpG dinucleotides in exon 21 of KIF21A increases susceptibility to mutational events [13]. Other than exon 21, only 2 other exons in the 38-exon gene have been reported to harbor mutations: exon 8 (p.M356T) and exon 20 (p.E944Q, p.M947V, p.M947T, p.M947R) [3]. The lack of mutations in other exons of KIF21A and the lack of other types of mutations (other than missense) may be because such mutations are lethal or because they underlie a phenotype thus far not associated with the gene. The normal function of the KIF21A protein includes the transport of membranous organelles, protein complexes, and mRNAs to specific destinations within the cell in a microtubule- and ATP-dependent manner. These functions are essential for normal morphogenesis and functioning of the cell [15]; however, missense KIF21A mutations only appear to significantly affect the orbit, causing widespread orbital dysinnervation [16].

Other clinical forms of CFEOM with a known genetic basis are CFEOM2 (OMIM 602078) and CFEOM3 (OMIM 600638). CFEOM2, the rarest CFEOM phenotype, is a recessive disorder that was first mapped to 11q13 in consanguineous families [17] and later found to be secondary to homozygous mutations in the hindbrain transcription factor paired-like homeobox 2A (PHOX2A) [18]. PHOX2A knockout animal models reveal that the gene is responsible for development of the oculomotor and trochlear cranial nerve nuclei [19,20]. The CFEOM2 phenotype is characterized by bilateral large-angle exotropia, ptosis, miosis, and ophthalmoplegia with abnormal synkinesis [4,21]. CFEOM3 is CFEOM that does not meet the classic criteria for CFEOM1 or CFEOM2. CFEOM3 can be unilateral or bilateral, is often autosomal dominant, and can have variable penetrance [4,22,23]. A family with autosomal dominant CFEOM is considered to be a CFEOM3 pedigree even if one or more members has (have) classic CFEOM1 if at least one affected family member does not meet the criteria for CFEOM1 [4,23]. Most CFEOM3 families have mapped to 16qter [23-25] and are due to heterozygous mutation in tubulin beta-3 (TUBB3), a gene involved in microtubule dynamics, kinesin interactions, and axon guidance [22]. Unlike patients with CFEOM1 or CFEOM2, patients with TUBB3-related CFEOM3 can have extraorbital neurologic findings as well [2]. In some instances, CFEOM3 can be caused by heterozygous missense KIF21A mutations [23].

CFEOM1 is considered to be an autosomal dominant fully penetrant condition. Although KIF21A is the only gene associated with CFEOM1 to date, up to 40% of sporadic CFEOM1 cases do not have identifiable mutations in KIF21A [3]. Among the possibly genetic causes are mutations in a KIF21A promotor, mutations in PHOX2A and/or TUBB3, or dominant or recessive mutations at a different locus. If a recessive cause for CFEOM1 exists, one would expect it to occur more commonly in CFEOM1 patients from large consanguineous families [26]. In the current study, we perform candidate gene testing in CFEOM1 patients from consanguineous families to explore the possibility of a recessive cause for the CFEOM1 phenotype.

Institutional board approval was granted for this study. Only probands with CFEOM1 who were from consanguineous families were invited to participate in the study. Enrolled patients, who had no known relationship to each other, had complete orthoptic and ophthalmic examination as well as 5 ml venous blood sampling for candidate gene testing and were referred to one of the authors (A.O.K.) from 2005 to 2010. Cyclopentolate 1% was used for dilation and cycloplegic refraction. When affected relatives were available and willing, they were examined as well. The candidate genes KIF21A, PHOX2A, and TUBB3 were directly sequenced. Briefly, polymerase chair reaction products from all exons of KIF21A (NM_017641), PHOX2A (NM_005169), and TUBB3 (NM_006086.2) were sequenced using the ABI Prism Big Dye Terminator v3.1 Cycle Sequencing Kit as described by the manufacturer. Results were exported in one of several formats for visualization and sequence was analyzed using SeqMan 6.1 (Lasergene 6 software package) [22,27]. Primers used for KIF21A are shown in Table 1, for PHOX2A are shown in Table 2, and for TUBB3 are shown in Table 3 and are as previously published [22].

Pedigrees for the 5 patients are shown in Figure 1. All 5 probands had classic CFEOM1 without pupillary abnormality as defined above. One proband had an affected sibling with CFEOM1 (family 3 in Figure 1) and 2 probands each had an affected sibling with CFEOM3 (unilateral CFEOM ; families 4 and 5 in Figure 1). Clinical features of the probands are summarized in Table 4 and the typical proband phenotype is shown in Figure 2 (patient 1 from Table 4). No patient had significant extra-orbital disease.

No proband had mutations or polymorphic variations in KIF21A, PHOX2A, or TUBB3.

Five CFEOM1 probands from consanguineous families were assessed in this study, none of whom had significant extra-orbital disease. Two were sporadic cases, one had a sibling with CFEOM1, and 2 each had a sibling with CFEOM3. No proband from this unique CFEOM1 cohort harbored mutations in KIF21A, TUBB3, or PHOX2A. Rather than being an autosomal dominant phenotype, CFEOM1 in our cohort was almost certainly related to homozygous mutations in a locus (or in loci) that to date has (or have) not been associated with the condition.

The 3 previously-reported CFEOM1 families from Saudi Arabia were not consanguineous and all harbored heterozygous missense KIF21A mutations [6,9]. Two families had autosomal dominant inheritance and both harbored the most common KIF21A mutation reported worldwide, p.R954W [9]. The third family, of Jordanian ancestry, exhibited apparent autosomal recessive inheritance with atypical abnormal pupils but in fact harbored heterozygous p.R954L KIF21A mutation with parental germline mosaicism [6]. In the current series, none of the 5 CFEOM1 patients harbored mutations in known CFEOM genes. Two cases were sporadic and 3 had affected siblings. For one, the sibling also had CFEOM1 (family 3 from Figure 1). For the other 2, the each had an affected sibling with CFEOM3 (families 4 and 5 from Figure 1). These latter 2 families would be considered by some authors as CFEOM3 pedigrees [4,23].

Studies of consanguineous families are more likely to uncover recessive cause for a given phenotype if a recessive cause exists because of parental shared recent ancestry. Although every individual is a heterozygous carrier for mutated alleles that would potentially cause recessive disease in the homozygous (or compound heterozygous) state, it is unlikely that the individual's spouse will carry the same disorder unless they are related [28]. Thus studies of exclusively consanguineous families with a specific phenotype offer a unique opportunity to uncover a recessive cause for the phenotype if a recessive cause exists. Our study confirms the existence of recessive CFEOM1. There may be one or more such loci, each of which may be a separate gene or locus that regulates pathways in known genes associated with CFEOM. Whether the 2 families that included a sibling with CFEOM3 (families 4 and 5 from Figure 1) are considered CFEOM3 families or families with CFEOM1 probands, the observed phenotype is likely related to a recessive cause that has not yet been described.

In summary, although most CFEOM1 is due to heterozygous missense KIF21A mutations, there exists at least one additional autosomal recessive cause for the phenotype. This information is useful in the genetic counseling of sporadic KIF21A-negative CFEOM1 patients. It is hoped that further ascertainment and study of CFEOM1 patients from consanguineous families will uncover the novel locus(loci).

Funding for this study was from the Department of Genetics, King Faisal Hospital & Research Center; Research Department, King Khaled Eye Specialist Hospital