Discussion

The CCDDs provide an excellent example of the power of current genetic techniques to advance our understanding of certain complex clinical disorders. This heterogeneous amalgam of congenital ocular motility disorders coalesced 50 years ago around the concept of undefined abnormalities in the development of extraocular muscles. Subsequent anatomic, neuroradiologic, and genetic information has largely left in place the clinical grouping, but replaced the central concept with one of developmental abnormalities in ocular motility mechanisms in the brainstem, peripheral cranial nerves, and orbit that are neurogenic in origin. This group of disorders proved very difficult to classify by clinical criteria alone, but much more progress has been made since it has become possible to study groups of genetically homogenous individuals.

There appear to be at least three pathways by which congenital abnormalities in the human ocular motility system occur in these patients with genetically defined disorders (Table 1). The first involves the failure of cranial nerve nuclei to develop normally. Perhaps the best current example of this mechanism is the HOXA1 clinical spectrum in which abducens nerve nuclei fail to develop as part of the loss of rhombomere 5 during development in the absence of a normal HOX cascade. The second mechanism involves genetic defects that lead to abnormal axonal transport of molecules necessary for normal guidance of developing peripheral neurons to the correct extraocular muscles and possibly for the normal functional development of these muscles. The best current example of this mechanism is CFEOM1 with mutations in KIF21A causing abnormalities in axonal transport and leading to loss of ocular motor axons and possibly abnormal termination of all peripheral nerves within the orbits. CFEOM3 is a clinically similar disorder but genetically heterogeneous, implying that other genes may play similar roles in orbital innervation. The third mechanism occurs with abnormalities in the development of upper motor neurons in the brainstem, sometimes with associated lower motor neuron abnormalities as well. The best-understood current example of this mechanism would be individuals with HGPPS who have mutations in ROBO3 resulting in the loss of decussation of a number of developing brainstem neural pathways, including those in the pontine paramedian reticular formation responsible for horizontal gaze.

Today, we can finally define at least some CCDD syndromes by the responsible gene as we build a montage of genes acting during development of the brainstem and orbits (Table 1). We now have some idea about what can go wrong in the setting of certain specific mutations, and by inductive reasoning, what role these genes play in the development of the normal human efferent visual system. Curiously, an intensive evaluation of the CCDD group has not yet yielded any implication that genes responsible for the development of the extraocular muscles themselves are involved. Rather, anatomic abnormalities in extraocular muscles noted in this group of disorders seem to be secondary to denervation and dysinnervation. Further study, of course, may change this situation.

Thus far, we have no complete explanation for the variability within each genetically defined disorder. At times, it is clear that individuals with identical mutations of the same gene nevertheless have a somewhat different phenotype, as described with mutations of KIF21A, ROBO3, and HOXA1. It is now clear that a single genetic abnormality can result in a fairly broad range of phenotypic presentations. For example, KIF21A mutations can cause CFEOM1 or CFEOM3, both of which may present with variable, bilateral ocular motility restriction and may occur with or without DRS. Likewise, HOXA1 mutations cause deafness in most, but not all, affected individuals, autism in some, and a range of cardiovascular and cerebrovascular malformations in all individuals who have been studied appropriately.

It also seems clear now, even within this limited set of genetic syndromes, that different mutated genes can result, at least in part, in the same phenotype. DRS, for example, occurs with heterozygous mutations of KIF21A and SALL4 and homozygous mutations of HOXA1. The cause of DRS is relatively well understood in each setting. HOXA1 mutations probably result in the loss of rhombomere 5, including the absence of abducens nucleus development. In contrast, KIF21A mutations cause abnormalities in directing the abducens nerve successfully to the ipsilateral lateral rectus muscle. The diagnostic feature of globe retraction on attempted adduction is probably the result of lateral rectus denervation by the abducens nerve and dysinnervation by a nearby branch of the oculomotor nerve.

Some variants of DRS are associated with somatic mutations involving the face, neck, ears, and extremities. Such an association is understandable in the setting of HOXA1 mutations, given that HOXA1 is the first gene in the HOX cascade that is responsible for the development of orientation and certain major anatomic features of the early fetus. In fact, it is somewhat surprising that HOXA1 mutations result in such a relatively benign syndrome given the role of HOXA1 at the beginning of the HOX cascade. However, consanguineous families with HOXA1 homozygous mutations have a greater than normal spontaneous abortion rate, implying that the result of these mutations may not always be benign.

The role of HOXA1 is complex because it seems that this gene is not expressed in the petrous bone, and yet the petrous bone fails to elaborate the bony hearing and vestibular structures in most individuals with homozygous HOXA1 mutations. It seems likely that abnormal development of the petrous bone results from the absence of appropriate developmental signals from abnormal neuroectoderm and brainstem nearby. The cause of cerebrovascular and cardiovascular abnormalities in patients with HOXA1 mutations seems somewhat more straightforward because HOXA1 may actually be active in the early development of the vascular system. The HOXA1 clinical spectrum resembles early thalidomide poisoning, bringing up the possibility that thalidomide interferes with the early HOX cascade. Thus, genetic knowledge of these uncommon congenital ocular motility syndromes can increase our understanding of teratogenic phenomena as well.

HGPPS brings up further considerations regarding development during childhood. The ocular motility abnormalities of HGPPS are congenital. However, the severe scoliosis associated with the syndrome is not present at birth; rather, it begins in early childhood and progresses rapidly. ROBO3 activity has not been documented to date in the developing spine. A direct effect of this gene’s activity remains possible, but an alternative explanation is that the tone of the powerful paraspinous muscles is asymmetric in affected individuals because the absence of ROBO3 prevents decussation of fibers in the brainstem and/or spinal cord important for equalization of paraspinous muscle tone. The action of gravity across the spine would conceivably result in accelerating scoliosis once a mild bend or torque occurs in the spine. Whatever the actual mechanism, it is possible that better understanding of the scoliosis that occurs with ROBO3 mutations will aid in understanding other early scoliosis syndromes and even the fairly common sporadic scoliosis in pubescent females.

Patients with homozygous ROBO3 mutations lack decussation of the major corticospinal tracts and medial lemniscus sensory tracts, an unanticipated and truly profound neuroanatomic abnormality. Curiously, these individuals in general are intellectually normal, implying that these major motor and sensory tracts find the correct termination, permitting grossly normal cortical activity despite being in the wrong hemisphere. The implication of this observation is that these developing fiber tracts respond to direction signals sequentially, with a response (or lack of it) to one signal potentially having little or no effect on the response to the next signal.

It is obvious that development of the human brainstem and development of the human efferent visual system do not depend exclusively on the few genes identified in the CCDD syndromes discussed here. It is quite likely, therefore, that more autosomal dominant and recessive clinical syndromes are yet to be identified under the CCDD rubric. Indeed, there have been many reports over the last 50 years of unusual congenital ocular motility syndromes that were clearly familial, and some of these are surely genetic in origin. The power of current genetics may finally permit us to identify the responsible genes.

See also: Abnormal Eye Movements due to Disease of the Extraocular Muscles and Their Innervation; Cranial Nerves and Autonomic Innervation in the Orbit; Extraocular Muscles: Extraocular Muscle Anatomy; Extraocular Muscles: Functional Assessment in the Clinic.

  • Contact
  • Category: Eye diseases