Ординатура / Офтальмология / Английские материалы / Ocular Periphery and Disorders_Dartt, Bex, Amore_2011

trajectory results in a significantly slower saccade than when the saccade occurs alone.

The linkage between blinking and saccadic eye movements indicates that blinking interacts with the brainstem circuits that generate saccadic eye movements. This interaction is apparent when reflex blinks increase the speed of saccadic eye movements in individuals with abnormally slow saccades, initiate saccade oscillations, and alter the speed of vergence eye movements where the eyes are moving in opposite directions. The neurophysiological mechanism that may underlie these effects is the activity of omnipause neurons, which control fixation, and are associated with reflex blinks. Tonically active omnipause neurons gate saccadic eye movements. Omnipause neurons cease discharging immediately before and during all saccadic eye movements and reflex blinks. Microstimulation of omnipause neurons blocks reflex blinks as well as saccadic eye movements. The evolutionary origins of eyelid control functionally intertwine blinking and saccadic eye movements so that blinks frequently occur with saccadic eye movements, blinking modifies the trajectory of eye movements, and blinking and saccadic eye movements utilize some of the same brainstem neuronal circuits.

Modifiability of the Blink System

In order to respond to challenges to the eyelid’s protective function, the neural control of blinking exhibits significant adaptive plasticity. The most common challenge to the protective function of blinking is the development of dry eye. Cornea irritation created by tear film inadequacy rapidly initiates several changes. First, the trigeminal system becomes more excitable, so that a given blinkevoking stimulus elicits a bigger blink than before cornea irritation. In addition to increased blink excitability, the trigeminal reflex blink circuit begins to ‘oscillate’ in response to a blink-evoking stimulus (Figure 3). In this condition, a single stimulus evokes a reflex blink followed

by blink oscillations, one or more additional large blinks. Blink oscillations occur at a constant interblink interval. These modifications compensate for dry eye in at least two ways. The reduced threshold for evoking a blink produces more frequent blinking so that blinks can occur before significant disruptions of the tear film. The larger blinks increase the amount of meibum released over the tear film. Increasing the thickness of the oily meibum layer superficial to the aqueous layer reduces aqueous evaporation. Disruptions of these normally adaptive mechanisms may underlie neurological disorders involving the eyelids, benign essential blepharospasm (BEB) and hemifacial spasm (HFS).

We have an outline of the neural basis for these adaptive modifications of the blink reflex initiated by eye irritation. The adaptive changes occur in trigeminal blink circuits, but not at OO motor neurons or reticular neurons receiving bilateral trigeminal inputs. If just one eye experiences irritation, stimulating the ophthalmic branch of the trigeminal nerve on the irritated side elicits adaptive blink modifications in both eyelids. Stimulating the contralateral trigeminal nerve associated with the unaffected eye, however, elicits normal, unadapted blinks in both eyes. Consistent with this result, spinal trigeminal neurons in the corneal-evoked blink circuit discharge before blink oscillations, and this discharge correlates with the pattern of OO activity producing the blink oscillation. Although the adaptive processes clearly involve the trigeminal system, the cerebellum is also essential for blink adaptation. Blink-related neurons in the cerebellum are active with blink adaptation, and lesioning the cerebellum blocks this form of motor learning. A hypothesis for the function of this brainstem–cerebellum circuit is that the cerebellum recognizes an error signal produced by eye irritation and then initiates modifications in spinal trigeminal blink circuits to compensate for the eye irritation. These adaptive modifications in the trigeminal blink circuits may involve long-term potentiation (LTP)- and long-term depression (LTD)-like processes.

As discussed in the next section, the cerebellum appears to be an important element in creating the symptoms of dystonic movement disorders such as BEB.

Benign Essential Blepharospasm and

Hemifacial Spasm

BEB is a focal dystonia characterized by involuntary bilateral spasms of lid closure, trigeminal reflex blink hyperexcitablity, and photophobia (i.e., excessive sensitivity to light). BEB frequently begins with a complaint of ocular discomfort. The available evidence indicates that BEB arises from the confluence of a predisposing and a precipitating factor. Although the predisposing factor has not been identified, the genetic basis for other forms of dystonia suggests that the predisposing factor for BEB is genetic. The precipitating factor appears to be ocular irritation. Consistent with eye irritation as the precipitating factor is that BEB characteristics appear to be an exaggeration of the normally adaptive response to dry eye. For most individuals, the spasms of lid closure in BEB are rapid, repetitive contractions of the OO muscle. This spasm pattern is equivalent to shortening the interblink interval of the blink oscillations developed in response to dry eye (Figure 3). BEB patients exhibit trigeminal reflex blink hyperexcitability. One adaptive response to dry eye is to elevate trigeminal reflex blink excitability, although this increase accompanying dry eye is not as profound as occurs with BEB. The trigeminal hyperexcitability with BEB is sufficient to explain the photophobia. Consistent with the exaggeration of dry eye hypothesis, photophobia is present with dry eye although the light sensitivity is not as debilitating as with BEB. Thus, the exaggerated dry eye hypothesis proposes that BEB begins with the onset of dry eye or significant eye irritation. The nervous system initiates blink modifications to compensate for this irritation, but the genetic predisposition prevents the nervous system from recognizing that the adaptive changes corrected the ocular irritation. The nervous system responds by further increasing these adaptive modifications until the normally compensatory mechanisms becomes so maladaptive as to create the BEB syndrome.

Although genetic modifications probably create the predisposing environment for BEB, the genetics underlying this focal dystonia are not yet clear. Most investigators argue that there is an autosomal dominant transmission with reduced penetrance in BEB. One challenge to linking genetics to specific types of dystonia, however, is that individuals with the same genetic mutation may exhibit very different forms of dystonia. For example, individuals with the DYT1 mutation responsible for the most common form of generalized dystonia may exhibit generalized dystonia, focal dystonia, or may be asymptomatic. Individuals with generalized dystonia, focal dystonia, and

asymptomatic DYT1 carriers all exhibit a similar pattern of abnormal brain activity. Positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) scans reveal hyperactivity in the basal ganglia, primary motor cortex, supplementary motor area, putamen, thalamus, and cerebellum. In patients with BEB, these areas are also active, but the focus of abnormal activity is in the brainstem and cerebellum. Abnormal cerebellar activity with BEB supports the hypothesis that abnormal blink adaptation involving the cerebellum is critical in the development of BEB. Consistent with the hypothesis that a genetic predisposing factor affects adaptation or motor learning, individuals with BEB exhibit exaggerated LTPlike plasticity of blink circuits relative to normal subjects.

There are two clear examples in which exaggeration of the adaptive processes initiated by eye irritation or dry eye leads to BEB. A small number of individuals with Bell’s palsy develop blepharospasm. The ocular irritation produced by incomplete eye closure, the precipitating factor, initiates unabated modifications that produce the BEB syndrome. As predicted by the exaggerated dry eye hypothesis, implanting gold weights in the paretic (weakened) eyelid, enabling nearly complete closure of the paretic eyelid, reduces eye irritation and allows a resolution of BEB signs. Combining predisposing and precipitating factors can also create a BEB-like syndrome in rats. In this model, a chemical lesion reducing approximately 30% of the dopamine neurons in the substantia nigra pars compacta elevates the excitability of the trigeminal blink reflex circuits. This increased trigeminal excitability acts as the predisposing factor. Transient eye irritation produced by crushing a facial nerve branch providing a portion of the OO innervation acts as the precipitating factor. The reduced eyelid mobility created by this procedure causes eye irritation that initiates adaptive blink modifications. Although the eye irritation resolves following nerve regrowth, rats continue to exhibit spasms of lid closure caused by repetitive bursts of OO activity and hyperexcitable trigeminal reflex blinks. Thus, the evidence indicates that BEB occurs in individuals genetically predisposed to the disorder who experience a precipitating condition, ocular irritation. The ocular irritation initiates a series of normally adaptive modifications. In the presence of the predisposing condition which creates an abnormal environment for motor learning, these modifications become exaggerated to create the signs of BEB.

HFS begins as spontaneous, unilateral spasms of eyelid closure. Over a period of weeks to months, the spasms expand to include the rest of the facial muscles on that side of the face. Another characteristic of HFS is synkinesis in which there is an involuntary activation of multiple muscles that normally do not act together. An example of synkinesis is that stimulating the supraorbital branch of the trigeminal nerve would strongly activate the mentalis muscle, as well as the OO, in HFS patients. Unlike BEB,

HFS is unilateral and involves most of the ipsilateral facial muscles in addition to the OO. There is strong evidence, however, that HFS also arises from a combination of predisposing and precipitating factors that disrupt normal motor learning.

The accepted precipitating factor for HFS is arterial compression at the root entry zone of the facial nerve. The most common blood vessels affecting the facial nerve in HFS are the anterior inferior cerebellar artery, the posterior inferior cerebellar artery, or the vertebral artery. Microsurgical decompression of the facial nerve typically reduces or eliminates the spasms and synkinesis of HFS so that spasms disappear in 64% and synkinesis in 53% of HFS patients within the first week following surgery. After 2–8 months, 90% of patients are spasm or synkinesis free. If arterial compression of the facial nerve alone is responsible for these signs of HFS, however, then decompression surgery should eliminate the spasms and synkinesis in less than 2–8 months. To understand the basis for this delay, it is important to consider what neural modifications might occur in response to pulsatile arterial compression of the facial nerve. A strong argument that facial nerve compression alone is insufficient to cause HFS is that 15–25% of the population exhibits arterial compression of the facial nerve, but do not develop HFS. This observation shows that HFS requires a predisposing factor as well as the precipitating factor, arterial compression of the facial nerve.

Pulsatile arterial compression of the facial nerve can generate many modifications in brainstem neural circuits. Compression injury to motor neuron axons effectively weakens facial muscles, which initiates adaptive modifications. Repetitive antidromic activation of facial motor neurons by pulsatile compression of axons can alter facial motor neuron excitability or motor neuron excitability may increase because of facial motor neuron axotomy, severing the facial nerve fibers. Repetitive orthodromic activation of facial muscles by pulsatile compression of motor neuron axons can lead to a reorganization of sensory trigeminal circuits. Simultaneous orthodromic (electrical impulses traveling in the normal direction) activation of muscles that normally do not act together, for example, mentalis and OO, causes trigeminal primary afferent sensory signals from mentalis and OO muscle contraction to reach second-order trigeminal neurons synchronously. This abnormal pairing of sensory inputs can restructure trigeminal receptive fields so that the second-order trigeminal neurons respond strongly to inputs from both mentalis and OO activity instead of weakly to one and strongly to the other. Pulsatile activation of the facial nerve can also augment reflex circuit excitability because of synchronous activation of circuit elements. Second-order trigeminal neurons receiving synchronous afferent inputs innervate facial motor neurons that are already depolarized by antidromic activation.

This activity pattern can strengthen trigeminal inputs onto facial motor neurons in a spike timing-dependent plastic- ity-like manner.

It is clear that pulsatile activation of the facial nerve can produce blink modifications, but these changes should not cause HFS by themselves. The eyelid system normally modifies itself so as to perform appropriately in the face of changes in the motor system or sensory inputs. For example, creating unexpectedly large blinks by adding weights to the upper eyelids initiates a rapid reduction in the trigeminal drive onto OO motor neurons. Similarly, chronic, repetitive facial nerve stimulation, such as occurs with pulsatile facial nerve compression, reduces blink amplitude. Thus, pulsatile facial nerve compression is insufficient to cause spasms of lid closure because the blink system will modify itself to prevent spasms of lid closure. The significant number of humans not experiencing HFS, but exhibiting arterial compression of the facial nerve, further supports this interpretation. These observations indicate that HFS, like BEB, requires a predisposing factor to develop spasms. Like BEB, an autosomal-dominant genetic mutation with low penetrance may provide the predisposing condition, which disrupts normally adaptive processes to create a pathological condition.

Further Reading

Bour, L. J., Aramideh, M., and de Visser, B. W. (2000). Neurophysiological aspects of eye and eyelid movements during blinking in humans. Journal of Neurophysiology 83: 166–176.

Evinger, C., Manning, K. A., and Sibony, P. A. (1991). Eyelid movements. Mechanisms and normal data. Investigative Ophthalmology and Visual Science 32: 387–400.

Fukuda, H., Ishikawa, M., and Okumura, R. (2003). Demonstration of neurovascular compression in trigeminal neuralgia and hemifacial spasm with magnetic resonance imaging: Comparison with surgical findings in 60 consecutive cases. Surgical Neurology 59:

93–99; discussion 99–100.

Manning, K. A. and Evinger, C. (1986). Different forms of blinks and their two-stage control. Experimental Brain Research 64: 579–588.

Nielsen, V. K. and Jannetta, P. J. (1984). Pathophysiology of hemifacial spasm: III. Effects of facial nerve decompression. Neurology 34: 891–897.

Rambold, H., Sprenger, A., and Helmchen, C. (2002). Effects of voluntary blinks on saccades, vergence eye movements, and saccade–vergence interactions in humans. Journal of Neurophysiology 88: 1220–1233.

Sibony, P. A. and Evinger, C. (1998). Normal and abnormal eyelid function. In: Miller, N. R. and Newman, N. J. (eds.) Walsh and Hoyt’s Clinical Neuro-Ophthalmology, vol. 1, pp. 1509–1594. Baltimore, MD: Williams and Wilkins.

Sparks, D. L. (2002). The brainstem control of saccadic eye movements.

Nature Reviews Neuroscience 3: 952–964.

VanderWerf, F., Brassinga, P., Reits, D., Aramideh, M., and Ongerboer de Visser, B. (2003). Eyelid movements: Behavioral studies of blinking in humans under different stimulus conditions.

Journal of Neurophysiology 89: 2784–2796.

VanderWerf, F., Reits, D., Smit, A. E., and Metselaar, M. (2007). Blink recovery in patients with Bell’s palsy: A neurophysiological and behavioral longitudinal study. Investigative Ophthalmology and Visual Science 48: 203–213.

D M Noden, Cornell University, Ithaca, NY, USA

ã 2010 Elsevier Ltd. All rights reserved.

Glossary

Mesenchyme – Embryonic cells with a fibroblast-like appearance, surrounded by extracellular matrix, lacking tight junctions with their neighbors, and often capable of undergoing extensive migratory movements. These can be of several different embryonic origins, and include cells that will contribute to many lineages.

Morphogenesis – It includes those processes that establish the correct locations and three-dimensional organization of tissues and organs. This includes the proper positioning of extraocular muscles around the globe and their attachments to the sclera and orbital skeleton.

Myoblasts – Mitotically active cells committed to the skeletal muscle lineage but not yet expressing muscle-specific proteins such as desmin and myosins, which generally are not evident until after myoblasts fuse to form multinucleated myofibers. Myotome – The several regions of each somite that contain progenitors of skeletal muscle progenitors. Neural crest – Mesenchymal cells that are derived from neural fold tissues and that move peripherally along well-delineated pathways to form neurons and glia of the peripheral nervous system and, in the head region, the connective tissues of the midface and branchial regions.

Paraxial mesoderm – Early embryonic cells that are located beside and beneath the developing brain and spinal cord, and include the precursors of most skeletal muscles.

Introduction

Muscles that move and stabilize the eye have been highly conserved during vertebrate evolution. While a few remarkable adaptations have occurred, such as co-opting of the dorsal (superior) oblique muscle to generate protective heating for the brain in some fishes, these muscles have retained an anatomical organization linking axes of the body and the eye that arose hundreds of millions of years ago.

Considering their ancient status, it is logical to assume that the early development of extraocular muscles would similarly be well conserved among different species, and

therefore amenable to comparative analyses that supplant the absence of direct examination in mammals, including humans. However, compared to trunk and limb muscles, our understanding of the origins of ocular muscles and the mechanisms that initiate then maintain their development is at best fragmentary.

Myogenesis of skeletal muscles is a lengthy process, with several parameters continuing to be function-dependent throughout the life of an animal. Primary myogenesis spans the period during which populations of myoblasts – committed, mitotically active muscle progenitors – arise, emigrate to their sites of differentiation, fuse to form multinucleated innervated myofibers, and establish intimate connections with connective tissues. This population forms a scaffolding, including the delineation of global and orbital domains, within which secondary myogenesis occurs. During secondary myogenesis stages, previously sequestered latent myoblasts are activated to proliferate and differentiate, forming more than 90% of the myofibers present in mature muscles and generating region-specific specialized fiber types that in most species are present before or soon after birth.

Origins of Extraocular Muscles

Striated (skeletal) muscles throughout the body arise within paraxial mesoderm, which is located in close apposition to the embryonic brain and spinal cord. Exceptions to this are the avian striated ciliary muscle that is of neural crest origin, and possibly the striated muscles of the esophageal wall; however, both of these are involuntary. Among voluntary muscles, some of the more ventrally located branchial muscles arise from lateral mesoderm that is contiguous with paraxial mesoderm.

Many early accounts of head myogenesis placed the embryonic origin of some eye muscles, especially the lateral rectus, in the same category as branchial (pharyngeal) arch muscles, and ascribed both to a lateral mesoderm origin. These claims were based on the sites at which muscle condensations are first grossly evident in the embryo. However, with the advent of mapping methods and assays for early muscle-specific gene expression patterns, separate and distinct sites of origin for all eye muscles within preotic (i.e., located rostral, in front of the developing inner ear) paraxial mesoderm was confirmed (Figure 1).

The sites of origin of extraocular muscles parallel the sites of emergence of the three cranial motor nerves that

innervate them. However, in contrast to axial and branchial muscles, sites of myogenesis are not congruent with locations of motor nerve emergence. Indeed, each of these cranial nerves needs to elongate considerably through peripheral tissues before initial contacts with target muscles are established. Some axons, such as those of the abducens nerve, must extend longitudinally beneath the brain to reach their target lateral rectus muscles, while others, such as the oculomotor nerve fibers, project perpendicular to the floor of the brain before decussating to approach their several target muscles.

In extant vertebrates, head paraxial mesoderm constitutes a sparse population of mesenchymal cells (Figure 2). This contrasts with the situation in the neck and trunk regions, where paraxial mesoderm first forms somites, which are segmentally arranged, cuboidal aggregates of epithelial cells. As each somite matures, it becomes delineated into distinct myogenic (myotome) and connective

tissue-forming (sclerotome) regions. The most cranial somite is located beside the hindbrain, immediately caudal to the otic vesicle, and paraxial mesoderm rostral to this site fails to form epithelial tissues and lacks overt evidence of segmentation.

Head paraxial mesoderm is present adjacent to the prospective eye-forming regions of the rostral neural plate, but is largely displaced caudally as the optic vesicles emerge and expand lateral to the diencephalon. In the midline just in front of the notochord, this population of loose paraxial mesoderm cells is contiguous with a sparse and spe- cies-variable population of prechordal mesenchymal cells. Mapping experiments in avian embryos have shown that prechordal mesoderm contributes to the genesis of extraocular muscles innervated by the oculomotor nerve (Figure 3), but it is not known whether this contribution is exclusive of or supplementary to that of paraxial mesoderm. It is not known if the same occurs in mammalian embryos.

Determination of Eye Muscle Precursors

Head paraxial mesoderm contains progenitors for many tissues in addition to skeletal muscle. These include cartilages and bones associated with the braincase, loose connective tissues such as meninges and adipocytes, and endothelial cells. In contrast to somites, wherein these progenitor populations are largely segregated, it appears based on mapping studies that these diverse precursors are either intermingled or contiguous in head mesoderm.

The significance of this lies in the problem of generating diverse lineages. Somite cells are held in fixed positions relative to the dorsal and ventral parts of the adjacent neural tube (hindbrain and spinal cord) and overlying surface epithelium, all of which provide combinations of positive and negative regulators of early myogenesis and skeletogenesis.

A further complication – and one essential for the development of all craniofacial musculoskeletal systems – is the presence of a large, later-arising population of mesenchymal cells called the neural crest (Figure 2). Derived from neural folds either during (mammals) or shortly following (birds) closure of the cephalic neural tube (brain), these cells acquire a mesenchymal phenotype and quickly move peripherally, mostly atop underlying paraxial mesoderm.

Neural crest cells from the rostral midbrain level move rostrally and caudally around the optic stalk and posterior part of the optic vesicle, then spread outwardly as the vesicle is transformed into the optic cup. After delineation of the lens from the lens placode, crest cells secondarily invade the space created anterior to the lens and establish the posterior epithelium (endothelium) and stromal populations of the cornea.

In the trunk, members of the wnt family of growth factors are secreted by surface ectoderm and provide essential positive stimuli for muscle differentiation. The same are released by head surface epithelium, but here their effects are to retard myogenesis of branchial muscles. Arriving neural crest cells separates branchial muscle progenitors from the source of these negative effects and, augmented by the release of additional myogenesispromoting factors, allows myogenesis to progress.

The extent to which eye muscle progenitors follow a similar scenario is unclear. Some extraocular precursors, particularly the lateral rectus progenitors, are embedded deep within paraxial mesoderm and are therefore quite distant from both surface ectoderm and, at early stages, neural crest cells. This deep location places the lateral rectus precursors close to the neural epithelium at the level of the future metencephalon (pons).

Several experiments have established that this location provides essential cues for lateral rectus formation. When newly formed trunk paraxial mesoderm cells were excised, before they had formed somites, then grafted into the head, in place of prospective lateral rectus mesoderm cells, the transplanted cells formed a muscle that expressed molecular markers unique to the lateral rectus and established proper anatomical connections with the braincase and sclera. Small changes in the location of the implants resulted in grafted cells contributing to the dorsal oblique or branchial arch musculature. Thus, the sites within head paraxial mesoderm at which each muscle primordium forms and is specified as to its identity are highly localized.

Placing a barrier between the brain and paraxial mesoderm at this region does not prevent myogenesis, but the developing muscle cells lack molecular features that define their specific identity. Together, these experiments suggest that a rich tableau of local signals is necessary for early eye muscle differentiation, with both general myogenic and individual eye muscle-specific components.

Molecular Signatures and Muscle

Differentiation

All skeletal myoblasts use members of a closely related set of muscle-specific transcription factors to promote and coordinate their differentiation. Two of these, myf 5 and myoD, are cross-activating regulators that are among the earliest muscle-specific genes expressed. The upstream regulatory components of these genes are body regionand muscle group-specific, and serve to integrate the diverse micro-environments surrounding each muscle group with a highly shared set of outcomes, for example, activation of genes for desmin, myosins, muscle-specific actins, and junctional receptors.

Expression of myf 5 then myoD genes in eye muscle precursors generally is slightly later than expression in trunk axial muscles, but is simultaneous with that of branchial muscles (Figure 4). Expression of these regulatory genes coincides with the onset of aggregation of most muscle precursors (Figure 5), although it is not known whether these aggregates represent the totality of muscle precursors or only a subset of primary myoblasts.

By these criteria, extraocular muscles appear similar to other head and also to trunk and appendicular muscles. However, as additional features of trunk and head myogenic regulatory networks have been identified, the number of differences has exceeded the similarities, and a heretofore underlying complexity has been revealed (Table 1). This area of investigation is rapidly expanding, and rather than detail each gene currently described, a few examples of categories of differences among muscle groups will be presented.

Pax3 is a regulatory gene expressed in axial and appendicular muscle precursors, and null mutations of this transcription factor (e.g., Splotch mutation) result in severe depletion of trunk muscles. However, it is not expressed in head muscle precursors, and null mutations have no discernable effect on branchial or extraocular muscles. Another pronounced difference is in the hepatocyte growth factor (HGF) – cMet growth factor-receptor complex, which is functionally required for the correct dissemination and differentiation of appendicular and tongue muscle precursors. Again, this pathway has no known role in eye and branchial myogenesis, even though HGF is expressed in and around the precursors of branchial arch, lateral rectus, and both oblique muscles.

The latter example further illustrates the considerable heterogeneity among extraocular muscles. The lateral rectus is particularly enigmatic, being the only head muscle that expresses the ubiquitous trunk paraxial mesoderm marker paraxis, and together with the dorsal oblique, the transcription factor lbx1, which is present in trunk hypaxial (thoracic and abdominal wall) muscle precursors.

A further complexity arises from the often-changing patterns of gene expression during the early stages of head paraxial mesoderm development. The transcription factor pitx2, which is a key mid-level participant in the integrated formation of left-right asymmetry for the heart and mid-gut, is initially expressed symmetrically and ubiquitously throughout head paraxial mesoderm (Figure 6). However, a day later, during early myogenesis stages, its expression becomes more restricted but includes the first branchial arch, lateral rectus, and both oblique muscles in addition to periocular neural crest cells. Another regulatory gene, Tbx1, which is located in the region of chromosome 22 wherein deletions cause the DiGeorge syndrome, is similarly expressed over a wide domain of mesoderm (and pharyngeal endoderm) before becoming restricted to branchial arch and the lateral rectus muscles.

At present the significance of these spatially and temporally dynamic expression patterns is unknown. It is possible that early expressions presage the later focal appearance of certain muscles, but it is equally plausible that each of these genes has multiple and distinct functions associated with each stage.

As extraocular muscles mature, they exhibit a progression of fiber types, evidenced by changes in the myosin

isoforms and related contractile and energy metabolism proteins expressed. Emergence of these complex patterns requires a series of interactions among developing myofibers, surrounding connective tissues, and innervation. In the avian embryo, the primary myofibers of most extraocular muscles express embryonic slow myosin isoforms. However, one muscle, the quadratus nictitans, which is homologous to retractor bulbi muscles and is innervated

Allantois

Left Right

(a)(b)

Figure 6 At stage 8 ((a), dorsal view) Pitx2 is expressed symmetrically throughout head paraxial (and lateral) mesoderm populations, but only on the left within trunk mesoderm. By stage 21 ((b), 4 days) it is restricted to a subset of eye muscles, the first and second branchial arch muscle masses, and periocular mesenchyme.

by the accessory abducens nerve, the myofibers are either completely fast myosin expressing, as in the quail embryo, or mixed, as in the chick embryo (Figure 7). To explore the basis for these distinct, species-specific patterns, periocular neural crest cells of the chick were replaced with comparable populations from a quail embryo. The quadratus muscle in these chimeric embryos exhibits the quail donor phenotype (Figure 7), indicating that initial differentiation of fiber types requires interactions among myofibers and encompassing connective tissues.

Muscle Morphogenesis

Except for muscles that remain closely associated with the vertebral column and skull, all muscle progenitors leave their sites of origin in paraxial mesoderm and disperse into peripheral tissues. Body wall muscles do so together with sclerotome-derived connective tissue precursors, and maintaining these close spatial relations is essential

for the morphogenesis of thoracic and abdominal muscles. For appendicular myogenesis, lateral myotome-derived cells move in sequential waves of primary and secondary myoblasts into nearby limb buds, where they form longitudinal bands of future dorsal and ventral groups before segregating into individual muscles. In the hindlimbs, some of these undergo secondary dispersal to form muscles of the perineal region.

Branchial muscles are comparable to body wall muscles in that they initially exhibit a constant juxtaposition with the precursors of their connective tissues, which in the head are all derived from neural crest cells, and also with the motor nerves that innervate them. This maintained registration permits continuous exchange of signals among all three components during all stages of muscle differentiation and morphogenesis.

This constant contiguity has most dramatically been demonstrated for the trapezius muscle, whose precursors arise among caudal branchial arch mesodermal populations and secondarily shift caudally to attach to the scapula. Mapping studies in both bird and mouse embryos revealed that the neural crest-derived connective tissues move with, and perhaps somewhat in advance of, these myoblasts and indeed contribute to the scapula. This recapitulates an ancestral condition in which the forelimb girdle articulated with the back of the skull, as is still present in fish.

Again, however, the extraocular muscles exhibit a developmental scenario unlike any other muscles. As illustrated in Figures 5 and 8, these muscle precursors move towards, then around, the equatorial region of the developing eye to assume their definitive locations. During this process, each muscle leaves the company of surrounding mesoderm cells and becomes fully encompassed by neural crest cells, which secondarily penetrate the muscle mass and form internal (e.g., endomysium) as well as surrounding (perimysium, fascia, and tendon) connective tissues.

These periocular crest cells need not have originated at the same axial level as the muscles. For example, the lateral rectus muscle, the neural crest cells that will form

interactions among contiguous progenitor populations, as occurs for branchial musculoskeletal systems.

The mechanisms by which aggregates of extraocular muscle primordia change both absolute and relative positions remain enigmatic. There is no precedence elsewhere in the embryo for condensations of cells moving actively through surrounding tissues. However, several lines of evidence support a model based on passive displacement of eye muscle primordia. As was shown in Figure 2, the interface between neural crest and myogenic paraxial mesoderm is initially a flat plane. Changes in the relative positions of the eye due to flexures and differential growth of the brain and expansion of the optic cup introduce distortions in this plane, but the extent to which this might affect individual eye muscles has been difficult to define.

In screening for a wide range of gene expression patterns, several were found that coincided with the patterns of movements taken by some extraocular muscles (Figure 9). These reveal a set of localized distortions of the neural crest-mesoderm interface. Finger-like projections of paraxial mesoderm expand dorsally and caudally around the optic cup, becoming interdigitated with periocular neural crest populations and passively bringing the dorsal and ventral oblique muscle primordia to their definitive