Evidence for Rectus Extraocular Muscle Pulleys in Rodents

Abstract

purpose. Extraocular rectus muscle (EOM) pulleys are important determinants of orbital biomechanics in humans. In this study, the authors evaluated orbital connective tissue morphology, specifically characterizing rectus muscle pulleys, in the rat, a species with laterally placed eyes, afoveate vision, and a less complex visuomotor repertoire than primates.

methods. Adult rat orbits were paraffin processed and serially sectioned for histochemical and immunohistochemical staining. Frozen sections of enucleated globes with intact EOMs and associated connective tissue were also studied with myosin immunohistochemistry and histochemistry for the mitochondrial enzyme, nicotinamide adenine dinucleotide (NADH)-tetrazolium reductase, to delineate the orbital layer relationship with the pulley tissue.

results. Focal condensations of collagenous connective tissue were found in relationship to the rectus muscles in the equatorial Tenon’s fascia, similar to those described as human recti muscle pulleys. The fibroelastic pulley rings were coupled to adjacent EOM pulleys by bands containing collagen and elastin. The coupling of pulleys to the orbital walls was significantly less than that previously described in humans. As in humans, there was a dual insertion of rodent rectus muscles, with the orbital layer inserting on the muscle pulley and the global layer attaching to the sclera.

conclusions. The data support the presence of structures in the rat orbit that are the morphologic equivalent of the human rectus pulley system. Although rodent and human pulleys were similar in many respects, there were species-specific properties that may relate to established differences in orbital anatomy and/or visuomotor behavior. These data extend the rectus muscle pulley concept to rodents and may provide insight into pulley structure–function relationships.

The oculomotor system is one of the best understood of the somatic motor systems. However, until recently, the intrinsic properties of the peripheral components of this system have been ignored. ¹ It is now clear that neural and mechanical models of eye movements can no longer neglect properties of the extraocular muscles (EOMs) and the orbital fasciae, which influence the EOM paths in the orbit. 

Radiographic studies in monkeys and humans first suggested that the rectus muscle bellies are stabilized relative to the orbit during shifts in gaze. ^{2 3} The stability of EOM paths during eye movements was confirmed in magnetic resonance imaging (MRI) studies by Miller. ⁴ It was then hypothesized that the orbital fascia connecting the muscle to the orbital wall could act as a pulley, thereby preventing muscle sideslip during ocular rotations. ^{4 5} Recent anatomic studies of the human orbit indicate that the pulleys consist of rings of collagen and elastin encircling each rectus muscle near the equator of the globe. Pulleys are tightly coupled to the orbital wall by musculofibroelastic septa. ⁵ Similar orbital connective tissue organization also has been found in monkeys. ⁶  

Improvements in MRI resolution have allowed direct visualization of rectus muscle pulley tissue relative to other orbital structures. ^{7 8} By defining the functional origins of rectus EOMs, pulleys are a direct determinant of muscle actions. Computer simulations of orbital mechanics have shown that binocular alignment and eye movement kinematics are highly sensitive to rectus pulley location. ⁹ Moreover, pulleys apparently shift along the muscle axis, altering the point of functional origin to meet the constraints of different classes of eye movements. ⁸ A mechanism by which EOMs coordinate movement of both the eye and pulley was proposed by Demer et al. ⁸ as the active-pulley hypothesis. Data in support of this hypothesis show that the orbital layer of each EOM inserts into the pulley tissue, to optimally position it during globe rotations, whereas the global layer inserts into the sclera as the effector of movements of the eye proper. 

Previous studies have shown that whereas EOM fiber types are largely conserved across mammalian species, orbital layer fiber types exhibit significant specialization, with mitochondrial content increasing in higher mammals. ¹⁰ Because the orbital layer’s properties correlate with the increasing complexity of visual sensory and oculomotor control systems, we predicted that rectus muscle pulleys would also exhibit simpler morphology in species with less complex visuomotor demands. To investigate this prediction and to better understand the structure–function correlation of the rectus muscle pulleys, the present study examined the morphology of orbital connective tissue organization in rat in relation to the EOM and its layers. 

Methods

These studies used adult Sprague-Dawley rats (Harlan, Indianapolis, IN). All animal procedures were approved by the Institutional Animal Care and Use Committee and were in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Gross dissection was performed on two rat orbits in situ after the rats were asphyxiated with carbon dioxide. An additional 10 rat orbits were studied using paraffin histology and 8 enucleated specimens were used for frozen-section immunohistochemistry and histochemistry. 

Paraffin Histology

Rats were anesthetized with intraperitoneal injection of ketamine HCl and xylazine HCl. Animals were then perfused transcardially with physiological saline followed by 10% formalin. The rat orbits were dissected en bloc, retaining the relationship between periorbital tissues and the bony orbit. The orbits were then decalcified in formic acid-formaldehyde solution for 48 hours to soften the orbital bone for sectioning. Subsequently, they were dehydrated through graded solutions of alcohol and xylene and infiltrated with paraffin, using a vacuum chamber to remove air bubbles. Serial 8-μm sections, coronal and longitudinal, were cut and mounted on albumin-coated superfrost glass slides. Masson’s trichrome stain was used to define skeletal muscle, smooth muscle and collagen, and alternate sections were stained with van Gieson’s stain to delineate elastin. Smooth muscle was confirmed by immunohistochemistry with a fluorochrome-tagged smooth muscle–specific actin antibody (Accurate Chemical Co., Westbury, NY) in 10 μg/ml concentration. Slides were examined and photographed under a microscope (BX-50; Olympus, Lake Success, NY) equipped with bright-field, differential-interference contrast and fluorescence optics, a video camera (MTI/Dage, Paramus, NJ), and a computer (G3 Macintosh; Apple Computer Corp., Cupertino, CA) equipped with a frame grabber (Scion, Frederick, MD). 

Immunohistochemistry and Histochemistry

Rats were asphyxiated using carbon dioxide. Fresh specimens of globe with intact EOMs, associated connective tissue, and the harderian gland were obtained by orbital dissection, freeze protected using 30% sucrose, and snap frozen using isopentane cooled in liquid nitrogen. Cryostat sections (8 μm thick) transverse to the globe axis were collected and mounted on coated slides (Vectabond; Vector Laboratories, Inc., Burlingame, CA). To delineate orbital and global layers, immunohistochemistry was performed using primary antibody to developmental myosin heavy chain (1:25 dilution, Novacastra Laboratories, Burlingame, CA), ¹¹ with localization by fluorochrome-tagged secondary antibody. Sections were examined under the microscope using fluorescence optics. The developmental myosin highlights the orbital layer, selectively labeling the proximal and distal segments (i.e., away from the neuromuscular junction band) of both orbital singly and multiply innervated fiber types. ^{12 13 14} Adjacent sections were examined for nicotinamide adenine dinucleotide (NADH)-tetrazolium reductase activity, a mitochondrial marker. The orbital layer fibers have a high mitochondrial content and can be distinguished by their location and darkly stained, small-fiber appearance. ¹⁵  

Results

In this study, rat orbits were examined to identify any specializations of the orbital connective tissue in relation to the rectus muscles that could function as pulleys. The rat orbit is different from the primate orbit, in that it has incomplete bony walls. The six EOMs that are seen in the human orbit are present in the rat, in addition to a retractor bulbi muscle. Conspicuous by its extensive intraorbital extent is the harderian gland. Although this gland fills much of the rodent orbit, influencing the structural organization of the orbital connective tissue, its exact function is not known. ¹⁶  

Gross Morphology on Orbital Dissection

Approaching the orbit anteriorly, the posterior Tenon’s fascia was seen to extend to the anterior orbital rim or, in its absence, to the harderian gland or the lateral periorbital muscle. The attachment to the orbital rim was particularly thick medially at approximately 3 mm posterior to the medial rectus insertion. At this level, the medial rectus muscle was seen to run through a dense connective tissue ring. Posteriorly, only loose attachments to the harderian gland were seen. When the globe was passively rotated, the muscles bent at the level of the thick equatorial Tenon’s fascia, suggesting that this tissue stabilizes the rectus muscle bellies relative to the orbit. 

Histologic Findings

Whole 8-μm-thick coronal and longitudinal sections of the entire orbit were cut from the cornea to the retrobulbar region. Sets of adjacent sections were stained with Masson’s trichrome and van Gieson’s stain, respectively. Adjacent coronal sections through the equatorial region of the globe are shown in Figure 1 . In the periequatorial region, the rectus muscles were surrounded by dense connective tissue, forming encircling sleeves. The sleeves do not all enter the plane of Figure 1 . Figure 2 shows each rectus muscle sleeve at its point of maximum development. The medial rectus sleeve exhibited the densest connective tissue development. Higher magnification views of the medial rectus sleeve (Fig. 3) show the dense mass of collagen (approximately two times the thickness of sclera), the interspersed elastin fibrils (stained black with van Gieson’s), and smooth muscle cells, confirmed with an antibody to smooth muscle actin. These components were concentrated in association with the orbital surface of the rectus EOMs. The rectus muscle sleeves exhibited progressive thinning posterior to the equator (Figs. 4 5) . In sections posterior to the globe–optic nerve junction, the rectus muscles were surrounded by either thin sleeves, containing small amounts of collagen and elastin and no smooth muscle cells, or the sleeves were altogether absent (Fig. 5A) . In the middle of each rectus muscle sleeve, the collagen was associated with densely packed elastin fibrils (Fig. 4) . 

Adjacent muscle sleeves were interconnected by fibroelastic bands, resulting in globe-encircling connective tissue at the equator (Fig. 1) . In addition to the interconnections of all four rectus muscle sleeves, there were some fibroelastic extensions from the anteriormost parts of the medial and lateral rectus muscle sleeves to the bony medial orbital wall and the lateral fibromuscular membrane that delimits the lateral orbit, respectively. At other sites, we observed only loose attachments of muscle sleeves to the harderian gland. The harderian gland proper was enveloped in a thin fibroelastic capsule, with extensions to both the perimuscular sleeves and intermuscular sheaths. The slips of the retractor bulbi muscle exhibited only very thin fibroelastic connective tissue, mainly on the orbital surface of each slip. 

Smooth muscle was found as a circumferential band in the equatorial Tenon’s fascia. This band began as lateral and medial extensions of the striated fibers of the levator palpebrae superioris muscle. One extension passed laterally and inferiorly, lateral to the lateral rectus pulley and the inferior oblique muscle (Fig. 1) . The medial band passed inferomedially to blend with the medial rectus pulley. Anteriorly, the smooth muscle band was continuous with superior and inferior palpebral smooth muscles. Posteriorly, smooth muscle was continuous up to the point of the globe–optic nerve junction. A separate sheet of smooth muscle was seen in the retrobulbar region, effectively forming a fibromuscular capsule in the inferolateral orbit, where the bony orbital wall is missing. This smooth muscle extended deep into the orbital apex. Smooth muscle at this site corresponds to the Müller’s orbital muscle described in humans. 

From these data, we infer that the globe, the harderian gland, and the EOMs in rats are supported by posterior Tenon’s fascia. In the retrobulbar region, Tenon’s fascia is very thin, permitting free movement of the muscle bellies and the optic nerve during globe rotation. By contrast, in the periequatorial region there are dense fibroelastic muscle sleeves. The muscle sleeves are stabilized by dense fibroelastic extensions to the adjacent EOM sleeves and peripheral posterior Tenon’s fascia, which, in turn, is supported by smooth muscle and elastin. Anteriorly, the sleeves have some fibrous extensions to the orbital wall or, in its absence, to the peripheral fibrous tissue. This patterned orbital connective tissue organization directly corresponds to the previously described primate rectus muscle pulleys. Thus, in the rat, the rectus muscle sleeves and their extensions represent rectus muscle pulleys. 

Orbital Layer of EOM in Relation to the Pulley Connective Tissue

The orbital layer of the rectus EOMs was identified by small fiber size, immunoreactivity of fibers for developmental myosin heavy chain (an orbital layer marker), and NADH-tetrazolium reductase enzyme activity, indicative of the high mitochondrial content of orbital layer fibers. The orbital layer fibers thus identified constituted nearly 50% of rectus muscle fibers. Neither the rectractor bulbi nor the levator palpebrae superioris exhibited an orbital muscle layer, and neither had a muscle pulley. We observed that the orbital layer ended before the myotendinous junction and the scleral insertion. The distal reduction in thickness and loss of the orbital layer corresponded with the anterior–posterior extent of the connective tissue sleeves or rectus muscle pulleys (Fig. 6) . As the orbital layer thinned and the pulley connective tissue thickened, there was an intermingling of muscle fibers and connective tissue, strongly suggesting that the orbital layer inserts into the pulley (Fig. 7) . 

Discussion

A re-examination of the human orbital connective tissue arrangement, in the context of EOM path stability relative to the orbit, led to the concept of fibroelastic pulleys associated with human rectus muscles. ⁵ Pulley position in the orbit is stereotypic, as a result of intercoupling between adjacent pulleys and with the orbital walls. Abnormal location of the pulleys is a major cause of strabismus. ^{17 18 19} Therefore, the specialized organization of orbital connective tissue is of immense importance in ocular kinematics. 

In this study, we examined the organization of orbital connective tissue in relation to EOMs in the rat, a species with laterally placed eyes and an oculomotor repertoire simpler than that of primates. We have shown that the equatorial Tenon’s fascia exhibits thickenings consisting of dense collagen interspersed with elastin. The fascia was organized to form perimuscular sleeves that are structurally analogous to primate rectus muscle pulleys. Rat pulleys were loosely coupled to the surrounding harderian gland but had dense fibroelastic extensions to adjacent pulleys. Extensions from the pulleys to the orbital rim were limited to the anteriormost aspect of the pulley and contained little elastin or smooth muscle. In the posterior orbit, the rectus muscles exhibited only thin connective tissue, comparable to an epimysium. The two EOM layers had divergent insertions, with the orbital layer inserting on the pulley and the global layer attaching to the sclera. 

Demer et al. ^{5 6} first reported the presence of dense fibroelastic rectus muscle pulleys in the primate orbit. Our data demonstrate that rat pulleys are located in a similar orbital position, at the equator of the eye, and the muscle sleeves progressively thin down to simple muscle ensheathments in the posterior orbit. In humans and monkeys, and now shown in rats, the medial rectus pulley is the best developed of the four pulleys, and there is substantial coupling between adjacent pulleys by dense fibroelastic bands. Although the connective tissue organization of the rat orbit largely parallels that of humans and monkeys, there are some differences. Human and monkey studies describe dense and organized collagen fibrils suited to high internal rigidity and abundant elastic fibrils essential for extensibility in the pulley tissue. ^{5 6 20} We found less dense collagen and less interspersed elastin in the rat pulley structure, showing that there are species differences in the internal organization of the pulleys. Human pulleys incorporate well-developed musculofibroelastic struts that suspend them from the orbital walls. Such highly developed suspensory struts were not observed in the rat, although the limited fibroelastic extensions to the orbital wall may serve the same purpose in rodents, which have a narrower oculomotor range and lower muscle forces. The presence of the massive harderian gland, the incomplete bony orbit, and the consequent need for a protective globe retraction reflex in rodents may contribute to the limited development of pulley struts in rodents. Collectively, these data establish that rectus muscle pulleys are a feature of the mammalian orbit that is conserved, regardless of the placement of the eyes (frontal or lateral) or presence of retinal specialization (foveate or afoveate). Findings also reinforce the notion that pulley constituents may exhibit variable organization in different species. ⁶ Finally, our data also show that tight interpulley coupling is a highly conserved feature and suggest that this aspect of pulley morphology may be more significant than previously appreciated. Interpulley coupling may provide stability relative to the orbit sufficient to achieve the kinematic function of the pulleys. 

Two smooth muscle bands in the equatorial Tenon’s fascia, running from the superior rectus pulley to the lateral rectus pulley and from the medial rectus to the inferior rectus pulley, have been described in both humans and monkeys. ⁶ These bands constitute Müller’s peribulbar muscle and exhibit a rich autonomic innervation. In the rat orbit, smooth muscle distribution was similar along the lateral levator aponeurosis, but the rat did not have the extension that connects the medial and inferior rectus pulleys. In humans, to this medial band of smooth muscle has been attributed a dynamic role in fine tuning the vertical recti pulley positions in convergence. ⁶ Its underdevelopment in the rat orbit may be related to the absence of vergence eye movements. The distribution of smooth muscle in the rat orbit has been described previously by Page. ²¹ Although Page did not qualify a distinct peribulbar muscle, he showed that the inferior palpebral muscle and the superior palpebral muscle occupy a wide circumferential course, almost enclosing the globe. We also noted the equatorial smooth muscle band (that is equivalent to the human peribulbar muscle) to continue into the lids as the palpebral muscle. 

Demer et al. ⁸ recently advanced the notion that the rectus muscle pulleys must be actively translated along the muscle axis during eye movements. The active-pulley hypothesis is based on evidence that the two layers of the EOM represent functionally distinct compartments, with the orbital layer inserting into the pulley tissue and the global layer continuing forward to insert onto the sclera. This arrangement allows degrees of coordination and potential independence in movement of the pulley and eye and has received support from MRI studies. ⁸ We evaluated the structural basis of the active-pulley hypothesis in rats, using the established orbital muscle layer criteria of small muscle fiber size, immunoreactivity of distal fiber segments for developmental myosin heavy chain, ^{12 13 14} and oxidative enzyme staining for the high mitochondrial content of orbital fibers, ¹⁵ to define the extent of the orbital layer with a high level of precision. This approach identified similar compartmentalization of muscle layer insertions in rodent EOM, suggesting that rodent eye movements also depend on the active-pulley model. However, at the level of resolution of histologic technique, we cannot exclude the possibility that some orbital fibers may still act directly on movements of the globe (e.g., through fiber branching, myomyous junctions, or insertions into connective tissue that merges with the global layer). 

There is a potentially important correlation between orbital layer fiber morphology and species-specific differences in pulley tissue development. Although the six EOM fiber types are conserved in mammals, orbital layer fibers exhibit higher mitochondrial content in those species with more complex visuomotor behavior. ¹⁵ By contrast, global layer fiber types are virtually identical across species. ^{10 15} Because pulley development appears to follow the same pattern, the more highly developed pulley struts and broader range of required pulley anterior–posterior translations in primates may require orbital fibers with higher oxidative capacity and therefore higher fatigue resistance. The rat therefore has correspondingly less developed rectus muscle pulleys and orbital layer fibers with considerably lower mitochondrial content than do primates. 

Functionally, the rat is an afoveate animal, with a much smaller field of binocular vision and a relatively poorly developed orbital layer, when compared with humans. Rodent rectus muscle pulleys must be evaluated in the context of lateral placement of the eyes, visual acuity, and eye movement repertoire. Eye movements in the rat differ in several respects from those in humans. ^{22 23 24 25 26 27} Compensatory eye movements (vestibuloocular reflex and optokinetic reflex) are more important in rats than are voluntary eye movements, whereas voluntary movements of the eyes are exclusively saccadic (afoveate saccades), with maximum amplitude of approximately 20°. ^{22 24} Smooth pursuit and vergence are absent in rat, and voluntary saccades differ from those in foveates, in that they are typically accompanied by target-directed head movements. Techniques for accurate recording of eye movements in small rodents are only now in development, ²⁸ and it is not yet clear whether eye movement mechanics are similar to those in better-studied primates. We suggest that differences in pulley structure between rats and humans should be viewed in relationship to these functional differences between the two species. Subsequent comparative studies should address the relationship between species variations in pulley morphology and oculomotor requirements, such as maintenance of commutativity of the oculomotor plant and implementation of Listing’s law. ^{8 29 30 31}  

In conclusion, our findings extend the rectus muscle pulley concept to rodents. These data establish that muscle pulleys are a phylogenetically conserved and central feature of mammalian orbital mechanics. Moreover, comparative anatomic analysis of rectus muscle pulleys suggests a correlation between pulley complexity, orbital structure, and visuomotor behavior. 

 Supported by grants from the Knights-Templar Eye Research Foundation, Chicago, IL (SK); Grants R01-EY09834 (JDP) and P30-EY11370 from the National Institutes of Health; the Evenor Armington Fund, Cleveland, OH (JDP); and a Departmental Grant and a Senior Scientific Investigator Award (JDP) from Research to Prevent Blindness. JDP is the Carl F. Asseff, MD, Professor of Ophthalmology.

 Submitted for publication November 28, 2000; revised March 5, 2001; accepted March 29, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Photomontages of adjacent coronal, 8-μm-thick sections of rat orbit near the equator of the globe demonstrating the pulley tissues. (A) Masson’s trichrome stain demonstrates collagen sleeve (blue) around the medial rectus muscle (arrow). (B) van Gieson’s stain demonstrates the elastin in black (arrow). Arrowheads: Smooth muscle. IO, inferior oblique; IR, inferior rectus; MR, medial rectus; LR, lateral rectus; RB, retractor bulbi; SR, superior rectus, LPS, levator palpebrae superioris; SO, superior oblique; HG, harderian gland.

Figure 2.

 

Coronal 8-μm-thick sections taken through the most developed part of each of the rectus muscle pulleys: superior (A), lateral (B), medial (C), and inferior (D). A thick collagen sleeve appears, particularly on the orbital side of each rectus muscle (arrows). The pulley collagen is densest for the medial and superior recti and is relatively less developed for the inferior rectus. The retractor bulbi (RB) and the levator palpebrae superioris (LPS) have a very thin sleeve.

Figure 3.

 

Adjacent coronal, 8-μm-thick orbital sections at the level of the medial rectus pulley. (A) Masson’s trichrome stain demonstrates a thick collagen sleeve (blue) on the orbital aspect of the medial rectus muscle. (B) van Gieson’s stain demonstrates elastin (black) interspersed in the collagen. (C) High-magnification view of (A) shows smooth muscle cells (sm) scattered in the pulley tissue. (D) Immunofluorescent stain with smooth muscle actin antibody confirms the presence of smooth muscle cells (bright green). Arrowheads denote thickness of the pulley (p). Scale bar in (D) also applies to (A) and (B).

Figure 4.

 

High-power photomicrographs of interrupted coronal sections showing the medial rectus pulley. The sections represent periequatorial-to-retrobulbar locations in order (A–D, E–H). In each pair (e.g., A, E; B, F), adjacent sections were stained with Masson’s trichrome stain (A–D) and van Gieson’s elastin stain (E–H). Dense elastin fibrils were observed in the pulley tissue (E–G, black fibers). Also note the thinning of the orbital layer (o) distally, as pulley tissue thickens. g, global layer.

Figure 5.

 

(A) Photomontage of low-power views of a coronal 8-μm-thick section taken through the rat orbit at the globe–optic nerve junction stained with van Gieson’s elastin stain. Scant connective tissue is present around the muscles and the extensive gland. (B) Longitudinal section stained with Masson’s trichrome stain. The perimuscular connective tissue (arrow) is thickest on the orbital side of the superior rectus anteriorly and markedly thins down posteriorly. See Figure 1 for abbreviations.

Figure 6.

 

Interrupted serial coronal frozen sections taken through the medial rectus muscle and its pulley (anterior to posterior from left to right). Adjacent sections were stained with fluorochrome-tagged developmental myosin heavy chain antibody (A–D) and for NADH-tetrazolium reductase activity (E–H). The orbital layer is distinguished by positive fluorescent staining in (A–D) and by darker staining band on the orbital surface in (E–H). Only some orbital fibers stained positively for the developmental myosin antibody in (D). Although myosin expression patterns vary along the length of orbital fibers, leading to this inconsistent staining pattern, orbital fibers are readily distinguished by intense NADH-tetrazolium activity in adjacent section (H). The thinning of the orbital layer anteriorly corresponds to the thickening of the pulley tissue. ct, pulley thickness indicated by white bars.

Figure 7.

 

(A) Photomicrograph of cross-section of the inferior rectus (IR) showing orbital layer fibers inserting into the pulley collagen (arrowheads). (B) Photomicrograph of longitudinal 8-μm-thick section showing the superior rectus (SR) and its pulley. Note that the orbital fibers of the muscle insert into the thickest part of its associated connective tissue, which constitutes the SR pulley (arrow). LPS, levator palpebrae superioris; RB, retractor bulbi; O, orbital layer; G, global layer.

The authors thank Anita Merriam, Alan Lee, and Jennifer DeVecchio for technical assistance; John Stahl, John Leigh, and Francisco Andrade for helpful discussions and for reading an earlier version of the manuscript; and Joseph Demer for assistance with histologic technique and data interpretation.