Human Extraocular Muscles: Unique Pattern of Myosin Heavy Chain Expression during Myotube Formation

Abstract

purpose. To study the myosin heavy chain composition of the human extraocular muscles (EOMs) during development.

methods. EOMs from human fetuses of 8 to 22 weeks of gestation were studied with immunocytochemistry and gel electrophoresis. Antibodies specific against nine isoforms of myosin heavy chain (MyHC) were used in serial frozen sections.

results. The developing EOMs had a delayed time course of myotube formation and a unique composition and distribution of MyHCs compared with human limb skeletal muscle. The primary myotubes coexpressed two developmental isoforms of MyHCI from the earliest stages. The third developmental MyHCI delineated the future orbital layer at 10 to 12 weeks of gestation. MyHC-slow tonic also appeared early, whereas MyHCα -cardiac and MyHC-extraocular, important components of adult EOM, were never detected at the gestational ages studied.

conclusions. The developmental features of the EOMs differed significantly from those reported for limb muscles of the corresponding ages. It is clear that the knowledge of limb muscle development does not fully apply to more specialized muscles, such as the eye muscles. The extreme complexity displayed by the EOMs probably reflects their distinct embryonic origin, innervation, and regulatory program of myogenesis.

The human body contains more than 300 muscles that differ widely in size, shape, strength, and resistance to fatigue. Still, all muscles are basically constructed in a similar way and consist of long multinucleated cells called muscle fibers, surrounded by connective tissue in which nerves and blood vessels are embedded. ^{1 2 3} The muscle fibers of the limb and trunk are divided into two major classes: slow-twitch, or type I, fibers and fast-twitch, or type II, fibers. The latter are often further subdivided into type IIa and IIb fibers. ^{2 3} However, there are several classification systems reflecting the complexity in structure and function of the different muscle fibers. ²  

The functional diversity among muscle fibers is primarily due to their myosin heavy chain (MyHC) composition. The MyHC exists in multiple isoforms and contains the adenosine triphosphatase (ATPase) and actin-binding sites and thereby dictates contractile velocity and contraction force. ^{3 4} The major MyHC isoforms present in mature human muscles are the slow MyHCI in type I fibers, the fast MyHCIIa in type IIa fibers, and, in contrast to other species, the fast MyHCIIx in type IIb fibers. ^{3 5}  

Muscles of the craniofacial region are structurally and functionally more specialized than the limb muscles. ^{6 7 8 9} The extraocular muscles (EOMs) represent a highly specialized group. ^{10 11} The fibers of these muscles differ from those of the limb muscles in that they are smaller, loosely arranged, and belong to very small motor units and are organized into two distinct layers: the orbital and global layers. ¹¹ The EOMs are considered to be among the fastest muscles in mammals, and their repertoire of fast MyHCs includes an isoform that is specific to the EOMs. This isoform was initially described at the mRNA level in the rat ¹² and has tentatively been identified in human EOMs by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). ¹³ In addition, the EOMs also contain slow tonic, ^{14 15} α-cardiac, ⁹ and developmental MyHC isoforms. ^{12 16}  

Most of our current knowledge about muscle development is based on studies of limb muscles, ¹⁷ which originate from the somites and are innervated by spinal nerves. The development of human limb musculature occurs in a biphasic manner by fusion of mononucleated myoblasts into multinucleated myotubes. Primary myotubes initially express MyHCembryonic (MyHCemb), MyHCfetal, and MyHCI. ^{17 18 19} The secondary myotubes give rise to most of the muscle fibers in adult tissue and express MyHCemb and MyHCfetal but not MyHCI. ^{17 18 19} Three different slow isoforms (here designated MyHCI/1^st, MyHCI/2^nd, and MyHCI/3^rd + IIa) are sequentially expressed in the primary myotubes. ¹⁹ The first MyHCI isoform appears at 6 to 8 weeks of gestation (wg). By 10 wg the second MyHCI isoform can also be detected. The third MyHCI isoform is additionally expressed in primary myotubes from 14 wg. ¹⁹ As the primary and secondary myotubes mature, expression of the developmental MyHCs decreases, and later on they exclusively express slow and/or fast MyHC isoforms. ^{18 19 20}  

Knowledge about the developmental events that give rise to more specialized muscles is scarce, particularly in human. The purpose of this study was to provide a comprehensive view of the development of the human EOMs, with special emphasis on the pattern of MyHC expression during myotube formation. 

Materials and Methods

Human extraocular and limb muscles were obtained from legal abortions at approximately 8, 10, 12, 14, 18, and 22 wg, with the approval of the Ethics Committee of the Medical Faculty, Umeå University, and in compliance with the Declaration of Helsinki. Gestational age was dated from the first day of the last menstrual period and was further confirmed by ultrasound before abortion in most cases. The samples were rapidly frozen in propane chilled in liquid nitrogen and stored at −80°C until use. Serial cross sections (5μ m thick) were cut in a cryostat (Reichert–Jung, Leica, Heidelberg, Germany). Specimens of the EOMs and of the limb muscles collected from the same one (8 wg) to two fetuses were examined at each gestational age. Limb muscle samples from an additional two to three fetuses of similar ages were also examined. 

Sections were processed for immunocytochemistry with previously characterized monoclonal antibodies (mAbs), ^{19 21 22 23 24 25 26 27} each recognizing distinct MyHC isoforms (Table 1) . The specificity of mAb 4A6 against MyHCextraocular was further assessed in human tissue with immunocytochemistry. This mAb did not react with adult limb muscle (MyHCI, MyHCIIa, MyHCIIx), heart muscle (MyHC α-cardiac), fetal limb muscle (MyHCemb, MyHCfetal), muscle spindles (MyHCif ¹³ ) or chicken anterior latissimus dorsi muscle (MyHCsto). 

The tissue sections were processed for immunocytochemistry, as previously described, ²⁵ using the indirect peroxidase–antiperoxidase complex (Dako, Copenhagen, Denmark) technique to visualize bound antibody. Incubation with the appropriate primary antibody was performed for 1 hour at 37°C. Development of peroxidase was obtained by applying a solution containing 1 mg/ml diaminobenzidine and H₂O₂ for 5 to 10 minutes, followed by rinsing in running water. Incubation with the primary antibody was omitted in control sections. No staining was observed in these sections. 

Whole-muscle extracts were prepared from frozen samples of adult EOM and 12-wg and 20-wg fetal limb muscles. ²⁸ Thick sections (30 μm) were cut from a 10-wg specimen and freeze dried, and the EOMs were microdissected under a stereomicroscope (Wild, Heerbrugg, Switzerland). From the 14- and 22-wg EOM samples, two 10-μm-thick frozen sections were used. SDS-PAGE was performed ²⁹ (Mini Protean II; Bio-Rad, Glattbrug, Switzerland). The gels were run for 22 hours at 75 V, with the lower two thirds of the gel unit surrounded by a 16°C water bath. The gels were then silver stained ³⁰ and photographed. 

Results

Immunocytochemistry

The mAbs 4A6 against MyHCextraocular and F88 against MyHCα -cardiac did not stain the (EOMs) at 8, 10, 12, 14, 18, and 22 wg, although they stained adult samples (not shown). 

Eight Weeks of Gestation.

The EOMs were identified as tissue condensations containing cells of various sizes and shapes stained with anti-MyHCemb (Figs. 1 2A ), anti-MyHCfetal, anti-MyHCI/1^st, and MyHCI/2^nd (Figs. 2A 2B 2C 2D) . It was difficult to visualize the same individual muscle cells from section to section, but in each section, all cells appeared to be labeled by these mAbs (Fig. 2) . 

Ten Weeks of Gestation.

The EOMs had a more condensed appearance (Fig. 3) , and most of the muscle cells appeared to be typical myotubes of various sizes. All myotubes were stained with anti-MyHCemb, anti-MyHCfetal, anti-MyHCI/1^st, and anti-MyHCI/2^nd (Fig. 3A 3B 3C 3D) , although heterogeneously with the two latter monoclonal antibodies. Staining with anti-MyHCI/3^rd + IIa (Fig. 3E) was weak or absent in most of the myotubes, but in the periphery of the muscles there were a few myotubes of various sizes that were strongly stained, both with anti-MyHCI/3^rd + IIa, anti-MyHCI/1^st, and anti-MyHCI/2^nd. The antibody against MyHCIIa did not label any myotubes at this age, and therefore the staining obtained with anti-MyHCI/3^rd + IIa revealed the presence of MyHCI/3^rd. Anti-MyHCsto (Fig. 3F) stained some myotubes strongly, mostly in the interior part of the muscles and some lightly, predominantly in the periphery. 

Twelve Weeks of Gestation.

All myotubes were labeled by anti-MyHCemb (Fig. 4A ), anti-MyHCfetal, anti-MyHCI/1^st, and anti-MyHCI/2^nd (not shown). Staining with anti-MyHCI/3^rd + IIa and anti-MyHCsto (Fig. 4B 4C) showed that the myotubes were organized into orbital and global layers. Nearly all myotubes in the orbital layer were strongly stained with anti-MyHCI/3^rd + IIa, whereas the myotubes in the global layer were either unstained or very lightly stained. The staining pattern observed with anti-MyHCsto was opposite that of anti-MyHCI/3^rd + IIa, and it revealed myotubes of either large or small diameter. These differences in diameter could be observed in several sections and did not appear to be contraction artifacts. Anti-MyHCIIa did not label any myotubes, and therefore the staining observed with anti-MyHCI/3^rd + IIa corresponded to MyHCI/3^rd. 

Fourteen Weeks of Gestation.

Anti-MyHCfetal (Fig. 5A ) revealed two distinct populations of myotubes: moderately stained, generally larger, myotubes and strongly stained, generally smaller, myotubes found in close apposition to the larger myotubes. Anti-MyHCI/1^st (Fig. 5B) and anti-MyHCI/2^nd (not shown) exclusively labeled the myotubes moderately stained with anti-MyHCfetal. The smaller myotubes described earlier were interpreted as typical secondary myotubes because of their close apposition to the larger primary myotubes and because of their MyHC content. In contrast to that observed earlier in gestation, anti-MyHCI/3^rd + IIa (not shown) also stained lightly all myotubes in the interior of the EOMs. The myotubes stained with anti-MyHCsto (Fig. 5C) corresponded to those stained with anti-MyHCI/1^st and anti-MyHCI/2^nd. 

Eighteen Weeks of Gestation.

The total number of myotubes had increased greatly. Small, secondary myotubes found in proximity with many of the large, primary myotubes were very strongly stained with both anti-MyHCemb (Fig. 6A ) and anti-MyHCfetal (Fig. 6B) . In the global layer, the larger myotubes were unstained or very lightly stained with anti-MyHCfetal, whereas in the orbital layer all myotubes were stained, although to different degrees. The larger myotubes in both layers corresponded to those labeled by the anti-MyHCI/1^st (Fig. 6C) , anti-MyHCI/2^nd (not shown), and anti-MyHCsto (Fig. 6E) . Anti-MyHCI/3^rd + IIa (Fig. 6D) labeled all myotubes; however, the myotubes in the global layer tended to be lighter than those of the orbital layer. Staining with anti-MyHCIIa (Fig. 6F) mainly occurred in the orbital layer and was generally weak. Some of the myotubes labeled by anti-MyHCIIa were also labeled by the slow mAbs. Otherwise, no clear pattern of correspondence was discernible. 

Twenty-Two Weeks of Gestation.

The myotubes were nearly uniform in size and were clearly arranged in fascicles. As far as could be assessed, all myotubes were still labeled by anti-MyHCemb (Fig. 7A ), although with slight heterogeneity. The variable staining intensity seen with anti-MHCfetal (Fig. 7B) at 18 wg was maintained at this age. Anti-MyHCI/1^st (Fig. 7C) , anti-MyHCI/2^nd (not shown), and anti-MyHCsto (Fig. 7E) labeled an evenly dispersed subpopulation of myotubes. These myotubes corresponded to those displaying lower or absent staining with anti-MHCfetal. All myotubes were stained with anti-MyHCI/3^rd + IIa (Fig. 7D) but to different degrees. The moderately to lightly stained myotubes corresponded to those labeled by anti-MyHCI/1^st, anti-MyHCI/2^nd, and anti-MyHCsto. The strongly stained myotubes were in many cases also strongly stained by anti-MyHCIIa (Fig. 7F) . Clear differences in the proportion of the myotubes strongly stained with anti-MyHCIIa and anti-MyHCI/3^rd + IIa were noted among the muscle fascicles. 

Limb Muscles

The reactivity of the primary myotubes to anti-MyHCI/1^st, anti-MyHCI/2^nd, and anti-MyHCI/3^rd + IIa in the limb muscles was, in general, as previously described. ¹⁹ Some variation in staining intensity with anti-MyHCI/3^rd + IIa was noticed among different muscles of the same limb. 

SDS-PAGE

The electrophoretically separated MyHC patterns of the EOMs at 10, 14, and 22 wg differed from that of adult EOM as well as from those of developing limb muscle (Fig. 8) . MyHCemb was the predominant isoform in developing EOMs, and MyHCI was detected at all ages, although only in small amounts at 10 to 14 wg. At 22 wg, MyHCfetal and MyHCIIa were clearly identifiable. MyHCextraocular, seen in adult EOM extracts, could not be detected in the developing muscles. 

Discussion

The present study showed that the developing EOMs differ fundamentally from the limb muscles in their time course of developmental milestones, pattern of MyHC expression (Table 2) , and cytoarchitecture. 

Developmental Milestones

Myotube formation in the EOMs appeared to have a later onset and to progress at a slower pace than in limb muscles. In the limb muscles, primary myotubes are rather homogenous in size, and they form between 8 and 10 wg. ^{18 19} Secondary myotubes are present in limb muscles from 10 wg. ^{18 25} Typical doughnut-shaped myotubes were first seen at 10 wg in the EOMs. Formation of primary myotubes was in progress at least until 12 wg; many individual, small myotubes containing MyHCI/1^st and 2^nd were seen at 12 wg, and typical secondary myotubes were first seen at 14 wg. 

MyHCI Isoforms

The pattern of MyHC expression during development of the EOMs was unique and had very striking features. The primary-generation myotubes in the EOMs coexpressed MyHCI/1^st and MyHCI/2^nd from the earliest stages of development. This is in clear contrast with the biphasic pattern reported for primary myotubes of limb, which initially express only MyHCI/1^st. ¹⁹ This fundamental difference between the EOMs and the limb muscles cannot be assigned to uncertainty in the estimation of the age of human fetal samples, because signs of early myogenesis were seen in the EOMs at 8 weeks of gestation. The onset of expression of MyHCI/3^rd at 10 wg, was also very early in EOMs, when compared with the respective onset at 14 wg in limb muscle. ¹⁹ Notably, this isoform was not present in all slow myotubes, as in limb muscle. Rather, it had a unique distribution, delineating the future orbital layer. To our knowledge, no similar regional pattern of myotube organization, based on differences in the content of MyHCI isoforms, has been reported previously, at these very early stages. Interestingly, the orbital layer predominantly contains fast fibers in the adult EOM. Although most, if not all, of those fast fibers arise from secondary myotubes, it is likely that these special primary myotubes labeled by MyHCI/3^rd play a role in establishing the identity of the future orbital layer. 

MyHCsto

The acquisition of MyHCsto early in development was a unique feature of the EOMs. Most of the myotubes in limb muscle do not express this isoform. The onset of MyHCsto expression at 10 wg in the EOMs coincided with that reported for the special myotubes in limb muscles that later form the muscle spindles. ²⁵ However, the number of myotubes initially expressing this MyHC in the EOMs was much higher than would be expected if they were muscle spindle precursors. ²⁵ Furthermore, at 10 wg these myotubes were preferentially located in the future global layer, and by 14 wg all slow–primary myotubes had acquired MyHCsto. These myotubes are likely to correspond to the slow, nontwitch fibers of the adult EOMs, which contain MyHCsto. ^{10 14}  

MyHC α-Cardiac

MyHC α-cardiac could not be detected in the developing EOMs, but it has been detected with immunocytochemistry in sporadic fibers in adult EOMs. ⁹ However, the total amount of MyHC α-cardiac in whole muscle extracts has been shown by SDS-PAGE to be below resolution level. ¹³ The fibers containing MyHC α-cardiac in the adult EOMs are a subset of the fibers that express both MyHCI and slow tonic ⁹ (Kjellgren et al., unpublished results, 1998). Because no such myotubes or fibers were present at 8 to 22 wg, it is likely that the induction of MyHC α-cardiac expression is a late event in the development of the EOMs. A subset of the myotubes and fibers expressing MyHCsto at 22 wg could further differentiate and later acquire MyHC α-cardiac. Alternatively, some of the fast myotubes and fibers could switch to the slow program later and acquire MyHCI, slow tonic, and α-cardiac. In human limb muscle, the special myotubes that form the muscle spindles are the only ones that ever express MyHC α-cardiac, and they begin to do so at 14 wg. ²⁵  

MyHCextraocular

In the present study, MyHCextraocular could not be detected in the developing EOMs with immunocytochemistry or SDS-PAGE. In the rat, the mRNA coding for MyHCextraocular has been reported to appear during the first 8 days after birth. ³¹ The induction of this isoform may also occur late in human EOM development. However, the possible late induction of MyHCextraocular is in striking contrast with the fact that MyHCIIa, another fast isoform, made its appearance in the EOMs at 18 wg, much earlier than reported for limb muscles. ¹⁹  

MyHCIIb and MyHCIIx

In contrast to other species, it has been shown that in humans ⁵ type IIb fibers contain MyHCIIx, and the gene for MyHCIIb has only recently been identified in humans. ³² Whether the MyHCIIb gene is expressed in any human muscles, including the EOMs, is still under investigation. MyHCIIx can be identified in human muscle fibers by immunocytochemistry, using an mAb that labels all MyHCs except MyHCIIx. Because all myotubes and fibers in the EOMs coexpress more than one MyHC isoform, this antibody is of no use. In our hands, MyHCIIx was not found in a large number of EOM samples analyzed with SDS-PAGE, whereas this isoform was detected in adult limb muscle samples (not shown). However, MyHCIIx could be present in the EOMs in amounts below the levels of detection by SDS-PAGE. 

Acquisition of the Adult Phenotype

To achieve the adult pattern of MyHC expression (Kjellgren et al., unpublished results, 1998), additional fiber types must emerge in the EOMs. At 22 wg all fibers containing MyHCI/1^st, 2^nd, and 3^rd coexpressed MyHCsto. In the adult EOMs a small percentage of fibers containing MyHCI have no MyHCsto (Kjellgren et al., unpublished results, 1998). The purely slow fibers in the adult may differentiate from fast myotubes that acquire a slow phenotype. Alternatively, they may differentiate from the population of myotubes containing both MyHCI and slow tonic at 22 wg. In limb muscles, the primary myotubes can switch to a fast phenotype, and the secondary myotubes can change to a slow phenotype late in gestation, depending on innervation. ¹⁷ In the EOMs, the fibers containing MyHCsto are known to be multiply innervated, ¹⁰ and therefore adjustments in innervation pattern occurring late in gestation could further influence the final phenotypes. Finally, in the adult EOMs, the expression of MyHCfetal is almost completely localized to the orbital layer (Kjellgren et al., unpublished results, 1998), and at 22 wg there was already a tendency for this isoform to be downregulated in the myotubes of the global layer. A similar pattern of elimination of MyHCfetal has been reported in the developing rat EOMs. ³¹  

The EOM Allotype

The EOMs are so unique in every respect that they are classified as a separate muscle allotype. The term allotype was introduced by Hoh and Hughes ³³ to describe classes of skeletal muscles with distinct MyHC gene repertoires. To date, three allotypes of skeletal muscle have been identified: limb–trunk, masticatory, and EOMs. ³⁴ Differences among the three allotypes are likely to reflect their respective embryonic origin, regulatory programs of myogenesis, and innervation. The limb muscles arise from the somites, whereas the EOMs are derived from the somitomeres, which are the mesoderm anlagens that form in the head and along the entire neural tube before the appearance of the somites. ¹⁷ Recent data indicate that myogenesis in the head is under a different regulatory program, because Pax 3 (Splotch)/myf 5 null mice have no skeletal muscle in the trunk and limbs, whereas the muscles of the head are apparently unaffected. ³⁵ Finally, the EOMs have very complex innervation, including both motoneurons, with discharge rates that are an order of magnitude higher than those of spinal cord motoneurons, and multiple neuromuscular junctions along the length of the slow tonic fibers. ¹¹  

Future Studies

The present work provides a framework for the comprehension of some of the developmental aspects that give rise to the mature EOM phenotype. Further studies are needed to fully elucidate the mechanisms orchestrating this high complexity. In particular, the role of innervation and extracellular matrix in the establishment of the adult phenotype should be investigated. 

 Supported by grants from The Medical Faculty, Umeå University, The Swedish Society for Medical Research, The Swedish Medical Research Council (12X-03934), and The National Health and Medical Research Council of Australia. YH was a student at Umeå University Biomedical Graduate School supported by the Swedish Foundation for Strategic Research.

 Submitted for publication March 18, 1999; revised December 17, 1999; accepted January 5, 2000.

 Commercial relationships policy: N.

 

Table 1.

 

Summary of Data on the Antibodies Used

Antibody	Specificity	Short Name	Studies
F1.652*	MyHCembryonic	Anti-MyHCemb	21
NCL-MHCn, †	MyHCfetal	Anti-MyHCfetal	22
A4.840*	First developmental MyHCI	Anti-MyHCI/1st	19
A4.951*	Second developmental MyHCI	Anti-MyHCI/2nd	19
N2.261*	Third developmental MyHCI, MyHCIIa	Anti-MyHCI/3rd+IIa	19
A4.74*	MyHCIIa	Anti-MyHCIIa	21
F88, ‡	MyHCα-cardiac	Anti-MyHCα-cardiac	23
ALD19, §	MyHCslow tonic	Anti-MyHCsto	24, 25
4A6	MyHCextraocular	Anti-MyHCeom	26, 27

*  Purchased from Developmental Studies Hybridoma Bank, Department of Biological Sciences, University of Iowa, Iowa City.

†  Purchased from Novocastra Laboratories, Newcastle-upon-Tyne, UK.

‡  Gift from Jean J. Léger, Institut National de la Santé et de la Recherche Médicale, Unité 249, Montpellier, France.

§  Gift from Donald A. Fischman, Cornell University, Ithaca, NY.

Figure 1.

 

Cross section showing the superior rectus (a), the lateral rectus (b), the inferior rectus (c), the medial rectus (d), and the superior oblique (e) stained with anti-MyHCemb at 8 wg.

Figure 2.

 

Cross sections of the lateral rectus muscle at 8 wg stained with anti-MyHCemb (A), anti-MyHCfetal (B), anti-MyHCI/1^st (C), and anti-MyHCI/2^nd (D). Note that all muscle cells were already labeled both by anti-MyHCI/1^st and 2^nd at this early stage.

Figure 3.

 

Serial cross sections of the inferior rectus muscle at 10 wg. All myotubes were labeled by anti-MyHCemb (A), anti-MyHCfetal (B), anti-MyHCI/1^st (C), and anti-MyHCI/2^nd (D). Staining with anti-MyHCI/3^rd + IIa (E) primarily occurred in the periphery of the muscle in myotubes that were also strongly stained with anti-MyHCI/1^st and 2^nd (arrows). Anti-MyHCsto (F) stained heterogeneously.

Figure 4.

 

Serial cross sections of the superior oblique muscle at 12 wg stained with anti-MyHCemb (A), anti-MyHCI/3^rd+ IIa (B), and anti-MyHCsto (C). Staining with anti-MyHCI/3^rd + IIa and anti-MyHCsto clearly showed that the myotubes were organized into orbital and global layers, respectively. Note that the orbital layer completely encircles the global layer in the oblique muscles. Anti-MyHCsto revealed two populations of myotubes: large diameter (long arrows) and small diameter (arrowheads).

Figure 5.

 

Serial cross sections of EOM at 14 wg stained with anti-MyHCfetal (A), anti-MyHCI/1^st (B), and anti-MyHCsto (C). Anti-MyHCfetal labeled all myotubes. Strongly stained myotubes (arrowheads), which generally were smaller in size, were found in close apposition to moderately stained myotubes. Anti-MyHCI/1^st and anti-MyHCsto exclusively labeled the latter myotubes (long arrows).

Figure 6.

 

Serial cross sections of EOM at 18 wg stained with anti-MyHCemb (A), anti-MyHCfetal (B), anti-MyHCI/1^st (C), anti-MyHCI/3^rd + IIa (D), anti-MyHCsto (E), and anti-MyHCIIa (F). The boundary between the orbital (orb) and the global (glo) layers is indicated by a dashed line in (D). Anti-MyHCfetal stained heterogeneously, but all small myotubes in proximity with larger myotubes were strongly stained (arrowheads). The larger myotubes, some of which were unstained, corresponded to those labeled by anti-MyHCI/1^st (long arrows). The myotubes in the orbital layer were generally more strongly stained with anti-MyHCI/3^rd + IIa than those in the global layer (D). Staining with anti-MyHCIIa mainly occurred in the orbital layer, where most of the myotubes were weakly stained, but sporadic myotubes were strongly stained.

Figure 7.

 

Serial cross sections of the rectus superior muscle at 22 wg. All myotubes were still labeled by anti-MyHCemb (A) whereas anti-MyHCfetal (B) stained with great heterogeneity. Anti-MyHCI/1^st (C) and anti-MyHCsto (E) labeled the myotubes displaying lower or absent staining with anti-MyHCfetal (long arrows). These myotubes also corresponded to those that were moderately stained with anti-MyHCI/3^rd + IIa (D). Many of the myotubes that were strongly stained with anti-MyHCI/3^rd + IIa were also strongly stained with anti-MyHCIIa (F, arrowhead).

Figure 8.

 

Electrophoretically separated MyHC patterns of EOMs at 10 wg, 14 wg, and 22 wg and adult and of limb muscles at 12 wg and 20 wg. MyHCextraocular was not detected in the fetal EOMs. E, MyHCemb; IIa, MyHCIIa; EOM, MyHCextraocular; I, MyHCI.

Table 2.

 

Summary of the General Staining Pattern of the Primary and Secondary Myotubes in the Developing EOMs and Limb Muscles

Extraocular Muscles
Weeks’ Gestation*	Staining Pattern
		Orbital Layer	Global Layer
8 (primary myotubes forming)	MyHCemb+, MyHCfetal+ MyHCI/1st+, MyHCI/2nd+	MyHCemb+, MyHCfetal+ MyHCI/1st+, MyHCI/2nd+
10 (typical primary myotubes present)	MyHCemb+, MyHCfetal+ MyHCI/1st+, MyHCI/2nd+ some myotubes MyHCI/3rd+IIa+	MyHCemb+, MyHCfetal+ MyHCI/1st+, MyHCI/2nd+ some myotubes MyHCsto+
12 (primary myotubes only)	MyHCemb+, MyHCfetal+ MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa+, MyHCsto−	MyHCemb+, MyHCfetal+ MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa−, MyHCsto+
14 (typical secondary myotubes appear)
Primary myotubes	MyHCemb+, MyHCfetal ± MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa+, MyHCsto±	MyHCemb+, MyHCfetal± MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa±, MyHCsto+
Secondary myotubes	MyHCemb+, MyHCfetal+	MyHCemb+, MyHCfetal+
18
Primary myotubes	MyHCemb+, MyHCfetal ±/+/− and MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa+, MyHCsto+/±, or MyHCIIa ±/+	MyHCemb+, MyHCfetal−/± and MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa±, MyHCsto+, or MyHCIIa± (very few)
Secondary myotubes	MyHCemb+, MyHCfetal+ MyHCI/3rd+IIa±/+ or MyHCIIa±	MyHCemb+, MyHCfetal+ MyHC I/3rd+IIa ± MyHCIIa± (very few)
22 (myotube size quite homogeneous)	MyHCemb+/±, MyHCfetal±/+/− and MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa±, MyHCsto+/±, or MyHCIIa+, MyHCI/3rd+IIa+	MyHCemb+/±, MyHCfetal−/± and MyHCI/1st+, MyHCI/2nd+ MyHCI/3rd+IIa±, MyHCsto+, or MyHCIIa+, MyHCI/3rd+IIa+

Limb Muscles
Weeks’ Gestation	Staining Pattern
6–8 (primary myotubes forming)	MyHCemb+, MyHCfetal± MyHCI/1st+
10–12 (secondary myotubes forming)
Primary myotubes	MyHCemb+, MyHCfetal± MyHCI/1st+, MyHCI/2nd+
Secondary myotubes	MyHCemb+, MyHCfetal+
14
Primary myotubes	MyHCemb+, MyHCfetal ± MyHCI/1st+, MyHCI/2nd+, MyHCI/3rd+IIa+
Secondary myotubes	MyHCemb+, MyHCfetal+
≥20
Primary myotubes	MyHCemb±/−, MyHCfetal±/− MyHCI/1st+, MyHCI/2nd+, MyHCI/3rd+IIa+
Secondary myotubes	MyHCemb+, MyHCfetal+, MyHCIIa+ or MyHCI/1st+, MyHCI/2nd+, MyHCI/3rd+IIa+

*  Developmental stage of myotubes is shown in parentheses.

 In limb muscles, MyHCsto present only in future muscle spindles, from 10 weeks gestation. ²⁵

The authors thank Mona Lindström for skillful technical assistance and to the staff at the University Hospital for their help in collecting the fetal muscle samples.