Development of the neuromuscular junction in extraocular muscles of white Leghorn chicks


  • Scott A. Croes,

    1. Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, Nevada
    Search for more papers by this author
  • Christopher S. Von Bartheld

    Corresponding author
    1. Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, Nevada
    • Department of Physiology and Cell Biology, Mailstop 352, University of Nevada School of Medicine, Reno, NV 89557
    Search for more papers by this author
    • Fax: 775-784-6903


Relatively little is known about the development of the neuromuscular junction of extraocular muscles (EOMs). In recent years, chicks have been increasingly used as a developmental model in ophthalmological research. To utilize this model system for understanding the development and plasticity of the extraocular motor system, we investigated the structural changes that occur at the developing neuromuscular junction in the chick between embryonic day 14 (E14) and posthatch day 2 (P2). Axons and nerve terminals were visualized with fluorescent neurofilament antibodies and motor endplates with rhodamine-conjugated α-bungarotoxin. Nerve fibers and endplates were colabeled within the same tissue samples. Motor endplates (density, length, width, and area) were measured and numbers of axons per neuromuscular junction were counted using confocal and conventional microscopy. In P2 chicks, densities of motor endplates were significantly greater in the superior oblique muscle when compared with the superior rectus and lateral gastrocnemius muscle. EOMs showed a two- to threefold larger area of motor endplate size as compared to gastrocnemius muscle. Motor endplate size also differed among EOMs with the superior oblique muscle having endplates with a larger area than those of the superior rectus muscle. The period of synapse elimination was similar between EOM and gastrocnemius muscle. Synapse elimination began at about E18 and was completed by P2. By describing the normal morphological changes in developing EOMs, this study provides a baseline for future work to elucidate underlying molecular mechanisms that regulate EOM innervation and strength. © 2005 Wiley-Liss, Inc.

Extraocular muscles (EOMs) are highly organized striated muscles that differ from limb musculature by their diversity of fibers (Spencer and Porter, 1988), their functional properties (Nelson et al., 1986), and their susceptibility to certain diseases (Ruff, 2002). Chick embryos and hatchlings have proven to be a very useful developmental model system in vision research. They are particularly well suited for eye muscle studies with their relatively large eyes, well-developed eye muscles, accessibility during development, and a highly conserved extraocular motor system (Heaton and Wayne, 1983). This research model can be further used to explore novel treatments for neuromuscular disorders that affect the EOMs, such as myasthenia gravis (autoimmune disorder) and strabismus (improper oculomotor innervation or neuromuscular junction development) (Ticho, 2003; Chilton and Guthrie, 2004). However, in order to assess the effects of pharmacological or molecular manipulations, baseline information is required about the structural changes that occur at the developing neuromuscular junction (NMJ) in this species.

During normal NMJ development in vivo, when a motor axon contacts a muscle fiber, the axon induces the formation of a motor endplate (acetylcholine receptor cluster) by the release of agrin (Sanes and Lichtman, 1999). After this initial innervation, additional terminals from other motor axons often converge at the same neuromuscular junction, resulting in the motor endplate being innervated by more than one axon (referred to as multiple innervation or hyperinnervation). The extent of multiple innervation has been found to differ between muscles and between fiber types in the same muscle (Brown et al., 1976; Pockett, 1981; Jansen and Fladby, 1990; Keller-Peck et al., 2001). Developmental changes that occur during motor endplate formation in mammals have been well documented for skeletal muscle (Bennett and Pettigrew, 1974; Steinbach, 1981; Sanes and Lichtman, 1999). In general, the mammalian NMJ undergoes dramatic changes in structure and function within the first few postnatal weeks. During this time, in a process called synapse elimination, all but one of the motor axon branches is withdrawn or eliminated, leaving each endplate innervated by a single axon (Jansen and Fladby, 1990). Loss of multiple innervation appears to be due to retraction of some terminal branches from each muscle fiber without changes in the number of motor axons innervating the muscle as a whole (Brown et al., 1976; Balice-Gordon and Thompson, 1988; Sanes and Lichtman, 1999). Thus, during synapse elimination, the size of motor units decreases, but not the number of motor axons.

Research on skeletal muscle NMJ formation and synapse elimination in avian models has primarily focused on the hindlimb (Pockett, 1981; Dahm and Landmesser, 1988, 1991) and wing muscles (Smith and Slater, 1983). Interestingly, there are variations in multiple innervation and synapse elimination depending on muscle location and time of innervation (Pockett, 1981). Synapse elimination has been shown to occur prior to hatching in some muscles and within weeks after hatching in others. Thus, the time period of NMJ formation and synapse elimination of EOM cannot be deduced by extrapolation from data on skeletal muscle, but rather has to be determined empirically.

The regression of multiple innervation in avian EOM has, to our knowledge, only been studied by Holt and Sohal (1978). They reported that most of the NMJs of superior oblique muscle in duck were multiply innervated from embryonic day 18 through day 27 (time of hatching) but, by 1 week after hatching, endplates were innervated by only a single axon. In the EOM of chick, it is not known when the regression of multiple innervation takes place, nor have there been any studies that describe the developmental morphological changes associated with the motor endplates. Since chicks have become a prime model for oculomotor development (Steljes et al., 1999; Rind and von Bartheld, 2002; Chen et al., 2003; Chen and von Bartheld, 2004; Chilton and Guthrie, 2004) and the chicken genome is currently being mapped (Burt, 2004), a detailed analysis of EOM innervation in this species is warranted. We have therefore quantified the developmental changes that occur at the neuromuscular junction of EOM (superior rectus and superior oblique) between embryonic day 14 (E14) and posthatch day 2 (P2). The specific aims of this study were to document changes in motor endplate morphology (length, width, area), to determine the motor endplate density (number/mm3 of muscle tissue), to estimate the total number of endplates and muscle fibers in EOMs, and to describe the time period of synapse elimination. Changes in the NMJ of skeletal (lateral gastrocnemius) muscle were also quantified as a means of comparison to extraocular muscle.


Fertilized White Leghorn chicken eggs were obtained from a local supplier (California Golden Eggs) and incubated in a force-draft incubator at 37–38°C. For developmental studies, eggs were incubated until the desired age and staged according to Hamburger and Hamilton (1951) at the time of sacrifice. Date of hatching was designated posthatch day 0 (P0). All experimental procedures were approved by the Institutional Animal Care and Use Committee of the University of Nevada in Reno, Nevada. Primary antibodies included monoclonal antineurofilament 200 and antineurofilament 68 from Sigma (St. Louis, MO) and synaptic vesicle antibody SV2 (supernatant) from the Developmental Studies Hybridoma Bank (Department of Biological Sciences, University of Iowa, Iowa City, IA). Biotinylated secondary antibody (horse antimouse) was from Vector Laboratories (Burlingame, CA). Tetramethylrhodamine-conjugated α-bungarotoxin (Rh-αBTX) and Alexa 488 were purchased from Molecular Probes (Eugene, OR).

Tissue Preparation for Fluorescence Analysis

At the time of sacrifice, chicks were anesthetized with sodium pentobarbital (Nembutal, 250 mg/chick) and perfused transcardially with 4% paraformaldehyde in phosphate-buffered saline (PBS). Both lateral gastrocnemius muscles and eye muscles (superior rectus and superior oblique) were carefully dissected out and placed into PBS. The fixed muscle samples were cryoprotected with 30% sucrose in PBS overnight, embedded in optimal cutting temperature (OCT) compound, and stored at −20°C. It should be noted that muscle samples were not postfixed. Many protocols call for this step, but we found superior antibody labeling when postfixing of samples was omitted. Serial longitudinal sections (30 μm thick) parallel to the orientation of the muscle fibers were cut on a cryotome (Leica CM 3050) and every 5th section for EOM and every 10th section for gastrocnemius muscle were collected. Sections were mounted on coated microscope slides (Vectabond-coated Superfrost/plus; Fisher Scientific, Pittsburgh, PA). Frozen sections were air-dried for ∼ 45 min, hydrated for 15 min with PBS, then treated with 10% bovine serum albumin (BSA) and 0.3% Triton X-100 for 1 hr.


All procedures were performed at room temperature unless otherwise stated. Antineurofilament antibodies (antineurofilament 68 and 200) were diluted 1:100 and synaptic vesicle antibody supernatant (SV2) was diluted 1:10, both with PBS. Sections were incubated for 24 hr with primary antibodies and then washed with PBS (three times, 30 min each). The appropriate biotinylated secondary antibody was applied for 1 hr and washed (three times, 30 min each). Fluorophore, Alexa 488 (Molecular Probes), was applied for 1 hr and washed (three times, 10 min each). Tetramethylrhodamine-conjugated α-bungarotoxin (Rh-αBTX) was applied for 45 min at a concentration of 2 μg/ml. Sections were washed and coverslipped with Aqua mount (Lerner Laboratories, Pittsburgh, PA) and stored in the dark at 4°C until analysis.

Data Collection

Histological material was obtained from 20 chicks (3–5 chicks per time point; E14, E16, E18, P0, P2). When muscles from different animals in the same experimental condition were analyzed, we first determined whether there were statistically significant differences within the group before combining data. In no case did data from individual muscles within an age group differ significantly from each other. Therefore, unless otherwise stated, “n” in the text refers to the number of endplates or neuromuscular junctions examined.

Chick gastrocnemius and EOM contained both en plaque and en grappe motor endplates. The larger, more compact en plaque endplates were easily distinguished from the en grappe endings, which appeared as small clusters of beads (each 2–5 μm in length) scattered along the muscle fibers. It was impossible always to distinguished the en grappe endings from background fluorescence. For these reasons, the morphological analysis was restricted to the larger, more compact, and clearly identifiable en plaque receptors. The total number of these endplates per muscle was estimated in P2 chicks by counting all en plaque endplates in every 5th (eye muscle) or 10th (gastrocnemius muscle) section collected. This value was then multiplied by 5 (eye muscle) or 10 (gastrocnemius muscle) to calculate the total number of endplates. The muscle volume was determined by the Cavalieri point counting method (Howard and Reed, 1998) using a point grid and a Nikon Optiphot microscope equipped with an electronic microcator to measure movements of the stage in the z-axis. Volume measurements were taken from every 5th section of the eye muscle and every 10th section in gastrocnemius muscle. The en plaque endplates measure about 5 μm in the z-axis in longitudinal sections. Analysis at high magnification (60× objective) and the use of an unbiased counting frame revealed that fragments at the upper or lower edge of the tissue sections are rare (less than 5%), probably because the knife blade rarely slices muscle fibers longitudinally. Furthermore, calibration by the two-section comparison method (Williams et al., 2003), using section pairs of 15 and 30 μm thickness, indicated that our estimates of total numbers are within 10–15% accurate.

Utilizing an upright fluorescent microscope (Nikon Eclipse E 600), the number of motor endplates and number of axons per NMJ were determined with 40× and 100× oil immersion lenses, respectively. To determine the extent of multiple innervation, the number of axonal branches per NMJ was determined within each muscle by randomly selecting at least 50 NMJs among the collected sections. By analyzing nerve fibers within the muscle, one can determine multiple innervation, while quantification of polyneuronal innervation (more than one motor neuron contributes to the innervation of one neuromuscular junction) would require whole-mount visualization of the contributing motor axons. Dual-color images of endplates and NMJs were acquired with a laser scanning confocal microscope (Model 1024; Bio-Rad, Hercules, CA) using a 60× oil immersion objective with a constant pinhole setting to preserve the thickness of the confocal plane. For dual-color imaging, Rh-αBTX and green fluorescence-conjugated secondary antibody labels were excited sequentially using the 568 and the 488 nm excitation lines of the krypton/argon laser. The laser intensity and gain settings were adjusted on a per-image basis. Full-frame 1,024 × 1,024 pixel images of endplates and nerve terminals were collected. Data for the length, width, and area of endplates were calculated using Simple PCI software (Thousand Oaks, CA).

Morphometry of Muscle Fibers

For morphometric analysis, two muscles each, superior rectus, superior oblique, and lateral gastrocnemius from P2 chicks were used. The muscles were postfixed in 4% paraformaldehyde, dehydrated in a graded ethanol series, and embedded in paraffin (Paraplast Plus; Oxford Labware, St. Louis, MO). Serial cross-sections (12 μm thick) were taken at the mid belly of each muscle and collected on saline-coated slides. Sections were deparaffinized, stained with thionin, dehydrated, and coverslipped with di-n-butyl-phthalate-xylene (DPX; BDH Laboratory Supplies, Poole, U.K.). Utilizing an upright fluorescent microscope (Nikon Eclipse E 600) equipped with Nomarski optics, we examined muscle fibers at 100× oil immersion and measured the average diameter of the muscle fibers by using a calibrated reticule. For diameter measurements, two muscles were examined, and 20 muscle fibers were randomly chosen from each muscle. The total number of muscle fibers per cross-section was determined by taking random samples from two muscles and the average number of muscle fiber per sample area was calculated. The total cross-sectional area was determined by point counting and the number of muscle fibers was calculated as described in Geuna et al. (2000).

Statistical Methods

All statistical analyses were conducted using SigmaStat software (Jandel, San Rafael, CA). Data are reported as the mean ± standard error of the mean, and n, unless otherwise stated, refers to the number of endplates or neuromuscular junctions analyzed. Statistical significance was evaluated by Student's paired t-test or one-way analysis of variance (ANOVA). For statistical significance, we used a confidence level of P < 0.05.


Motor Endplate Morphology

Our description refers to the en plaque endplate, not the en grappe type, unless indicated otherwise. To characterize the development of the motor endplates, we examined the endplate morphology (length, width, and area) in the extraocular muscle (superior rectus, superior oblique) and in lateral gastrocnemius muscle in chicks ranging in age from E14 through P2. Analysis of motor endplates stained with Rh-αBTX showed that there was a progressive rate of growth in all parameters measured from E14 through P2 (Fig. 1, Table 1). The length of the gastrocnemius muscle motor endplate increased from 3.9 ± 0.06 (SEM) to 6.9 ± 0.09 μm, width increased from 1.3 ± 0.03 to 3.5 ± 0.05 μm, while the endplate area increased from 5.1 ± 0.10 to 23.5 ± 0.34 μm2. In EOM, both the width and the length of the endplates increased considerably from E14 to P2, resulting in an elongated shape. The length of the superior rectus muscle motor endplate increased from 6.6 ± 0.15 to 17.1 ± 0.27 μm, width increased from 2.1 ± 0.07 to 3.4 ± 0.05 μm, and the area increased from 13.8 ± 0.47 to 57.4 ± 1.07 μm2. The length of the superior oblique muscle motor endplate increased from 7.5 ± 0.16 to 21.0 ± 0.44 μm, width increased from 2.5 ± 0.07 to 4.0 ± 0.09 μm, while the area increased from 18.8 ± 0.62 to 79.9 ± 1.68 μm2.

Figure 1.

Confocal images of motor endplates in lateral gastrocnemius, superior rectus, and superior oblique muscles from E14 to P2. Acetylcholine receptors of motor endplates (red) are labeled with Rh-αBTX. The endplates increased in length, width, and area as the chick matured. Scale bar = 10 μm.

Table 1. Length, width, and area of motor endplates in chick extraocular muscles and lateral gastrocnemius muscle
Agea and muscleLengthbWidthbAreabnc
  • a

    Age in days and staged according to Hamburger and Hamilton (1951).

  • b

    Values: mean ± standard error of the mean (SEM).

  • c

    n represents number of endplates.

E14 (St 40)    
 Gastrocnemius3.9 ± 0.061.3 ± 0.035.1 ± 0.10121
 Superior rectus6.6 ± 0.152.1 ± 0.0713.8 ± 0.4783
 Superior oblique7.5 ± 0.162.5 ± 0.0718.8 ± 0.62111
E16 (St 42)    
 Gastrocnemius4.7 ± 0.072.2 ± 0.0510.1 ± 0.20130
 Superior rectus8.1 ± 0.163.1 ± 0.0723.9 ± 0.53150
 Superior oblique8.8 ± 0.182.8 ± 0.0724.7 ± 0.66116
E18 (St 44)    
 Gastrocnemius4.8 ± 0.072.4 ± 0.0411.6 ± 0.16181
 Superior rectus9.2 ± 0.183.1 ± 0.0728.0 ± 0.60158
 Superior oblique10.4 ± 0.173.1 ± 0.0532.3 ± 0.66225
P0 (St 46)    
 Gastrocnemius6.9 ± 0.083.1 ± 0.0521.2 ± 0.33186
 Superior rectus13.5 ± 0.263.6 ± 0.0747.0 ± 1.00151
 Superior oblique17.4 ± 0.373.6 ± 0.0761.9 ± 1.56151
 Gastrocnemius6.9 ± 0.093.5 ± 0.0523.5 ± 0.34222
 Superior rectus17.1 ± 0.273.4 ± 0.0557.4 ± 1.07247
 Superior oblique21.0 ± 0.444.0 ± 0.0979.9 ± 1.68144

We then compared the normal endplate structure between the EOM and gastrocnemius muscle at P2 (Fig 2, Table 1). The width of the endplates was similar between the three muscles (∼ 3.4 to 4.0 μm); however, the length was significantly greater in both the superior rectus (17.1 ± 0.27 μm) and the superior oblique muscle (21.0 ± 0.44 μm) as compared with the gastrocnemius muscle (6.9 ± 0.09 μm; P < 0.05). The motor endplate area was also found to be significantly larger in the EOM as compared to the gastrocnemius muscle (P < 0.05). The endplate area was twice as large within the superior rectus muscle (23.5 ± 0.34 vs. 57.4 ± 1.07 μm2) and three times as large within the superior oblique muscles (23.5 ± 0.34 vs. 79.9 ± 1.68 μm2). The size of the receptor clusters also differed between EOM, with the superior oblique muscle containing motor endplates with a significantly greater length (21.0 ± 0.44 μm) and thus area (79.9 ± 1.68 μm2) than those in the superior rectus muscle (length, 17.1 ± 0.27 μm; area, 57.4 ± 1.07 μm2; P < 0.05).

Figure 2.

Comparison of motor endplate area, length, and width in gastrocnemius (n = 222), superior rectus (n = 247), and superior oblique (n = 144) muscles of P2 chicks. Motor endplates of EOMs have significantly greater area and length than those in the gastrocnemius muscle (P < 0.05). Comparison between EOMs reveals that the superior oblique muscles have endplates with both greater area and length than the superior rectus muscles (P < 0.05). Error bars indicate SEM.

Examination of the motor endplates shows a developmental change in acetylcholine receptor (AChR) distribution (Fig. 1). At E14, the AChRs tend to be clustered in oval plaques with a relatively uniform density. Over the next 9 days (E14 to P2), these plaque-like endplates increase in size and become perforated. Many motor endplates take on a pretzel-like morphology, with the regions that maintain the AChRs becoming the pretzel-shaped branches of the mature singly innervated NMJ previously described for NMJs of skeletal muscles (Balice-Gordon et al., 1990; Balice-Gordon and Lichtman, 1993; Sanes and Lichtman, 1999; Marques et al., 2000). As described below in synapse elimination, during this time period (E14 to P2), inputs were eliminated, leaving only a single axon with a nerve terminal that appeared to correspond with the pretzel-like endplate spatially.

Neuromuscular Innervation

In order to determine the time period of synapse elimination, we quantified the number of axons innervating a motor endplate of extraocular muscle and gastrocnemius muscle in chicks ranging in age from E16 through P2. A motor endplate was considered to be multiply innervated when two or more axons were in contact with an endplate (Fig. 3A). While motor endplates with single axon innervation were observed at all ages (Fig. 3C and D), there was a progression from multiple to single axon innervation (Fig 4, Table 2). At age E16, 60% or more of the endplates of each selected muscle (superior rectus 74%, superior oblique 60%, gastrocnemius 72%) were multiply innervated; of these, the majority were innervated by two axons. During the subsequent 7 days, the proportion of multiply innervated endplates decreased rapidly to 2.4% or less by P2. On E18, multiple innervation decreased to less than 50%. In those endplates that were multiply innervated, one of the axons seemed to be distinguished by a larger diameter (Fig. 3B and C). At P0 and beyond, the endplate was only rarely innervated by more than one axon. Motor endplates with three or more axons were not detected at P0 or P2. Thus, we conclude that the process of synapse elimination is ongoing at E18 and is completed by 2 days after hatching.

Figure 3.

Confocal images of motor endplates (red) and nerve terminals (green) in chick extraocular muscles (superior rectus and superior oblique) and lateral gastrocnemius muscles. A: Superior oblique muscle at E18 showing two axons innervating a single endplate (multiple innervation). B: Retraction bulbs (thin axon terminating in a bulb; arrow number 3) were evident, a common feature during synapse elimination. B and C: Two axons innervate the junction of a superior rectus muscle (E18). Note that axon number 2 is relatively small compared with axon number 1. Superior rectus (D) and gastrocnemius (E) muscles with single axon innervation at P2. F: Axon innervating en grappe receptors (arrows) in the superior oblique muscle (P2). Axons and nerve terminals (both green) are labeled with antineurofilament antibodies and anti-SV2 antibodies, respectively, and acetylcholine receptors of motor endplates (red) are labeled with rhodamine-conjugated α-bungarotoxin. Scale bars = 10 μm; scale bar for A is the same for C, D, and F.

Figure 4.

Regression of multiple innervation with age from E16 to P2 in superior rectus, superior oblique, and lateral gastrocnemius muscles. Symbols indicate the relative percentage of neuromuscular junctions innervated by more than one axon. Relative percentages were determined by examination of at least 50 junctions from each of the selected muscles obtained from three to five animals at each age. Axons/nerve terminals were visualized with fluorescent neurofilament antibodies and motor endplates with rhodamine-conjugated α-bungarotoxin, colabeled within the same tissue samples.

Table 2. Regression of multiple innervation in chick extraocular muscles and lateral gastrocnemius muscle
Agea and muscleMean number of axons/endplateAxons/endplate (relative %)nb
  • a

    Age in days and staged according to Hamburger and Hamilton (1951).

  • b

    n represents number of neuromuscular junctions sampled.

E16 (St 42)      
 Gastrocnemius1.7828.063.09.0 150
 Superior rectus1.8626.060.014.0 150
 Superior oblique1.7436.
E18 (St 44)      
 Gastrocnemius1.6052.036.012.0 150
 Superior rectus1.2876.020.04.0 150
 Superior oblique1.4466.
P0 (St 46)      
 Gastrocnemius1.0496.04.0  150
 Superior rectus1.1288.012.0  150
 Superior oblique1.1684.016.0  150
 Gastrocnemius1.0298.21.8  386
 Superior rectus1.0297.62.4  416
 Superior oblique1.00100.0   102

Motor endplates with single axon innervation have an axon with a relatively large diameter, but thinner axons can be observed within the vicinity. We observed bulb-like endings of these thin axons near the endplate (Fig. 3B) that appear to be retraction bulbs. This is similar to what has been described in skeletal muscle of rat (Riley, 1977), mouse (Nguyen et al., 1998), and humans (Gramsbergen et al., 1997), as reviewed by Bernstein and Lichtman (1999).

Estimate of Number of Endplates

The total number of en plaque motor endplates was estimated for the superior rectus, superior oblique, and the gastrocnemius muscle at P2. As previously discussed, only the clearly identifiable en plaque endplates were counted. The average number of these endplates for the superior oblique muscle was 20,664 ± 1,328; for the superior rectus, 11,473 ± 808; and for the gastrocnemius, 106,490 ± 12,209.

Estimate of Motor Unit Size

Motor unit sizes (number of myofibers per neuron) were compared between the superior rectus, superior oblique, and the gastrocnemius muscles at P2. The average number of myofibers per muscle for the superior oblique muscle was 6,241 ± 190; for the superior rectus, 8,050 ± 218; and for the gastrocnemius, 63,161 ± 3,144. The number of motor neurons that innervate the chick superior oblique is known to be about 900 (Hatton and von Bartheld, 1999); for the chick superior rectus, about 500–700 (Steljes et al., 1999); and for the lateral gastrocnemius in a mammalian species (cat), about 280 (Weeks and English, 1985). Thus, the motor unit size for the superior oblique is about seven myofibers per neurons, the superior rectus is 11–16 myofibers per neuron, and the gastrocnemius is 225 myofibers per neuron (assuming that chicks have a similar number of motoneurons innervating the lateral gastrocnemius as cats).

Motor Endplate Density

Motor endplate densities (en plaque endplates/mm3) were compared between the superior rectus, superior oblique, and the gastrocnemius muscles at P2 (Fig. 5). The superior oblique muscle had a significantly higher density of endplates (3,442 ± 179) than either the gastrocnemius (1,937 ± 121) or the superior rectus muscle (2,2890 ± 156; P < 0.05). The density of motor endplates was similar between the gastrocnemius and superior rectus muscles. In the 2-day-old chick, the myofiber diameters of the EOMs are similar to those of the gastrocnemius (superior oblique: 8.7 ± 0.45 μm; superior rectus: 7.2 ± 0.41 μm; gastrocnemius: 7.0 ± 0.37 μm).

Figure 5.

Comparison of motor endplates density (motor endplates/mm3) between the gastrocnemius (n = 6) muscle, superior rectus (n = 6) muscle, and superior oblique (n = 4) muscle of P2 chicks. Superior oblique muscles have significantly higher endplate densities than the gastrocnemius and superior rectus muscles. n represents number of chicks. Error bars indicate SEM.


The present study was undertaken to describe the developmental changes associated with the neuromuscular junctions of EOM in White Leghorn chicks from E14 through P2.

Synapse Elimination

To determine the time period of synapse elimination for chick EOM, we analyzed regression of multiple innervation from E16 to P2. Multiple innervation was prevalent (60% or greater at E16) in all muscles analyzed until E18, at which time it decreased rapidly over the following 5 days. By P2, the muscles of the gastrocnemius, superior rectus, and superior oblique were predominantly (97.6–100%) innervated by a single axon indicating the completion of synapse elimination.

The time course of synapse elimination in chick occurs earlier when compared with duck or various mammalian species. Holt and Sohal (1978) found that in duck the majority of motor endplates within the superior oblique were multiply innervated at the day of hatching and decreased to single fiber innervation by the end of the first week. In our study in chicks, only 16% of the motor endplates within the superior oblique (12% within the superior rectus) were multiply innervated at the time of hatching and endplates were rarely innervated by more than one axon 2 days later (P2). Our investigation showed that the incidence of multiply innervated endplates of the chick gastrocnemius declined to less than 2% by 2 days after hatching, which is much earlier when compared to various limb muscles of mammals (Brown et al., 1976; O'Brien et al., 1978; Greensmith and Vrbova, 1991). Previous mammalian studies have reported the frequency of multiple innervation to be about 15% at 13–14 days postnatal in the rat gastrocnemius (Bennett et al., 1986), soleus (Brown et al., 1981), and rectus femoris (Tweedle and Stephens, 1981).

The unique early pattern of synapse elimination in chicken as compared with other species may relate to their precocious movement. The shift from multiple to single axon innervation has been correlated with an increase in muscle activity. In the soleus muscle of rats, the regression from polyneuronal innervation (Brown et al., 1976) coincides with the development of the adult type of walking pattern (Westerga and Gramsbergen, 1990). Stimulation of the soleus muscle in rat leads to an acceleration of the regression process (O'Brien et al., 1978), whereas a decrease in activity, e.g., by paralyzing the muscle with α-bungarotoxin (Duxson, 1982; Greensmith and Vrbova, 1991) or by tendinotomy of the Achilles' tendon (Benoit and Changeux, 1975; Riley, 1978), delays the regression from polyneuronal to single motor axon innervation. During the hatching process, the chick uses a number of complex movements (limb and head) to free itself from the shell. Within a few hours, the chick is able to walk and accurately peck for seeds. Such complex activity patterns indicate an advanced muscular maturity and thus one would expect an earlier maturation of neuromuscular innervation.

The earlier completion of synapse elimination in the EOM of chick compared to duck may relate to their feeding behavior. Chickens are ground-feeding birds that require an accurate aim to obtain food (mainly small seeds), whereas ducks are water-feeding birds that tend to sift through the water for food (mainly plants). Differences have been found in neuronal circuitry mediating visuomotor behavior, suggesting increased oculomotor control for activities such as focusing on grains and accurate pecking movements (Stelling and McVean, 1988; Uchiyama, 1989). Therefore, we speculate that the NMJs of the chick EOMs need to mature precociously in order to establish such feeding behaviors.

Motor Endplate Number

We estimate the total number of en plaque endplates within the P2 chick superior oblique muscle to be about 20,500. The total number of endplates in this muscle at age P14 has previously been reported to be about 50,000 (Sohal et al., 1985). The difference is likely due to the type of receptors counted. In our study, only the clearly identifiable en plaque endplates were counted from longitudinal sections, whereas Sohal et al. (1985) counted all endplates from cross-sections, presumably including both en plaque and en grappe receptors. Chick superior oblique muscle contains about 6,000 muscle fibers (see our results) and is innervated by about 900 trochlear motoneurons (Hatton and von Bartheld, 1999). Thus, while we cannot draw conclusions about the en grappe endplates, our data indicate that trochlear motoneurons supply on average about seven muscle fibers with en plaque endplates. Since the superior oblique (P2) muscles contain about 6,000 muscle fibers, it is estimated that the trochlear motoneurons supply about 3.3 en plaque endplates per muscle fiber, and they contribute, on average, to 23 en plaque endplates in this muscle.

Motor Endplate Density

The motor endplate densities (motor endplates/mm3) were compared between the superior rectus, superior oblique, and gastrocnemius muscles at P2. The superior oblique muscle endplate density was 1.78 times greater than in the gastrocnemius muscle and 1.50 times greater than in the superior rectus muscle. The density of endplates per mm3 between the gastrocnemius and superior rectus muscles was surprisingly similar. It is well known that in the mature animal, the cross-sectional areas of individual myofibers in the gastrocnemius are significantly larger when compared to the EOMs; thus, we expected the endplate density of the gastrocnemius to be significantly less than that of the superior rectus. However, in a 2-day-old chick, the myofiber diameters of the EOM (superior oblique, superior rectus) are very similar (7–9 μm) to those of the gastrocnemius. Thus, at this stage in development, the similarity in cross-sectional area of individual myofibers together with the relatively large total number of endplates as described above for the gastrocnemius muscle appears to be responsible for the unexpected similarities of endplate densities between the gastrocnemius and superior rectus muscles.

The higher density of endplates in the superior oblique muscle could have implications for future molecular manipulations of the EOMs. First, assuming that each NMJ requires a certain number of signaling molecules, the superior oblique muscle would need a higher dose due to the higher NMJ density. Second, after peripheral nerve repair, it has been shown that children have better nerve recovery than do adults (Hallin et al., 1981). This could be due to either a shorter distance for nerve regeneration or to differences in the plasticity of the nervous system in children (Lundborg, 1988). Ma et al. (2002) compared differences of NMJs between juvenile and adult skeletal muscle and postulated that the higher NMJ density in juvenile muscle provides more NMJs for reinnervation and thereby a shorter distance for the nerve to sprout to find them.

Motor Endplate Morphology

It is commonly thought that there are two main types of motor endplates that have been described for the extrafusal muscle fibers in vertebrates, the larger en plaque and the smaller en grappe (Tiegs, 1953; Hess, 1970). In aves, en grappe and en plaque endings have been studied in various skeletal muscles (Hess, 1970), but nothing has been reported about the endplate types of EOM. Here, we report that the EOM of chicks contains both en plaque (Fig. 3A–E) and en grappe (Fig. 3F) endplates, similar to what has been described in human EOM (Oda and Hiroshi, 1988; Spencer and Porter, 1988).

Our study shows that with maturation, endplates of each muscle increased in size. We found the length and width increased between 60% and 168%, while the area increased up to 360%. Previous research has demonstrated that increases or decreases in muscle fiber diameter correlate with the size of the NMJ area (Balice-Gordon et al., 1990). While developmental changes in muscle fiber diameters were not examined in this study, it seems reasonable to assume that the muscle fibers also increased in size with age.

The difference in motor endplate size of EOM compared to skeletal muscle as seen in our study as well as studies in other animal species raises the question as to why the muscle fibers of EOM would have larger endplates. One possible explanation relates to differences in electrical activity and the safety factor (SF) for neuromuscular transmission. The SF for neuromuscular transmission is defined as SF = EPP/(RP − EAP), where EPP is the endplate potential amplitude, RP is the resting membrane potential, and EAP is the action potential threshold (Banker et al., 1983; Kaminski et al., 1997; Ruff and Lennon, 1998). Another way to describe the SF is to say that it is the difference between the actual EPP amplitude and the EPP amplitude required to trigger a muscle fiber action potential (Floeter, 1999). Several aspects contribute to the safety factor, including the quanta of acetylcholine (ACh) released, AChR conduction properties, AChR density, density of voltage-gated sodium channels, and acetylcholine esterase activity (Boonyapisit et al., 1999). The firing rates of EOM fibers are higher than extremity fibers, which places an additional electrical stress on the EOM synapses. The increased electrical stress results in a greater influence on sodium channel inactivation, AChR desensitization, and affects the efficiency of neuromuscular transmission (Ruff and Lennon, 1998; Ruff, 2002). Thus, the benefit of a relatively large endplate size of en plaque fibers of EOM may be that it increases the safety factor for neuromuscular transmission.

Interestingly, there were differences in motor endplate size between different extraocular muscles. The motor endplates of the superior oblique muscles were significantly larger than those of the superior rectus muscles. This was surprising and we currently can only speculate about its relevance. With the diversity of eye movements and discharge patterns of the oculomotor neurons, it may relate to the stimulation rate of the superior oblique muscle. If the superior oblique muscle is stimulated at a higher rate than the superior rectus muscle, the larger motor endplates could increase their safety factor for neuromuscular transmission.

In humans, there is postnatal shaping of fiber characteristics, which are likely to be critical for appropriate function of the eye movement system. Porter et al. (1995) proposed that there may be a critical period for eye muscle development during the first 3–6 months after birth, during which time the eye muscles acquire the structural/functional characteristics demanded by binocular vision. The maturation of the extraocular muscle soon after hatching makes the chick a particularly useful model system. For example, one could test trophic factors or other pharmacological agents on chick EOM to improve potentially the strength of the neuromuscular junction (Rind and von Bartheld, 2002; McLoon and Christiansen, 2003; Chen and von Bartheld, 2004). The early postnatal period up to 3 years of age may represent a window of opportunity to apply trophic factors as a treatment of weak eye muscles (strabismus) in humans. Animal studies to address this hypothesis are currently in progress. By describing the developmental changes that occur at the NMJ, we provide a baseline that can be used in future molecular manipulations of EOM in White Leghorn chick embryos/hatchling as an experimental model.


The authors thank Will Hatton and Keith Murray for technical assistance in the use of the confocal microscope and Jim Kenyon and Robert Ruff for critical comments. The synaptic vesicle (SV2) antibody was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the National Institute of Child Health and Human Development (NICHD) and maintained by the Department of Biological Sciences, University of Iowa.