New Perspectives About Human Laryngeal Muscle: Single-Fiber Analyses and Interspecies Comparisons

Background  In companion studies on canine and rodent laryngeal muscle, we observed that (1) muscle fibers in both the canine and rodent posterior cricoarytenoid (PCA) muscles have a slower myosin heavy-chain (MyHC) isoform profile than those in the thyroarytenoid (TA) muscle; (2) the muscle fiber composition of PCA and TA muscles in canines and rodents is complex given the presence of so-called hybrid fibers (fibers coexpressing various combinations of MyHC isoforms); (3) the types and proportions of hybrid fibers are both muscle specific and, in some cases, region specific; and (4) the MyHC isoform profile of canine laryngeal muscle appears to be slower than that of rodent laryngeal muscle, suggesting the possibility that larger mammals have a slower MyHC isoform profile.

Objectives  Given the findings of these companion studies and the fact that very little is known about the MyHC isoform composition of laryngeal muscle fibers, the primary objectives of this study were to determine (1) the types of MyHC isoforms found in the human PCA and TA muscles, (2) if there were regional differences in MyHC isoform composition, (3) if hybrid fibers commonly occur in human laryngeal muscle, and (4) if the MyHC isoform profile of human laryngeal muscle is slower than that of canine and rodent laryngeal muscle.

Results and Conclusions  The findings of this study clearly demonstrate that both the PCA and TA muscles in humans express 3 types of MyHC isoforms (ie, slow type I, fast type IIA, and fast type IIX MyHC isoforms). At the single-fiber level, there were distinct regional differences and hybrid fibers were a common occurrence. Finally, the data demonstrate that the PCA and TA muscles of humans have a slower MyHC profile than that found in either canine or rodent laryngeal muscle.

THE INTRINSIC laryngeal muscles have been traditionally classified into 3 groups: (1) the abductors that open the glottis, (2) the adductors that close the glottis, and (3) the tensors that regulate vocal cord length and tension. More recent investigations^1-4 have indicated that each laryngeal muscle is more complex than this traditional viewpoint. For example, the posterior cricoarytenoid (PCA) appears to be involved in complex functions that include stabilizing the arytenoid with tonic activity, glottic opening, abduction in unvoiced speech, and activity during high-pitched voice.

Given the relationship between myosin adenosine triphosphatase activity and various contractile properties, previous investigators^5-11 have used myofibrillar adenosine triphosphatase techniques to examine the muscle fiber composition of laryngeal muscle and make inferences about function. For instance, Sanders et al⁹ reported that the canine PCA muscle contains 3 distinct neuromuscular compartments. They described the vertical compartment as 24° from true vertical, inserting on the lateral aspect of the muscular process of the arytenoid, and composed of 65% fast type II muscle fibers. Sanders et al⁹ reported that the oblique compartment is oriented 44° from vertical, inserts on the superior aspect of the muscular process of the arytenoid, and contains 77% type II (fast) muscle fibers. The horizontal region is oriented 63° from vertical, inserts on the medial aspect of the muscular process of the arytenoid, and is physiologically slower than the 2 other compartments, with only 59% type II muscle fibers.⁹ This difference in muscle fiber profile suggests that the different compartments of the PCA might perform distinctive motions during phonation and inspiration.⁵

Although the types of histochemical techniques used by previous investigators^5-11 have provided important information about the muscle fiber composition of laryngeal muscle, there are certain inherent limitations in these methods that fail to accurately account for the diversity of muscle fiber types. As a result, there has been a growing interest in examining the myosin heavy-chain (MyHC) isoform composition of laryngeal muscle via immunohistochemical or electrophoretic techniques.^12-19 In companion studies,^18,19 we developed techniques for determining the MyHC isoform composition of single fibers from canine and rodent laryngeal muscle. Using these techniques, we observed that (1) muscle fibers in both the canine and rodent PCA muscles have a slower MyHC isoform profile than those in the thyroarytenoid (TA) muscle; (2) the muscle fiber composition of PCA and TA muscles in canines and rodents is complex given the presence of so-called hybrid fibers (fibers coexpressing various combinations of MyHC isoforms); (3) the types and proportions of hybrid fibers are both muscle specific and, in some cases, region specific; and (4) the MyHC isoform profile of canine laryngeal muscle appears to be slower than that of rodent laryngeal muscle, suggesting the possibility that larger mammals have a slower MyHC isoform profile.

Given the findings of the companion studies and the fact that little is known about the MyHC isoform composition of laryngeal muscle fibers, the primary objectives of this study were to determine (1) the types of MyHC isoforms found in the human PCA and TA muscles, (2) if there are regional differences in MyHC isoform composition, (3) if hybrid fibers commonly occur in human laryngeal muscle, and (4) if the MyHC isoform profile of human laryngeal muscle is slower than that of canine and rodent laryngeal muscle.

The findings of this study clearly demonstrate that both the PCA and TA muscles in humans express 3 types of MyHC isoforms (ie, slow type I, fast type IIA, and fast type IIX). At the single-fiber level, there were distinct regional differences and hybrid fibers were a common occurrence. Finally, the data demonstrate that the PCA and TA muscles of humans have a slower MyHC profile than that found in either canine or rodent laryngeal muscle.

Five human PCA and 4 TA muscles were surgically excised in patients with squamous cell carcinoma. The age of the patients ranged from 60 to 65 years. Care was taken to ensure that the muscles studied were distant from and unaffected by neoplastic involvement (ie, all PCA and TA samples were taken from muscles that were contralateral to the tumor and had normal function prelaryngectomy). Before this experiment, approval was obtained from our institutional review board.

Each PCA muscle was divided into horizontal and oblique regions, whereas each TA muscle was divided into the medial and lateral regions. Unlike the canine PCA muscle, the human PCA muscle demonstrates only 2 distinct compartments, as previously reported by Sanders et al.²⁰ Fresh PCA and TA muscles were removed and placed immediately into a glycerol-relaxing solution that contained 2-mmol/L EDTA, 1-mmol/L magnesium chloride, 4-mmol/L adenosine triphosphate, 10-mmol/L imidazole, and 100-mmol/L potassium chloride (pH 7.0). The samples were stored in a −20°C freezer until used for sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis. Approximately 40 single-muscle fibers (total number of fibers sampled, 775) were microdissected from each sample using a dissecting microscope (×50) and fine Dumont forceps. Each fiber segment was approximately 2 to 3 mm in length.

Each single-muscle fiber segment was placed into sample buffer,²¹ heated to 95°C for 2 minutes, and ultrasonically agitated for 30 seconds. A 2-stage gel was then poured in which the separating gel contained 8% acrylamide, 0.6% bisacrylamide, 29% glycerol, 0.4% SDS, 0.2-mol/L Tris, and 0.1-mol/L glycine. The stacking gel consisted of 4% acrylamide, 0.2% bisacrylamide, 30% glycerol, 70-mmol/L Tris (pH 6.7), 4-mmol/L EDTA, and 0.4% SDS. Polymerization of a gel was initiated using N,N,N,N-tetramethylenediamine (0.05%) and ammonium persulfate (10%). The gels were immersed in a running buffer (0.1-mol/L Tris, 0.15-mol/L glycine, and 0.1% SDS) and run at a constant voltage of 275 V for 24 hours at 4°C. The gels were then silver stained using a Silver Stain Plus Kit (BioRad, Hercules, Calif) and scanned for MyHC isoforms using an automated densitometer (Molecular Dynamics, Sunnyvale, Calif).

The MyHC isoforms were separated by SDS-PAGE and then transferred onto nitrocellulose membranes. The primary monoclonal antibodies (MAbs) used in this study were BA-D5 (slow type I MyHC), SC-71 (fast type IIA), BF-35 (exclusive of fast type IIX, reacting with slow type I and fast type IIA), and MY-32 (fast type IIX, type IIA). The MAbs BA-D5, SC-71, and BF-35 were produced from cell lines purchased from American Type Culture Collection, Rockville, Md. These cell lines were originally created and characterized by Schiaffino and colleagues.^22,23 The MAb MY-32 was obtained from Sigma (St Louis, Mo). The blots were stained using a Vector VIP substrate kit (Burlingame, Calif). Transfer conditions, concentrations of primary and secondary antibodies, and time of incubation have been described previously.¹⁷

The MyHC isoform compositions of the PCA and TA muscles shown in Figure 1 and Figure 2 were analyzed using a 1-way analysis of variance. The fiber type distribution data shown in Figure 3 and Figure 4 were analyzed using a χ² test. For all analyses, statistical significance was defined as P≤.05.

Three MyHC isoforms (type IIX, type IIA, and type I) were separated in the human PCA and TA muscles (Figure 5, Figure 6, and Figure 7). The fastest migrating band in the human PCA and TA laryngeal muscles comigrated with the rodent slow type I MyHC isoform (Figure 6 and Figure 7). In addition, the MAb (BA-D5) specific for the rodent slow type I MyHC protein isoform reacted with this band (Figure 5). Hence, this band was identified as the slow type I MyHC isoform. The next fastest migrating band found in the PCA and TA muscles comigrated with the rodent fast type IIA MyHC isoform (Figure 6 and Figure 7). With respect to Western blots, this band reacted with MAb SC-71 (specific for fast type IIA; Figure 5) and BF-35 (specific for all but the fast type IIX MyHC isoform). On the basis of these criteria, this band was identified as the fast type IIA MyHC isoform. The slowest band migrated just above the fast type IIA band (Figure 6 and Figure 7) and reacted with the MY-32 MAb (specific for all fast type MyHC isoforms; Figure 5) but did not react with BF-35 (Figure 5). Previous studies^22,23 have shown that BF-35 reacts with all rodent MyHC isoforms except the fast type IIX. This order of migration is consistent with that found in human hindlimb musculature. Based on these criteria, we classified this isoform as the fast type IIX MyHC isoform.

Figure 3 demonstrates that there was a substantial difference between the single-fiber MyHC isoform compositions of the horizontal and oblique regions of the PCA muscle. Most fibers in the horizontal region of the PCA muscle expressed the slow type I MyHC isoform (≈75%), whereas most fibers (≈60%) in the oblique region were fast (Figure 3 and Figure 6). There was also a significant difference in the relative proportions of hybrid fibers found in the 2 different regions. In the horizontal region, approximately 10% of the fibers coexpressed various combinations of MyHC isoforms (Figure 3 and Figure 6). Most of these fibers coexpressed the slow type I and fast type IIA MyHC isoforms. In contrast, hybrid fibers accounted for approximately 40% of the total fiber population in the oblique region (Figure 3 and Figure 6). Most of these fibers coexpressed combinations of either I and IIA or IIA and IIX MyHC isoforms (Figure 6).

The single-fiber MyHC isoform compositions for the medial and lateral regions of the TA muscle are shown in Figure 4. Although there were significant regional differences in the proportions of fiber types, the types of fibers found in the medial and lateral regions of the TA were strikingly similar. In the medial region, there were approximately equal proportions (≈25%-30% of the total pool of fibers) of slow type I, fast type IIA, and hybrid IIA/IIX fibers. In contrast, the predominant fiber type in the lateral region of the TA was the IIA/IIX fiber type, which represented approximately 40% of the total population of fibers.

Several studies^24-27 have shown a good correlation between the MyHC isoform composition of a given muscle or muscle fiber and its corresponding maximal shortening velocity. As a consequence, the MyHC isoform composition of a muscle or muscle fiber has been used as a "physiological marker" that provides insight about the functional properties of that muscle or muscle fiber. Given this background, there has been an evolving interest in studying the MyHC isoform composition of laryngeal muscle.^12-19 However, little is known about the MyHC isoform composition of human laryngeal muscle. Within this context, there are 4 key findings of this study. First, we observed that human laryngeal muscle expresses 3 types of MyHC isoforms and that the relative proportion of each isoform is both muscle and region specific. Second, at the single-fiber level, each laryngeal muscle contained so-called hybrid fibers. Third, we did not observe a MyHC isoform analogous to that of the fast type IIL MyHC isoform found in rodent laryngeal muscle. Finally, human laryngeal muscle appears to have a slower MyHC profile than that of canine or rodent laryngeal muscle.

Recently, Shiotani et al¹⁶ examined the whole muscle MyHC isoform composition of a number of human laryngeal muscles. Shiotani et al reported that human laryngeal muscles express 3 different types of MyHC isoforms (ie, slow type I, fast type IIA, and fast type IIB). Currently, it is believed that the electrophoretic band previously identified as the human fast type IIB MyHC is, in fact, the fast type IIX MyHC isoform.²⁸ Hence, the fast type IIB MyHC isoform identified by Shiotani et al should be categorized as the fast type IIX isoform. Within this context, there are some important similarities between our whole muscle data and those of Shiotani et al. For instance, we also found 3 different types of MyHC isoforms in both the PCA and TA muscles. In addition, both the data of Shiotani et al and those reported herein indicate that the PCA muscle has a MyHC isoform profile that is slower than that of the TA muscle.

Although Shiotani et al did not examine regional differences in MyHC isoform composition, the findings of the present study demonstrate that there are regional differences in both the PCA and TA muscles. However, the regional differences appear to be more pronounced in the PCA muscle. This finding is in direct contrast to canine laryngeal muscle, where it has been observed that the greatest regional differences occur in the TA muscle.¹⁹

Although the whole muscle analyses used by Shiotani et al¹⁶ provide a global overview of the MyHC isoform composition of laryngeal muscle, there are important issues that can only be resolved by single-fiber analyses (ie, immunohistochemical, electrophoretic). In the past, it has been assumed that a given muscle fiber could only express one type of MyHC isoform. However, studies^29-34 during the past 5 to 10 years have shown that hybrid fibers commonly occur in some muscles. Within this context, it would be inappropriate to assume that the large proportions of the fast type IIX MyHC isoform found in both the medial and lateral regions of the TA muscle were confined to fibers that only expressed this isoform. Rather, the fast type IIX MyHC isoform in these regions was found primarily in hybrid fibers where the fast type IIX MyHC isoform was coexpressed with the fast type IIA MyHC isoform (Figure 4).

As previously noted, we observed in companion studies that both canine¹⁹ and rodent¹⁸ laryngeal muscles contained so-called hybrid muscle fibers. One of the primary objectives of this study was to test the hypothesis that human laryngeal muscle is also composed of hybrid fibers. As shown in Figure 3 and Figure 4, both the PCA and TA muscles had significant proportions of hybrid fibers. The primary type of hybrid fiber found in the PCA was the I/IIA fiber. In contrast, both the medial and lateral regions of the TA muscle contained substantial proportions of IIA/IIX hybrid fibers. In a previous study,¹⁷ we noted that the canine lateral cricothyroid muscle had a substantial proportion of hybrid fibers. Based on the present findings and those of the 2 companion articles, it is clear that hybrid fibers are a common rather than novel occurrence in laryngeal muscles.

In a previous study, Shiotani et al¹⁶ reported the presence of a band that migrates between that of the slow type I and fast type IIA MyHC isoforms. These investigators noted that the migration pattern of this band was analogous to that of the type IIL MyHC isoform found in rodent laryngeal muscle. The findings from recent studies^14,35 suggest that the rodent type IIL MyHC isoform is really an extraocular MyHC isoform. Importantly, Shiotani et al found that this unidentified band did not react with a MAb specific for the extraocular MyHC isoform. Hence, it is unclear whether this unidentified band represents a product of protein (ie, myosin) degradation or a MyHC isoform unique to human laryngeal muscle.

In contrast to the findings of Shiotani et al, we did not observe a band between the slow type I and fast type IIA MyHC isoform bands (Figure 6 and Figure 7). There are several differences between our study and that of Shiotani et al that might explain this discrepancy. First, the muscle samples examined in the present study were frozen within approximately 15 to 30 seconds following extirpation. Hence, our results are not obfuscated by the possibility of protein degradation as are those of Shiotani et al, who obtained specimens from cadavers (within 24 hours of death). Interestingly, Shiotani et al found this additional band in only 2 of 6 specimens. Second, our specimens were harvested from individuals undergoing laryngectomy for squamous cell carcinoma. Although the muscles we analyzed were contralateral to the involved side, it is possible that the innervation, mechanical loading, or extracellular milieu of these muscles was altered in such a way as to suppress the expression of such an isoform. In contrast, the specimens examined by Shiotani et al were obtained from cadavers that were absent of disease related to the larynges. Clearly, further studies are needed to confirm or reject the presence of a fourth MyHC isoform in human laryngeal muscle. If there is a fourth MyHC isoform in human laryngeal muscle, it must represent a small proportion of the total myosin pool given that Shiotani et al did not quantitate its presence.

The MyHC isoform compositions of the PCA and TA muscles from human, canine, and rodents are shown in Figure 8 and Figure 9. This data set clearly demonstrates that, for both the PCA and TA muscles, the MyHC isoform profiles are slowest in the human and fastest in the rat. Moreover, the rodent laryngeal muscle appears to be unique in several ways compared with both human and dog laryngeal muscle. For instance, rodent laryngeal muscles express the so-called type IIL MyHC isoform (ie, embryonic MyHC isoform), whereas this isoform appears to be absent in both human and canine samples. In addition, both the human PCA and TA muscles express significant proportions of the slow type I MyHC isoform, whereas this MyHC isoform is almost completely absent in the rodent PCA and TA muscles. Also, human laryngeal muscle does not express the fast type IIB MyHC isoform, whereas this isoform is the most abundant MyHC isoform found in both the rodent PCA and TA muscles. Collectively, these significant discrepancies between human and rodent laryngeal muscle raise the issue of whether rodent laryngeal muscle is a suitable analog for studying issues related to the MyHC isoform composition of human laryngeal muscle.

Despite differences in the types and proportions of fiber types across species, there are some general patterns that are common to human, canine, and rodent laryngeal muscle. First, in each species, the PCA muscle appears to have a slower MyHC profile than the TA muscle. Second, in both canine and human laryngeal muscle, the horizontal region of the PCA muscle appears to have a slower MyHC isoform profile than the other regions of this muscle. Interestingly, the dominant fiber type is the slow type I fiber in the horizontal regions of the PCA muscles in both the human and canine. With respect to the TA muscle, the medial regions in both humans and canines appear to have a slower MyHC isoform profile than the lateral region. Finally, human, canine, and rodent PCA and TA muscles contain so-called hybrid fibers. Hence, these data suggest that hybrid fibers occur across a broad spectrum of mammals.

As shown in Figure 3, the MyHC isoform compositions of the 2 regions (horizontal and oblique) of the PCA are notably different from one another. The horizontal region has a high proportion of slow type I fibers (approximately 70% of the total population). These fibers are thought to be nonfatigable and have slow contractile properties. These physiologic properties might be helpful in counterbalancing the pull of the cricothyroid muscle (via the vocal ligament) in high-pitched phonation. Conversely, the oblique region of the PCA was found to express mainly the fast type IIA and IIX MyHC isoforms. Since these isoforms can produce high contraction speeds, they would be very effective at quickly opening the glottis during inspiration.

The MyHC isoform expression in the 2 regions of the human TA (lateral and medial) also varied, although less so than in the PCA muscle. Although there was no regional difference found with respect to the slow type I and fast type IIA isoforms, the lateral region was found to express a higher percentage of the fast type IIX than the medial region. It is postulated that the lateral portion of the TA is involved in reflex laryngeal closing and thus would need to produce the fast contraction compatible with the fast type IIX MyHC isoform. This quick closure would be useful in preventing foreign bodies from entering the trachea.

There are 4 major findings derived from this study. First, 3 MyHC isoforms (slow type I, fast type IIX, and fast type IIA) were identified in the human PCA and TA laryngeal muscles. Second, the 2 regions of PCA muscle have very different MyHC compositions, with the more lateral oblique belly having faster characteristics. The 2 regions of TA muscle have subtle, yet meaningful, differences in MyHC composition, the lateral nonvocalis portion expressing faster isoforms than the medial vocalis portion. Third, many PCA and TA fibers simultaneously coexpressed 2 or more MyHC isoforms. Finally, there appear to be distinct differences between the MyHC isoform compositions of human and rodent laryngeal muscles, suggesting that rodent laryngeal muscle may not be a good analog for exploring issues related to the MyHC isoform composition of human laryngeal muscle.

Accepted for publication December 22, 1999.

Reprints: Vincent J. Caiozzo, PhD, Medical Sciences I B-152, Department of Orthopaedics, College of Medicine, University of California, Irvine, CA 92717 (e-mail: vjcaiozz@uci.edu).