Unique fiber phenotype composition and metabolic properties of the stapedius and tensor tympani muscles in the human middle ear

The middle ear muscles have vital roles, yet their precise function in hearing and protection remains unclear. To better understand the function of these muscles in humans, the morphology, fiber composition, and metabolic properties of nine tensor tympani and eight stapedius muscles were analyzed with immunohistochemical, enzyme‐histochemical, biochemical, and morphometric techniques. Human orofacial, jaw, extraocular, and limb muscles were used as references. The immunohistochemical analysis showed that the stapedius and tensor tympani muscles were markedly dominated by fibers expressing fast contracting myosin heavy chain MyHC‐2A and MyHC‐2X (79 ± 6% vs. 86 ± 9%, respectively, p = 0.04). In fact, the middle ear muscles had one of the highest proportions of MyHC‐2 fibers ever reported for human muscles. Interestingly, the biochemical analysis revealed a MyHC isoform of unknown identity in both the stapedius and tensor tympani muscles. Muscle fibers containing two or more MyHC isoforms were relatively frequently observed in both muscles. A proportion of these hybrid fibers expressed a developmental MyHC isoform that is normally absent in adult human limb muscles. The middle ear muscles differed from orofacial, jaw, and limb muscles by having significantly smaller fibers (220 vs. 360 μm2, respectively) and significantly higher variability in fiber size, capillarization per fiber area, mitochondrial oxidative activity, and density of nerve fascicles. Muscle spindles were observed in the tensor tympani muscle but not in the stapedius muscle. We conclude that the middle ear muscles have a highly specialized muscle morphology, fiber composition, and metabolic properties that generally showed more similarities to orofacial than jaw and limb muscles. Although the muscle fiber characteristics in the tensor tympani and stapedius muscles suggest a capacity for fast, fine‐tuned, and sustainable contractions, their difference in proprioceptive control reflects different functions in hearing and protection of the inner ear.


| INTRODUC TI ON
In the middle ear, the transmission of sound vibrations and pressure changes to the cochlea is modulated by two muscles, the tensor tympani muscle, which is attached to the malleus shaft, and the stapedius muscle, which is attached to the stapes. Activation of the stapedius muscle, the smallest muscle in the human body, decreases the movements of the stapes in the oval window of the cochlea, while activation of the tensor tympani muscle pulls the malleus medially, increasing the tension of the tympanic membrane. In several mammals, both the stapedius and the tensor tympani muscles contract in response to auditory stimuli (Borg, 1972;Eliasson & Gisselsson, 1955;Murata et al., 1986), while in humans, the stapedius muscle reflex is considered to be the dominant acoustically evoked reflex pathway to protect the inner ear from loud noises (Mukerji et al., 2010). The reflex is triggered by loud sound (70-100 dB), which causes muscle contraction dampening the vibration of the stapes and decreasing the intensity of sound transmitted to the cochlea by about 15 dB. In contrast, the human tensor tympani muscle seems not to be activated by external sound but rather by tactile stimulation of facial areas, swallowing, phonation, and most commonly, as part of the startle reaction . It has also been attributed to regulating pressure in the middle ear (Sadé & Ar, 1997).
Although various roles have been assigned to the stapedius and the tensor tympani muscle, the functions of these muscles in humans are still largely unclear and often speculative (Mukerji et al., 2010).
In muscles, the myosin heavy chain (MyHC) is the major component of the myosin motor protein which converts chemical energy into mechanical force. This motor protein exists in different isoforms, and the type of isoforms in a muscle determines its contractile properties. In humans, at least eight well-defined MyHC isoforms are present, each encoded by a specific gene (Schiaffino & Reggiani, 2011). Three of these are identified in adult human limb muscles, one slow myosin isoform, MyHC-1 (I/β), coded by the MyH7 gene, and two fast isoforms, MyHC-2A and MyHC-2X, coded by the MyH2 and MyH1 gene, respectively. Muscle fibers containing slow MyHC-1 or fast MyHC-2A have a relatively higher mitochondrial oxidative capacity and capillary supply than MyHC-2X fibers, which, on the other hand, have higher contraction velocity and power during short activities (Schiaffino & Reggiani, 2011). A high proportion of muscle fibers co-expressing mixtures of different MyHC isoforms, that is, hybrid fibers, give the muscles a wide range of physiological properties.
Despite the vital roles of the middle ear muscles in protecting the inner ear, the morphology and physiological properties of these muscles, such as power, speed of shortening, and fatigue resistance, are largely unknown. To better understand the functions of the human middle ear muscles, this study aimed to investigate the muscle morphology, MyHC composition, oxidative capacity, and vascularization of the human stapedius and tensor tympani muscles.

| Approval of the study
Autopsy collection was done according to Swedish laws and regulations on autopsy and transplantation and in agreement with the declaration of Helsinki. Consent was given by the Swedish National Board of Health and Welfare (Dnr 5254-17,784).

| Muscle samples
The middle part of almost the entire musculus tensor tympani (five females, four males, mean age: 59.8, range: 28-95 years) and musculus stapedius (three females, five males, mean age: 68.4, range: 28-95 years) were obtained postmortem from previously healthy subjects. For comparison, muscle samples were obtained from the orofacial, jaw, and limb muscles of five previously healthy subjects (range: 15-75 years). The orofacial samples were taken from the middle part of the zygomaticus major and the palatopharyngeus muscle (three subjects). The jaw muscle sample was taken from the middle superficial part of the anterior masseter muscle, and the samples from the limb were obtained from the middle superior parts of both the biceps brachii (long head) and the vastus lateralis. For analysis of MyHC isoforms by gel electrophoresis, one additional muscle, the oblique superior extraocular muscle (EOM), was used. All muscle specimens were obtained 1-2 days postmortem, a delay acceptable for obtaining reliable fiber typing (Eriksson et al., 1980). The samples were mounted for transverse sectioning in optimal cutting temperature (OCT) compound (Tissue-Tek; Miles Inc.), rapidly frozen in liquid propane, chilled with liquid nitrogen, and stored at −80°C until further processing.

| Immunohistochemistry
Serial muscle cross-sections, 5-6 μm thick, were cut in a cryostat at −20°C and mounted on glass slides. Immunohistochemical staining was performed using well-characterized monoclonal antibodies (mAbs) and modified standard immunohistochemical techniques. For antibody specificity and concentrations, see Table 1. In brief, the sections were immersed in 5% normal nonimmune donkey serum (Jackson ImmunoResearch Laboratories, Inc.) for 15 min and after that, rinsed in 0.01 M phosphate-

| Morphometric analysis
The entire or nearly the whole muscle cross-section area in both the stapedius and tensor tympani muscle samples was scanned in  >200 fibers from each muscle sample. The quantification of capillarization of the middle ear muscles was limited to six tensor tympani and three stapedius muscles owing to a lack of muscle tissue.
The analysis of muscle fiber capillarization was based on 2505 fibers in the tensor tympani muscle (a mean of 418 fibers per subject, range: 105-537) and 1808 fibers in the stapedius muscles (a mean of 363 fibers per subject, range: 255-497). In a normal human limb muscle, capillarization of 50 fibers is sufficient to make a robust analysis of the capillary network (McCall et al., 1998).

| Variability in muscle fiber diameter
Variability in fiber area was expressed for each fiber type as the coefficient of variation (CV). The CV was calculated according to the formula standard deviation x 1000/mean fiber area.

| Capillary variables
Capillary density (CD) was calculated as the total number of capillaries per mm 2 in the cross-section of muscle fibers. To analyze the number of capillaries around each fiber (CAF), all capillaries within a distance of 5 μm from each individual muscle fiber were included.
Capillaries related to each individual fiber cross-sectional area (CAFA) were calculated according to the formula CAF/fiber crosssectional area × 10 3 .

| Biochemical methods
Frozen muscle cross-sections (10-40 μm thick) from five tensor tympani and five stapedius muscles, one EOM, and one biceps brachii muscle were placed in a Laemmli sample buffer (Bio-Rad Laboratories AB). Myosin heavy chain isoforms were analyzed in 8% sodium dodecyl sulfate (SDS)-glycerol gels containing 30% glycerol (Talmadge & Roy, 1993). Protein fractions were separated in the Bio-Rad mini-Protean 3-cell system (Bio-Rad Laboratories). The upper buffer contained 10 mM 2-mercaptoethanol (Blough et al., 1996). The gels were silver-stained for 24 h at 70 V and scanned in a soft laser densitometer (Molecular Analyst software; Bio-Rad Laboratories) to determine the relative content of MyHC isoforms. For immunoblotting, the separated proteins were transferred to nitrocellulose sheets and exposed to mAbs A4.951 and A4.74.

| Muscle spindles
Muscle spindles were in transverse muscle sections defined as groups of small-sized muscle fibers surrounded by a connective tissue capsule.

| Statistical analysis
All statistical tests were performed using SPSS version 23 (IBM SPSS Statistics; IBM Corporation). Normality in the distribution of data was tested using the Shapiro-Wilk test. Comparisons between more than two groups were made using one-way ANOVA, and significance between individual groups was confirmed using the Tukey post hoc test. Differences between the two groups were made using the independent sample t-test for variables with normal distribution and the Mann-Whitney U test for non-normally distributed variables. p ≤ 0.05 was considered significant 3 | RE SULTS

| General muscle morphology and immunoreaction of antibodies
The tensor tympani and the stapedius muscles were characterized by a population of relatively less densely packed small-sized fibers, often with a round rather than a polygonal shape. However, the interindividual and intramuscular variability in fiber size and fiber form was noticeable in both muscles ( Figure 1). Muscle spindles were observed in three of the tensor tympani muscle samples, while no muscle spindles were found in the stapedius muscles ( Figure 2). The density of nerve bundles in the muscle cross-section was observed to be higher in the tensor tympani and stapedius than in the limb muscles ( Figure 3).   Figure 1).

| Frequency of fibers expressing neonatal myosin
Muscle fibers co-expressing MyHC-neo were found in six of the eight analyzed stapedius muscles ( Figure 2). The average percentage of fibers containing MyHC-neo in the stapedius muscle was 3.3 ± 3.2%. MyHC-neo was not co-localized with any particular MyHC fiber type.

| Mitochondrial oxidative activity
Muscle fibers expressing slow MyHC-1 generally exhibited strong NADH-TR activity. The majority of fibers expressing fast MyHC-2A or MyHC-2X showed a moderate to strong NADH-TR staining intensity, which was generally stronger than in the corresponding fibers in limb muscles (Figure 1).

| Mitochondrial oxidative activity
The NADH-TR staining activity in the tensor tympani fibers was in the same range as in the stapedius muscle. Whereas fibers containing MyHC-1 generally showed strong staining activity, fibers expressing MyHC-2A or MyHC2X displayed a moderate to strong NADH-TR staining intensity, which generally was stronger than in limb muscles (Figure 1).

| Capillarization
The average CD for all tensor tympani muscles was 428 ± 92 capillaries per mm 2 . The mean CAF was 0.84 ± 0.24, where fibers expressing MyHC-1 and MyHC-1/2A showed the highest CAF values (0.99 ± 0.22 and 1.02 ± 0.36, respectively). The CAFA mean value F I G U R E 3 Cross-sections from a middle ear muscle, the tensor tympani (a and b), and one limb muscle, the biceps brachii (b and d), showing the density of nerve fascicles (a and c) and capillaries (c and d). Panel c is scanned in a lower magnification (20×). The sections are immunostained for laminin in the basement membrane (red) and MyHC-1 in fibers (green). Nerves and capillaries are marked white. Note the high density of nerve bundles in the tensor tympani muscle (a) compared with the lack of nerve fascicles in the area from the biceps brachii muscle (c). Note also the higher density of capillaries in the muscle cross-sections of the tensor tympani muscle (b) compared to the limb muscle (d). Scale bars 50 μm.

| Myosin heavy chain isoforms revealed by gel electrophoresis
The biochemical analysis revealed three major MyHC isoforms in the  Figure 5, arrows) was observed in all analyzed stapedius muscles (12.8 ± 7.4%) and three of the five analyzed tensor tympani muscles (1.6 ± 2.3%), but was not found in the biceps brachii or EOM muscles. The pattern of MyHC isoforms in the gels is shown in Figure 5.

| Age and gender
When the data from the tensor tympani and stapedius muscles were divided according to subjects age, <60 years and ≥60 years, the mean proportion of hybrid MyHC-1/2A and MyHC2A/2X fibers was significantly higher in the older group than in the younger group. No other statistical observations were found. Neither were any differences observed between genders.

| Muscle fiber phenotypes in orofacial, jaw, and limb muscles
The limb and the orofacial and jaw muscles showed a MyHC fiber type composition as previously reported for these muscles (Pontén & Stål, 2007;Stål et al., 1987Stål et al., , 1990Stål et al., , 2003Stål & Lindman, 2000;Staron et al., 2000). The biceps brachii and vastus lateralis muscles had an even distribution of relatively large, densely packed angulated fibers, about an equal proportion of which expressed MyHC-1 and MyHC-2 isoforms. The zygomaticus major muscle and palatopharyngeus were predominated by less densely packed, smallsized rounded fibers containing MyHC-2 isoforms. In contrast, the samples of masseter muscle were predominated by MyHC-1 fibers and a relatively large proportion of small-sized MyHC-2X fibers (Table 2; Figure 6).

| Comparison between middle ear, cranial, and limb muscles
The tensor tympani muscle had a significantly higher proportion of fibers containing MyHC-2 isoforms compared to the stapedius muscle (86% vs. 79%, p = 0.04) (Figure 4). No significant differences were observed between the two middle ear muscles for fiber area and capillary parameters with the exception that the fiber area for MyHC-1/2 fibers was significantly larger in the tensor tympani muscle ( Figure 4; Table 3).
Both the stapedius and tensor tympani muscles generally shared more morphological similarities with the orofacial than with the jaw and limb muscles, i.e., the fibers were mainly composed of loosely packed and very small-sized MyHC-2 fibers with high NADH-TR activity (Table 2; Figure 4). However, the middle ear muscles differed also from orofacial, jaw and limb muscles by having significantly higher variability in fiber size (CV) and more extensive capillarization per fiber area (CAFA) (Figure 4).

TA B L E 3
Muscle fiber area (μm 2 ), the number of capillaries around fibers (CAF), and the number of capillaries around fibers relative to its cross-sectional area (CAFA) of fibers expressing different MyHC isoforms in the stapedius and tensor tympani muscles.

F I G U R E 4
Comparison between human middle ear, cranial, and limb muscles. Bar graphs showing fiber area (a), variability in fiber area (CV) (b), percentage of MyHC-2 fibers (c), capillary density (CD) (d), number of capillaries around fibers (CAF) (e), and capillaries around fibers relative to its cross-sectional area (CAFA) (f) in the stapedius (orange) and tensor tympani (blue) muscles. For comparison, data for two different orofacial muscles, the zygomaticus major and palatopharyngeus, one masticatory muscle, the masseter, and two limb muscles, biceps brachii and musculus vastus lateralis, are included. Values are expressed as means and ± standard deviations. Statistical differences (p < 0.05) to the stapedius muscle are marked (*), and to the tensor tympani muscle are marked ( †). Note the small fiber area (a), high variability in fiber size (b), high frequency of MyHC-2 fibers (c), and a high number of capillaries per fiber area (f) in both middle ear muscles.

| DISCUSS ION
This is the first detailed study describing muscle characteristics of the human stapedius and tensor tympani muscles. The investigation revealed that the middle ear muscles have distinctive contractile, metabolic, and proprioceptive properties to meet the complex demands of auditory perception and the protection of inner ear structures. Although both muscles have many characteristics in common, they are also individually unique (Table 4).
An interesting finding was that the stapedius and tensor tympani muscles had one of the highest proportions of fastcontracting MyHC-2 fiber types reported in any human muscle. Moreover, the gel electrophoresis showed a relatively high amount of MyHC-2X protein, indicating that the muscles are even more powerful and fast-contracting than revealed from the immunohistochemical analyzes. Further support for very rapid contractions is the observation of a protein band of unknown identity at the level of myosin in the gel-electrophoretic separation. Smerdu and Cvetko have previously reported an unidentified band at the same level in human laryngeal muscles, which are considered to produce extremely fast movements (Smerdu & Cvetko, 2013). They showed that the laryngeal muscles contain the transcript for MyHC-2B, the fastest MyHC isoform in mammals, but as in our study, they could not identify the protein in the gels using an antibody directed against MyHC-2B. Given the short reflex time the ear muscles have in responding to loud sounds to protect the cochlea, it is reasonable to speculate that the stapedius muscle may contain an undetected, very fast protein isoform.
Further investigations are required to ascertain whether human middle ear muscles contain a MyHC-2B-like isoform or another unique superfast MyHC isoform.
Another interesting observation was the presence of fibers coexpressing developmental MyHC-neo in the middle ear muscles. This MyHC isoform has been reported as a normal feature throughout adult stages in human masticatory, orofacial, and EOM muscles but not in limb muscles (Ciciliot & Schiaffino, 2010;Sartore et al., 1987;Stål et al., 1994;Stål & Lindman, 2000). Although the role of MyHCneo in the fibers is unclear, it has been speculated that neonatal myosin has low contraction velocity and is specialized in contractions against a low load (Schiaffino et al., 2015).  (Granberg et al., 2010;Stål & Lindman, 2000).
The predominance of high oxidative fast-contracting fibers with the ability for fine-tuned movements supports the stapedius muscles widely accepted role as a protector of the cochlea in the inner ear against damaging sound levels. In excessive sound vibrations, the acoustic stapedius reflex contracts the muscle and tilts the stapes away from the fenestra vestibuli, which stiffens the ossicular chain.
This acoustic reflex function explains why paralysis of the muscle causes hyperacusis (Rubini et al., 2020). The acoustic stapedius reflex is also reported to be involved in improved speech recognition at low-frequency moderate noise levels (Aiken et al., 2013) as well as degraded speech recognition at high presentation levels (Shehorn et al., 2020). Further studies regarding the acoustic reflex feedback to the stapedius muscle are needed to confirm these theories.
Compared with the stapedius muscle, considerably less is known about the function of the tensor tympani muscle since no valid method of detecting its contraction exists. The muscle acts as an antagonist to the stapedius muscle and is proposed to form a functional unit with the levator veli palatini collaborating in the ventilation of the TA B L E 4 Muscle characteristics and known orofacial, pharyngeal, and middle ear factors influencing the activity of the human middle ear, facial, and limb muscles. middle ear (Kierner et al., 2002). The muscle probably play an active part in middle ear pressure regulation and clearance since the tensor tympani is activated simultaneously with the levator veli palatini opening the Eustachian tube (Sadé & Ar, 1997). Even though contractions of the tensor tympani muscle have been audiometrically shown to result in low-frequency mixed hearing loss (Wickens et al., 2017), there is evidence that the tensor tympani in humans is not activated by loud sounds from the environment and has non-acoustic functions (Salomon & Starr, 1963). Several reports agree that the muscle is activated when exposed to self-generated auditory stimuli such as swallowing, speaking, or chewing and nonauditory stimuli in the anticipation of loud sounds (Edmonson et al., 2022).
The tensor tympani tendon has been suggested to protect against outward displacement of the ossicular chain (Hüttenbrink, 1989).
However, we have previously reported that the ligament by itself has very little effect in limiting the outward movement of the tympanic membrane and malleus (Rönnblom et al., 2021). In the same study, we found the combination of a negative pressure created by a finger being extracted from a wet ear canal and a simultaneous counteracting reflexive force by the tensor tympani muscle could cause an isolated malleus fracture with an intact tympanic membrane. Thus, the high proportions of fibers with fast, relatively powerful, and sustainable contractions in the human tensor tympani muscles, together with the presence of muscle spindles contributing to a modulated stretch reflex response, highlights its well-suited role in protecting the ossicular chain and the inner ear from the excessive outward movement of the tympanic membrane and the ossicular chain.
The presence of muscle spindles in the tensor tympani muscle, but the lack of typical spindles in the stapedius muscle, highlights the different roles of these muscles. As the muscle spindle is a sensory receptor that primarily transmits changes in the length and speed to the central nervous system, which in turn coordinates adequate motor neuron activity to resist muscle stretching, the tensor tympani muscles seem better adapted for flexible sensory performance than the stapedius muscle. The difference in proprioceptive control of the muscles may be attributed to the fact that the stapedius muscle is involved in reflex movements that rapidly pull the stapes sideways to protect the inner ear from loud noises while the tensor tympani muscle pulls the tympanic membrane rapidly inward but in a more varied and fine-tuned movement pattern.
We conclude that the human middle ear muscles have a highly specialized muscle morphology and fiber phenotype composition with specific contractile, metabolic, and proprioceptive properties that have more similarities with human orofacial than with jaw and limb muscles. The high proportion of fast-contracting MyHC-2 fibers with pronounced mitochondrial oxidative capacity, the relatively high presence of hybrid fibers, and the extensive capillarization and density of nerve fascicles in both the stapedius and tensor tympani muscles show that these muscles are adapted to perform rapid finetuned contractions with relatively high endurance. However, despite these similarities, they have different functions in the middle ear.
The stapedius muscle, which lacks ordinary muscle spindles, has a muscle fiber composition adapted to act rapidly in a fine-tuned manner stiffening the ossicular chain through the acoustic reflex.
In contrast, the human tensor tympani muscle which contain muscle spindles, perform various non-acoustic functions in a fast and precise fashion and assists in ventilation of the middle ear. Stretch reflexes triggered by muscle spindles most likely prevent excessive outward movements of the tympanic membrane avoiding disruption of the ossicular chain or damage to the inner ear.

ACK N OWLED G M ENTS
We thank Dr. Mona Lindström and Mrs. Anna-Karin Olofsson for their excellent laboratory work. This work was supported by the Region of Norrbotten (NLL-968473).

CO N FLI C T O F I NTE R E S T S TATE M E NT
The authors declare that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.