The Sensory Innervation of the Human Pharynx: Searching for Mechanoreceptors



The coordinate neural regulation of the upper airways muscles is basic to control airway size and resistance. The superior constrictor pharyngeal muscle (SCPM) forms the main part of the lateral and posterior walls of the pharynx and typically is devoid of muscle spindles, the main type of proprioceptor. Because proprioception arising from SCPM is potentially important in the physiology of the upper airways, we have investigated if there are mechanical sensory nerve endings substitute for the muscle spindles. Samples of human pharynx were analyzed using immunohistochemistry associated to general axonic and Schwann cells markers (NSE, PGP 9.5, RT-97, and S100P), intrafusal muscle fiber markers, and putative mechanical sense proteins (TRPV4 and ASIC2). Different kinds of sensory corpuscles were observed in the pharynx walls (Pacini-like corpuscles, Ruffini-like corpuscles, spiral-wharves nerve structures, and others) which are supplied by sensory nerves and express putative mechanoproteins. No evidence of muscle spindles was observed. The present results demonstrate the occurrence of numerous and different morphotypes of sensory corpuscles/mechanoreceptors in human pharynx that presumably detect mechanical changes in the upper airways and replace muscle spindles for proprioception. Present findings are of potential interest for the knowledge of pathologies of the upper airways with supposed sensory pathogenesis. Anat Rec, 296:1735–1746, 2013. © 2013 Wiley Periodicals, Inc.


The information about muscle stretch, that is, a part of the sense of proprioception, from the cephalic muscles is not only important in regulating the masticatory force and oromotor behaviors but also in the response of important reflexes related to speech, swallowing, cough, vomit, or normal breathing (for a review see Miller, 2002). Therefore, the coordinate neural regulation of the upper airways muscles is basic to control airway size and resistance.

The superior constrictor pharyngeal muscle (SCPM) forms the main part of the lateral and posterior walls of the pharynx, and is intimately associated with the normal movement of the upper airways during speech, swallowing, and respiration. SCPM is doubly innervated by the branches of the pharyngeal plexus, derived from the glossopharyngeal and vagal cranial nerves, and a small contribution of facial cranial nerve (see Shimokawa et al., 2004). It is organized into functional fiber layers, as indicated by distinct motor innervation as well specialized muscle fibers distribution (Mu and Sanders, 2007), and in normal human subjects SCPM usually exhibits phasic expiratory activity during both wakefulness and sleep (see for references Kuna, 2000).

The proprioceptors are specialized sensory receptors that detect stretch or tonicity stimuli within muscles, and the primary proprioceptors in striated muscles are the muscle spindles, although some others exist localized at the myotendinous junction (Zelena, 1994). Some years ago Bossy and Vidic (1967) doubted of the occurrence of muscle spindles in the human pharynx, and Seto (1963) in his book “Studies on the sensory innervation (human sensibility)” did not describe the occurrence of muscle spindles in the muscles of this region. Consistently, the pharyngeal muscles have been regarded to be free from spindles muscles (Kuehn et al., 1990; Liss, 1990), and no other kinds of proprioceptors (Golgi tendon organs, Pacinian corpuscles) have been described in the pharynx. Interestingly, the proprioception from the head muscles seems to be solely derived by innervation arising from the trigeminal nerve (see for a review Lazarov, 2007), and the pharyngeal muscles have not innervation from this cranial nerve. Nevertheless, Sengupta and Sengupta (1978) described muscle spindles in the inferior constrictor pharyngeal muscle of the crab-eating monkey (Macaca irus), and typical muscle spindles were found in other muscles of the oropharyngeal region (Kuehn et al., 1990; Liss, 1990).

The proprioception arising from SCPM is potentially important in the physiology of the upper airways, whereas typical muscle spindles are absent in the pharyngeal muscles. Thus we examined if typical muscle spindles are substituted by other sensory formations able to detect stretch. Therefore, we have conducted a morphological and immunohistochemical study focused on the detection of muscles spindles or morphologically differentiated mechanoreceptors in human SCPM. Moreover because putative mechanoproteins have been detected in mechanoreceptors (see Del Valle et al., 2012; Chen and Wong, 2013), including muscle spindles and other putative proprioceptors (Calavia et al., 2010; Simon et al., 2010) we have investigated the occurrence associate to nerves of some of them, in particular ASIC2 and TRPV4. Acid-sensing ion channels (ASICs) are a H+-gated group of the voltage-insensitive ion channels, and some members of this superfamily, especially ASIC2, may function as mechanosensors or are required for mechanosensation in a diverse range of species and cell types (Lumpkin and Caterina, 2007; Holzer, 2009; Tsunozaki and Bautista, 2009). On the other hand, TRPV4 is a polymodal Ca2+-entry channel member of the vanilloid (V) subfamily of transient receptor potential (TRP) ion channels. It not only works as a mechanosensor including detection of osmotic stimuli but also responds to warm temperatures, acidic pH, and several chemical compounds (Eid and Cortright, 2009; Nilius and Owsianik, 2011). Both ASIC2 and TRPV4 have been localized in both mechanoreceptive sensory neurons and at the site of mechanotransduction in nerve terminals, that is, the mechanreceptors (see for a review and references Lumpkin et al., 2010; Del Valle et al., 2012; Chen and Wong, 2013), and ASICs are the principle channels in skeletal muscle afferents (Gautam and Benson, 2013).

This study was aimed to investigate the morphological and molecular basis of the mechanoreceptive-proprioceptive innervation of the human SCPM, and might serve as a baseline for future studies in the pathology of upper airways with a hypothesized neurological pathogenesis, like obstructive sleep apnea (Guilleminault et al., 2005; Guillaumet and Ramar, 2009; Levy et al., 2012).



Samples from pharynx were obtained during removal of organs for transplantation from subjects that died in traffic incidents (Hospital Universitario Central de Asturias, Oviedo). They were taken from three females (age-range, 36 and 51 years), and four males (age-range, 26–60 years) free of known neurological disease. The material was obtained in compliance with Spanish Laws, and the guidelines of the Helsinki Declaration II. The tissue samples were cleaned in cold saline-solution, fixed in 10% formaldehyde in 0.1M phosphate buffer saline (PBS) at pH 7.4 for 24 hr at 4°C, dehydrated and routinely embedded in paraffin. From each pharynx the lateral and posterior walls of the oropharynx, and the lateral walls of the hypopharynx were sampled. Samples consisted of pieces 1–2 cm2 containing the whole thickness of the pharyngeal walls. Sections at 10 μm, or 30 μm, thickness perpendicular or parallel to the pharynx surface were obtained, mounted on gelatin-coated microscope slides and processed for immunohistochemistry.


Indirect peroxidase antiperoxidase immunohistochemistry was used, and the procedure was developed as follows: sections were deparaffinized and rehydrated, 10 rinsed in 0.05M HCl Tris buffer (pH 7.5) containing 0.1% bovine serum albumin and 0.1% Triton X-100. Thereafter the endogenous peroxidase activity (3% H2O2) and non-specific binding (10% fetal calf serum) were blocked, and the sections incubated overnight in a humid chamber at 4°C with primary antibodies (Table 1). The antibodies were diluted in a solution of Tris–HCl buffer (0.05M, pH 7.5) containing 0.1% bovine serum albumin, 0.2% fetal calf serum and 0.1% Triton X-100. After incubation with the primary antibodies, sections were rinsed in the same buffer and incubated with Dako EnVision System labeled polymer-HR anti-rabbit IgG or anti-mouse IgG (DakoCytomation, Denmark) for 30 min at room temperature. Finally, sections were washed and the immunoreaction visualized using 3-3'-diaminobenzidine as a chromogen. For control purposes, representative sections were processed in the same way as described above using non-immune rabbit or mouse sera instead of the primary antibodies, or omitting the primary antibodies in the incubation. Moreover, additional experiments using pre-absorbed antibodies for ASIC2 and TRPV4 (5 µg of the blocking peptide in 1 mL of the antibody working solution) were carried out (blocking peptides are in Table 1). To ascertain structural details, sections were slightly counterstained with hematoxylin & eosin.

Table 1. Primary antibodies used in the study
  1. a

    Glostrup, Denmark.

  2. b

    Poole, England, UK.

  3. c

    Mannheim, Germany.

  4. d

    Thermo Scientific, Freemont, CA, USA.

  5. e

    Seattle, WA, USA.

  6. f

    Cambridge, UK.

  7. g

    Billerica, MA, USA.

  8. ASIC2: acid-sensing ion channel 2; NSE: neuron specific enolase; PGP 9.5: Protein gene product 9.5; RT-97: Phosphorylated 200 kDa neurofilament protein; TRPV4: transient receptor potential channel vanilloid 4. Anti-ASIC2: rabbit polyclonal antibody raised against a synthetic peptide from the extracellular domain of mouse ASIC2 conjugated to an immunogenic carrier protein. Catalogue LS-C93915. Blocking peptide: Lifespan Biosciences LS-PB156. Anti-TRPV4: rabbit polyclonal antibody raised against a synthetic peptide derived from the cytoplasmic n-terminus conserved in mouse, human and rat TRPV4 conjugated to immunogenic carrier protein. Catalogue ab63003. Blocking peptide: Abcam ab166832.

General neural markers
NSE (clone BBS/NC/VI-H14)Mouse1:1000Dakoa
PGP 9.5Rabbit1:1000Biogenesisb
S-100 proteinRabbit1:1000Dakoa
S100P (clone 4C4.9)Mouse1:1000Thermo Scientificd
Putative mechanoproteins
ASIC2Rabbit1:200Lifespan Biosciencese
Intrafusal muscle fibers
Myosin heavy chains (clone A4.1025)Mouse1:100Millipore Corg

Double Immunohistochemistry

Sections were processed for simultaneous detection of NSE and S100 protein, RT-97 and S100 protein, NSE and ASIC2, NSE and TRPV4. The antibodies against NSE, PGP 9.5, and RT97 were used to label axons; antibodies against S100 protein to label Schwann cells, and Schwann-related cells (see Vega et al., 2009). Double immunostaining was performed on 10-µm and 30-µm thick deparaffinized and rehydrated sections. Non-specific binding was reduced by incubation for 30 min with a solution of 1% bovine serum albumin in tris buffer solution (TBS). The sections were then incubated overnight at 4°C in a humid chamber with a 1:1 mixture of anti-NSE and anti-S100P antibodies, or anti-RT97 and anti-S100P antibodies; anti-NSE or anti-RT97 antibodies with anti-ASIC2 or anti-TRPV4 antibodies; and anti-S100 antibody with anti-ASIC2 or anti-TRPV4 antibodies. After rinsing with TBS, the sections were incubated for 1 hr with Alexa fluor 488-conjugated goat anti-rabbit IgG (Serotec, Oxford, UK), diluted 1:1000 in TBS containing 5% mouse serum (Serotec), then rinsed again and incubated for another hour with Cy™ 3-conjugated donkey anti-mouse antibody (Jackson-ImmunoResearch, Baltimore, MD) diluted 1:50 in TBS. Both steps were performed at room temperature in a dark humid chamber. Finally, to ascertain structural details sections were counterstained with DAPI (10 ng/mL). Sections were then washed, dehydrated, and mounted with Entellan®. Triple staining was detected using a Leica DMR-XA automatic fluorescence microscope coupled with a Leica Confocal Software, version 2.5 (Leica Microsystems, Heidelberg GmbH, Germany) and the images captured were processed using the software Image J version 1.43 g Master Biophotonics Facility, Mac Master University Ontario (; see also the legend of Supporting Information). Controls of the specificity of the immunoreaction developed were performed as for simple immunohistochemistry.


The antibodies against ASIC2 and TRPV4 labeled both nervous (NSE, RT-97, PGP9.5, and S100 protein) exclusively labelled the axons and the Schwann cells, as well as nerve-related structures. S100 protein also labelled the Schwann-related cells of the motor endplates (Musarella et al., 2006). Also, the antibody A4.1025 used to evidentiate myosin heavy chains in muscle spindles, selectively labeled intrafusal fibers in the samples of rectus femoris muscle but never in sections of SCPM (data not shown). The antibodies against for ASIC2 and TRPV4 were detected in both nervous and non-nervous tissues of the pharynx; in particular ASIC2 and TRPV4 immunoreactivities were variably detected in the pharyngeal mucosa, some muscle fibers, blood vessels and submucosa glands (data not shown).

General Data

Nerves in the human pharynx were detected in all the histological layers, and abundant nerve bundles or isolated nerve fibers were observed in the connective tissue of the adventitia, in the connective tissues within the muscular layer, and in the submucosa. In the muscular layer, the nerve bundles have different sizes, and run mainly parallel to the direction of the muscle fibers, but they frequently anastomosed (Fig. 1a). In the connective tissue among muscle fibers different morphotypes of nerve structures resembling “corpuscles” were also regularly found (Fig. 1b). Moreover, nerves were observed forming perivascular and periglandular plexuses (Fig. 1c). Interestingly, in two cases we observed clusters of 2–3 neurons forming small microganglia in the submucous of the pharynx. The neurons displayed immunoreactivity for neuronal markers (Fig. 2a) and the perineuronal cells expressed S100 protein immunoreactivity (Fig. 2b), thus indicating they are satellite glial cells. These ganglia composed of few neurons are sometimes found in the upper segments of the digestive tract (Sbarbati and Osculati, 2007).

Figure 1.

The intramural nerve apparatus of the pharynx was evidenced using an antibody against the general neuronal marker PGP 9.5. It forms nerve bundles (a; nb, arrows), different morphotypes of sensory corpuscles (b, arrows), and perivascular plexuses (c). bv: blood vessels. Scale bar = 50 µm.

Figure 2.

Small microganglia formed by 2–3 neurons were occasionally found in the adventitia of the pharynx. The neuronal somata and the axons were immunoreactive for NSE (a) whereas the perineuronal cells and the Schwann cells displayed S100 protein immunoreactivity (b). n: neurons. Scale bar: 30 µm.

Pacini-Like Corpuscles in the Adventitia and Raphe of the Pharynx

The identification of Pacini-like sensory corpuscles was based on the morphology, the arrangement of the cells forming it, and the expression of specific axonic and Schwann cells markers (see Vega et al., 2009). In the adventitia of the pharynx, structures resembling morphologically Pacinian corpuscles were regularly found. Mostly were rounded capsular formations structures containing a variable number of axons (2–8) surrounded by a more or less developed inner core S100 protein positive (Fig. 3a,b). The axons of these structures are RT-97 positive (Fig. 3c) thus suggesting they are supplied by large mechanosensory neurons (Lawson and Waddel, 1991). Other kinds of capsulated or non capsulated sensory corpuscles were not found in the adventitia. Similarly as for the adventitia in the vicinity of the raphe in the posterior wall of the pharynx polyinnervated Pacini-like corpuscles were also found (Fig. 4).

Figure 3.

Pacini-like corpuscles were found in the adventitia of the pharynx. These corpuscles were polyinnervated. Serial sections showed the arrangement of axons (a and c) and the inner core cells (b). Scale bar: 100 µm.

Figure 4.

Pacini-like corpuscles were observed the vicinity of the raphe of the posterior wall (arrows in a) of the pharynx. Serial sections showed the arrangement of axons (a) and the inner core cells (b and c). Picture in a shows transverse (left arrow) and longitudinal (right arrow) section of two Pacini-like corpuscles (details are in c and ). Scale bar: 100 µm.

Corpuscle-Like Structures Within the Pharyngeal Muscles

Within the muscular layer of the pharynx there were numerous corpuscle-like structures, identified on the basis of their morphology and the immunohistochemical characteristics of the elements forming it. Nevertheless, as for other viscera, there is not a regular or common pattern of arrangement of the elements forming these structures, and they cannot be described following the morphological criteria of the cutaneous sensory corpuscles. Some time they were oval or elongated in shape, apparently capsulated, supplied by a single axon branching within them (Fig. 5a–c); some others resembled elongated or round Ruffini-like corpuscles, of different sizes, with a very complex arrangement of the axons (Fig. 5d,e); some others displayed three-like morphology with organization with a very variable pattern of arborisation (Fig. 5f) or resembled Meissner-like corpuscles (Fig. 5g); also simple elongated formations contacting muscle fibers were found (Fig. 5h). Double immunostaining confirmed these observations (Fig. 6).

Figure 5.

Different morphotypes of sensory corpuscles in the walls of the human pharynx. Scale bar: 30 µm.

Figure 6.

Sensory formations, or corpuscles from the human pharynx labeled with anti-S100 protein antibody conjugated with Alexa fluor 488 (green fluorescence), and with anti-NSE antibody conjugated with Cy™3 (red fluorescence). The images show the relation between axons and glial cells in these sensory structures. Scale bar: 25 µm.

Spiral-Wharves Nerve Structures

A regular finding in all subjects analyzed was the presence in the lateral walls of the pharynx of nerve profiles with spiral-wharves shape (Figs. 7 and 8). They were primarily observed in the connective septa close to the raphe, and were parallel to the muscle fibers of the SCPM. These structures were very long, formed by a few number of axons covered of Schwann cells, and isolated from the surrounding tissues by a capsule (Fig. 7). Interestingly, the axons were immunoreactive for RT-97 (Fig. 8a,b), thus indicating they originates from intermediate or large sensory neurons which are regarded as mechanoceptive and proprioceptive (see Lawson and Waddell, 1991). The disposition of the cell elements forming these spiral-wharves nerve formations can be confirmed using laser confocal microscopy (Fig. 8c).

Figure 7.

In the walls of the pharynx large, spiral-wharves nerve structures were regularly found. They consisted of axons and Schwann cells, capsulated and isolated from the surrounding tissues. Squares in a and b are enlarged in c, d, and e. Scale bar: 100 µm for a and b; 20 µm for c–e.

Figure 8.

RT-97 positive nerve fibers (arrows) in the spiral-wharves nerve structures (a,b), and the arrangement of axons and Schwann cells within them as demonstrated with laser confocal microscopy. Image in c is a reconstruction of confocal image planes obtained from a 30-µm section. Scale bar: 80 µm for a and b; 20 µm for c.

Putative Mechanoproteins Associated to Nerves and Corpuscle-Like Structures

We used antibodies for ASIC2 and TRPV4 as putative mechanical sense proteins. Nerve profiles similar to those described in previous sections were found displaying immunoreactivity for TRPV4 (Fig. 9) and ASIC2 (Fig. 10). Regarding TRPV4 using light microscopy the immunoreactivity showed images similar to the axons (Fig. 9a–c), whereas using laser confocal microscopy the immunoreactivity for TRPV4 was positive in periaxonic cells, presumably Schwann cells and fibroblasts (Fig. 9d,e), as well as in the axons (Fig. 9f).

Figure 9.

Immunohistochemical detection of TRPV4 in different kinds of mechanoreceptors in the walls of the human pharynx. Figures d to f they show positivity for TRPV4 (green fluorescence), and NSE (red fluorescence) in sensory corpuscles. TRPV4 is primarily found in the axons. Scale bar: 50 µm for a–c; 30 µm for d–f.

Figure 10.

Immunohistochemical detection of ASIC2 in different kinds of mechanoreceptors in the walls of the human pharynx. Figures c–h they show positivity for ASIC2 (green fluorescence), and NSE (red fluorescence) in different morphotypes of sensory corpuscles. ASIC2 is primarily found in the axons. Figures i–k they show positivity for ASIC2 (green fluorescence), and NSE (red fluorescence) in the axon of a Pacini-like corpuscle present in the pharyngeal adventia. Scale bar: 70 µm for a; 30 b to k.

On the other hand, ASIC2 immunoreactivity was regularly present in the nerves supplying the pharynx (Fig. 10a) and in spiral-wharves structures present in the connective septa or contacting muscle fibers (Fig. 10b). In the different corpuscle-like structures, ASIC2 partially co-localized with neuronal markers (Fig. 10c–h), including the axons of the Pacini-like corpuscles found in the adventitial peripharyngeal tissue (Fig. 10i–k).


This study was designed to investigate the microscopic sensory innervation of the human pharynx, especially the proprioceptive innervation regarded as a specific part of the mechanoreception. Although the study included the whole thickness of the pharynx walls, the observations were focused on SCPM.

The motor innervation of the human pharynx muscles, including the regional variations in the density of innervations, is well-known at present (van Lunteren and Strohl, 1986; Mu and Sanders, 2007). Conversely the studies on the sensory innervation are very scarce and centered in the sensory intraepithelial fibers supplying the pharyngeal mucosa (Yoshida et al., 2000; Hayakawa et al., 2010). As far as we known no studies exists about the sensory innervation of the muscular layer in spite of the capital role of the pharyngeal muscles in the function of the upper airways.

It is classically admitted that the pharyngeal muscles of most mammals, including man are devoid of the main type of propioceptores, that is, muscle spindles, although they are present in some other muscles of the region like genioglossus, tensor veli palatini, palatoglossus, and some muscles of the tongue (Bossy and Vidic, 1967; Sengupta and Sengupta, 1978). In this study, we have not found any morphological or immunohistochemical evidence for the presence of muscle spindles in SCPM, and did not demonstrate immunoreactivity for specific markers of the intrafusal fibers (Liu et al., 2002). Therefore, we confirm the absence of typical muscle spindles in human SCPM. According to Lazarov (2007) the proprioception of the cephalic muscles depends on the trigeminal nerve. As the pharynx does not receive innervations from this cranial nerve it would lack classic proprioceptive innervation. Thus, the pharynx would have other sensory structures substitute from muscle spindles.

Our data show that the human SCPM has a dense supply of sensory corpuscle-like structures, including Pacini-like, Golgi-like and Ruffini-like corpuscles, and a wide range of other morphologically unclassified sensory formations. In all the cases, they were well differentiated from the surrounding tissues, and displayed immunoreactivity for neuronal and Schwann cell markers. Nevertheless, these presumptive sensory formations cannot be functionally assimilated to those present in other locations of the body, especially in the skin, although based on the morphology alone they should be mechanoreceptors. In any case remains to be established if these possible mechanoreceptors can to replace and to act the same way as the muscle spindles.

An exciting finding of our study was the identification in the wall of the pharynx of a special kind of nerve structure which we have denominated spiral-wharves nerve formations. They are very long, located near the insertion of SCPM, and may change their length according to contraction of SCPM. Based on these criteria we consider that the so-called spiral-wharves nerve formations are good candidate to be stimulated during the contraction-elongation of SCPM. In supporting the sensory nature of these formations was that they are supplied by RT-97 immunoreactive nerve fibers, and display immunoreactivity for the putative mechanoprotein ASIC2. RT-97 immunoreactive axons are regarded as originated from large mechanosensitive and proprioceptive sensory neurons (Lawson and Waddell, 1991).

But the main help to identify the above pharyngeal sensory corpuscles as mechanoreceptors, and therefore candidate to proprioceptors, was the occurrence in these structures of the putative mechanoproteins TRPV4 and ASIC2. At present members of the degenerin/epithelial sodium channels (DEG/ENaC; especially acid-sensing ion channels, ASICs), the TRP channel, and the two-pore domain potassium (K2P) channel super-families are being considered as putative mechanotransducer channels (see Lumpkin and Caterina, 2007; Tsunozaki and Bautista, 2009; Del Valle et al., 2012; Chen and Wong, 2013). Thus, they are expected to be expressed in the mechanosensory neurons and/or in the mechanoreceptors where the mechanotransduction occurs. Both TRPV4 and ASIC2 have been detected in mechanosensory neurons and different kinds of mechanoreceptors (see for a review Del Valle et al., 2012), and ASICs play key roles in muscle afferents (Gautam and Benson, 2013). Importantly, TRPV4 and ASIC2 have been detected in the neurons of the nodose ganglia (Zhang et al., 2004; Page et al., 2005; Lu et al., 2009) that participate in the sensory innervation of the pharynx.

All together the results of this study demonstrate that the human SCPM muscle is innervated by different kinds of mechanoreceptors, which presumably also works as proprioceptors and participates in the neural control of the upper airways. Interestingly, these mechanoreceptors express two putative mechanoproteins ASIC2 and TRPV4. These results can be of interest because abnormal sensory responses have been found some pathological conditions as in the upper airway of obstructive sleep apnea patients (Guilleminault et al., 2005; Guillaumet and Ramar, 2009; Levý et al., 2012), or Parkinson' disease (Mu et al., 2012).