Intralaryngeal neuroanatomy of the recurrent laryngeal nerve of the rabbit

Abstract

We undertook this study to determine the detailed neuroanatomy of the terminal branches of the recurrent laryngeal nerve (RLN) in the rabbit to facilitate future neurophysiological recordings from identified branches of this nerve. The whole larynx was isolated post mortem in 17 adult New Zealand White rabbits and prepared using a modified Sihler's technique, which stains axons and renders other tissues transparent so that nerve branches can be seen in whole mount preparations. Of the 34 hemi-laryngeal preparations processed, 28 stained well and these were dissected and used to characterize the neuroanatomy of the RLN. In most cases (23/28) the posterior cricoarytenoid muscle (PCA) was supplied by a single branch arising from the RLN, though in five PCA specimens there were two or three separate branches to the PCA. The interarytenoid muscle (IA) was supplied by two parallel filaments arising from the main trunk of the RLN rostral to the branch(es) to the PCA. The lateral cricoarytenoid muscle (LCA) commonly received innervation from two fine twigs branching from the RLN main trunk and travelling laterally towards the LCA. The remaining fibres of the RLN innervated the thyroarytenoid muscle (TA) and comprised two distinct branches, one supplying the pars vocalis and the other branching extensively to supply the remainder of the TA. No communicating anastomosis between the RLN and superior laryngeal nerve within the larynx was found. Our results suggest it is feasible to make electrophysiological recordings from identified terminal branches of the RLN supplying laryngeal adductor muscles separate from the branch or branches to the PCA. However, the very small size of the motor nerves to the IA and LCA suggests that it would be very difficult to record selectively from the nerve supply to individual laryngeal adductor muscles.

Introduction

The recurrent laryngeal nerve (RLN) provides the motor innervation to all intrinsic laryngeal muscles except the cricothyroid and is responsible for adjustments in glottic aperture during the respiratory cycle. The pattern of respiratory-related discharge observed from whole RLN recordings is complex, exhibiting activity either exclusively in inspiration or during both respiratory phases (Green & Nail, 1955; Sica et al. 1984, 1985). As a result interpretation of whole RLN motor activity in relation to control and function of laryngeal abductor and adductor muscles is difficult. Recordings obtained from motor units in the RLN show that the RLN contains functionally diverse groups of motoneurones (Sica et al. 1985; Ryan et al. 2002). It is assumed that these motoneurones are destined for different muscles groups on the basis of discharge pattern and the known electromyographic activity profiles and function of laryngeal abductor and adductor muscles. RLN motor units with post-inspiratory discharge are presumed to innervate laryngeal adductor muscles (Bartlett et al. 1973; Ryan et al. 2002) while units with inspiratory-related activity are assumed to supply the posterior cricoarytenoid muscle (PCA), the principal abductor of the vocal cord. These assumptions may not be entirely valid because (i) the action and functions of individual laryngeal muscles are incompletely understood, (ii) laryngeal adductor activity is not confined to expiration (Insalaco et al. 1990) and (iii) the PCA exhibits tonic expiratory activity on which phasic inspiratory activity is superimposed (Sherrey & Megirian, 1980).

Care must be taken when interpreting recordings obtained from motor units in the main trunk of the RLN, as we cannot be certain of their final destination within the larynx. Zhou et al. (1989) used a novel approach to studying the motor control of laryngeal muscles by recording separately from the intralaryngeal branches supplying the abductor and adductor muscles. We intend to perform similar experiments in our decerebrate rabbit preparation (Ryan et al. 2002). However, in a complex structure such as the larynx, gross dissection is not always a reliable means of tracing the neural pathway from extramuscular to intramuscular terminal branches, as the latter are very small and hard to distinguish from blood vessels and connective tissue, even with the aid of a dissecting microscope. Furthermore, when tracing nerve branches deeper into the muscles they innervate, preserving one branch often necessitates unwanted destruction of other motor and/or sensory nerves to the same structure.

The primary aim of this study is to characterize accurately the anatomical distribution of intralaryngeal branches of the RLN so that for future studies of electrophysiological terminal branches of this nerve we could be confident of the final destination of motor axons while recording their activity. In addition, we sought to determine whether the branching pattern of intramuscular nerves was consistent with compartmentalization of laryngeal muscles in the rabbit, as recent evidence suggests that in both dog and human, individual laryngeal muscles (Drake et al. 1993; Sanders et al. 1993a, 1994a, b; Bryant et al. 1996) and in particular the PCA, are organized into anatomically and functionally distinct neuromuscular compartments. Wu & Sanders (1992) have refined a relatively old histological technique, Sihler's stain, to investigate intramuscular branching of motor nerves without the problems associated with the use of microdissection or serial histological sections. Modified Sihler's stain renders the whole specimen relatively translucent while counterstaining its nerve supply (Wu & Sanders, 1992) and has the advantage of preserving the three-dimensional structure of the whole specimen.

Methods

Seventeen adult New Zealand White rabbits weighing 2.5–4.5 kg were killed using a lethal overdose of sodium pentobarbitone (200 mg kg⁻¹; Sagatal, Rhone Merieux, Ireland) administered intravenously through a marginal ear vein. A ventral midline neck incision was made to expose the trachea and larynx. The large veins of the neck were divided and heparanized saline (500 mL) was infused at 90 mmHg pressure through cannulae inserted into both common carotid arteries. The whole larynx, including a portion of the trachea approximately 5 mm long, was isolated post mortem and prepared using a modified Sihler's staining technique in seven stages. Reagents were obtained from Sigma Aldrich, UK, unless otherwise stated.

1. Fixation.

The whole larynx was immersed immediately after removal in 10% unneutralized formalin for 2 weeks.

2. Maceration and de-pigmentation.

The fixed specimens were removed from the formalin and washed in distilled water for 30 min before being placed in a solution containing 3% aqueous potassium hydroxide (KOH) with three drops of 3% hydrogen peroxide (H₂O₂) added to every 100 mL of solution. They were allowed to incubate in this solution at room temperature. The solution was changed every second day initially, and at least once a week thereafter, or whenever the solution became cloudy. Maceration continued for 3 weeks until the tissues became translucent.

3. Decalcification.

After washing in distilled water for 30 min, the laryngeal specimens were transferred into Sihler's I solution for 2 weeks to decalcify the specimens. Sihler's I solution was prepared as follows: one part glacial acetic acid, one part glycerin in six parts 1% aqueous chloral hydrate. The specimens were transferred to fresh solution every 2–3 days.

4. Staining.

Following decalcification, each larynx was washed in distilled water for 30 min and incubated in Sihler's II solution. Sihler's II solution was prepared using one part Ehrlich's Haematoxylin (Lennox Laboratory Supplies, Ireland), one part glycerin, and six parts 1% aqueous chloral hydrate. The solution was changed once per week. Each specimen was regularly examined under a dissecting microscope. The staining procedure was carried out for at least 3 weeks or until the large nerves within the specimens turned dark purple and the terminal nerve branches were well stained.

5. Destaining.

Each larynx was washed in distilled water (30 min) then immersed in Sihler's I solution to remove excess stain. We found that 30 min exposure time is adequate, with longer periods resulting in loss of fine detail due to excessive loss of stain from fine terminal nerve branches.

6. Clearing.

Before clearing, specimens were washed in distilled water (30 min). Clearing involved immersing the tissues in increasing concentrations of glycerin (40%, 60%, 80% and 100%) in dark conditions. The specimen remained in 40% and 60% glycerin for 2 days, and one day each in 80% and 100% glycerin.

7. Trimming.

Each specimen was divided along the midline and the 34 hemi-laryngeal preparations were examined under a dissecting microscope (Wild M50), trans-illuminated with a variable fibre-optic light source, and carefully dissected to examine the intralaryngeal course of the RLN and its terminal branches. All laryngeal muscles were dissected from their cartilaginous attachments and examined as whole-mount preparations. Drawings and photomicrographs were taken during the dissecting process.

Results

Thirty-four hemi-laryngeal preparations were examined and, of these, 28 were adequately stained. The RLN approached the larynx as a single trunk and entered it through a groove between the oesophagus and trachea on the posterior tracheal surface. The intralaryngeal neuroanatomy of the RLN revealed the following innervation pattern for individual laryngeal muscles.

Pca

As the RLN entered the larynx, passing over the cricothyroid joint, the nerve supply of the PCA branched medially from the main trunk (Fig. 1). In 23 of 28 PCA specimens, one single branch arose from the main trunk of the RLN to innervate the PCA. Upon entering the muscle this branch subdivided into two or three smaller intramuscular nerve branches (Fig. 2). Because of the extensive arborization of these intramuscular nerve branches, no clear evidence was found to indicate any discrete compartmentalization of the intramuscular nerve supply. In only five of the PCA specimens examined (four of which were located on the right-hand side of the larynx) two or three distinct branches arose from the RLN main trunk to supply the PCA (not shown).

Fig. 1.

Innervation of the intralaryngeal muscles by the recurrent laryngeal nerve (RLN). This photograph shows a view of the branching anatomy of the main trunk of the RLN (A) to supply the posterior cricoarytenoid muscle (B), the interarytenoid muscle (C) and the lateral cricoarytenoid (D). The muscle nerve branches supplying the thyroarytenoid proper (E) and pars vocalis (F) neuromuscular compartments of the thyroarytenoid (E) are also shown. The RLN, its terminal branches and the muscles they supply have been dissected from their respective laryngeal insertions.

Fig. 2.

Posterior view of the left posterior cricoarytenoid muscle (PCA). The PCA (A) has been removed and squashed to demonstrate the intramuscular branches of the PCA. In this photograph only one branch from the recurrent laryngeal nerve (RLN, E) supplying the PCA was identified. The PCA nerve bifurcates upon entering the muscle, forming three distinguishable intramuscular branches. A communicating twig (B) between the PCA (A) and the branch to the interarytenoid muscle (C), and the remainder of the RLN main trunk destined for laryngeal adductor muscles (D) are also shown.

Interarytenoid muscle (IA)

The nerve supply to the IA arose from the main trunk of the RLN rostral to the branch(es) supplying the PCA (Fig. 1). In three specimens the nerve supply to the IA originated from the medial branch supplying the thyroarytenoid muscle (TA). The IA branch ran diagonally, upwards and medially, along the anterior border of the PCA and next to the cricoarytenoid joint before entering the inferior lateral border of the IA (Figs 1 and 2). In 25 of 28 cases the IA branch ran as a single trunk before dividing into two fine filaments prior to its entry into the IA (Figs 1 and 3). However, in three of 28 laryngeal preparations two small filaments arose from the RLN main trunk but quickly rejoined, continuing as a single branch and separating into two twigs before innervating the IA (not shown). These twigs further subdivided upon entering the muscle (4–5 intramuscular nerve branches), forming a complex branching anatomy of the intramuscular nerve branches (Fig. 3).

Fig. 3.

Posterior view of the right interarytenoid muscle (IA). A single nerve branch arising from the recurrent laryngeal nerve bifurcates (B) to form two small neural filaments close to its insertion into the IA (A). The intramuscular branching anatomy of these filaments forms a complex anastomatic network.

Lateral cricoarytenoid muscle (LCA)

After providing branches to supply the PCA and IA, the RLN continued cranially, either as a single trunk or two parallel divisions (medial and lateral), to the superior border of the cricoid, where it turned sharply anteromedially. As it approached the superior border of the cricoid, one or two very fine twigs branched from the RLN and travelled laterally to innervate LCA (Fig. 4). However, in two specimens one of the two twigs supplying LCA arose from the IA branch (Sanders et al. 1994a). The most consistent feature of the intramuscular neuroanatomy of the LCA innervation was division into 2–5 smaller terminal nerve branches, which exhibited extensive arborization throughout the muscle (Fig. 4).

Fig. 4.

Posterior view of the right lateral cricoarytenoid muscle (LCA). The LCA (A) receives two small filaments (B), one arising from each of the two medial and lateral divisions (C) of the recurrent laryngeal nerve destined for the thyroarytenoid muscle. The intramuscular branching anatomy demonstrates 2–5 further subdivisions of the filaments innervating the muscle.

TA

The remaining fibres of the RLN innervated the TA. Two separate neuromuscular compartments were identified, one supplying the pars vocalis, the other branching extensively to supply the remainder of the TA (Figs 5 and 6).

Fig. 5.

Anteromedial view of the recurrent laryngeal nerve (RLN) branches to the thyroarytenoid (TA), lateral cricoarytenoid (LCA) and interarytenoid (IA) laryngeal adductor muscles. The adductor branches of the RLN supplying the TA and LCA approach superior to the arytenoid cartilage (E). Two or three small twigs are seen branching from the RLN to supply the LCA (C). The RLN continues to supply two larger branches (A and B) to the TA muscle. Branch A divides to diffusely innervate the TA muscle proper whereas branch B medial to the arytenoid cartilage innervates the pars vocalis. Also shown is the nerve supply to the IA muscle (D).

Fig. 6.

Anteromedial view of the recurrent laryngeal nerve (RLN) branches to the thyroarytenoid muscle (TA). Before innervating the TA the adductor branch of the RLN bifurcates (C). One supplies the TA proper (A) dividing further into smaller intramuscular branches that extend rostrally through the TA towards the base of the epiglottis. A second branch (B) courses anteriorly and medially along the border of the arytenoid cartilage (removed) to innervate the pars vocalis.

Communicating branches between individual intralaryngeal nerves

Communicating twigs between individual intralaryngeal nerves often occurred, the most common of which was that between the primary branch supplying the PCA and IA (Fig. 2). Others included a communication between terminal nerve endings supplying the TA and IA (n = 2) and the IA and LCA (n = 2). A common communicating branch in one specimen between the PCA, IA and TA muscles (Fig. 1) was also found. Except in one preparation no evidence of any neural connection between the intramuscular neural plexus supplying the IA and the internal branch of the superior laryngeal nerve (iSLN Nordland, 1930; Mu et al. 1994) was found.

Discussion

Our results clearly demonstrate that the RLN provides separate and anatomically distinct branches to laryngeal abductor (PCA) and adductor muscles (TA, IA, LCA) and verify that in the rabbit at least, it is possible to record from branches of the RLN and be confident that one is separately recording motor neurones destined for abductor as opposed to adductor muscles. However, because of the very small size of the motor nerves to the IA and LCA it would be very difficult to record selectively from the nerve supply to individual adductor muscles.

The distribution of intramuscular branches of the PCA neither confirms nor rules out the possibility that this muscle is composed of two or three anatomically and functionally separate compartments. In the majority of specimens examined a single branch from the RLN main trunk supplied the PCA. The intramuscular distribution of this branch was such that it subdivided into two or three smaller intramuscular nerve branches whose terminal endings exhibited extensive arborization branching irregularly throughout the muscle. We could not distinguish separate compartments on the basis of this branching pattern. This contrasts with other studies in dog (Diamond et al. 1992; Drake et al. 1993; Sanders et al. 1993a) and human (Sanders et al. 1994a) PCA muscles, where the PCA was organized into separate bellies or compartments each receiving distinct innervation from primary branches of the RLN supplying this muscle (Diamond et al. 1992; Drake et al. 1993; Sanders et al. 1993a, 1994a). It is difficult to conclude with certainty whether a muscle is divided into neuromuscular compartments based only on anatomical studies of neural branching pattern. The division of the PCA into distinct compartments in other species is based on a greater body of evidence, including histochemical (Sanders et al. 1993a) and functional (Diamond et al. 1992; Sanders et al. 1993a) evidence.

A powerful yet under-utilized technique is to obtain selective electromyographic (EMG) recordings from different regions of the muscle. Woo & Bortoff (1987) described different EMG responses in the medial and lateral divisions of rat PCA during various anaesthetic and respiratory conditions. Sanders et al. (1994b), investigating the pattern of vocal fold motion evoked by direct electrical stimulation of different muscle bellies within the PCA, found that each belly is capable of moving the arytenoid in different ways. We have previously shown in the rabbit that there are two distinct types of RLN motor units with inspiratory discharge, phasic inspiratory and tonically active units whose activity is inhibited during inspiration (Ryan et al. 2002). Similar populations of inspiratory motor units have been demonstrated in the cat (Murakami & Kirchner, 1972; Sica et al. 1985). Given that the PCA is the principal abductor of the vocal cords and is the only laryngeal muscle innervated by the RLN that exhibits inspiratory activities under normal conditions (Suzuki & Kirchner, 1969), it is presumed that both these inspiratory RLN motor units innervate the PCA. A multi- or single unit EMG recording from the PCA would determine whether these different motor units are located in distinct anatomical regions or compartments of the PCA.

In relation to laryngeal adductor muscles, we clearly identified separate branches arising from the main trunk that travelled separately to innervate the TA, LCA and IA. The motor innervation of the IA in all cases arose from the ipsilateral RLN alone and ran along the rostral border of the PCA. Previous investigations on the innervation pattern of the IA have been performed on specimens from human subjects. One study (Mu et al. 1994) found that all human IA muscles studied receive a ‘bilateral’ nerve supply from both RLNs together with nerve twigs derived from the descending division of the iSLN. However, Nordland (1930) found that 18 of 19 human IA muscles were exclusively innervated by the iSLN. We documented only one rabbit hemi-laryngeal specimen where the IA received a communicating branch between the ipsilateral iSLN and RLN. In all remaining laryngeal preparations each IA received an ipsilateral nerve supply from the RLN only.

The LCA plays an important role in adducting the vocal cords in phonation and reflex glottic closure (Tanaka & Tanabe, 1986) but may also produce laryngeal abduction (Stroud & Zwiefach, 1956). While these differing functions might be reflected in separation of the LCA into functionally different neuromuscular compartments we and others (Sanders et al. 1993b) have found the LCA to comprised a single neuromuscular compartment with a homogeneous motor nerve plexus (Sanders et al. 1993b). We found the nerve branches supplying the LCA arose from the RLN usually as two separate fascicles or rarely as two twigs, one from the IA branch (Wu & Sanders, 1992; Sanders et al. 1993b) and one from the RLN main trunk.

The TA had two distinct neuromuscular compartments (Sanders et al. 1993c; Wu et al. 1994; Inagi et al. 1998), the pars vocalis and the TA proper. We did not identify any other prominent source of or communicating anastomoses with the neural supply to the TA contrary to observations made in other species (Dilworth, 1921; Kambic et al. 1984; Wu et al. 1994). For instance, the presence of communicating anastomoses between the RLN and SLN has been demonstrated several times in studies of the human larynx (Dilworth, 1921; Kambic et al. 1984; Wu et al. 1994; Sanudo et al. 1999). Wu et al. (1994) found that the communicating nerve is composed of fascicles that are organized into sensory and intramuscular motor nerve branches. The sensory component supplies the subglottic area. The intramuscular branch usually joined the RLN within the TA. These investigators propose that the human communicating nerve is ‘an extension of the external superior laryngeal nerve that innervates the vocal cord’ connecting the cricothyroid and TA muscles. We were unable to identify this communicating nerve in the rabbits. The only other mammal in which this nerve has been demonstrated was the dog (Dedo & Ogura, 1965). The function of this nerve may perhaps be related to vocalization or phonation, functions well ascribed to the TA. Both species in which the communicating nerve has been identified (human and dog) produce vocal sounds. Phonation is not, however, a well-developed function in rabbits, and this may explain why we failed to identify a communicating branch in our laryngeal specimens.

In conclusion, Sihler's stain is a powerful tool providing valuable information regarding the neuroanatomy of the larynx. The primary benefit arising from this study is that we now know the precise anatomical distribution of RLN branches to abductor and adductor muscles. This will be a valuable asset when performing future investigations to examine the respiratory role of different laryngeal motor neurone pools in relation to their known function of regulating airway patency and expiratory airflow. However, with regard to the organization of the PCA into discrete neuromuscular compartments, our results using neuroanatomical tracing techniques neither confirm nor refute their presence in the rabbit.