Newly identified axon types of the facial nerve unveil supplemental neural pathways in the innervation of the face

Abstract

Introduction: Neuromuscular control of the facial expressions is provided exclusively via the facial nerve. Facial muscles are amongst the most finely tuned effectors in the human motor system, which coordinate facial expressions. In lower vertebrates, the extracranial facial nerve is a mixed nerve, while in mammals it is believed to be a pure motor nerve. However, this established notion does not agree with several clinical signs in health and disease.

Objectives: To elucidate the facial nerve contribution to the facial muscles by investigating axonal composition of the human facial nerve. To reveal new innervation pathways of other axon types of the motor facial nerve.

Methods: Different axon types were distinguished using specific molecular markers (NF, ChAT, CGRP and TH). To elucidate the functional role of axon types of the facial nerve, we used selective elimination of other neuronal support from the trigeminal nerve. We used retrograde neuronal tracing, three-dimensional imaging of the facial muscles, and high-fidelity neurophysiological tests in animal model.

Results: The human facial nerve revealed a mixed population of only 85% motor axons. Rodent samples revealed a fiber composition of motor, afferents and, surprisingly, sympathetic axons. We confirmed the axon types by tracing the originating neurons in the CNS. The sympathetic fibers of the facial nerve terminated in facial muscles suggesting autonomic innervation. The afferent fibers originated in the facial skin, confirming the afferent signal conduction via the facial nerve.

Conclusion: These findings reveal new innervation pathways via the facial nerve, support the sympathetic etiology of hemifacial spasm and elucidate clinical phenomena in facial nerve regeneration.

Keywords: Axon quantification; Facial muscles; Facial nerve; Facial palsy; Hemifacial spasm; Sympathetic fibers; proprioception; sensory feedback.

Graphical abstract

Fig. 1

Mixed axonalpopulation within the human extracranial facial nerve. A. Schematic illustration of the harvesting site for the extracranial facial nerve. Cross-sections of the extracranial facial nerve are stained using antibodies against neurofilament (red) and antibodies against choline acetyltransferase (green). Axons displaying a double signal (anti-NF and anti-ChAT) demonstrate the predominant motor axon population. Axons with only an anti-NF signal are non-cholinergic. The overall axon number within the main trunks was 12,343 ± 1231, whereby 1622 ± 453 (15%) are noncholinergic. B. The magnified image (x60 magnification) of the cross-section of the facial nerve demonstrates clustered non-cholinergic fibers (dashed line), which are seen resembled in every single fascicle of the extracranial facial nerve. The image represents an overlapping green (anti-ChAT) and red (anti-NF) signals.

Fig. 2

Axon quantification of the facial nerve and its distal branches in the rat. (A, B) Absolute numbers of the cholinergic and noncholinergic axons within the facial nerve and its respective branches (n = 12). (C) Quantification of axons within the facial nerve and its distal branches from twelve animals. The axons within cross-sections of the nerves stained with antibodies against NF/ChAT were quantified, whereby noncholinergic axons were labeled only with anti-neurofilament and cholinergic axons showed double signal with anti-NF and anti-ChAT. The number of noncholinergic axons was almost equally distributed along all distal branches: 4.6 ± 1.2% in the temporal branch, 5.2 ± 1.6% in the zygomatic branch, 5.9 ± 1.8% in the buccal branch, 6.9 ± 2% in the marginal mandibular branch, 12.9 ± 5.6% in the cervical branch. Overall, the main trunk of the facial nerve consisted of 6.1 ± 0.32% noncholinergic fibers. The auricular posterior nerve, known for its partial afferent innervation of the auricula, contained 28.2±% of noncholinergic fibers, which were present in a separate nerve fascicle. Data is presented as means ± standard deviation.

Fig. 3

Analysis of different axon types in the facial nerve. (A), (B) Quantitative analysis of the different axonal components in the facial nerve in the rat. ChAT + for somatic efferent axons, CGRP + for afferent axons and TH + for sympathetic axons. Data are presented as means ± standard deviation. Cross-sections of the main trunk of the facial nerve stained with anti-NF, anti-ChAT (C), anti-NF and anti-TH (D), anti-NF and anti-CGRP antibodies (E). Magnified images of the cross-sections using anti-NF, anti-ChAT and bright-field light (F), anti-NF, anti-TH and bright-field light (G), anti-NF and -CGRP antibodies (H) and bright-field light show that ChAT fibers form the majority within the facial nerve.

Fig. 4

New paradigm for the central representation of the facial nerve. (A) Schematic interpretation of the central origins for different fiber types in the facial nerve. Three different fiber types from the facial nerve were traced back to their neural sources. The facial nucleus is the origin of the somatic efferent axons innervating the facial muscles. The geniculate ganglion is the origin of the afferent fibers from the facial nerve. The superior cervical ganglion is the origin of the sympathetic fibers. (B) The overall number of the sympathetic fibers and labeled cell bodies in the superior cervical ganglion did not statistically differ: 124 ± 49 vs. 120 ± 30 respectively (paired t-test, p = 0.18, n = 5). (C) Number of anti-ChAT positive axons and the labeled motoneurons in the facial nucleus did not statistically differ: 4803 ± 100 somatic efferent axons vs. 4837 ± 218 motoneurons (paired t-test, p = 0.48, n = 14). (D) The number of anti-CGRP positive fibers within the facial nerve corresponded with the number of cell bodies labeled within the geniculate ganglion but showed a difference with statistical significance: 230 ± 81 vs. 194 ± 48 (paired t-test, p = 0.035, n = 5).

Fig. 5

Innervation pattern of the facial muscles via the facial nerve. (A) Illustration of the harvested muscles. Left: levator auris longus (LAL); middle: dilator nasi (MDN); right: levator labii superioris (LLS) muscles. The facial muscles were harvested in nine rats for the whole-mount staining after deafferentation procedure. (B) Immunofluorescent images of the facial muscles using anti-ChAT, anti-NF and alpha-bungarotoxin. Anti-ChAT and alpha-bungarotoxin signals demonstrate the cholinergic nature of the neuromuscular innervation. (C) Immunofluorescent images using the antibodies anti-NF, anti-CGRP and alpha-bungarotoxin in LAL, LLS and MDN muscles harvested after deafferentation. All muscles show CGRP-positive signals within the axonal bundles (arrows). (D) Immunofluorescent images of LAL, LLS and MDN muscles after deafferentation using anti-NF, anti-TH antibodies and alpha-bungarotoxin. Whole-mounts indicated the presence of the TH-positive fibers within all three facial muscles (arrowheads).

Fig. 6

Innervation of the skin and genal vibrissae via the facial nerve. (A) Schematic illustration of the harvesting region of the facial skin containing genal vibrissae. (B, C) Immunofluorescent images of the skin harvested on the unoperated side using anti-NF, anti-CGRP and phalloidin at 10x (B) and 20x (C) magnification. The images show CGRP-positive axonl bundles within the skin indicating their afferent nature. (D, E) Immunofluorescent images of the facial skin harvested from the denervated side. (D) Denervated skin show a dense innervation pattern of CGRP-positive axons after the deafferentation procedure (arrows). (E) Immunofluorescent staining using anti-NF, anti-TH and phalloidin antibodies show only a few TH-positive axons witihin the skin vessels, suggesting sympathetic innervation of the vessels (dashed line).

Fig. 7

Facial nerve conducts afferent signals from the genal vibrissae (A) Illustration of the experimental setting for electrophysiological measurements. The buccal branch of the facial nerve was exposed and placed onto the hook-electrode. The genal vibrissae and the surrounding skin was stimulated by slight touch of vibration. The measurements from the unoperated (B) and operated (C) sides after applying stimuli to the ipsilateral genal vibrissae in the following order: touch vibrissae, touch skin, vibrate vibrissae, vibrate skin. (D) Illustration of applying a stretch stimulus to the dilator nasi muscle: after the complete isolation of the muscle from the surrounding tissue, the tendinous muscle part of the dilator nasi muscle was looped using a vessel loop. Sharp and short stretches were applied to the vessel loop to induce a stretch response. (E) No afferent signals were obtained from the buccal branch either after applying the stretch or nociceptive stimuli. The average amplitudes measured in the unoperated rats and rats after deafferentation did not statistically differ (F): 0.86 ± 0.66 µV and 0.76 ± 0.27 µV respectively for touching the vibrissae (p = 0.61; Mann-Whitney test), 1.96 ± 0.96 µV and 1.49 ± 0.92 µV for touching the emergence point of the vibrissae (p = 0.62; unpaired T-test), 1.69 ± 0.88 µV and 1.24 ± 0.44 µV for vibrating the genal vibrissae (p = 0.25; unpaired T-test), 3.87 ± 2.51 µV and 2.29 ± 1.62 µV for vibrating the emerging point of the genal vibrissae (p = 0.35; Mann-Whitney test).