Establishment of Facial Nerve Injury Rat Model for Idiopathic Facial Paralysis Research

The present protocol establishes a facial nerve injury rat model using microscopy to investigate the diagnostic and therapeutic mechanisms of idiopathic facial paralysis.

Idiopathic facial paralysis is the most common type of facial nerve injury, accounting for approximately 70% of peripheral facial paralysis cases. This disease can not only lead to a change in facial expression but also greatly impact the psychology of patients. In severe cases, it can affect the normal work and life of patients. Therefore, the research on facial nerve injury repair has important clinical significance. In order to study the mechanism of this disease, it is necessary to carry out relevant animal experiments, among which the most important task is to establish an animal model with the same pathogenesis as human disease. The compression of the facial nerve within the petrous bone, especially the nerve trunk at the junction of the distal end of the internal auditory canal and the labyrinthine segment, is the pathogenesis of idiopathic facial paralysis. In order to simulate this common disease, a compression injury model of the main extracranial segment of the facial nerve was established in this study. The neurological damage was evaluated by behavioral, neuroelectrophysiological, and histological examination. Finally, 50 g constant force and 90 s clamp injury were selected as the injury parameters to construct a stable idiopathic facial paralysis model.

As a type of peripheral facial paralysis, idiopathic facial paralysis is characteristic of unknown etiology, acute onset, and self-limiting course^1,2. The etiology and pathogenesis of idiopathic facial paralysis are still uncertain³. At present, there are various treatment methods for facial paralysis⁴, and the diversity of treatments reflects the lack of optimal treatment options. Using cellular and molecular biology techniques to study the mechanism of facial nerve injury is the foundation for establishing effective treatment methods for facial paralysis. Therefore, a suitable and stable facial nerve injury model is particularly important.

At present, there is no standard method for establishing a facial nerve injury model. The current preparation methods include virus inoculation⁵, transection⁶, cold stimulation⁷, and compression⁸ methods. It is believed that viral infection, neurotrophoblastic vasospasm, autoimmune inflammation, etc., may all cause local ischemia, degeneration, and edema of the facial nerve^9,10,11. Moreover, all of the above factors can cause compression of the main trunk of the facial nerve in the narrow bony facial nerve canal^12,13. In addition, the most common peripheral nerve injuries identified during surgical procedures were compression and contusion¹⁴. Based on the above theories and clinical phenomena, we believe that preparing the facial nerve injury model through compression injury is more reasonable. However, most of the current methods for implementing compression injuries do not provide quantitative parameters of force and time. In this study, we quantified the force and duration of compression injury to improve the reproducibility of the established model.

All the animal experiments were approved and supervised by the Animal Ethics Committee of Xinhua Hospital Affiliated to Shanghai Jiaotong University School of Medicine (XHEC-F-2023-061). Sprague-Dawley male Rats, 200-300 g, were used for the present study. The animals were obtained from a commercial source (see Table of Materials). The rats were randomly divided into four groups (n = 10): Sham surgery group, 30-s injury group, 60-s injury group, and 90-s injury group.

1. Induction of anesthesia and animal preparation

1. Wear the following personal protective equipment (PPE): surgical mask, surgical gloves, disposable gown.
2. Weight the rats and anesthetize them with ketamine hydrochloride at a dose of 50 mg/kg by intraperitoneal (i.p.) injection. Administer meloxicam (5 mg/kg; i.p.) for perioperative analgesia. Confirm the depth of anesthesia using a toe pinch.
3. Apply ophthalmic ointment on both eyes to prevent drying.
4. After anesthesia, place the rats in the prone position and fix the head so that the left side of the face is up. Shave off the hair behind the left ear and disinfect the skin. Cover the rat with the sterile surgical drape.

2. Establishing a local crush injury model of the extracranial trunk of the facial nerve

NOTE: Sterilize all equipment before use. All operations were performed in the operating room.

1. Make a 2 cm long longitudinal incision behind the left ear, and dissect the skin and subcutaneous tissue to separate the natural gap between facial and cervical muscles.
2. Use micro tweezers and micro scissors to fully dissociate the facial nerve trunk between the stylomastoid foramen and the parotid gland, with an exposed length of approximately 1 cm.
3. Use peripheral nerve quantitative injury forceps (Table of Materials) to clamp the facial nerve trunk to cause injury. Locate the injury site 0.5 cm away from the stylomastoid foramen. Apply an injury intensity of 50 g and injury time of 30 s, 60 s, and 90 s, respectively.
4. Suture the subcutaneous and skin with silk thread. Disinfect the incision.
5. For rats in the sham-surgery control group, cut the skin and subcutaneous tissue after anesthesia and expose and separate the corresponding main trunk of the facial nerve. Next, sutured the incision immediately.
6. Monitor the animal's health, maintain sternal recumbency, and keep it in warm conditions.
7. Move the rat back to the housing cage after the rat is conscious.

3. Behavioral testing

NOTE: The facial nerve function of the rats was evaluated before surgery and 48 h after surgery (Figure 1). The scores of blink reflex, palp movement, and nasal tip position were calculated¹⁵. The higher the total score, the more severe the degree of facial nerve injury (Table 1).

1. Blink reflex (BR):
   1. Attach an 18 G needle to a 2 mL syringe and blow air into the rat's eye from a distance of 2 cm. Observe the eyelid movement and closure.
   2. Score per the following criteria: No significant difference on both sides: 0 points; Delayed closure of the affected side compared to the healthy side: 1 point; Inability to close the affected eyelid: 2 points.
2. Vibrissae movement (VM):
   1. Count the bilateral tentacle movements of rats within 30 s.
   2. Score per the following criteria: No significant difference in bilateral tentacle movement: 0 points; The movement of the affected side's whiskers is weaker than that of the healthy side: 1 point; Loss of whisker movement on the affected side: 2 points.
3. Nasal tip position.
   1. Center nose tip: 0 points; Nose tip leaning towards the healthy side: 1 point.
      ​NOTE: A total score of 0 points indicates normal, 1-2 points indicate mild facial paralysis (paresis), 3-4 points indicate obvious facial paralysis (paresis), and 5 points indicate complete facial paralysis¹⁵.

4. Neuroelectrophysiological detection

NOTE: Facial electrography (ENoG) was performed before the injury, immediately after surgery, and 48 h after surgery (Figure 2, Table 2, Table 3, and Table 4).

1. Place the grounding electrode under the skin of the left lower limb.
2. Insert the recording electrode (bipolar concentric needle electrode) into the injured side of the tentacle muscle, with a penetration depth of 5 mm.
3. Place the stimulation electrode (concentric needle electrode) on the facial nerve membrane. Stimulate the proximal and distal ends of the injured facial nerve separately.
4. Use a square wave pulse current with a frequency of 1 Hz, a wave width of 0.1 ms, and a filtering range of 10-3000 Hz.
5. Use stimuli of 2 mA, 5 mA, 10 mA, 15 mA, and 20 mA to induce the generation of compound muscle action potential.
6. Record the latency (Lm) and peak amplitude (Am) of the M-wave.
   NOTE: The M-wave refers to the first and most obvious waveform recorded. The point where the waveform leaves the baseline is the starting point of the waveform. The distance from the starting point of the baseline to the starting point of the waveform is Lm. The distance between the highest and lowest peaks of the waveform is Am.
7. Allow a 5-min interval between each stimulation to ensure nerve recovery.

5. Histological examination

1. After completing the electrophysiological tests, use micro tweezers to lift the nerve and micro scissors to cut the nerve specimen. The specimen includes the facial nerve trunk from the injury point to the parotid gland, which is the nerve fiber at the distal end of the injury point, with a total length of about 0.5 cm.
2. Fix the nerve specimen in 4% paraformaldehyde for 24 h and prepare paraffin-embedded sections¹⁶.
3. Stain the sections with the hematoxylin-eosin staining (H&E) method¹⁶ and acquire images at 100x and 400x magnification using an optical photographic microscope (Figure 3).
   NOTE: After removing the nerve specimen and suturing the skin under anesthesia, the rats were euthanized by pentobarbital sodium (150 mg/kg; i.p.).

Behavioral testing
Before surgery, the scores of blink reflex, palp movement, and nasal tip position were 0 points in all experimental rats, indicating that all rats had intact facial nerve function. In the facial nerve function assessment 48 h after surgery, it was found that the individual scores of rats in each injury group were increased. Moreover, the total score increased gradually with the prolongation of facial nerve injury time (Table 1).

The facial activity records of rats in each group 48 h after surgery showed that the facial symmetry gradually disappeared with the extension of injury time. In the 30 s injury group, the faces of rats were basically symmetrical. In the 60 s injury group and 90 s injury group, the nasal tip of the rats shifted towards the healthy side (right side). This phenomenon was more pronounced in the 90-s injury group than in other groups (Figure 1).

Neuroelectrophysiological detection
Before surgery, electrophysiological tests found that stimulating the mastoid foramen of the facial nerve could stably generate M-waves (Figure 2A,D,G). Immediately after surgery, when stimulating the proximal end of the injury site, M waves were recorded in the 30-s injury group, but both Lm and Am were prolonged and decreased. In the 60-s and 90-s injury groups, no M waves were recorded. When stimulating the distal part of the injury site, M-waves could be recorded stably in all three injury groups, and the waveform, Lm, and Am were consistent with those before surgery (Figure 2B,E,H). At 48 h after surgery, there were significant differences between the injury groups (Figure 2C,F,I). In the 30-s injury group, M waves were recorded in both proximal and distal stimulation. Moreover, there was no statistically significant difference in Lm before and after surgery (see Table 2). In the 60-s injury group, M waves were not recorded under proximal stimulation. Under distal stimulation, 70% of rats did not record M waves. A significantly weakened compound muscle action potential (CMAP) was recorded in 30% of rats. The Lm and Am were significantly prolonged and decreased, respectively (see Table 3). For the 90-s injury group, neither proximal nor distal stimulation showed CMAP (see Table 4).

Pathological examination
H&E staining showed that there was no damage to the facial nerve specimen in the sham surgery group, and the morphology of the facial nerve was complete. In the 30-s injury group, there was no significant axonal collapse or demyelination of the distal nerve at the damaged site, and the morphology of axons and myelin sheaths was similar to that of the control group. The deeply stained nuclei were still evenly distributed, and there was no significant increase in number. In both the 60-s and the 90-s injury groups, axonal collapse and demyelination changes were observed in the distal nerves at the damaged site, with visible nerve fiber swelling and normal interruption of axons and myelin sheaths. Under a high-power microscope (400x), it was observed that the number of deeply stained nuclei in the specimen was significantly increased, and a large number of round nuclei appeared, which were considered infiltrating macrophages. The original flattened Schwann cell nucleus also became significantly enlarged. The above pathological changes were more pronounced in the 90-s injury group than in the 60-s injury group (Figure 3).

Figure 1: Representative photographs taken 48 h after surgery showing the facial features of different groups of rats. (A-D) Anterior: displays the front part. (E-H) Inferior: displays the lower part. Healthy control: Sham surgery group. 30-s: 30-s injury group. 60-s: 60-s injury group. 90 -s: 90-s injury group. Please click here to view a larger version of this figure.

Figure 2: Representative waveform of CMAP induced both proximally and distantly to the nerve injury point in each group at different times. (A,D,G) CMAP generated by stimulation at the stylomastoid foramen before injury.(B,E,H) CMAP generated by distal and proximal stimuli immediately after injury. (C,F,I) CMAP generated by distal and proximal stimuli 48 h after injury. In the B-C, E-F, and H-I, the upper two rows of waveforms were generated by distal stimulation, while the lower three rows of waveforms were generated by proximal stimulation. 30-s: 30-s injury group. 60-s: 60-s injury group. 90-s: 90-s injury group. Please click here to view a larger version of this figure.

Figure 3: Representative H&E staining images of resected facial nerves in each group at 48 h after injury. Please click here to view a larger version of this figure.

Table 1: Rat facial nerve function score at 48 h after surgery (n=10).

Table 2: Latency (Lm) and Amplitude (Am) of M-waves at different time points in the 30-s injury group (n = 10). * indicates a statistical difference between groups (P < 0.05), ** indicates a significant difference between groups (P < 0.01). All results were compared with preoperative.

Table 3: Latency (Lm) and Amplitude (Am) of M-waves at different time points in the 60-s injury group (n = 10). Δ indicates that only 30% of rats (n = 3) exhibit M-waves. / indicates that no waveform has appeared. * indicates statistical difference between groups (P < 0.05), ** indicates significant difference between groups (P < 0.01). All results were compared with preoperative.

Table 4: Latency (Lm) and Amplitude (Am) of M-waves at different time points in the 90-s injury group (n = 10). / indicates that no waveform has appeared.

It is necessary to study the repair mechanism of facial nerve injury in patients with idiopathic facial paralysis¹⁷. The degree of facial nerve injury model should meet the following requirements. Firstly, the degree of facial nerve injury should not be too mild, such as Sunderland Grade 1^st degree¹⁸, which can completely self-repair without drug intervention. Secondly, it should not be too severe, such as Sunderland 5th degree, which requires surgical intervention. Thirdly, the degree of facial nerve injury should be stable and uniform. Therefore, it is important to integrate behavioral, electrophysiological, and histological tests as the basis for model selection.

The factors affecting the strength of crush injury are the magnitude of the force and the duration of the action. In this study, we used self-made peripheral nerve quantitative injury forceps, with the force set as quantitative (50 g) and the action time as variables (clamping for 30 s, 60 s, and 90 s, respectively). We scored and assessed the degree of facial paralysis by observing the blink reflex, vibrissae movement, and nasal tip position of the rats. This is similar to the House Brackmann (H-B) scale commonly used in clinical assessment¹⁹. We found that the average total score of facial nerve function evaluation was 1.2 ± 0.92 points in the 30-s injury group, indicating mild facial nerve paralysis. When the injury time was 60 s, the average total score was 3.2 ± 1.23 points, and the rats showed obvious facial paralysis symptoms, such as the disappearance of the blink reflex, decreased tentacle activity, and changes in nasal tip position. In the 90-s injury group, all rats showed complete facial paralysis of 5 points.

The neuro electrophysiological detection showed that stimulation of the proximal end immediately after injury indicated weakened conduction in the 30-s injury group, while stimulation of the proximal and distal ends after 48 h indicated partial recovery of conduction function. It meant that there was partial axonal destruction in the damaged facial nerve at the moment of injury. However, this damage was relatively mild, and the neuroconductive function could partially self-repair. Therefore, the 30-s injury group model was not conducive to observing the effect of therapeutic intervention on facial nerve repair. For the 90-s injury group, no waveform was observed at the proximal end of the stimulation immediately after the injury, indicating that all axons in the local area of the injury had lost conduction function. However, no waveform was observed at the distal end of the stimulation 48 h later, indicating that all nerves at the distal end of the stimulation point had undergone Wallerian's degeneration within 48 h, and nerve conduction function had been completely lost. At this time, the waveform could not naturally appear at the proximal end of the stimulation. Therefore, this model was suitable for idiopathic facial paralysis research. For the 60-s injury group, we found that the detected electrophysiological effects were not stable and were not suitable for subsequent research.

The tissue morphology examination found that there was almost no significant change in the 30 s injury group 48 h after the injury, indicating that the degree of damage was mild and had been basically repaired, in line with Sunderland Grade 1st degree. Both the 60-s injury group and the 90-s injury group exhibited the typical early pathological changes of Wallerian degeneration²⁰. Among them, the 90-s injury group showed more significant changes, such as axonal disintegration, demyelination, Schwann cell proliferation, and macrophage infiltration, indicating a more severe degree of damage in this group. All injury groups did not damage the nerve outer membrane and fasciculus; therefore, there was no Sunderland fourth to fifth-degree injury.

Although this animal model could accurately reflect the characteristics of facial nerve paralysis, it could not fully approximate the actual incidence of facial paralysis in humans. Compared to other preparation methods, the force and time of facial nerve compression were quantified in this study, and the model preparation was repeatable. This animal model will be helpful in the diagnosis and treatment of idiopathic facial palsy.

The authors declared that no competing conflicts of interest exist.

This work was supported by project grants from the National Natural Science Foundation of China (82203637) and the Science and Technology Development Foundation of Nanjing Medical University (NMUB20210220).