Oxaliplatin is a third-generation platinum compound that focuses on treating cancer patients [1, 2]. It has demonstrated effectiveness across a broad spectrum of human tumors, including colorectal, pancreatic, and other malignant tumors [1, 3] However, therapy involving oxaliplatin induces acute and chronic neurotoxicity in nearly all patients [4]. Adverse effects such as flushing, edema, breathing difficulties, gastrointestinal disorders, and sensory symptoms like mechanical and cold allodynia, numbness, itching and tingling in the hands and feet are commonly experienced [5, 6]. Among the side effects, the peripheral neuropathy symptoms are the most prominent reason that patients discontinue chemotherapy [7].

Efforts to mitigate this chemotherapy-induced peripheral neuropathy (CIPN) involve strategies such as reducing cumulative doses or using tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors (SNRIs), and gabapentinoids. However, these treatments may also lead to additional side effects and have not been fully proven effective for treating CIPN [8, 9]. While various medications and alternative methods have been researched for preventing and treating peripheral neuropathy, many are withdrawn or discontinued due to insufficient efficacies or emergent side effects [7]. There exists a need for safe yet effective interventions against CIPN in clinics.

Vagus nerve stimulation (VNS) is a non-pharmacological approach to solve the problem. The vagus nerve is the longest cranial nerve in the human body, serving as a vital bidirectional conduit between the body and the brain [10, 11]. Consisting of 20% efferent fibers and 80% afferent fibers, the vagus nerve transmits signals to various brain regions including the locus coeruleus (LC), a region contributing to pain modulation via norepinephrine (NE) [12-16]. The VNS has been used for alleviating various neurological symptoms such as epilepsy and treatment-resistant depression [13, 15, 17-19] Furthermore, researches have investigated the therapeutic effects of the VNS against inflammation [20] and Alzheimer's disease [21]. Studies also indicated that the VNS exerts pain-relieving effects through diverse pathways [22, 23]. In studies related to the CIPN, the chronic as well as short-term VNS therapy demonstrated pain-relieving effects in animal models of chronic pain [24, 25].

Although the analgesic effects of the VNS have been investigated as such, the method used in those studies were cervical VNS that involves surgical interventions. To avoid the risk of surgical complications, recent approaches have adopted non-invasive VNS techniques as alternatives, indirectly stimulating the vagus nerve using devices placed on the skin over the neck or ear [26]. However, the response of the pain processing system to these alternative approaches has not been elucidated, and analgesic effect have not been tested against CIPN. In this study, stimulation targeted the auricular branches of the vagus nerve, specifically focusing on the cymba concha area [27]. We measured the analgesic effects of the aVNS and explored the possible mechanisms. Regarding the mechanism of aVNS, various transmitters have been observed to play a role [13, 28]. This study aimed to investigate the potential mechanism of pain relief through the adrenergic pathway using short-term aVNS in an acute CIPN model.

Animal

Sprague-Dawley (SD) rats (7~8 weeks old, 180~210 g) were obtained from DBL (Chungcheongbuk-do, Korea) and housed in animal facility with access to food and water ad libitum. The rooms were maintained on a 12 hour light/dark cycle and kept at 23±2°C. All procedures involving animals were approved by the Institutional Animal Care and Use Committee of Kyung Hee University (KHUASP(SE)-23-516) and conducted following the ethical guidelines of the International Association for the Study of Pain [29]. At the end of the study, the animals were euthanized with CO₂ gas. In all the experiments, animals were randomly assigned to a group, and the experimenter was blinded to the group assignment.

Oxaliplatin administration

Oxaliplatin (Wako Pure Chemical Industries, Osaka, Japan; Tocris Bioscience, Bristol, England) was dissolved in a 5% glucose solution at a concentration of 2 mg/ml and administered via intraperitoneal injection at a dose of 6 mg/kg [30, 31] The control group received the same intraperitoneal injection of a 5% glucose solution as a vehicle.

Behavioral experiments

Rats underwent a week acclimation period in the animal facility before starting the experiments. They were placed on a plastic box (20×20×14 cm) over a mesh chamber for at least 30 minutes for adaptation.

For evaluating cold allodynia, acetone (20 μl) was applied to the right hind paw using a pipette tube after the acclimation period. Behavioral responses were observed for 40 seconds, scoring different reactions (0~3 points). Responses to acetone were scored as follows: 0, no response; 1, quick withdrawal, flick or stamp of the paw; 2, prolonged withdrawal or repeated flicking of the paw; 3, repeated flicking of the paw with licking directed at the ventral side of the paw [32]. Tests were repeated three times and mean score was calculated.

Mechanical allodynia was assessed by applying Von Frey filaments (bending forces to 0.4, 0.6, 1.0, 2.0, 4.0, 6.0, 8.0 and 15.0 g; equivalent in log units: 3.61, 3.84, 4.08, 4.31, 4.56, 4.74, 4.93 and 5.18; Stoelting, IL, USA) to the center of the right hind paw after the acclimation period. The 50% paw withdrawal threshold (PWT) was calculated using Dixon's up-down method and Chaplan's calculation, with a cut-off value of 15 g [33, 34]. Animals with a baseline PWT of less than 15g during initial screening were excluded from further studies.

Non-invasive aVNS treatment

Animals were anesthetized using 2% isoflurane with O₂ and N₂O. The cymba concha area, where the auricular branch of the vagus nerve is innervated [27], was stimulated using a USB Multifunction 2 channel Arbitrary waveform generator (Digistim-2, ALA Scientific Instruments, Inc., NY, USA). Custom-made magnetic electrodes, composed of round-shaped neodymium magnets with a diameter of approximately 2 mm, were attached to alligator clips and positioned to ensure stimulation solely on the cymba concha and the dorsum of the ear. The stimulation parameters included a fixed 5 voltage, square wave form, with different frequencies (2, 20, or 100 Hz). The stimulation was applied to left ear for 20 minutes in each group. The sham control group underwent the same procedure without electrical stimulation.

Antagonist treatment

To investigate the involvement of spinal noradrenergic receptor in the effect of aVNS, either of prazosin, idazoxan, or propranolol was administered to CIPN animals before the aVNS treatment. Prazosin (30 μg), an α₁ adrenergic antagonist, was dissolved in 20% dimethyl sulfoxide (DMSO). Idazoxan (50 μg), an α₂ adrenergic antagonist, and propranolol (50 μg), a non-selective β-adrenergic antagonist, were dissolved in normal saline. All reagents were administered by intrathecal injection based on the previous studies [31, 35].

In the experiments, rats treated with oxaliplatin were randomly divided into groups. After intrathecal injection under isoflurane anesthesia, the animals were returned to homecage and given 30 minutes waiting period. Subsequently, another round of anesthesia was applied, and aVNS stimulation was conducted for 20 minutes. The stimulation parameter used was the 20 Hz of 5 V square wave form, which exhibited the most effective analgesic effect. Behavioral experiments were conducted one hour after the aVNS treatment.

Immunohistochemistry (IHC)

After inducing anesthesia using isoflurane, perfusion was performed with phosphate-buffered saline (PBS) and 4% formalin, followed by brain removal from the skull. The brains were immersed in 4% paraformaldehyde for 4 hours, then stored at 4°C until they sank in 15% and 30% sucrose consecutively. The brains were frozen using OCT in isopropanol, and cryo-sectioning was performed at a thickness of 30 μm. The tissues were mounted on glass slides and left overnight. After washing three times with PBS for 30 minutes, the tissues were incubated with 0.3% BSA and 0.1% Triton X-100 in PBS and treated with each of primary antibody of c-Fos (ab190289, Abcam) and tyrosine hydroxylase (TH) (MAB318, Sigma Aldrich) diluted at 1:1000 ratio for 24 hours. Following four times washing with PBS for 30 minutes, the tissues were incubated with a secondary antibody Alexa Fluor 546 (A11010, Invitrogen) and Alexa Fluor 488 (A11001, Invitrogen) diluted each of 1:400 and 1:800 for 2 hours. The tissues were then washed four times with PBS for 30 minutes. Finally, Vectashield with DAPI (H-1200-10, Vector Laboratories) was applied for mounting and DAPI staining, and cover glasses were placed. The slides were stored at -20°C [36]. The processed sections were stored at -20°C prior to imaging. Fluorescent images were quantified using QuPath v0.4.4 after setting regions of interest (ROIs) by TH antibody of intensity and comparing them with the rat brain atlas researched by Paxinos et al [37-39].

Statistical analysis

Prism 10.1.0 (GraphPad Software, USA) was used for statistical analysis and graphical works. The experimental data were presented as mean±S.E.M. For all analyses, p<0.05 was considered significant. Statistical tests were performed using two-way analysis or one-way analysis of variance with post hoc Dunnentt's test, Sidak's test, or unpaired t-test.

Successful induction of pain symptoms in CIPN model animals

To induce CIPN, oxaliplatin (6 mg/kg) was intraperitoneally administered to experimental animals [31]. Fig. 1A shows the experimental schedule. The mechanical allodynia test was performed utilizing von Frey filament stimuli and calculating 50% PWT, while the cold allodynia test was performed by applying a 20 μl acetone drop to the paw and scoring the behavioral responses. The rats treated with oxaliplatin developed both mechanical and cold allodynia, as shown by a significant decrease in PWT (Fig. 1B) and an increase in cold score (Fig. 1C). The control group treated with a vehicle solution showed no significant changes. In the oxaliplatin-treated animals, pain behaviors gradually developed starting from day 1 and continued to increase over time until day 5. These results are in line with previous research, indicating that the most intense pain occurs after day 5 post-oxaliplatin administration [40, 41]. Based on these findings, aVNS treatment and behavioral experiments were conducted on the 5th day, targeting the peak of pain and sustained discomfort.

Alleviating pain symptoms of CIPN with various frequencies of aVNS treatment

CIPN animals were randomly assigned to separate groups and treated with aVNS or sham treatment. The aVNS groups were treated with frequency of either 2, 20, or 100 Hz, aiming to assess the extent and duration of pain relief according to the aVNS frequency. The control group experienced a sham treatment in which the animals were given the same procedure except for the electrical stimulation. The results showed that the aVNS with either frequency exerted analgesic effect on the CIPN symptoms. Among the stimulation parameters, the 20 Hz stimulation exhibited the most significant pain-relieving effect both in the mechanical and cold allodynia, while the 2 and 100 Hz stimulation showed slight pain alleviation (Fig. 2C). In the 20 Hz group, both mechanical and cold allodynia were significantly relieved at all the time points measured (1, 2, and 4 hours after the aVNS). In the 2 and 100 Hz groups, a weak pain-relieving effect on the mechanical allodynia was observed at 2 hours after the aVNS, but the effect vanished at 4 hours. The analgesia on the cold allodynia was appeared 1 hour after the aVNS and disappeared thereafter (Fig. 2D). The overall analgesic efficacies of 2 and 100 Hz aVNS were transient and less prominent compared to 20 Hz aVNS.

Activation of LC of CIPN animals in response to 20 Hz aVNS treatment

VNS is known to activate the LC, and it is known that an activated LC contributes to endogenous analgesic effects [42-44]. To verify whether the aVNS could activate the LC, c-Fos expression in the LC was measured using IHC after 20 Hz aVNS. Preparation of brain samples was performed 1 hour after aVNS treatment after behavioral experiments in oxaliplatin- or vehicle-treated animals. In the analysis of IHC slices, the ROI of the LC was set according to the intensity level based on the staining of the TH antibody in the tissue (Fig. 3A), and the cells in the ROI were detected based on the DAPI staining (Fig. 3B). Ratio of the cells expressing c-Fos to the total number of cells in the ROI was calculated for each slice sample (Fig. 3C). The data were then averaged across each animal subjects to determine whether there were differences between groups in individual level (Fig. 3D). It was observed that the ratio of the c-Fos expressed cells in the LC was higher in the aVNS-treated oxaliplatin group compared to the sham-treated oxaliplatin group and sham-treated vehicle group (Fig. 3C, D) [45]. This indicates that aVNS treatment activated the LC in CIPN animals.

The mediation of the aVNS-induced analgesic effect by spinal β-adrenergic receptors

Activation of the LC induces the endogenous analgesic effects via release of norepinephrine and ensuing activation of adrenergic receptors. To elucidate the mechanism by which the analgesic effect occurs when CIPN animals were treated with 20 Hz aVNS, the behavioral tests were performed in the presence of the antagonists of adrenergic receptors [46]. CIPN animals were randomly assigned to groups and treated with either of prazosin, idazoxan, propranolol or vehicle solutions 30 minutes before aVNS treatment (Fig. 4A). The experimental results showed that vehicle solutions such as 20% DMSO and NS had no effect on the analgesic effect of 20 Hz aVNS, Prazosin and idazoxan also failed to block the analgesic effect of 20 Hz aVNS, However, propranolol blocked the analgesic effect of 20 Hz aVNS, shown by lack of the changes in PWT and cold score (Fig. 4B~E). These results indicate that β-adrenergic receptors contribute to the analgesic action of 20 Hz aVNS treatment in CIPN animals.

This study initially demonstrated that a single intraperitoneal administration of oxaliplatin (6 mg/kg) induced peripheral neuropathic pain characterized by mechanical and cold allodynia in rats (Fig. 1). Subsequently, animals were treated with aVNS with various stimulation frequencies, revealing that the most effective parameter for pain relief was 20 Hz (Fig. 2). IHC analysis confirmed that the 20 Hz aVNS treatment could induce a significantly higher expression rate of the c-Fos in the LC compared to the sham and vehicle group (Fig. 3). This finding aligns with previous studies showing that the LC can be activated in response to VNS [15, 28]. Finally, in line with previous studies [47], it was demonstrated that the β-adrenergic receptor, but not the α₁ and α₂ receptors, was involved in the analgesic effects of 20 Hz aVNS treatment (Fig. 4).

VNS suppresses pain

Previous studies have suggested that aVNS can produce analgesic effects [48]. Guo and Gharibani [49] utilized various parameters to confirm the analgesic effects VNS or aVNS on visceral pain in rats. Chao et al. [50] observed analgesic effects of the aVNS on the chronic constriction injury (CCI) model. These previous studies employed varying conditions, each showing different analgesic outcomes. For example, Guo and Gharibani [49] used 5, 25, or 100 Hz frequency stimulation, while Chao et al. [50] used 2~10 Hz frequency stimulation. The other stimulation parameters were also varied (different pulse width, train on- and off- time, duty cycle, intensity adjustment, and polarization of current). In this study, a frequency of 20 Hz was crucial for producing analgesic effects of aVNS. The frequency specificity of aVNS has been explored in various previous studies regarding to the parameters. Labiner and Ahern [51] reported typical effective frequency of vagus nerve stimulation therapy in treating depression and epilepsy is between 20 and 30 Hz. In terms of LC-activating aVNS, 20 Hz aVNS was found to increase dynamic functional connectivity of LC more effectively than 1 Hz aVNS [52]. Also, the discrepancy in analgesic efficacy observed in this study compared to the others might be attributed to differences in the stimulation conditions, such as using bipolar current, varying voltage levels, and adjustment of voltage [25, 53]. These differences in conditions may account for the varied effects observed [28, 50, 54, 55].

The analgesic effects of VNS are mediated by the adrenergic system

The involvement of adrenergic activation in the endogenous pain modulation has been extensively studied. It is known that the LC is a primary site for NE secretion in the brain, and the activation of LC is crucial for analgesic effects [56]. Research findings suggest that LC, when activated through VNS, serves as a major source of NE [42, 44, 56, 57]. Previous studies have also observed that VNS induces the expression of c-Fos in the LC, indicating that the LC was activated in response to VNS [58, 59]. Based on this, this study corroborated the activation of the LC by confirming the expression of c-Fos in the LC in response to the 20 Hz aVNS treatment. Activated LC neurons may have increased NE levels in the nervous system and influenced adrenergic receptors activation in circuits involved in pain transmission.

Analgesic effect of aVNS on the CIPN involves specific subtype of adrenergic receptor in the spinal cord

The involvement of α₁, α₂, and β receptors in the analgesic effect mediated by NE has been widely acknowledged [31, 40, 60, 61]. We observed that even after injecting α₁ and α₂ receptor antagonists into the spinal cord, the aVNS still could induce an analgesic effect on the pain symptoms in the CIPN animals. Meanwhile, pretreatment of the β receptor antagonist completely blocked the analgesic effect of the aVNS in CIPN animals. Although alpha adrenergic receptors are well known to be involved in spinal analgesic system, previous studies have demonstrated that spinal beta-adrenergic receptors mediate neuropathic pain reduction [62-64]. Blockage of the analgesic effect only by β-adrenergic receptor antagonist but not by α₁ and α₂ antagonists suggests the selective activation of adrenergic receptor subtypes in response to specific stimulation parameters, i.e., aVNS using 20 Hz bipolar 5 V square wave in the current study. Indeed, a previous study found that low-voltage stimulation methods primarily activate β receptors [25, 28]. These results should not be interpreted to indicate that α- adrenergic receptors are not activated by the increased NE levels in this condition. Our claim is that the α receptors are not involved in the aVNS-induced analgesic effect in the oxaliplatin-induced CIPN symptoms. Activation of the spinal β-adrenergic receptor was critically involved in the analgesic effect. Since the propranolol used in the experiment is a non-selective β-adrenergic receptor antagonist, additional research is needed to determine a role of the β₁ or β₂ receptor subtypes in the aVNS-induced analgesic effect and to exclude a possible side effects [65].

Limitation

Although the present study demonstrated the activation of LC in response to aVNS and involvement of spinal adrenergic receptor in the aVNS-induced analgesic effect, it could not exclude the possibility that directly secreted NE could act within the spinal cord [66]. The vagus nerve, being the longest nerve in the body and responsible for regulating various organs, presents multiple pathways. While this study primarily focused on the LC in the brain as the source of descending pathway-induced analgesia, there are other potential pathways to consider. For instance, the vagus nerve contains nerve fibers expressing tyrosine hydroxylase, responsible for synthesizing dopamine and NE [66]. According to previous human studies, cervical and thoracic vagus nerve contains sympathetic nerve fibers, and these catecholaminergic components could play a role in functional effects elicited by VNS [67]. Alternatively, the vagus nerve signals could promote NE release through the activation of adrenal gland [68, 69].

Conclusion

The pain symptoms of CIPN are often intractable and cannot be attenuated with conventional analgesic techniques. As it often causes discontinuation of chemotherapy, it is important to recognize that the pain symptoms should be carefully managed during the chemotherapy. Our study suggests that non-invasive aVNS can be a novel method for managing pain symptoms in CIPN patients. Future researches should focus on improving analgesic efficacy by in-depth studying the mechanisms underlying the activation of endogenous pain modulatory system by aVNS. It is also needed to investigate the safety of this technique so that it can be widely applied in clinical settings.

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NRF-2022R1F1A1072901 to HY, RS-2023-00302281 to GC, RS-2023-00262810 to SKK).

HY is employed by Neurogrin Inc., which was founded by GC and SKK. Neurogrin Inc. holds the patent application related to the contents of this article (10-2023-0168520 in Korea). The remaining authors (ISB and SC) declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.