Sensory afferent neurons are arguably the longest cells in the body; each consists of a soma at the dorsal root ganglia (DRG), a peripheral axon projecting to the end organ, a central axon terminating in the dorsal horn of spinal cord, and a stem axon forming a T-junction with the peripheral and central axons (Lawson, 2005; Abraira and Ginty, 2013). Blocking the afferent neural transmission to the spinal cord has been extensively applied in the clinic for managing various sensory-related disorders, especially many types of chronic pain arising from sensitized afferent input from peripheral tissues (Epstein and Palmieri, 2012; Verrills et al., 2016). In particular, patients suffering from nociceptive pain, postherpetic neuralgia, peripheral neuropathy, musculoskeletal pain, and visceral pain report significant pain relief from treatments that block afferent drives from the spinal nerves (Eldabe et al., 2018; Huygen et al., 2020; Berger et al., 2021; Kim et al., 2021; Chapman et al., 2023). Migraines and headaches can also be managed by blocking afferent drives from the trigeminal ganglia pathway (Messlinger and Russo, 2019). In addition, blocking vagal afferent drive shows therapeutic efficacy in treating obesity and neuropathic cough (Roslin and Kurian, 2001; Canning et al., 2006).

Pharmacological approaches are implemented to block peripheral afferent drives in the clinics. Peripherally restricted kappa-opioid receptor agonists show analgesic efficacy in chronic visceral pain and neuropathic pain in animal models and clinical trials (Albert-Vartanian et al., 2016). Eluxadoline, a peripherally restricted small molecule drug with dual agonist/antagonist action on opioid receptor subtypes, has been approved by the FDA for treating visceral pain associated with irritable bowel syndrome (Lembo et al., 2016). Another dual-acting delta/kappa opioid receptor agonist CAV1001, has shown promising pain attenuating effects in the treatment of arthritis pain, neuropathic pain, and bone cancer pain (Hartrick et al., 2020). Also, cannabinoid receptor agonists activating only peripheral CB1 receptors (CB1Rs) have been reported to effectively alleviate multiple chronic pain conditions with limited side effects (Zhang et al., 2018). Peripheral nerve block (PNB) by local anesthetic injection effectively relieves acute pain (Shah et al., 2018). With advancement in catheter delivery of anesthetics and chemical wrapping to extend the anesthetic’s release, PNB has also been applied to treat chronic pain conditions like migraine headache, chronic pelvic pain syndrome, and chronic Achilles’ tendinopathy (Koo and O'Brien, 2011; Shauly et al., 2019; Chimenti et al., 2020).

Peripheral sensory tissue ablation is an irreversible procedure used to block the afferent drives. Electrical radiofrequency (RF) ablation removes target nerves and tissues by thermal damage, which is effective in suppressing chronic hip pain, chronic discogenic back pain, chronic thoracic and abdominal pain, and knee osteoarthritis pain (Davis et al., 2018; Kumar et al., 2019; Kapural and Deering, 2020; Kim et al., 2020; Singh et al., 2021). Chemical neurolysis uses substances like ethanol or phenol to irreversibly damage peripheral tissues, e.g., celiac plexus for managing chronic upper abdominal pain (Cornman-Homonoff et al., 2017; Sachdev and Gress, 2018) and sacral dorsal rami for sacroiliac joint pain (Nouer Frederico et al., 2021). In addition, cryoablation resorts to extreme low temperature (−60 to −100°C) to ablate peripheral tissues for treating pain (Yasin et al., 2020).

In addition to the pharmacological and tissue ablation approaches, electrical stimulation of peripheral nerves by either anodic direct current or charge-balanced alternating stimulation at kilohertz range can reversibly block afferent neural transmission, as thoroughly reviewed recently (Kilgore and Bhadra, 2014; Patel and Butera, 2018). Anodal block leads to imbalanced electro-chemical reactions at the electrode-tissue interface and is usually reserved as a research tool to selectively block myelinated axons (Sassen and Zimmermann, 1973; Whitwam and Kidd, 1975; Petruska et al., 1998). Charge-balanced kilohertz stimulation reversibly blocks peripheral nerves with rapid onset (within seconds) and modest carry-over effects (seconds to minutes after terminating the stimulus), and has demonstrated efficacy in clinical applications (Patel and Butera, 2018).

Kilohertz frequency stimulation typically ranges from 5 to 100 kHz and has been investigated for its ability to block different types of nerve fibers. Myelinated fast-conductive A-fibers are often blocked at lower amplitudes compared to slow-conductive unmyelinated C-fibers (Patel and Butera, 2018). Studies have shown that higher amplitudes are required to block unmyelinated C-fibers effectively, which may result in non-selective blocking when targeting mixed nerve populations (Patel and Butera, 2015).

Various animal models have been used to study the effects of kilohertz frequency (KHF) block. For instance, Bhadra et al. demonstrated the reversible block of the sciatic nerve in rats using high frequency alternating current with frequencies from 10 to 30 kHz at amplitudes of between 2 and 10 V. Their study showed that high frequency alternating current can induce a reversible nerve conduction block without causing long-term damage to the nerve fibers (Bhadra and Kilgore, 2005). Recent research by Pelot et al. extended these findings by quantifying the effects of KHF on the rat vagus nerve, revealing that block thresholds increase monotonically with frequency for both fast and slow nerve fibers, and that neural conduction can take tens of seconds to recover following certain KHF application. This suggests that while KHF can effectively block nerve activity, the recovery time may vary depending on the duration and amplitude of the applied signal (Pelot and Grill, 2020). Despite these promising results, there are important limitations to consider, kilohertz stimulation does not seem to offer very high selectivity in blocking unmyelinated C-fibers over myelinated A-fiber (Patel and Butera, 2015). This limited selectivity is a critical consideration in the development and application of electrical blocking techniques for therapeutic use. As reported in our recent study, sub-kilohertz electrical stimulation of the afferent somata in the DRG offers frequency-dependent transmission block of C-fiber and Aδ-fiber afferents in an ex vivo preparation for single-fiber recordings from colorectal afferents (Chen et al., 2022). The optimal blocking frequency (OBF) for C-fiber afferents is at 20–50 Hz, while the OBF for Aδ-fiber afferents is at 50–100 Hz. The underlying mechanism of transmission block is likely through activity-dependent conduction slowing, and thus requires supra-threshold stimulation (Chen et al., 2022). However, it remains undetermined whether electrical peripheral nerve stimulation (ePNS) at sub-kilohertz range could also achieve frequency-dependent afferent transmission block. In the current study, we harvested three different peripheral nerves from mice, i.e., the sciatic, saphenous, and vagal nerves, and conducted ex vivo single-fiber recordings to study the afferent transmission block by ePNS of the nerve trunk. We also performed complementary computational simulations in the NEURON environment to provide a mechanistic interpretation of the axonal transmission block by peripheral nerve stimulation.

All experiments were reviewed and approved by the University of Connecticut Institutional Animal Care and Use Committee. All the mice used in the following experiments were housed in pathogen-free facilities which are Public Health Service assured and American Association for Accreditation of Laboratory Animal Care accredited following the Guide for the Care and Use of Laboratory Animals Eighth Edition. Mice resided in individual ventilated caging systems in polycarbonate cages (Animal Care System M.I.C.E.) and were provided with contact bedding (Envigo T7990 B.G. Irradiated Teklad Sani-Chips). Mice were fed ad lib with either 2,918 Irradiated Teklad Global 18% Rodent Diet or 7,904 Irradiated S2335 Mouse Breeder Diet supplied by Envigo and supplied with reverse osmosis water chlorinated to 2 ppm using a water bottle. Nestlets and huts were supplied for enrichment. Rodent housing temperature was set for 73.5°F with a range from 70 to 77°F. Humidity was set at 50% with a range of 35 to 65%. Mice were housed with a maximum of 5 animals per cage. All animals were housed on a 12: 12 light–dark cycle. Animals were observed daily by the animal care services staff. Cages were changed every 2 weeks.

In mammalian peripheral nerves, the transmission of axonal action potentials (APs) in large, myelinated axons is robust, enabling sustained transmission of trains of impulses at 100 Hz or higher for hours (Prochazka and Gorassini, 1998). Consequently, electrical pulse stimulation to block the conduction of A-fiber axons typically requires the stimulus frequency in the kilohertz range, often exceeding 10 kHz (Kilgore and Bhadra, 2014; Patel and Butera, 2018). The frequency range effective for blocking A-fiber transmission has been established through either direct recordings of compound action potentials (CAP) from a population of A-fiber axons in whole-nerve configurations (Juan et al., 2014) or indirect measurement of evoked forces in muscles innervated by motor A-fiber axons (Dowden et al., 2010). The muscle force measurement is not applicable for studying unmyelinated C-fiber axons or thinly myelinated Aδ-fiber axons, as they do not directly influence muscle contraction. CAP recordings from unmyelinated C-fibers pose challenges due to their small signal amplitudes, limiting resolution for assessing subtle neuromodulatory effects. Unlike the monopole-like transmembrane currents from saltatory transmission in myelinated A-fiber axons, AP transmission in slow-conducting C-fibers generates dipole-like transmembrane currents that resulted in a significantly lower extracellular electrical field compared to the A-fiber currents (Plonsey and Barr, 2007). Thus, extracellular recordings of action potentials from Aδ- or C-fiber axons require close proximity of the recording electrode to the nerve axon, typically achieved by manually teasing nerve bundles into fine filaments approximately 10 microns thick (e.g., Chen et al., 2017a; Ilham et al., 2018; Liu et al., 2021). Recent studies by us and others implemented this teased fiber approach to record from split dorsal roots of rats (Chao et al., 2020) and mice (Chen et al., 2022), respectively. Both studies reported reversible blocking of AP transmission in Aδ- and C-fiber afferents by sub-kilohertz pulse stimulation of the dorsal root ganglion (DRG). Especially, frequency as low as 10 Hz is capable of blocking the transmission in C-fiber afferents.

In the current study, we focused on peripheral nerve trunks and demonstrated that ePNS from outside the epineurium can achieve similar transmission block of Aδ- and C-fiber axons in three mouse peripheral nerves: two spinal nerves and one autonomic nerve. The nerve blocking effect by sub-kilohertz ePNS aligns with our prior DRG stimulation study in three aspects (Chen et al., 2022). First, both DRG stimulation and ePNS require suprathreshold stimulation for nerve transmission block, as subthreshold stimulation without evoked action potentials had no significant effect on nerve conduction delay. Second, the range of stimulus frequency efficiently blocking conduction depends on axonal size, with the optimal blocking frequency (OBF) ranging from 10 to 50 Hz for unmyelinated C-fibers and 50–1,000 Hz for myelinated Aδ-fibers. Third, similar increases in conduction delay follow the onset of both ePNS and DRG stimulation until complete conduction block. The DRG was originally hypothesized as the block location due to unique anatomical structures of afferent neurons there, including stem axons, T-junctions, and somata. However, the exclusion of DRG and dorsal roots in the current nerve blocking study strongly indicates the nerve axons as the location of conduction block. In support, the OBF for Aδ-fibers is comparable between DRG stimulation and ePNS, and C-fibers are efficiently blocked by 10–50 Hz stimulation in both cases. It is worth noting that DRG stimulation at 100–1,000 Hz blocked over 50% of C-fiber afferents (Chen et al., 2022), whereas in current study ePNS in the same frequency range blocked no more than 5% of C-fiber axons. This suggests the differential excitability of C-fiber afferents to high frequency stimulation (100–1,000 Hz) at the spinal nerve versus DRG. Indeed, a needle electrode placed at epidural DRG was sufficient to evoke C-fiber afferent activities whereas a glass suction electrode was required to efficiently excite C-fiber axons protected by multiple connective tissue layers within the peripheral nerve trunk. It is speculated that the connective tissues within the nerve trunk may act as a low-pass filter, dampening the intensity of high-frequency stimulation. This could lead to inadequate activation of C-fiber axons with stimulation frequencies ranging from 100 to 1,000 Hz, potentially resulting in the absence of a conduction block.

The finding that stimulation as low as 10 Hz can block afferent transmission has significant clinical implications for pain management with peripheral neuromodulation. The conventional “Gate Control” theory first published by Melzack and Wall (1965) suggests that neurostimulators relieve pain by activating low-threshold, myelinated afferents, triggering paresthesia, a non-painful tingling sensation that masks nociceptive signals (Shamji et al., 2017). However, many patients receiving DRG stimulation do not require paresthesia to achieve pain relief; paresthesia-free patients reported comparable and even better therapeutical benefit from DRG stimulation as patients experiencing paresthesia (Verrills et al., 2019; Mekhail et al., 2020). Thus, it is possible that nociceptor signaling from unmyelinated C-fiber afferents can be blocked by DRG stimulation with clinically applied stimulus intensities, which activates the C-fiber afferents at the OBF to cause conduction block. In contrast, the ePNS has been reported as the validation of the “Gate Control Theory,” and paresthesia appears necessary for its pain-relieving effect (Weiner, 2003). This indicates that most C-fiber afferents are not activated by ePNS at clinically applied intensities. Both clinical observations and preclinical experimental studies indicate that C-fibers are more difficult to evoke by peripheral nerve stimulation than by DRG stimulation. Further research is warranted to investigate underlying anatomical and functional characteristics accounting for this threshold difference in evoking C-fiber afferents at the DRG versus the nerve trunk.

Both Aδ- and C-fiber afferents exhibit pronounced activity-dependent conduction slowing when stimulated at their respective OBF. This phenomenon has been observed in microneurographic studies on human peripheral afferents, which demonstrated progressive slowing in C-fibers using a 2-Hz stimulation protocol (Serra et al., 1999). A complementary computational modeling study suggests that the gradual increase in intra-axonal sodium concentration contributes to this conduction slowing (Tigerholm et al., 2013). With a 2-Hz stimulation, the increase in conduction delay (CD) typically plateaus at about 10% and is usually no more than 30% (Schmelz et al., 2000; Schmidt et al., 2002; Schmelz and Schmidt, 2009; Obreja et al., 2010). However, nerve conduction block was not assessed or reported in those microneurographic studies. In current study, we monitored CD every 2 s during the ePNS protocol that stimulates at frequencies much higher than 2 Hz, revealing a progressive increase in CD until conduction block. The maximum increase in CD (CDI_max) varies greatly between samples, reaching levels as high as ~100%. Furthermore, the notably higher level of CDI_max in blocked axons compared to unblocked ones suggests that the extent to which conduction delay can be significantly increased by a particular ePNS protocol may serve as a predictor of the likelihood of conduction block in the stimulated axon. It is worth noting that the ex vivo setting used in our current study may not fully represent the complex homeostatic mechanisms present in an in vivo context. These differences could potentially lead to divergent outcomes in terms of the effectiveness and duration of the conduction block. We acknowledge that the findings from our ex vivo study will require further validation through carefully designed in vivo experiments to confirm their translational relevance.

Our computational simulation of action potential initiation and propagation reproduces the experimental results of conduction block in both Aδ- and C-fiber axons, capturing the progressive increase in conduction delay (CD) following ePNS until transmission block. We heavily utilized components from our prior computational model of a mouse colorectal afferent, including various subtypes of sodium and potassium ion channels, the sodium, potassium-ATPase (NaKA), and the tracking of intra-axonal ionic concentration (Feng et al., 2015). Specifically, we simulated variations in intracellular Na⁺ and K⁺ concentrations resulting from transmembrane ionic current flow and axial ionic diffusion, showing significant changes following a 2-s stimulation train at 20 Hz for the C-fiber model and at 100 Hz for the A-fiber model. In the C-fiber model with a 1-micron diameter, Na⁺ and K⁺ concentrations exhibited a millimolar change per action potential. This aligns with a simple calculation assuming axonal membrane capacitance charging exclusively through transmembrane Na⁺ current and discharging exclusively through K⁺ current. Our modeling simulations strongly suggest that disruption of transmembrane Na⁺ and K⁺ concentration gradients underly the transmission block by ePNS. Notably, our model predicts a reduction in the sum of intracellular Na⁺ and K⁺ concentration following ePNS, caused by a net outward ionic flow through the NaKA due to the 3: 2 transport ratio of Na⁺ and K⁺ ions. This model-predicted reduction in intra-axonal osmolarity is supported by experimental observations from myelinated A-fiber axons undergoing prolonged 100 Hz high-frequency stimulation, which demonstrated a significant increase in the periaxonal space due to osmotically driven water diffusion (Trigo and Smith, 2015).

The current computational model provides one plausible mechanism of conduction block, i.e., the dysregulated transmembrane ionic gradients. When the extra- and intra-axonal ionic concentrations are kept constant, the same neuromodulation protocols failed to block the transmission in C-fiber and Aδ-fiber models (Figure 8). However, it is important to acknowledge several limitations of the current model, which potentially excludes the interrogation of other mechanisms that may contribute to conduction block. The discrepancy of the longer duration required for block in the model simulation than in the experimental observation clearly indicates the involvement of additional mechanisms in the transmission block process. First, our current model does not consider ionic and metabotropic mechanisms mediated by calcium ions. There is a significant increase in calcium concentration in peripheral axons following action potential conduction (Anderson et al., 2018), which likely play a modulatory role in AP conduction velocity and slowing. Future experimental and computational studies should focus on the role of calcium in stimulation-induced conduction block. Second, the potential reduction in intra-axonal resistivity due to the altered Na⁺ and K⁺ concentrations is not accounted for in the current model. This reduction could potentially play a role in expediting the conduction block by ePNS. Third, the model parameters for the C-fiber and Aδ-fiber are tuned to encode at low frequencies (no more than 50 Hz). Thus, the current model is not suitable to assess the blocking effect of high-frequency stimulation above 100 Hz, as it will not allow us to tease out the intrinsic model properties from other effects that account for the failed block by high-frequency stimulation. Finally, the current gating kinetics of the Na⁺ and K⁺ channels do not include time constants on the order of minutes, which is the time scale for recovery from conduction block. New equations for the voltage-gated ion channels are required for a focused study on the role of extra-slow channel inactivation on the scale of minutes in the transmission block.

The current study presents direct experimental evidence unveiling the reversible conduction block of Aδ- and C-fiber peripheral axons by ePNS within the sub-kilohertz range. More than 88% of Aδ- and C-fibers in three distinct peripheral nerves were blocked by at least one of the five stimulus frequencies tested: 10, 50, 100, 500, and 1,000 Hz. Conversely, no axon was blocked by 5-Hz stimulation. Notably, the efficacy of ePNS-induced blockage depends on the stimulation frequencies relative to the conduction velocities of the targeted nerve axons; slow-conducting C-fibers are optimally blocked by 10–50 Hz stimulation, while Aδ-fibers are effectively blocked by 100–1,000 Hz stimulation. ePNS leads to a progressive increase in conduction delay until transmission blockage, with blocked axons exhibiting significantly higher maximum conduction delay increase than unblocked ones. Additionally, complementary computational modeling of action potential transmission mirrors the blocking effects observed with extracellular pulse stimulation, indicating that disrupted trans-axonal ionic concentration gradients contribute to ePNS-induced conduction block. These findings offer a novel nerve-blocking mechanism that could be leveraged by peripheral neuromodulation methods to enhance therapeutic interventions for managing chronic pain and other neurological disorders.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was approved by University of Connecticut Institutional Animal Care and Use Committee. The study was conducted in accordance with the local legislation and institutional requirements.

SZ: Data curation, Methodology, Writing – original draft, Writing – review & editing, Formal analysis, Investigation. LC: Investigation, Methodology, Writing – review & editing. SL: Investigation, Methodology, Writing – review & editing, Formal analysis, Writing – original draft. AS: Investigation, Methodology, Writing – review & editing. JL: Investigation, Methodology, Writing – review & editing, Data curation. BF: Data curation, Methodology, Writing – review & editing, Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft.