A pathological missense mutation in the deubiquitinase USP5 leads to insensitivity to pain

Nociception involves the transduction of peripheral sensory information into electrical signals, which triggers the release of neurotransmitters and neuropeptides at spinal dorsal horn synapses (Colloca et al., 2017; Ellison, 2017). These then activate ascending neuronal pathways to the brain to induce an unpleasant sensation (Kuner, 2010). In response to inflammation and tissue injury, the pain axis becomes sensitized, in part due to the dysregulation of ion channels in sensory neurons and the spinal cord (SC) (Waxman and Zamponi, 2014; Bourinet et al., 2014).

Among these channels, voltage-gated Cav3.2 channels exhibit an increase in expression in a wide range of painful conditions (Harding and Zamponi, 2022), including during inflammation (García-Caballero et al., 2014; Watanabe et al., 2015), neuropathic pain (García-Caballero et al., 2014; Jagodic et al., 2007; Jagodic et al., 2008; Wen et al., 2010; Feng et al., 2019), visceral pain (Marger et al., 2011; Ozaki et al., 2018), and postoperative pain (Joksimovic et al., 2018). Due to their specific biophysical properties, these channels are well suited toward regulating the excitability of primary afferent fibers and SC interneurons, in addition to supporting low threshold neurotransmitter release at dorsal horn synapses. A key contributor to this upregulation of Cav3.2 channels is an aberrant increase in the expression of the deubiquitinating enzyme ubiquitin-specific protease 5 (USP5) (García-Caballero et al., 2014; Gadotti et al., 2015), which binds directly to the channel and reduces its ubiquitination (García-Caballero et al., 2014). This in turn leads to reduced channel degradation and consequently an increase in their plasma membrane expression levels, culminating in enhanced transmission of nociceptive information. Blocking the interaction of USP5 with the Cav3.2 channel, on the other hand, reduces channel deubiquitination and increases channel degradation, culminating in analgesia (Alonso-Carbajo et al., 2019; Joksimovic et al., 2018; Gadotti et al., 2015; García-Caballero et al., 2014).

Here, we report on a patient with a de novo heterozygous missense mutation in USP5 (p.R24W) that leads to insensitivity to acute pain. The mutated residue is located in the N-terminal Zn finger ubiquitin-specific protease domain, a unique structural feature of USP5 that engages in tight binding to the catalytic core (Gao et al., 2024).

Given that the USP5 amino acid sequence is 98.7% identical between humans and mice (Reyes-Turcu et al., 2008), we generated a heterozygous CRISPR mouse model with the corresponding point mutation to investigate how mutant USP5 alters pain transmission. We show that this mouse exhibits diminished nociception, thus phenocopying the human patient. We also demonstrate that the mutation confers resistance to inflammatory and chronic pain with these effects being influenced by the sex of the animals, and we provide evidence for a dominant-negative loss of function of USP5 modulation of Cav3.2 channels as the underlying molecular mechanism.

The patient was a boy, born 10 days prior to term by vaginal delivery (birth weight 3,898 grams and length 51 cm). His parents were non-consanguineous. He had older siblings who were healthy. His parents reported that he had hypotonia from birth with apneic periods and that he had loose stools from 4 mo of age. He was referred to the hospital at 6 mo with what was believed to be a floor of the mouth tumor, which was surgically removed, but the histology showed ulcerating granulation tissue and fibrosis. Later it was clear that the soft tissue destruction of his lips and tongue and fingertips was caused by self-mutilation. He eradicated his whole lower lip by biting, while his tongue became partly destroyed and locked to the floor of the mouth, and the oral commissures expanded laterally to the middle of the cheeks on both sides. He had further loss of teeth, developed micrognathia, and repeated CT scans showed progressive bone resorption in the lower jaw (Fig. S1). At 2 years of age, it was evident that he could not feel pain, and based on the clinical features, he received the hereditary sensory and autonomic neuropathy diagnosis.

Neurography performed at age 11 years showed a normal motor response, but no sensory responses at all from the following tested localizations: left side median nerve, left side ulnar nerve, left side radial nerve, both sides sural nerves, and left side superficial peroneal nerve. A somatosensory thermotest was performed at age 11 years for detection of heat/cold/heat-pain/cold-pain at the following tested localizations: right thenar, both feet, both proximal legs, and T3 left thorax. There was no temperature sensation at any tested location, except for some heat sensation in the thoracic area, where cold was not perceived. Normal electromyography results were obtained when testing both sides anterior tibial nerves. Neurocognitive tests performed at 10 years of age showed scores clearly lower than average. His head circumference was normal; at 11.5 years of age, it was 52.5 cm (10 percentile), no microcephaly. He continued to have loose, voluminous stools, which was likely due to malabsorption, since he was eating well. At 14 years of age, he developed achalasia and swallow problems, treated with peroral endoscopic myotomy, which helped. He had obstructive sleep apneas and used bilevel positive airway pressure during the nights. He developed visual loss due to corneal scarring in both eyes (neurotrophic keratitis); there was blindness in the right eye, and remaining vision 20% left eye when last tested at 15 years of age.

Already at 1 year of age his body length was below the 2.5 percentile. At 15 years of age, his length was 133.5 cm (25 cm below 2.5 percentile). His short stature had not responded to growth hormone supplementation nor testosterone treatment. His skeletal age corresponded to the biological age. He had kyphoscoliosis. At 15 years of age x-ray showed right convex curvature of the thoracic spine 46o at L1. The levoconvex curvature was 29^o of the upper thoracic spine and 15^o lumbosacral. The thoracolumbar kyphosis curvature was 59^o.

At 11 years of age, he had compartment syndrome in his left leg, which needed surgery with fasciotomy and rhabdomyolysis with creatine kinase levels at 80,000. He experienced several recurrent infections and osteomyelitis over the years. At 15 years of age, he had septic knee arthritis with a leg abscess. The leg abscess was treated with surgical drainage and systemic antibiotics, but then he got septic embolism of the lung and spleen, as well as abscesses, with destruction of the distal right femur. He died suddenly and unexpectedly at 16 years of age; he stopped breathing, collapsed, and was not possible to resuscitate. No autopsy was performed.

He was extensively genetically tested with repeated whole-exome sequencing and trio whole-exome sequencing filtered with various gene panels at different clinical laboratories with negative findings, especially no deleterious or suspected variants in any hereditary sensory and autonomic neuropathy-related genes were found. Microarray-based comparative genomic hybridization 180k did not show any copy number variants that could explain his phenotype, nor did specific testing such as multiplex ligation-dependent probe amplification for neuropathies including Charcot–Marie–Tooth disease. A careful, detailed, and expanded evaluation of the results from one of the trio whole-exome sequencing performed as part of a research setting showed one de novo occurring variant in USP5, detected in half of the reads in the affected patient and not observed in the parents. NM_001098536.1 (USP5):c.[70C>T];[ = ], p.(Arg24Trp) heterozygous. No other individual was registered with this specific USP5 variant in GnomADv.3.1.2. The nucleotide c.70C as well as the codon and protein domain are all highly conserved across species, and the variant was predicted to be deleterious by combined annotation dependent depletion, PolyPhen2, sorting intolerant from tolerant, and MutationTaster.

The parents agreed to both the extended genetic testing and the publication of their child’s clinical features.

To model the pathology of the pediatric patient and to understand the molecular basis for his pain insensitivity, a line of mice carrying the R24W mutation of USP5 on the C57BL/6J background was created using CRISPR technology as described in Materials and methods (Fig. 1 A). Experimental animals were generated by crossing WT mice with mice heterozygous for the mutation as determined by PCR, followed by restriction digestion (Fig. S2). Out of 796 pups genotyped (female and male), 64.4% were WT, and 35.6% were heterozygous for the mutation (R24W) showing a Mendelian inheritance, no difference was observed between male and female mice ratio (Fig. 1 B). These data indicate that the R24W mutation does not affect gestation or survival. R24W litter were same size (Fig. 1 C) and weight (Fig. 1 D) as WT animals, so no gross anatomical or developmental deficiencies were observed.

To investigate whether the point mutation in USP5 has a role in nociception, we measured paw withdrawal responses of naive WT and R24W via Von Frey hairs. Fig. 2 A shows that male R24W mice exhibit a higher mechanical withdrawal threshold compared with WT animals, indicating hyposensitivity of the R24W mice. In contrast, female animals did not show a difference in withdrawal thresholds between WT and R24W genotypes, indicating that there is a sex difference in the underlying cellular/molecular mechanisms.

We also examined nocifensive responses to application of the proalgesic agent capsaicin. Capsaicin is a potent irritant that causes burning pain in humans and acts on nociceptive primary afferents through the activation of transient receptor potential vanilloid 1 channels (Szolcsányi, 1977; Yang et al., 2015). Upon capsaicin injection into the hind paw, WT animals presented robust licking and shaking of the affected paw. In contrast, R24W mice displayed significantly reduced responses to nociceptive sensitization (Fig. 2 B). The decrease of nociceptive behavior was observed in both sexes but was more prominent in male compared with female R24W mice.

To test responses to acute inflammation (Tjølsen et al., 1992), formalin was injected into one hind paw of WT and R24W mice. The formalin test allows the study of two different processes; the first phase is associated with the activation of transient receptor potential ankyrin 1 channels located in nociceptors which become sensitized. The second phase is due to inflammation and central sensitization in the dorsal horn (McNamara et al., 2007). Formalin produced nocifensive responses both in the first and second phases in WT animals (Fig. 2, C and D), as characterized by the time animals spent licking and biting their paws. While male and female mice showed similar responses in phase 1, they appeared to be exacerbated in female mice in the second phase compared with their male counterparts (Fig. 2 D), and such sex differences are in fact observed in several other pain models (Rosen et al., 2017). Importantly, R24W animals of both sexes showed a significantly attenuated response to formalin in both phases (Fig. 2, C and D).

Overall, our findings reveal that R24W animals exhibit a reduced nociceptive response to acute pain, which, in the case of male mice phenocopies the patient’s condition.

We evaluated mechanical and thermal hypersensitivity in the CFA inflammatory pain model. CFA was injected unilaterally into the hind paw of WT and R24W animals, and their contralateral hind paw was used as internal control. Mechanical and thermal withdrawal thresholds were determined respectively, with a digital plantar aesthesiometer and a Hargreaves apparatus.

CFA induced a strong inflammatory response in both genotypes. In both male and female WT mice, CFA significantly decreased the mechanical paw withdrawal threshold (PWT) compared with the contralateral side (Fig. 3, A and C), indicating mechanical hypersensitivity. This effect was eliminated in male animals carrying the USP5 point mutation (Fig. 3 A), whereas R24W female mice did not show a reversal of mechanical withdrawal thresholds (Fig. 3 C). CFA-treated WT mice exhibited a reduction in thermal withdrawal latency to a radiant heat stimulus in the ipsilateral paw (Fig. 3, B and D). R24W mice of both sexes did not present a difference in thermal withdrawal latency (Fig. 3, B and D), suggesting that R24W mice did not develop thermal hypersensitivity. These data also indicate that there is a sex difference in pain modality with thermal, but not mechanical thresholds being affected by the R24W mutation in female mice.

Next, we investigated whether R24W mice exhibit different mechanical and thermal hypersensitivity in response to peripheral nerve injury. In both male and female WT mice subjected to partial sciatic nerve ligation (PSNL), there was a reduction in mechanical and thermal withdrawal thresholds on the ipsilateral compared with the contralateral side (Fig. 4, A–D). In contrast, male R24W mice exhibited no sensitization to mechanical (Fig. 4 A) or thermal stimuli (Fig. 4 B), consistent with insensitivity to neuropathic pain. On the other hand, female R24W mice showed normal behavioral sensitization to mechanical stimuli (Fig. 4 C), whereas sensitization to thermal stimuli was abolished (Fig. 4 D) like in the male subjects. This mirrors our observations in the CFA test with regard to a sex difference that is pain modality specific.

Finally, we previously reported (Antunes et al., 2024) an increase in USP5 protein levels in a model of chemotherapy-induced peripheral neuropathic pain—a common side effect affecting a significant percentage of cancer patients treated with cytostatic drugs such as oxaliplatin (Fukuda et al., 2017). To determine if the R24W mutation alters oxaliplatin-induced neuropathic pain, we treated WT and R24W mice chronically with oxaliplatin and tested them for mechanical hypersensitivity and cold allodynia. Consistent with our prior work, male and female WT mice developed mechanical hypersensitivity and cold allodynia following oxaliplatin treatment (Fig. 5). On the other hand, R24W mice did not develop mechanical or cold hypersensitivity (Fig. 5).

Together, these data indicate that R24W male animals do not exhibit mechanical, heat, or cold hypersensitivity in response to peripheral inflammation or nerve injury. In contrast, female R24W mice retain the ability to develop mechanical hypersensitivity in the CFA and PSNL models, but interestingly not in oxaliplatin-treated animals.

It is well established in various pain models that Cav3.2 channels play a critical role in the transmission of painful signals (Harding and Zamponi, 2022). During pain states, an increase in Cav3.2 protein levels and current is observed in primary afferent and SC neurons. We previously reported that this increase is due to an upregulation of USP5, which deubiquitinates and stabilizes the channels at the plasma membrane, thereby contributing to pain sensitivity.

Following intraplantar (i.pl.) CFA injection into WT and R24W mice, we collected tissue from the ipsilateral dorsal root ganglia (DRG) and SC. Fig. 6 A shows a representative western blot of SC tissue from male WT and R24W mice injected with PBS or CFA. An increase in USP5 expression was observed in WT animals injected with CFA compared with those injected with PBS; however, this effect was not as evident in R24W mice. Because western blots can be challenging to quantify, we measured USP5 and Cav3.2 levels in the ipsilateral side of DRG and SC tissue using ELISA. Fig. 6, B and C reveal a significant increase in USP5 and Cav3.2 expression in DRG of WT mice in CFA compared with PBS conditions, with no difference at the contralateral side (Fig. S3). However, this CFA-induced increase was greatly attenuated if not absent in R24W mice. The CFA effect on Cav3.2 levels was then also examined using whole-cell patch-clamp recordings from DRG neurons of CFA- or PBS-treated male WT and R24W mice (Fig. 6 D), which reflect Cav3.2 channel levels at the cell surface. Here, DRG neurons from CFA-treated mutant mice showed a significant decrease in whole-cell T-type calcium current density compared with WT, in agreement with the ELISA data of Fig. 6 C. We then performed ELISA analysis from SC tissue in analogy to the experiment shown in Fig. 6, B and C. As shown in Fig. 6, E and F, data from the SC essentially mirrored those obtained from DRG, overall indicating that R24W mice are resistant to CFA-mediated upregulation of USP5 and Cav3.2. This lack of upregulation correlates with the absence of responses observed during behavioral tests of the different pain models we examined.

In female mice, CFA also mediated an upregulation of USP5 and Cav3.2 in WT mice, and like in male subjects, this effect was abrogated in the R24W animals (Fig. S4). Interestingly, female R24W mice exhibited a basal elevation of Cav3.2 levels in DRG but not SC tissue compared with WT animals. It is possible that this difference to male animals may contribute to the observed sex differences of pain responses.

Since numerous ion channels expressed in DRG neurons are known to be dysregulated in pain states, we performed ELISA studies on DRG tissue from male mice to examine protein levels of three additional ion channels: Cav2.2 and HCN1, which have been shown to exhibit increased expression in CFA models (Alkhatib et al., 2019; Cho et al., 2009), and Cav2.1, another Cav channel involved in neurotransmission (Dolphin and Lee, 2020) but not commonly associated with afferent pain signaling. Cav2.1 protein levels were not altered by CFA treatment in either WT or R24W mice (Fig. S5 A). In PBS conditions, WT and mutant mice had the same Cav2.2 levels. CFA treatment increased Cav2.2 expression levels as previously reported, and after CFA injection, the two mouse lines also had statistically the same Cav2.2 levels (Fig. S5 B). However, a direct comparison of Cav2.2 in PBS vs CFA in R24W mice did not meet statistical significance. This apparent inconsistency is likely related to sample size. HCN1 expression is robustly increased in both WT and R24W mice, indicating that this upregulation is not USP5 dependent.

To determine whether the amino acid change from arginine to tryptophan in USP5 alters interacting residues, we performed in silico modeling based on the x-ray crystallographic structure of USP5. We found that arginine 24 (R24) is stabilized by hydrogen bonds with histidine 26 (H26), phenylalanine 50 (F50), and tryptophan 260 (W260), as depicted in Fig. 7 A. In contrast, the missense mutation of tryptophan 24 forms a Van der Waals interaction with histidine 26 (H26; Fig. 7 B), resulting in a predicted decrease in binding affinity (ΔΔG) of −0.199 kcal/mol. This then leads to a global structural perturbation that may in turn affect USP5 function.

To investigate whether R24W retains its deubiquitinase (DUB) activity, we used purified USP5 and R24W proteins and assessed their catalytic activity by using a ubiquitin-rhodamine 110 cleavage kinetic assay (Hassiepen et al., 2007). We observed that the R24W mutant exhibited a dramatically decreased DUB activity at every time point tested compared with the USP5 WT (Fig. 7 C). DUB activity was not entirely absent, but massively impaired in the R24W construct. For all intents and purposes, unless there is compensation, in a subject carrying a heterozygous mutation, this would lead to a 50% loss in overall USP5 DUB activity due to essentially zero activity of one of the two gene copies.

USP5 binds to Cav3.2 channels via the III–IV linker region of the channel, stabilizing the channels at the membrane by removing ubiquitin groups (García-Caballero et al., 2014). To determine whether the R24W mutant retains the ability to bind to Cav3.2 channels, we conducted binding assays using ELISA to test the affinity of WT and R24W recombinant proteins for the III–IV linker peptide. We found that R24W binds significantly more strongly to the Cav3.2 III–IV linker peptide (Fig. 7 D). This is important because this would allow mutant nonfunctional R24W USP5 to outcompete WT USP5 at the channel interaction site, thereby acting in a dominant-negative fashion.

Finally, to investigate whether the R24W mutation impacts transcription, we performed RNA sequencing to estimate the relative proportion of WT and R24W transcripts under PBS and CFA treatment conditions (Table S1). For PBS-treated CRISPR animals, the WT transcript accounted for 49.02 ± 0.47% of total expression, while the R24W transcript accounted for 50.98 ± 0.47%. CFA injection did not alter this ratio (48.94 ± 0.47% in WT and 51.06 ± 0.47% in R24W), suggesting that the reduction in USP5 expression observed in CFA-treated CRISPR animals is not due to a transcriptional mechanism but rather to translation efficiency or protein turnover. In WT animals, the relative expression of USP5 under CFA and PBS conditions was the same, which suggests that injection of CFA does not cause transcriptional upregulation of USP5 in WT animals. This stability in transcript levels also supports the premise that changes in USP5 protein levels are likely to be due to posttranscriptional regulation rather than transcriptional modulation.

Congenital insensitivity to pain is a rare (1 in 25,000 births) neurological condition in which patients cannot feel normally painful stimuli (Drissi et al., 2020). Afflicted individuals typically do not exhibit other abnormalities, but affected children often suffer from secondary injuries, such as bite wounds in their oral cavities, damage to the eyes due to rubbing, and unnoticed infections that can be fatal (Drissi et al., 2020). Mutations in several different genes have been associated with congenital insensitivity to pain, including PR domain zinc finger protein 12, zinc finger homeobox 2, neurotrophic receptor tyrosine kinase 1, and the pseudogene FAAH-OUT (Chen et al., 2015; Habib et al., 2018, 2019; Indo et al., 1996). Loss of function mutations in SCNA9, the gene encoding the sodium channel Nav1.7, has perhaps attracted the greatest level of attention (Dormer et al., 2023; Baker and Nassar, 2020), in part because there is a clear mechanistic explanation for the physiological effects of these types of variants (Cox et al., 2006).

We report here a male pediatric patient with a de novo heterozygous missense mutation in USP5 (R24W) that causes acute pain insensitivity. To study how mutant USP5 affects pain transmission, we generated a heterozygous CRISPR mouse model carrying the R24W mutation. These mice showed no gross anatomical or developmental deficiencies, with size and weight comparable with WT animals. In general, R24W mice were resistant to acute pain, which essentially phenocopies the patient. In addition, R24W mice (in particular male subjects) were resistant to chronic pain states induced by neuropathy or peripheral inflammation. We do not know whether the affected patient also showed resistance to such chronic pain conditions, as this situation was not encountered clinically.

Our data are overall consistent with our prior findings that link an interaction between USP5 and Cav3.2 channels to pain responses in inflammatory and neuropathic pain models (García-Caballero et al., 2014). Knockdown of USP5 by intrathecal delivery of short hairpin ribonucleic acid against USP5 reversed inflammatory and neuropathic pain (García-Caballero et al., 2014), as did the disruption of the USP5–Cav3.2 interaction with interfering peptides or small organic molecules (García-Caballero et al., 2014, 2022; Joksimovic et al., 2018; Gadotti et al., 2015). Our new data show that USP5 and Cav3.2 channel levels are increased in DRG and SC tissue in WT animals injected with CFA. In contrast, CFA did not cause an increase in USP5 or Cav3.2 channel levels or current densities in the R24W mice that lack one functional copy of the USP5 gene. The molecular mechanisms underlying the CFA-mediated increase in USP5 in WT animals and the absence of such an effect in mutant mice remain to be determined. However, our results indicate that WT/R24W transcript proportions remain unchanged in R24W mice, suggesting that the regulation of USP5 expression occurs at a posttranscriptional level rather than through transcriptional control.

Modeling of the mutation (R24W) in silico suggests that it induces a global structural perturbation that is likely to affect its function. Indeed, the DUB activity of R24W was dramatically attenuated, and at the same time, R24W USP5 appeared to more strongly interact with Cav3.2 channels, thus likely outcompeting WT USP5 to mediate an additional dominant-negative effect that renders the normative copy of USP5 less effective in modulating Cav3.2 channel levels. Indeed, we showed previously that a ∼80% reduction in USP5 protein levels in cath.-a-differentiated (CAD) cells treated with USP5 shRNA mediated increase in Cav3.2 ubiquitination levels (García-Caballero et al., 2014; Watanabe et al., 2015). Although we did not explicitly examine Cav3.2 ubiquitination levels in DRG neurons from WT and mutant mice, we expect a comparable loss of USP5 activity to our previous studies in CAD cells and thus increased Cav3.2 ubiquitination (and consistent with the reduced Cav3.2 current densities seen in our electrophysiological analysis). We have previously shown that SUMOylation of USP5 reduces its ability to interact with the Cav3.2 channel (Garcia-Caballero et al., 2019), but preliminary data (not shown) indicate that the SUMOylation levels of WT and R24W USP5 are similar and therefore unlikely to be responsible for the tighter interaction of mutant USP5 with the channel. Finally, we note that our data with the mutant mice parallel our findings with acute shRNA knockdown. Hence, the observed resistance to pain hypersensitivity in the R24W mice is unlikely due to a developmental effect but rather due to a reduced ability of USP5 to stabilize Cav3.2 channels. During our study, we generated a number of homozygous mice. We did not test these since they did not reflect the genotype of the patient. These mice were viable, which differs from what is known about homozygous USP5 knockout mice, which die at embryonic day 7. These data indicate that mutant USP5 retains some key physiological functions that are necessary for survival.

The observed sex differences fit with a number of previous studies in both humans and rodents. Sex differences in inflammatory and chronic pain have been reported. For example, women have a higher prevalence of inflammatory and chronic diseases (Rosen et al., 2017; Fillingim et al., 2009). Females produce a stronger proinflammatory immune response to tissue damage since circulating estrogen levels increase the release of proinflammatory cytokines by mast cells, macrophages, and T cells, which can affect the initiation and maintenance of neuropathic pain (Calippe et al., 2010; Smith et al., 2011; Gilmore et al., 1997). It has been reported that the TLR4 agonist LPS induces dose-dependent allodynia when injected into the SC of male but not female mice, in a testosterone-dependent manner (Sorge et al., 2011). Mac-1–saporin, a neurotoxin causing microglial depletion, reversed allodynia in male, but not female animals in the spared nerve injury model, an effect that has been linked to T cells (Sorge et al., 2015). In this context, it is interesting to note that USP5 is expressed in T cells (Ovaa et al., 2004; Xiao et al., 2023), where it regulates the production of effector cytokines. It is thus possible that the sex-dependent effects of the USP5 mutation may involve a neuro–immune interaction. For example, it is conceivable that mutant USP5 interferes with release of inflammatory mediators from female, but not male T cells. In this context, it is particularly interesting that these sex differences were pain modality specific, with only mechanical hypersensitivity being affected. We know from our previous work that USP5 is capable of regulating CFA-induced mechanical hypersensitivity even in female mice (Gadotti and Zamponi, 2018), hence the observed sex differences are likely related to the mutation per se rather than due to a differential ability of USP5 to regulate mechanical sensitivity in male and female mice. It is therefore possible that mutant USP5 may alter an immune process that selectively affects mechanosensitive neurons in female mice, or perhaps the processing of mechanosensitive inputs into the SC. We also cannot rule out the possibility that USP5 may regulate other types of ion channels, such as, for example, a mechanosensitive channel that may be expressed in a sex-dependent manner. Moreover, we note that basal USP5 and Cav3.2 levels in DRG tissue of female R24W mice appear to be elevated compared with WT animals, and this may perhaps also be a contributing factor. Sex differences that are pain modality specific are not without precedent, since we recently showed that different types of Cav3 calcium channels differentially regulate cold allodynia, but not heat or mechanical hypersensitivity in mice in male and female mice (Antunes et al., 2024). Be that as it may, because the patient described here is the only known case in the world with this specific mutation, we do not know what the phenotype of a female human carrier of the mutations might be.

In summary, our data show that the R24W mutant of USP5 critically modulates acute and chronic pain states in both humans and mice by acting as a dominant-negative regulator of Cav3.2 channel stability. USP5–Cav3.2 channel interactions are being actively explored as a target for the development of new analgesics, with promising results in a number of preclinical models (Garcia-Caballero et al., 2022). As with many therapeutics, it is often unclear whether data obtained in preclinical models are predictive of outcomes in human patients. The data presented here validate for the first time USP5 as a potential therapeutic target for chronic pain in humans.

Our study examined male and female animals, and sex-dimorphic effects are reported.

The R24W allele was created at the Molecular Biology Core of the University of Calgary using two ribonucleoprotein complexes consisting of Cas9 protein and specific single-guide RNAs (sgRNA1 and sgRNA2). These ribonucleoprotein complexes were simultaneously employed to cleave the genome at two targeted sites. A linear single-stranded DNA donor template, spanning 481 nucleotides with 100-nucleotide homology arms on each end, was utilized. The donor DNA sequence was designed with target sites located 130-base pairs upstream and downstream of the R24 site. The upstream target sequence (5′-TTG​GAT​GCG​AGA​GAC​GGA​GG-3′, forward orientation) and downstream target sequence (5′-CAG​GGT​GCA​TCG​CAA​GGA​GG-3′, reverse orientation) guided the editing process. The donor DNA sequence used for homology-directed repair was: 5′-GAG​GGT​GGA​GCG​CCA​GGA​GGC​AGG​CTG​GGT​CCA​ACT​GCC​GAT​CTC​GCG​CGT​CTT​GCC​TTT​AGC​GCA​TGC​GCA​CAC​TGA​GTC​GCG​CTC​ATT​GGA​TGC​GAG​AAT​TCG​AGG​GGG​GCG​GTA​CTA​GGT​ACA​GTG​GGA​GCT​GCT​GTG​TGA​GGA​GGA​GCT​GCT​GCC​GGT​GTC​ATG​GCG​GAG​CTG​AGT​GAA​GAG​GCG​CTG​CTG​TCA​GTG​TTA​CCG​ACG​ATC​CGT​GTC​CCC​AAG​GCG​GGA​GAC​TGG​GTC​CAT​AAA​GAC​GAG​TGC​GCT​TTC​TCT​TTC​GAC​ACG​CCG​GTA​AGC​CCA​TTC​CCC​ACG​CCC​GCG​ACG​ACC​ACG​ACT​TCC​TTC​CAT​TGC​CCT​GGT​CAT​TCG​GCC​AGG​CCT​GCA​AAC​CTT​GGG​CTA​CCG​CCT​CCA​TGC​GAT​GCA​CCA​TGG​GAC​TTG​TAG​TTT​TCT​CCA​TAC​CTC​TCT​GCT​TTT​GCT​TTT​CAT​TCT​CTG​TAG​CTG​TAG​TAT​GAC​TAC​CAC​TCC​CGG​AAG​CTA​CAG​CTC​TCG​CCT​TGC​C-3′. Verification of the R24W mouse genotype was conducted using PCR with the following primers: forward: 5′-TTC​GGC​CAT​CAG​TTG​TAC​CC-3′, reverse: 5′-TGT​AGC​TTC​CGG​GAG​TGG​TA-3′. One-cell stage mouse embryos were microinjected with CRISPR components, including Cas9, guide RNA, and donor DNA. The embryos were then allowed to develop to the blastocyst stage, after which they were implanted into pseudopregnant mice. Once knock-in mice were obtained, the presence of the R24W mutation was confirmed by sequencing. For subsequent breeding, we used the PAGE technique to check for the mutation in the litters (Ota et al., 2013), which is now a widely established method for genotyping mouse lines with single-point mutations. The resulting pups were analyzed after they were weaned; a representative gel is shown in Fig. S1. The body length of mice at 4 and 8 wk of age was measured from nose to anus using a ruler. Body weight was monitored with a digital scale from week 3 to week 8.

At experimental endpoints, animals were euthanized by isoflurane overdose. All i.pl. injections were performed into the ventral surface of the right hind paw. For behavioral experiments, the investigators were not blinded to the treatment conditions; however, the experimental groups were carefully randomized and balanced among different experimenters. For all experiments, mice were habituated in the laboratory for 60–90 min before any test.

A 20 μl solution of capsaicin (1.6 μg/paw in saline; M2028; Sigma-Aldrich) was injected i.pl. Animals were observed individually for 5 min following the capsaicin injection. The time spent licking the injected paw was considered indicative of nociception.

Formalin tests were conducted according to the method by Hunskaar et al. (1985) and as previously utilized by our lab (García-Caballero et al., 2014). Mice were injected i.pl. with 20 μl of a 2.5% formalin solution prepared in PBS (F1635; Sigma-Aldrich). After i.pl. injections, individual animals were immediately placed in observation chambers and monitored from 0 to 5 min (acute nociceptive phase) and from 15 to 30 min (inflammatory phase). The time spent licking or biting the injected paw was recorded by chronometer as a nociceptive response.

CFA (F5881; Sigma-Aldrich) was used to induce peripheral inflammation. 20 microliters of CFA were injected i.pl. 2 days after the CFA injection, mice were assessed for thermal withdrawal latency and mechanical withdrawal threshold. The contralateral paw assessment was used as a control.

To induce chronic neuropathic pain, a PSNL was performed. This involved placing a single tight ligature around one-third to one-half of the sciatic nerve’s diameter. Mice were anesthetized with isoflurane (5% induction and 2.5% maintenance) and monitored for respiration and spontaneous breathing throughout the procedure. The fur around the upper right leg was shaved and disinfected with 70% ethanol, and an incision was made using a surgical blade. The leg muscle was separated by blunt dissection to expose the sciatic nerve, which was then ligated just proximal to the trifurcation. In sham-operated mice, the nerve was exposed but not ligated (Malmberg and Basbaum, 1998). The incision was closed with 4.0 Vicryl sutures (three to four stitches). Heat support was provided during the short surgeries, and a heating pad was placed under half of the cage to assist with recovery from anesthesia. After surgery, mice were monitored for up to 1 h to ensure successful recovery. Mice were monitored daily for 7 days to ensure suture integrity and proper recovery from surgery. All mice healed well and were subsequently included in the behavioral experiments. Behavioral tests were performed before the surgery (baseline) and 21 days after (test).

Oxaliplatin (2623; Tocris Bioscience) was dissolved in a 5% glucose solution. Mice received intraperitoneal injections of oxaliplatin (3 mg/kg) or a vehicle (5% glucose solution) twice weekly for 3 wk. Mechanical and cold sensitivity tests were conducted on the mice before treatment (baseline: B) and immediately after the development of OIPN (time zero: T0). For this model, animals were tested first for the acetone test, followed by a von Frey test.

Mechanical sensitivity was measured using the 50% PWT in response to a series of von Frey filaments (North Coast Medical Inc.) following the up and down method (Chaplan et al., 1994). Mice were placed individually in small, enclosed testing arenas on top of a grid platform. A set of von Frey filaments with approximately equal logarithmic incremental bending forces were chosen (0.6, 1, 1.4, 2, and 4 g). Each trial began with a von Frey filament exerting a force of 0.6 g, applied perpendicularly to the plantar surface of the right hind paw for 2–3 s. An abrupt withdrawal of the foot during stimulation or immediately after the filament removal was recorded as a positive response. If there was a positive response, the next weaker filament was applied; if there was a negative response, the next stronger filament was used. This procedure was applied six times or 5 consecutive responses, with at least a minute of interval between the stimuli. The 50% PWT was calculated using the formula: $50% PWT=10^((X+kδ)/f)$⁠. X is the exact value (in log units) of the final Von Frey filament, k is the tabular value for the pattern of the last six positive/negative responses, and δ is the mean difference (in log units) between stimuli filaments (Chaplan et al., 1994). If an animal did not respond to the highest von Frey filament, the value was recorded as 4 g. For the OIPN model mechanical allodynia, hind paw withdrawal was quantified by applying a single von Frey filament. The von Frey hair (0.4 g) was applied with constant pressure for 5 s to the ventral surface of the right hind paw, in a total of 10 trials. The number of withdrawal reactions to the stimuli was counted and expressed as a percentage response (Martins et al., 2012).

Mechanical hyperalgesia was evaluated using a digital plantar aesthesiometer (Ugo Basile). Mice were individually placed in a small, enclosed testing arena on a grid platform. The aesthesiometer was positioned beneath the testing floor to accurately place the filament under the plantar surface of the ipsilateral hind paw. Each paw was tested three times per session, as described by García-Caballero et al. (2014).

Thermal hyperalgesia was assessed by measuring the latency of right hind paw withdrawal from a focused beam of radiant heat (infrared heat source setting = 30) using a Hargreaves apparatus (Ugo Basile). Mice were individually placed in a small, enclosed testing arena (20 × 18.5 × 13 cm) on a plastic glass floor. The heat source was positioned beneath the animal, directly under the plantar surface of the ipsilateral hind paw. Each mouse was tested three times. A cutoff time of 30 s was set to avoid tissue damage.

The cooling effect of evaporating acetone was used to quantify nociceptive behavior. 20 μl of acetone were applied to the ventral side of the right hind paw. The cumulative duration of flinching or licking of the hind paw was recorded for 40 s using a stopwatch (Colburn et al., 2007). The average response was calculated from at least two to three trials of the acetone test, with a minimum of 5 min between each trial.

DRGs were harvested from both sides of the fourth lumbar vertebra (L4) to the sixth lumbar vertebra (L6) of 6-wk-old C57Bl/6 male mice. The tissue was cut into small pieces and incubated at 37°C for 30 min in a solution containing 4 mg/ml collagenase (Gibco) and 40 μl/ml papain (Worthington) in Dulbecco’s modified Eagle’s medium (Gibco), supplemented with 10% heat-inactivated fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Gibco). This was followed by an additional 10-min incubation with 1 μg/ml DNase (Sigma-Aldrich). The digested DRG tissue was then washed three times with culture medium and mechanically dissociated into single cells. DRG neurons were plated on glass coverslips pre-treated with poly–D-lysine (Sigma-Aldrich) and laminin (Sigma-Aldrich) and maintained at 37°C in a 5% CO₂ incubator. Electrophysiological recordings were performed 12–24 h after plating.

Unilateral (right side) DRGs from the spinal lumbar section (L4–L6) were obtained from PBS- or CFA-injected WT or R24W animals. Small and medium neurons (<25 pF) were used for patch-clamp recordings using the external solution (in mM): 40 TEA-Cl, 65 CsCl, 20 BaCl₂, 1 MgCl₂, 10 HEPES, and 10 glucose, pH adjusted to 7.4. The internal solution contained (in mM): 140 CsCl, 2.5 CaCl₂, 1 MgCl₂, 5 EGTA, 10 HEPES, 2 Na-ATP, and 0.3 Na-GTP, pH adjusted to 7.3. Recordings were conducted using an EPC 10 amplifier connected to a personal computer with Pulse (V8.65) software (HEKA Elektronik). The junction potential of 2.8 mV was calculated using the Junction Potential Calculator in Clampex 11.2 (Molecular Devices) and was not corrected. Inward Ca²⁺ currents were induced by a 300-ms depolarization from a holding potential of −90 mV to a test potential of −30 mV.

SC tissue was lysed in a modified radioimmunoprecipitation assay buffer (in mM; 50 Tris, 100 NaCl, 0.2% [vol/vol] Triton X-100, 0.2% [vol/vol] NP-40, and 10 EDTA), pH 7.5 plus protease inhibitor cocktail (Complete 1x). Lysates were prepared by sonicating samples at 60% pulse for 10 s followed by centrifugation at 13,000 rpm for 15 min at 4°C. Samples were loaded on 10% Tris-glycine gel and resolved using SDS-PAGE. Samples were transferred to 0.45-mm polyvinylidene difluoride membranes by dry transfer using an Iblot2 machine (Invitrogen). Western blot assays were performed using anti–α-tubulin (1:2,000, 7291; Abcam) and anti-USP5 (1:250; 15158-1-AP; ProteinTech).

Experiments were carried out in 384-well black polypropylene microplates (Greiner). Recombinant WT-USP5 (1 nM, long isoform from Enzo) or recombinant R24W (1 nM; Bon Opus Biosciences) and ubiquitin-rhodamine 110 (200 nM; UBPBio) were incubated in the wells containing 20 mM Tris, pH 7.5, 30 mM NaCl, and 0.01% Triton-X (vol/vol) for a final volume of 60 μl. Fluorescence kinetics readings were immediately recorded for 40 min at 37°C using a Microplate reader (SpectraMax i3x) with excitation and emission wavelengths set to 485 and 530 nm, respectively.

Neutravidin-coated 384-well plates (Thermo Fisher Scientific) were incubated with 60 μl of blocking buffer (Tris 50 mM, pH 7.4, NaCl 150 mM, and BSA 1%) at room temperature (RT). The plates were washed three times with Tris buffer (Tris 50 mM pH 7.4 and NaCl 150 mM). 58.5 ng of biotinylated Cav3.2 III–IV linker (Genemed Synthesis, Inc.) were diluted in 15 μl Tris buffer, added to the wells, and incubated at RT for 1 h. The plates were washed three times with Tris buffer. WT-USP5 (Enzo Life Sciences, Inc.) or R24W-USP5 (Bon Opus Biosciences) recombinant proteins at concentrations of 0.00078 μg/μl in 15 μl of Tris buffer were added to all wells except for the negative control wells and incubated at RT for 1 h, followed by washing the wells three times. USP5 primary antibody (15158-1-AP; ProteinTech) was diluted in blocking buffer (1:5,000), added to the wells, and incubated at RT for 1 h. The wells were washed three times for 10 min each. HRP-conjugated secondary antibody (1:10,000, 211-032–171; Jackson ImmunoResearch) was added and incubated at RT for 1 h, followed by washing. QuantaBlu substrate plus stabilizer (9:1 ratio) (Thermo Fisher Scientific) was added to the wells and incubated at RT for 10 min while tumbling. QuantaBlu stop solution was then added to finalize the reaction. Readings were taken using a SpectraMax i3x plate reader at Ex/Em: 325/420 nm, with 10 readings per well on automatic settings.

The in silico structure of the USP5 was structure was generated as described previously (Garcia-Caballero et al., 2022) and was based on the known x-ray crystal structure (https://www.rcsb.org/structure/3IHP). A single-point mutation (arginine 24 to tryptophan) was introduced into the model, and interacting residues were analyzed using the mCSM-PPI2 computational tool (Rodrigues et al., 2019). Furthermore, the web server DynaMut was used to perform a visual analysis of deformation energies and atomic fluctuation in both proteins (Rodrigues et al., 2018). All structure figures were prepared in Discovery Studio Visualizer (Accelrys) and PyMol.

DRG (lumbar vertebra 3–6) and SC tissue (L3 to L6 level) were collected (ipsilateral and contralateral sides were separated) and stored in a −80°C freezer. Samples were homogenized in PBS with complete protease inhibitor cocktail (Sigma-Aldrich) and subjected to 2 freeze–thaw cycles to further break the cell membranes. Subsequently, the samples were centrifuged at 10,000 g for 10 min. The supernatant was quantified using the bicinchoninic acid assay. 10 μg of protein were diluted in 90 μl of PBS (1:10). Protein levels were measured using commercial kits (Cav3.2, MBS458074; MyBiosource; Cav2.1, MBS2887113; MyBiosource; Cav2.2, MBS455655; MyBiosource; USP5, ELI-39983m; LifeScienceMarket, and HCN1, ELI-39983m; LifeScienceMarket) following the manufacturer’s instructions.

Total RNA was extracted using the TRIzol reagent (15596026; Ambion life technologies). cDNA was synthesized from 1,000 ng of DNase-treated DRG RNA using a SuperScript IV VILO Master MIX (11756050; Invitrogen). PCRs were performed using KAPA2G Fast HotStart ReadyMix (KK5603; Roche). Using the following primers with overhang adapters to generate the amplicon:

forward primer: 5′-TCG​TCG​GCA​GCG​TCA​GAT​GTG​TAT​AAG​AGA​CAG​AGC​TGC​TGC​CGG​TGT​CAT​GG-3′, reverse primer: 5′-GTC​TCG​TGG​GCT​CGG​AGA​TGT​GTA​TAA​GAG​ACA​GGG​TGC​AGG​TAG​ACA​CGC​TGG-3′. PCR was cleaned up using Ampure XP beads (A63880; Beckman Coulter). The PCRs obtained were indexed, cleaned up, and sequenced using Illumina technology (RNA sequences are available at BioProject ID PRJNA1249093).

Animal experiments were approved by the Institutional Animal Care Committees of University of Calgary. All procedures were carried out in accordance with animal care regulations and policies of the Canadian Council on Animal Care and in agreement with the guidelines of the Committee for Research and Ethical Issues of the International Association for the Study of Pain.

All data were analyzed using Prism 8.4.3 (GraphPad Software). Comparisons between multiple sets of data with two variables were performed by ordinary two-way ANOVA, followed by Bonferroni’s multiple comparison test, with individual variances computed for each comparison. The P value was adjusted to count for multiple comparisons and family wise significance of 0.05 (95% confidence interval). Comparisons between two sets of data were performed using unpaired t test, two-tailed, with a statistical significance of P < 0.05. Three-way ANOVA, followed by Bonferroni’s multiple comparison test, was used to analyze the OIPN model, where we obtained multiple sets of data with three variables (genotype, treatment, and time of when animals were tested). The P value was adjusted to count for each comparison and family wise significance of 0.05 (95% confidence interval).

Supplementary data Figs. S1, S2, S3, S4, and S5 provide the additional data for clinical phenotype, validation of genotype, USP5 and Cav3.2 levels in SC and in tissues from female mice, and levels of other type of ion channels in WT and KI mice. Table S1 provides the mRNA transcript levels of Cav3.2 and USP5 in WT and mutant mice in naive and CFA conditions.

The data that support the findings of this study are available within the article and could be made available from the corresponding author upon reasonable request. The conducted research was not preregistered with an analysis plan in an independent, institutional registry.

We thank Mrs. Lina Chen for cell culture and genotyping and Dr. Magnhild Rasmussen for contributing clinical expertise. We thank Dr. Kenichi Ito at the Clara Christie Center for Mouse Genomics for help with generating the mutant mice. We thank the Centre Health for Genomics and Informatics at the University of Calgary for help with the Illumina sequencing and bioinformatic analysis. We thank the parents of the patient for their support in publishing this work.

G.W. Zamponi is supported by grants from the Canadian Institutes of Health Research and holds a Canada Research Chair. F.T.T. Antunes holds a fellowship from the Alberta Children’s Hospital Research Institute, and from the University of Calgary’s Eyes High program. M.Y. Ali is supported by a Mitacs Fellowship. E.K. Harding holds a Canadian Institutes of Health Research fellowship.

Author contributions: F.T.T. Antunes: data curation, formal analysis, investigation, methodology, validation, visualization, and writing—original draft, review, and editing. M.A. Gandini: conceptualization, data curation, formal analysis, investigation, project administration, and writing—original draft. A. Garcia-Caballero: conceptualization, formal analysis, investigation, methodology, validation, visualization, and writing—original draft. S. Huang: data curation, formal analysis, investigation, and visualization. M.Y. Ali: methodology and software. E. Gambeta: investigation and methodology. I.A. Souza: investigation and writing—review and editing. E.K. Harding: investigation and writing—review and editing. L. Ferron: formal analysis, methodology, and writing—review and editing. A. Stray-Pedersen: data curation, formal analysis, investigation, resources, and writing—original draft, review, and editing. V.M. Gadotti: conceptualization, data curation, formal analysis, supervision, and writing—review and editing. G.W. Zamponi: conceptualization, funding acquisition, project administration, supervision, and writing—original draft, review, and editing.