Potential sex differences in activation of pain-related brain regions in nonhuman primates with a unilateral spinal nerve ligation

Previous preclinical and clinical neuropathic pain studies have identified a number of brain regions that are activated with peripheral cutaneous stimulation utilizing in vivo neuroimaging (Wanigasekera et al., 2018; Da Silva and Seminowicz, 2019; Hama et al., 2021). Activated brain regions include the “lateral pain system,” which includes the insular cortex/secondary somatosensory cortex (Ins/SII), primary somatosensory cortex, and ventroposterior thalamus, and the “medial pain system,” which includes the (anterior) cingulate cortex (CC), medial thalamic nucleus and (anterior) Ins. The lateral pain system plays a role in mediating the sensory-discriminative aspects of pain whereas the medial pain system is associated with the affective-cognitive components of pain (Brooks and Tracey, 2005; Da Silva and Seminowicz, 2019). In rats, these brain regions show little or no responsiveness to innocuous stimuli before peripheral nerve injury (Da Silva and Seminowicz, 2019). However, after nerve injury, these brain regions showed an increased activation in previously innocuous stimuli. Thus, evoked brain activation in the neuropathic or sensitized the state could be utilized as a mechanism-based marker of neuropathic pain (Davis et al., 2020). The responsiveness of these activated brain regions to therapeutics, however, has yet to be extensively evaluated. While the brain regions activated in rodent models of neuropathic pain appear to parallel those activated in neuropathic pain patients, there are nonetheless key genetic, molecular, neuroanatomical, and neurophysiological differences between rodents and humans that could limit translatability of findings derived from rodents (Huggins et al., 2012; Dhopeshwarkar and Mackie, 2014; Chen and Hackos, 2015; Yezierski and Hansson, 2018; Abboud et al., 2021). By contrast, nonhuman primates are phylogenetically and neuroanatomically closer to humans compared with other preclinical nonhuman animal model species (Izpisua Belmonte et al., 2015). Reflecting differences in evolutionary pressures between rodents and primates, cortical development and volume of white matter significantly differ between these species and these differences have been suggested to underlie species differences in processing time and processing capacity, which in turn could lead to species-specific processing of pain and response to analgesics at the cortical level (Ventura-Antunes et al., 2013). One another potential barrier in translating preclinical findings to clinically useful treatments is the potential sexual dimorphism observed in humans with respect to pain processing. Sex differences in response to experimentally induced noxious stimuli, the prevalence of chronic pain states and possibly pain processing at the supraspinal level have been reported, but preclinical studies mainly utilize male nonhuman animals (Abboud et al., 2021; Kim et al., 2021). Responses of male and female rhesus macaques to acute noxious stimulation were similar (Negus et al., 2004). However, an examination of responses to somatosensory stimulation in chronic pain models has yet to be reported. Treatment with an analgesic dose of the gabapentinoid pregabalin in both neuropathic pain patients and a nonhuman primate model of neuropathic pain suppresses stimulus-evoked regional brain activation, suggesting a neurophysiological mechanism of pregabalin analgesia (Wanigasekera et al., 2018; Hama et al., 2021). Drugs that do not decrease brain activation in areas such as the Ins are also not analgesic, suggesting that modulating brain activation could be used to delineate analgesics from non-analgesics, further supporting brain activation as an objective mechanism-based marker of pain (Wanigasekera et al., 2018; Shidahara et al., 2019; Hama et al., 2021). While parallels in evoked brain activation and pharmacology have been observed between a macaque model of chronic constriction injury (CCI) of the sciatic nerve and patients with neuropathic pain, it is unclear if the findings from the CCI model generalize to clinical neuropathic pain states. A previous study describing nonhuman primate model of unilateral L7 spinal nerve ligation (SNL) demonstrated increased sensitivity of ipsilateral and contralateral feet to non-noxious cutaneous stimuli within days of SNL (Carlton et al., 1994). Terminal electrophysiological preparations in these macaques suggest the development of nerve injury-induced central sensitization, that is, enhanced responding of spinal dorsal horn neurons to previously innocuous stimuli (Carlton et al., 1994; Palecek et al., 2004). However, the effects of analgesics in awake SNL macaques, CNS neural activation, and long-term pain-related behaviors following SNL have yet to be reported. The present study characterized regional brain activation in nonhuman primates in response to a non-noxious stimulus, straight leg raise (SLR) following SNL over time. The macaque spinal nerve root ligation model in this study could reflect a sciatica-like pain state. The SLR is utilized to put the sciatic nerve under tension to evoke pain in patients with suspected sciatic nerve or nerve root compression (Summers et al., 2009). The effects of clinical analgesics on awake SNL macaques and brain activation under anesthesia were examined. Finally, to broaden the clinical applicability of the model and brain activation, the present study also examined the behavior and brain activation in male and female SNL macaques.

Eight Macaca fascicularis (four males and four females, age 2–3 years, body weight 2.4–3.5 kg; Eve Bioscience Co., Ltd., Wakayama, Japan) were utilized. The Hamamatsu Pharma Research, Inc. (Japan) Animal Care and Use Committee (HSTIRB-347) reviewed and approved the animal use protocol on August 3, 2021. Animal use protocol review and approval by the institutional animal care and use committee was based on (but not limited to): the rationale and purpose of the proposed use of experimental animals; a clear and concise sequential description of the procedures involving the use of experimental animals easily understood by all members of the committee; justification of the species and number of experimental animals proposed; description of the impact of the proposed procedures in the experimental animals’ well-being; appropriate sedation, analgesia and anesthesia; post-operative care and observation; rationale of the primary outcome measures; criteria and process for timely intervention, removal of experimental animals from the study, or euthanasia if extreme pain and distress are anticipated; method of euthanasia, if needed, and adequacy of training and experience of personnel in the procedures used and their roles and responsibilities. It should be noted that the drugs used to pharmacologically challenge pain behaviors in the current macaque model (i.e., duloxetine, pregabalin, and morphine; see the part of “Drugs” as below) are clinically used analgesics. The present study was performed in a dedicated nonhuman primate facility accredited by AAALAC International and care and housing were according to the Guide for the Care and Use of Laboratory Animals, 8 th ed (National Research Council, 2011). Housing rooms were under a 12-hour light/dark cycle and rooms were maintained at 20–26°C and 30–70% relative humidity. Individual macaques were housed in stainless steel cages, maintaining olfactory, auditory, and visual contact with neighboring macaques. Each macaque had free access to tap water and was given about 100 g/d standard non-human primate chow (Oriental Yeast Co., Ltd., Tokyo, Japan). In addition, macaques were given fresh fruits and vegetables at least once per week. For environmental enrichment, macaques were provided with manipulanda and received positive human interaction by the study staff and animal care staff. During the study, any animals showing signs of extreme stress, pain, or significant acute weight loss were to be euthanized and excluded. None of the animals were excluded in this study. Figure 1 illustrates the overall sequence of events for behavioral testing, surgery, brain imaging, and drug testing. Eight macaques underwent monkey chair restraint acclimation, habituation of restraint, and behavioral testing for 2 weeks, no more than 5 days per week and no more than 2 hours per day in monkey chair restraint. Macaques that did not readily habituate to chair restraint were removed and underwent habituation training the next day. The week before surgery (“Pre,” “Week -1”), baseline behavioral responses to a modified straight leg raise (SLR) were recorded. On week 0, three female and three male macaques underwent SNL surgery and one female and one male macaque underwent sham surgery. Overall study schedule. A total of six spinal nerve ligated (SNL) macaques were used, three males (1, 2, 3) and three females (A, B, C). During drug treatment cycles, awake straight leg raise was performed first. The following day, under anesthesia, brain activation, was performed with functional magnetic resonance imaging. By the end of the third treatment cycle, all six SNL macaques received all three treatments. All six SNL macaques received morphine on week 12. von Frey filaments testing was performed before and 2 weeks after SNL. Because of apparent spontaneous pain-related behavior immediately following SNL surgery, testing with von Frey filaments was performed 2 weeks after surgery. The SLR test appeared qualitatively painful 2 weeks after SNL. Thus, SLR testing and MRI were conducted beginning 4 weeks after surgery. On week 4, the modified SLR test was performed, with both awake and anesthetized macaques. Spinal nerve-ligated macaques were dosed per os (p.o.) and then the SLR test was performed in awake SNL macaques. One day later, SNL macaques were dosed with the same drug, and then under propofol anesthesia (13–20 mg/kg/h, Maruishi Pharmaceutical, Co., Osaka, Japan), and underwent functional magnetic resonance imaging (fMRI). On weeks 6 and 8, SNL macaques received a different treatment from the previous treatment, and then underwent testing and fMRI; at the end of week 8, all six SNL macaques received each of the three treatments. Sham-operated macaques underwent SLR testing and fMRI on weeks 4 and 8. On week 12, all SNL macaques were treated intramuscularly (i.m.) with morphine (Shionogi Pharma Co., Osaka, Japan), and then the SLR test was performed. The next day, all six SNL macaques were dosed with morphine again and then underwent fMRI. Following the final MRI, SNL macaques were euthanized. Macaques with sham surgery were returned to the colony. In this study, study staff were blinded to treatment during behavioral assessment and staff performing fMRI data analysis was also blinded to treatment. The lack of robust pain-related behavior and brain activation in the SNL macaques during von Frey filament stimulation led to the consideration of the SLR as the primary stimulus to evoke pain in awake macaques and to evoke activation of pain-related brain regions during fMRI. Inflammation and compression injury of lumbar spinal nerve roots, for example, via spinal stenosis, has been reported to lead to sharp and burning pain extending from the back to the lower extremities (Soar et al., 2022). Radicular pain on lumbar flexion/extension is reported by clinical patients (Hama et al., 2021); the SLR is used as a diagnostic tool to distinguish lumbar radicular pain from other peripheral neuropathies and low back pain. To evoke pain following SNL, a modified seated SLR test was developed based on the sitting SLR (Summers et al., 2009). The maximum SLR threshold for each macaque, established before the study, was not exceeded. Macaques were restrained and seated upright in a monkey chair with the lower legs resting over the edge of the seat ( Figure 2 ). The lower leg at rest was at 0° before testing. While grasping the ankle, such that there was no rotation at the ankle, and restraining the knee, the lower leg was extended at the knee. With the lower leg fully extended, the “leg angle” was 90°. At 90°, the entire leg, with the lower leg fully extended, was lifted until 160° relative to the lower leg at rest before testing (0°) ( Figure 2 ). Tightness of the facial muscles or resistance to further leg lifting signaled the end of leg lifting and the angle was recorded. The angle between the lower leg and 0° was measured with goniometers (Takumed, Sankei Kanko Co., Ltd., Osaka, Japan). Angles were measured to the nearest 10°. The maximum leg angle was 160° and testing stopped at 160°. A total of two leg angle measurements were taken for each leg, about 2 minutes apart. The unligated leg was first measured followed by the ligated leg. The averaged leg angle was converted to a 6-point “angle score”, at 20° intervals, with “1” being the lowest score (0–20°) and “6” being the highest score (150° to 160°). The modified straight leg raise test in awake macaques. For awake behavioral assessment, macaques were habituated to restraint to a monkey chair prior to measurement. With the macaque seated, the lower legs were over the edge of the seat and perpendicular to the floor (0°). At the start of the test, the ankle was grasped, such that the toes pointed perpendicular to the lower limb, and the knee held but allowed to rotate. The leg was slowly raised to a maximum of 160°, relative to the position of the lower leg before testing. A facial grimace or movement of the legs stopped the test and the angle of the leg was recorded. Two angle measurements were taken for each leg. The averaged leg angle was converted to a 6-point “angle score,” from 1 (0 to 20°) to 6 (150° to 160°). The tactile sensitivity of both feet was assessed with von Frey filaments (Hama et al., 2021) as described in Additional file 1 before and 2 weeks after SNL. Awake macaques were restrained in a monkey chair and a series of increasing force of von Frey filaments was applied to the center of the plantar foot, beginning with the lowest force filament. Each filament, in ascending order, was applied ten times to the plantar foot and the number of responses out of ten applications was recorded. Ligation of the L7 spinal nerve was performed as previously described (Carlton et al., 1994). The left L7 spinal nerve was ligated in three male macaques and three female macaques. One male and one female macaque received a sham surgery—exposure of the left L7 spinal nerve without ligation. Macaques were first anesthetized with an intramuscular injection of ketamine (20 mg/kg; Daiichi Sankyo, Tokyo, Japan) and then with continuous intravenous infusion of propofol (10 mg/kg/h; Maruishi Pharmaceutical Co.). The level of anesthesia was monitored during the surgery and additional ketamine was given when needed. After the skin of the lower back was shaved, the location of the left L7 spinal nerve was located using a C-arm fluoroscope (Siremobil Compact L, Siemens, Tokyo, Japan). The incision site to access the L7 nerve was marked with a permanent ink marker. The skin was cleaned with povidone-iodine. Using an aseptic technique, a 3 cm skin and subcutaneous tissue incision was made and the L7 nerve was exposed near the entrance to the vertebral foramen. The L7 spinal nerve was ligated tightly in two sites next to each other, ~2 mm apart, ~5 mm distal from the intervertebral foramen, using 2-0 silk thread (Matsuda Medical Industry Co., Ltd., Tokyo, Japan). The muscle layer was closed using 3-0 VSORB (Kono Seisakusho co., Ltd., Tokyo, Japan) and the skin was closed using 3-0 Nescosuture (Alfresa Pharma Corp., Osaka, Japan). Heating pads were used to maintain body temperature while macaques were under anesthesia. Respiratory rate, body temperature, and heart rate were monitored. Before awakening, macaques were given buprenorphine (0.03 mg/kg, i.m.; Otsuka Pharmaceutical Co., Tokyo, Japan) and then twice daily for at least 3 days. For postoperative infection control, enrofloxacin (5 mg/kg, i.m.; Bayer, Tokyo, Japan) was administered twice daily for at least 3 days. Following surgery, macaques were monitored in their home cages for recovery from anesthesia. After confirmation of wound healing, skin sutures were removed 2 weeks after surgery. Clinically used analgesics were assessed on SLR in macaques 4, 6, 8, and 12 weeks after SNL. On the day of drug testing, SLR tests were performed before and after p.o. administration of vehicle, serotonin-norepinephrine reuptake inhibitor duloxetine, or gabapentinoid pregabalin in a volume of 3 mL/kg. Approximately 1 hour after treatment, the SLR test was performed on SNL macaques. A cross-over design was used—by the end of the final dose cycle on week 8, each macaque received all three doses. At 12 weeks after SNL, morphine was given i.m. in a volume of 0.6 mL/kg in all SNL macaques. Awake SNL macaques were tested for ~10 minutes after morphine treament. One day after SLR testing, SNL macaques were administered the same drug, and then regional brain activation was examined with fMRI. Brain imaging was initiated about 1 hour after p.o. treatment and about 30 minutes after i.m. morphine administration. All drugs were prepared on the day of administration. The dose of pregabalin (Kemprotec, Ltd., Cumbria, UK) and the dose of duloxetine HCl (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) were 30 mg/kg each. Pregabalin and duloxetine were prepared in distilled water (Otsuka Pharmaceutical Co., Ltd., Tokushima, Japan). The opioid morphine HCl (6 mg/kg; 10 mg/mL, Shionogi Pharma Co., Ltd., Osaka, Japan) was used undilutedly. The dose of morphine used has been shown to ameliorate postoperative pain in awake macaques and suppress evoked activation of the Ins/SII and CC (Hama et al., 2018). In humans, peak morphine plasma concentration has been observed between 0.5 and 2 hours following i.m. administration (Laitinen et al., 1975). The dose of duloxetine has been shown to ameliorate non-noxious cold hypersensitivity and suppress non-noxious cold activation of the Ins/SII in oxaliplatin-treated macaques (Shidahara et al., 2019). The dose of pregabalin used has been shown to reduce stimulus-evoked activation of the Ins/SII, thalamus, and CC in macaques with a unilateral sciatic nerve CCI (Hama et al., 2021). In naïve, uninjured macaques, 1 hour following p.o. administration, the concentration of pregabalin in CSF is (mean ± SD) 0.3 ± 0.2 µg/mL ( n = 3, unpublished data). Significantly decreased spontaneous pain has been reported to be 1.5 hours after one oral dose of pregabalin (Jensen-Dahm et al., 2011). Duloxetine efficacy in painful diabetic neuropathy has been observed as early as the first day of treatment (Pritchett et al., 2007). Stimulus-evoked regional brain activation was visualized using a 1.5 T MRI (Signa Explorer, GE Healthcare, Milwaukee, WI, US). Macaques were sedated by continuous intravenous infusion of propofol (13–20 mg/kg/h; Maruishi Pharmaceutical Co.). Anesthesia was used to minimize movement during scanning and the dose of propofol used has little, if any, antinociceptive effect (Lichtner et al., 2018; Shirai et al., 2020). During the scans, animals were kept warm with heating pads and blankets. The anatomical MRI protocol consisted of a 3D-T1 Cube sequence (repetition time (TR)/echo time (TE), 55.1/8.9 ms; number of averages, 4; flip angle, 90°; field of view, 10 8 mm × 10 8 mm; matrix, 256 × 224; slice thickness/interval, 1.0/0.5 mm; number of slices, 168). Functional scan sequences consisted of field-echo, echo-planar imaging (TR/TE, 3000/45 ms; flip angle, 90°; field of view, 140 mm × 140 mm; matrix, 64 × 64; slice thickness, 2.0 mm; number of slices, 25). During one fMRI scan, macaques were in the supine position during block design stimulation ( Figure 3 ). One stimulus “set” was 30 seconds of an “OFF” stimulus followed by 30 seconds of an “ON” stimulus, and then 30 seconds of a “rest” interval (no stimulation). The upper leg was perpendicular to the MRI patient table and supported in place by a cushion. During the “OFF” stimulus, the lower leg was left in place resting on the cushion (corresponding to 0° in the awake seated state in Figure 2 ). During the “ON” stimulus, the lower leg was extended at the knee within 2–3 seconds until perpendicular to the MRI table (corresponding to 90° in the awake seated state in Figure 2 ). The foot was kept in a flexed position by grasping the ankle. The lower leg was held in the extended position for 30 seconds. For each “set,” 10 frames were acquired, for 50 frames per functional scan. Five stimulation sets were performed on the right unligated leg followed by the left ligated leg. One fMRI scan was ~2 hours in duration. Functional magnetic resonance imaging: SLR block design stimulation. Anesthetized macaques were supine on the magnetic resonance imaging (MRI) patient table. The thigh was supported with cushions such that the thigh was perpendicular to the MRI table. In the “OFF” position, the lower leg was flexed at the knee, such that the lower leg was parallel to the MRI table (this position would correspond to an angle of 0o, or an angle score of “0”, in awake macaques. See Figure 2 ). In the “ON” position, the lower leg was extended, within 2-3 seconds, such that the lower leg was perpendicular to the MRI table (this position would correspond to an angle of 90°, or score of “4”, in awake macaques. See Figure 2 ). Under anesthesia, during each session, macaques underwent a total of five sets of SLR stimuli. One SLR set consisted of a 30-second “OFF” period [lower leg at rest, parallel to the MRI table]), a 30-second “ON” period (lower leg raised to perpendicular to the MRI table and left in this position for 30 seconds) and a 30-second no stimulation (lower leg at rest). The SLR was performed first on the right unligated leg and then on the left ligated leg. SLR: Straight leg raise. All subsequent image analyses were conducted using SPM12 software (Wellcome Trust Centre for Neuroimaging, London, UK). The images were realigned and resliced onto the mean echo-planar imaging (EPI) image to correct head motion. The EPI images were co-registered to the corresponding T1-weighted anatomical image and normalized to a macaque brain template (Black et al., 2004). Stereotaxic coordinates were confirmed according to Horsley-Clarke’s stereotaxic coordinates (Horsley and Clarke, 1908). The resulting image was smoothed with a 4 mm × 4 mm × 4 mm full width at half-maximum Gaussian kernel. Voxel-wise statistical analysis with a threshold z-value greater than 1.96, at P < 0.05, was based on a general linear model. A fixed effects model was used for group analysis of data from three male and three female SNL macaques. Contrast (subtraction) maps were defined to isolate regions responsive to stimulation-related signals of the whole brain. Group mean contrasts were calculated between OFF stimulation and ON stimulation, as the OFF stimulus did not evoke significant activation (z-value less than 1.96). For drug treatment, group mean contrast maps were defined as (vehicle > post-drug treatment) to determine decreases in activation following drug treatment. Peak voxels were considered significant at a z-value greater than 1.96 ( P < 0.05, uncorrected for multiple comparisons, one-tailed t -test). One consideration of the study design was the availability of macaques and the use of the minimal number of macaques as possible with respect to the principles of the 3Rs (Replacement, Reduction and Refinement). A total of eight macaques (four females and four males) were available. Previous pharmacological and brain imaging studies with nonhuman primates have utilized 3–4 subjects and a cross-over design schedule such that each subject received all or most of the treatments, thereby minimizing the overall number of subjects needed for the study (Ogawa et al., 2016; Hama et al., 2021). Sex differences in human pain perception have been reported, but results from a rhesus macaque study suggested the minimal sex-related difference in pain perception (Negus et al., 2004); both composite statistics and statistics by sex ( n = 3 per sex) are reported in this study. The effects of drug treatment, before and after administration, on SLR scores were analyzed using Wilcoxon signed rank test. Comparisons between treatment groups were performed with Kruskal-Wallis test by ranks. The Friedman test was used to analyze SLR scores over time following SNL. Post hoc analysis, a comparison of SLR scores following SNL with pre-SNL SLR scores, was performed with Dunnett test for multiple comparisons (with Bonferroni correction for multiple comparisons). Statistical analyses were performed using SAS Analytical Pro version 9.4 (SAS Institute Japan, Tokyo, Japan) and EXSUS version 10.1 (EP Croit Corp., Tokyo, Japan). P values less than 0.05 were considered statistically significant. Unless otherwise indicated, data are expressed as the mean ± SD. All macaques were used in the study without dropout.

Female and male macaques gained weight normally and surgical incisions healed without complications following surgery ( Additional Table 1 ). Possible signs of pain-related behavior of the ligated leg were observed within days following SNL surgery, peaking at ~7 days after surgery and then diminishing and abating at ~13 days after surgery. Some of the non-quantified signs of pain-related behavior observed in macaques in their home cages included an apparent lack of ipsilateral weight bearing, weak grip strength of the ipsilateral toes, partial flexion of the ipsilateral knee, suggestive of reluctance to fully extend the lower leg, and an apparent dislike of brushing or stroking of any part of the ligated leg. None of these signs of pain was observed on the unligated leg or in the sham-operated macaques. By 2 weeks after SNL surgery, these signs were generally absent. Body weight before and after spinal nerve ligation At 2 weeks after surgery, neither SNL males nor SNL females demonstrated increased response to a range of von Frey filaments applied to the ligated or unligated foot ( Additional Table 2 ). Without robust changes in von Frey filament thresholds of the ligated side compared with the pre-ligation baseline, no further von Frey filament testing was performed. Foot responses to von Frey filaments before and 2 weeks after spinal nerve ligation Individual macaque responses to von Frey filaments and the mean (± SD) responses of male and female SNL macaques. Values from K-1815 and K-2137, sham-operated macaques, are shown for comparison. The lowest force von Frey filament (1 g) was applied to the center of the plantar foot 10 times and the number of responses was recorded. Filaments were then applied to the plantar foot in ascending order of force. Before SNL, both male and female macaques showed few responses to von Frey filament probing of either the left or right foot. When examined by sex, two weeks after SNL, no statistically significant increase in responding to von Frey filaments, on either the ligated or unligated foot, was observed (female: P = 0.6617; male: P = 0.8110). When the responses of both sexes were combined, there was no statistically significant increase in responding compared to before surgery ( P = 0.3857). Before SNL or sham surgery, all macaques of both sexes allowed their legs to be extended to the maximum 160° or a score of “6” ( Figure 4 ). Four weeks after SNL, significantly decreased mean SLR scores of the ligated leg were shown in female and male macaques and significantly decreased mean straight leg raise scores were observed at 6, 8 and 12 weeks after SNL. Over time, mean SLR scores in both male and female macaques tended to increase towards pre-SNL. When SLR scores in both male and female macaques were combined (total n = 6), overall mean SLR scores of the ligated leg following SNL were significantly decreased compared with pre-SNL over time ( P < 0.0005; data not shown). Combined mean SLR scores were significantly decreased at each time point, from week 4 through week 12 ( P = 0.0000), compared with pre-SNL. SLR scores before and after SNL or sham surgery. Each macaque was tested twice on each leg which was averaged. Angle scores of left and right legs are reported separately. Data are expressed as the mean ± SD. * P < 0.05, ** P < 0.01, *** P < 0.0001, vs . pre-SNL (“Pre”). No significant changes in angle scores were observed in the right (unligated) leg. SNL n = 3 males and 3 females. Sham n = 1 male and 1 female. SLR: Straight leg raise; SNL: spinal nerve ligation. When analyzed separately by sex, overall mean straight leg raise scores of males ( P = 0.0364) and females ( P = 0.0272) were significantly decreased compared with pre-SNL. For males, mean SLR scores were significantly decreased on weeks 4 ( P = 0.0066) and 6 ( P = 0.0266) compared with pre-SNL, and trends of decreased SLR scores were observed on weeks 8 ( P = 0.0789) and 12 ( P = 0.0548). For females, mean SLR scores were significantly decreased from week 4 through week 12 ( P = 0.0001). Mean SLR scores of the right unligated leg did not significantly change over time in female and male macaques compared with pre-ligation ( P > 0.05; Figure 4 ). Following sham surgery, neither female nor male macaques showed changes in SLR scores of the sham-ligated or right unligated leg over time ( Figure 4 ). Comparisons of SLR before and after each treatment were performed (i.e., within-group comparisons) combining all macaques ( n = 6). When female and male SLR scores of the ligated side were combined, no statistically significant differences were observed before and after treatment with vehicle ( P = 1.000), pregabalin ( P = 0.1250), or duloxetine ( P = 0.2500; Figure 5 and Additional Figure 1 ). By contrast, a statistically significant increase in SLR scores was observed after morphine treatment compared with before morphine treatment ( P < 0.0313; Figure 5 ). SLR of the left ligated leg of female and male macaques with spinal nerve ligation before and after drug administration. On weeks 4, 6, and 8, SLR testing was performed before and ~1 hour after treatment with vehicle, pregabalin, or duloxetine. On week 12, macaques were tested ~10 minutes after treatment with morphine. Morphine treatment increased left leg SLR scores compared with pre-treatment (* P < 0.05). Individual SLR scores, from both male and female macaques, are shown. The horizontal line is the mean. SLR: Straight leg raise. When SLR scores in male and female macaques were combined ( n = 6), a comparison of SLR after vehicle, pregabalin, duloxetine, and morphine treatment was performed (i.e., inter-group comparison). Compared with vehicle treatment, there were no statistically significant differences in SLR scores following pregabalin, duloxetine, or morphine treatment ( P = 0.0618; data not shown). Within-group analyses (i.e., before and after treatments) by sex were performed. Treatment in SNL male macaques ( n = 3) with either vehicle ( P = 1.000), pregabalin ( P = 0.5000), duloxetine ( P = 0.5000), or morphine ( P = 0.2500) did not significantly increase mean SLR scores ( Additional Figure 1 ). Treatment in SNL females ( n = 3) with either vehicle ( P = 1.0000), pregabalin ( P = 0.5000), duloxetine ( P = 0.5000), or morphine ( P = 0.2500) did not significantly increase mean SLR scores ( Additional Figure 1 ). No effect of drug treatment was observed on the contralateral leg; SLR scores were approximately “6” before treatment ( Figure 4 ). In male macaques with SNL, significant mean brain activation in the contralateral insular cortex/secondary somatosensory cortex (Ins/SII) and contralateral thalamus (Th) was observed during ligated SLR after vehicle treatment ( Table 1 and Figure 6 ). In male SNL macaques, there was no significant activation of the CC during SLR. Effect of treatment on left (ligated) straight leg raise-evoked brain activation in male macaques with SNL Group mean peak voxel z values and coordinates following modified left straight leg raise in male spinal nerve ligation (SNL) macaques. Brain imaging was performed 4, 6, and 8 weeks after SNL. Male macaques ( n = 3) were administered per os vehicle, pregabalin (30 mg/kg), or duloxetine (30 mg/kg) 1 hour prior to brain imaging. Morphine was tested 12 weeks after SNL. Macaques ( n = 3) were dosed intramuscularly with morphine (6 mg/kg) 30 minutes prior to brain imaging. Stereotaxic coordinates ( x , y , z ) are confirmed according to Horsley-Clarke’s stereotaxic coordinates. z -value > 1.96, P < 0.05. Voxels, number of voxels exceeding threshold z -value 1.96. Transaxial view of SLR evoked-regional brain activation in SNL male macaques 4, 6, and 8 weeks after SNL. The mean activation from three male macaques is shown. Series is ventral to dorsal, from top left to bottom right. During SLR of the left ligated leg, robust activation of the contralateral insula cortex/secondary somatosensory cortex and contralateral thalamus was observed. R: macaque’s right. See Table 1 for z-values. SLR: Straight leg raise; SNL: spinal nerve ligation. In vehicle-treated female SNL macaques, mean activation of the cingulate cortex (CC) and contralateral Ins/SII was observed during ligated SLR ( Table 2 and Figure 7A ). During SLR of the ligated leg, there was no significant activation of the contralateral thalamus ( Figure 7B ). Effect of treatment on left (ligated) straight leg raise-evoked brain activation in female macaques with SNL Group mean peak voxel z values and coordinates following left spinal nerve ligation in female spinal nerve ligation (SNL) macaques. Brain imaging was performed 4, 6, and 8 weeks after SNL. Female macaques ( n = 3) were administered per os vehicle, pregabalin (30 mg/kg), or duloxetine (30 mg/kg) 1 hour prior to brain imaging. Morphine was tested 12 weeks after SNL. Macaques were dosed intramuscularly with morphine (6 mg/kg) 30 minutes prior to brain imaging. Stereotaxic coordinates ( x , y , z ) are according to Horsley-Clarke’s stereotaxic coordinates. z -value > 1.96, P < 0.05. Voxels, number of voxels exceeding threshold z -value 1.96. Transaxial view of SLR evoked-regional brain activation in SNL female macaques 4, 6, and 8 weeks after SNL. (A) The mean activation from three female macaques is shown. Series is ventral to dorsal, from top left to bottom right. During SLR lifting of the left ligated leg, robust activation of the cingulate cortex (CC) and contralateral insula cortex/secondary somatosensory cortex was observed. R: macaque’s right. See Table 2 for z-values. (B) Coronal sections of mean activation of male and female brains during straight leg lift of the left ligated leg following vehicle treatment. Note the differential regional activation between male and female macaques. SLR: Straight leg raise; SNL: spinal nerve ligation. In both SNL female and male macaques, SLR of the right unligated leg did not evoke significant brain activation ( Table 3 ). Effect of right (unligated) straight leg raise on brain activation in macaques with SNL Mean peak voxel z values and coordinates following right, unligated straight leg raise in male and female spinal nerve ligation (SNL) macaques. Brain imaging was performed 4, 6, and 8 weeks after SNL. Macaques (males, n = 3; females, n = 3) were administered per os vehicle 1 hour prior to brain imaging. Note the lack of significant z -value voxels with lifting of the unligated leg ( z value < 1.96). Stereotaxic coordinates ( x , y , z ) are confirmed according to Horsley-Clarke’s stereotaxic coordinates. Vehicle-treated male SNL macaques showed activation of the contralateral Ins/SII and thalamus during SLR ( Table 1 and Figure 8A ). Pregabalin and duloxetine reduced mean contralateral thalamic activation, but not contralateral Ins/SII activation ( Table 1 ). Morphine treatment reduced SLR-evoked activation of the Ins/SII and thalamus ( Table 1 ). Contrast maps, comparison of activation during either pregabalin or duloxetine treatment with that of vehicle treatment, showed no change in evoked regional activation ( Table 4 ). Compared with vehicle, morphine treatment significantly reduced activation of contralateral Ins/SII and thalamus ( P < 0.05). Coronal sections of mean brain activation during straight leg lift of the ligated leg 4, 6, and 8 weeks after spinal nerve ligation. Male (A) and female (B) imaging began about 1 hour after treatment with vehicle, pregabalin, or duloxetine and ~30 minutes after intramuscular morphine administration. CC: Cingulate cortex; Ins/SII: insular cortex/secondary somatosensory cortex; R: right; Th: thalamus. See Table 4 for z-values. Contrast maps of brain activation during drug treatment and left SLR in male and female macaques with SNL Contrast maps (vehicle > drug) of mean peak voxels during treatment and left SLR. Spinal nerve ligated macaques were treated p.o. with either vehicle, pregabalin or duloxetine 1 hour before brain imaging. For morphine, macaques were treated intramuscularly 30 minutes prior to brain imaging. In SNL male macaques ( n = 3), activations following either pregabalin or duloxetine treatment were not different compared to vehicle treatment. By contrast, compared to vehicle treatment, morphine significantly reduced activation of both contralateral Ins/SII and thalamus. In SNL female macaques ( n = 3), compared to vehicle treatment, both pregabalin and duloxetine reduced CC activation but not contralateral Ins/SII activation. Compared to vehicle treatment, morphine treatment reduced both CC and contralateral Ins/SII activation in SNL female macaques. Stereotaxic coordinates ( x , y , z ) are according to Horsley-Clarke’s stereotaxic coordinates. z value > 1.96, P < 0.05. Voxels, number of voxels exceeding threshold z value 1.96. Vehicle-treated female SNL macaques showed activation of the CC and contralateral Ins/SII ( Table 2 and Figure 8B ). Both pregabalin and duloxetine treatment reduced mean activation of the CC but not the contralateral Ins/SII ( Table 2 ). Contrast maps showed significantly decreased CC activation during pregabalin and duloxetine treatment ( Table 4 ). Morphine treatment, compared with vehicle treatment, significantly reduced activation of contralateral Ins/SII and CC ( P < 0.05).

In both male and female SNL macaques, significantly decreased SLR scores on the ligated leg, suggestive of radicular-like pain, were observed, which persisted for at least 12 weeks following SNL. By contrast, SLR scores of the contralateral unligated leg and ipsilateral leg of sham-operated macaques were not decreased. Neither pregabalin nor duloxetine significantly increased SLR scores in SNL macaques. However, morphine, when male and female SLR scores were grouped together, significantly increased SLR scores. In both male and female macaques, SLR of the ligated leg activated the contralateral Ins/SII. In male SNL macaques, prominent activation was also observed in the contralateral thalamus. Pregabalin and duloxetine treatment in male SNL macaques reduced mean activation of the contralateral thalamus. In female SNL macaques, prominent activation was also observed in the CC. In female macaques, pregabalin and duloxetine reduced mean CC activation. Morphine reduced all regional brain activation in both male and female macaques. While suppression of activation in sex-specific brain regions, but not Ins/SII, was observed following either pregabalin or duloxetine treatment, this did not parallel antinociception, suggesting that Ins/SII is a key region mediating pain following SNL. Brain activation could be utilized as a marker of pain and suppression of activation of key brain regions could be utilized to guide the development of effective analgesics. A previous study described male cynomolgus macaque model of neuropathic pain noted robust hypersensitivity evoked with von Frey filaments and brushing of the ipsilateral plantar foot within a few days of SNL; in some macaques, hypersensitivity was observed in both ipsilateral and contralateral feet (Carlton et al., 1994). Carlton et al. (1994) also reported pain-related behaviors appearing within days of SNL (e.g., reluctance to place the ligated foot on the chair crossbar), which were also observed in the present study. It is not known if bilateral hypersensitivity to foot stimulation and symptoms persisted beyond the 14-day observation period reported by Carlton et al. (1994). The present study did not observe a statistically significant increased response to plantar foot stimulation with von Frey filaments following unilateral SNL. The filaments used in the current study ranged from 1 g to 26 g, whereas Carlton et al. used filaments in the range from 0.068 g to 15 g. Carlton et al. reported qualitative responses to the filaments rather than a raw number of responses as performed in the present study. Perhaps incorporating qualitative responses (e.g., mild, moderate, and robust; Carlton et al., 1994) would have revealed “responsiveness” to von Frey filament stimulation in this study. Significantly decreased ipsilateral SLR scores persisted for at least 12 weeks after SNL, with female SNL macaques tending to display lower SLR scores compared with that of males. Elsewhere, male rhesus macaques tended to be more sensitive to noxious heat, as demonstrated by hot water tail immersion, than female macaques, even after accounting for sex differences in stress and menstrual cycle (Negus et al., 2004). Based on their findings, Negus et al. (2004) suggested that human sex differences in pain perception are mainly due to “sociocultural factors” rather than innate factors. However, a “mild” genetic component has been reported to mediate sex differences in chronic pain conditions such as chronic back pain and experimental acute pain perception (Rosen et al., 2017; Freidin et al., 2021). Possible neural substrates underlying sex differences in pain perception include distinctive regional brain activation and differences in functional connectivity between brain regions, which have been reported in human experimental pain and clinical pain studies (Gupta et al., 2017; Kim et al., 2021). Sex differences in neural substrates of pain perception have also been suggested to be the basis of differential responses to analgesics (Gupta et al., 2017). Given sex differences in brain activation, SNL macaques could be utilized as a preclinical model to further elaborate neural mechanisms underlying sex differences in pain perception and also to examine potential sex differences in response to analgesics. Brain regions associated with the medial pain system activated by SLR in SNL macaques do not generally appear to be sex-specific. While CC activation was observed in SNL female macaques in the present study, CC activation has also been observed in some male macaque pain models, such as during the application of non-noxious pressure near a midline abdominal skin and muscle incision and during plantar stimulation with non-noxious von Frey filaments ipsilateral to a CCI (Hama et al., 2018, 2021). By contrast, in the oxaliplatin-induced model of cold hypersensitivity, no CC activation was observed in male macaques (Nagasaka et al., 2017). Likewise, evoked activation of components of the lateral pain system, such as the thalamus, does not generally appear to be sex-specific. In macaques with naturally occurring endometriosis, non-noxious pressure applied to the lower abdomen evoked bilateral thalamic activation (Yano et al., 2019). While the thalamus is a key relay center which receives nociceptive information from the spinal level and transmits nociceptive information to cortical regions involved in sensory discrimination, cognition and affective disorder, stimulus-evoked activation appears to be limited to macaque models of chronic pain rather than acute pain (Millan, 1999). In chronic pain models, it is possible that thalamic neurons, presynaptic neurons that synapse with thalamic neurons or both are sensitized due to a long-standing peripheral injury and are activated with non-noxious stimuli; the timing of a potential sensitization has yet to be fully elaborated using neuroimaging in macaques. However, sex differences in regional brain activation were identified in the current neuropathic pain model. Interestingly, discrete brain regions were sensitive to analgesics but were not associated with antinociception. In SNL macaques, no significant antinociceptive effect was observed following either duloxetine or pregabalin treatment. In males, decreased mean thalamic activation was observed following treatment with either duloxetine or pregabalin, compared with vehicle treatment. However, the decrease in thalamic activation was not significant compared with that of vehicle treatment according to contrast analysis. Persistent activation of both the contralateral thalamus and Ins/SII could underlie the lack of significant antinociception in SNL males despite duloxetine and pregabalin treatment. Likewise, in female macaques during either pregabalin or duloxetine treatment, CC activation was decreased but not contralateral Ins/SII. In other macaque pain models, the reduction of CC activation alone was not sufficient for antinociception (Hama et al., 2018). Previous studies in macaque pain models, demonstrated robust antinociception with suppression of Ins/SII activation, indicating the importance of targeting the Ins/SII for analgesia (Hama et al., 2018, 2021; Shidahara et al., 2019). To confirm the importance of the Ins/SII in males, perhaps microinjection into the Ins/SII, while the thalamus remains activated, could be antinociceptive (Nagasaka et al., 2017; Lee et al., 2022). In SNL macaques, perhaps inactivation of all activated brain regions to the extent that morphine decreases brain activation is necessary for significant antinociception. It is possible that an acute systemic dose of either duloxetine or pregabalin was not sufficient to reduce Ins/SII activation and pain and repeated administration is required. However, as mentioned earlier (see the Methods section), acute, meaningful neuropathic pain relief has been reported and significant efficacy has been observed across numerous rodent models of neuropathic pain following a single treatment of either drug (Hama and Borsook, 2005; Le Cudennec and Castagne, 2014). The gabapentinoids are ligands that bind to the alpha-2-delta subunit of voltage-gated calcium channels (Taylor and Harris, 2020). The varying efficacy of pregabalin, in terms of antinociception and suppression of brain activation, across nonhuman primate pain models suggests that the expression of alpha-2-delta is disease- or model specific. In fact, lumbar radicular pain is poorly managed with gabapentinoids, which implies low level or no expression of alpha-2-delta in key areas modulating pain perception such as the spinal cord dorsal horn (Gimenez-Campos et al., 2022). Changes in alpha-2-delta expression in both humans with neuropathic pain and in the current macaque SNL model acutely and over time have yet to be reported. Duloxetine enhances the efficacy of descending inhibitory pathways and increases synaptic inhibition by blocking norepinephrine and serotonin reuptake (Pritchett et al., 2007). Blocking norepinephrine and serotonin reuptake is also believed to alleviate affective disorders, but the analgesic effect of duloxetine is independent of changes in affect (Pritchett et al., 2007). Antinociception with decreased Ins/SII activation following a single dose of duloxetine has been described in oxaliplatin-treated macaques (Shidahara et al., 2019). The antinociceptive effect of duloxetine in the oxaliplatin-treated macaques paralleled a randomized, cross-over design clinical study of patients with oxaliplatin-induced peripheral neuropathy (Smith et al., 2013; Shidahara et al., 2019). In a patient-based survey, duloxetine was not very effective against radicular pain, particularly against dysesthesia/paresthesia (Nakajima et al., 2022). While increasing synaptic norepinephrine and serotonin in both patients and macaques with oxaliplatin-induced neuropathic pain is sufficient for pain relief, this does not appear to be sufficient to alleviate pain in radicular pain patients and SNL macaques. A potential source of sex differences in pain perception in humans could be regional differences in brain activation and functioning (Gupta et al., 2017; Lee et al., 2022). Differences observed between males and females, generalized across chronic pain states, include, for example, greater activation of the Ins in males compared with females (Gupta et al., 2017). In the present study, regional differences were observed—activation of the CC in SNL females and contralateral thalamus in SNL males. The activation of the CC in females suggests more “unpleasant” pain and could underlie the tendency of lower SLR scores compared with that of males (Brooks and Tracey, 2005). Perhaps activating the thalamus in males leads to more intense, localized pain. Further clinical data—of sex differences in pain perception to a particular chronic pain state—are needed to confirm the utility of the current brain imaging findings. Gupta et al. (2017) reported the tendency in neuroimaging studies to mix or collapse male and female subjects into one analysis, thereby obscuring any potential sex-related differences in regional activation and functioning. With respect to functioning, a sex difference in functional connectivity, that is, the interaction (or correlation) between brain regions associated with pain processing, and differences in brain “resting state” have also been described (Gupta et al., 2017). The present study did not explore functional connectivity as propofol anesthesia tends to disrupt functional connectivity (Lichtner et al., 2018). Thus, studies in awake subjects will be needed to examine a network (a “matrix”) beyond the examination of discrete pain processing regions. Somatosensory stimuli that tend to activate certain regions over others could also facilitate defining potential sex differences in pain perception. What is also lacking in both the preclinical and clinical literature is studies regarding sex differences in response to a given therapeutic. Preclinical macaque models, at defined stages of pain development, tested with the same stimulus, undergoing the same treatment, could be utilized in uncovering potential mechanisms mediating sex differences in pain processing and uncovering sex differences in treatment responses. There could be intercellular processes that mediate the observed sex differences as well as general regional brain activation which could be explored via invasive tissue sampling techniques such as transcriptomics (Xu et al., 2022). In conclusion, our findings in a nonhuman primate model of neuropathic pain suggest a differential activation of brain areas depending on sex. Meaningful differences in chronic pain perception between sexes could be related to differential brain activation and, furthermore, could underlie sex differences in response to analgesic treatment. The effects of reducing the activation of particular pain regions on overall awake pain using other stimuli has yet to be tested. With further elaboration, brain activation could be utilized in developing future strategies for the management of neuropathic pain.

Additional file 1 : Additional methods . Additional methods. Additional Table 1 : Body weight before and after spinal nerve ligation . Additional Table 2 : Foot responses to von Frey filaments before and 2 weeks after spinal nerve ligation . Additional Figure 1 : Effect of drugs on straight leg lift . Effect of drugs on straight leg lift. Straight leg lifts scores were obtained before and then after treatment with either vehicle, pregabalin, duloxetine or morphine. Symbols show individual responses to treatment. The clear bars show the mean of n = 3 per sex.