Contextual control of conditioned pain tolerance and endogenous analgesic systems: Evidence for sex-based differences in endogenous opioid engagement

The mechanisms underlying the transition from acute to chronic pain are unclear but may involve the persistence or strengthening of pain memories acquired in part through associative learning. Contextual cues, which comprise the surrounding environment where events occur, were recently described as a critical regulator of pain memory; both rodents and humans exhibit increased pain sensitivity in environments recently associated with a single painful experience. It is unknown, however, how repeated exposure to an acute painful unconditioned stimulus in a distinct context modifies pain sensitivity or the expectation of pain in that environment. To answer this question, we conditioned mice to associate distinct contexts with either repeated administration of a mild visceral pain stimulus (intraperitoneal injection of acetic acid) or vehicle injection over the course of three days. On the final day of experiments animals received either an acid injection or vehicle injection prior to being placed into both contexts. In this way, contextual control of pain sensitivity and pain expectation could be tested respectively. Both male and female mice developed context-dependent conditional pain tolerance, a phenomenon mediated by endogenous opioid signaling. However, when expecting the presentation of a painful stimulus in a given context, males exhibited conditional hypersensitivity whereas females exhibited endogenous opioid-mediated conditional analgesia. Successful determination of the brain circuits involved in this sexually dimorphic anticipatory response may allow for the manipulation of pain memories, which may contribute to the development of chronic pain states.


Introduction 45
Chronic pain development may involve the generalization and strengthening of acute pain memories 46 (Moseley & Vlaeyen, 2015). Neuronal encoding of painful stimuli, known as nociception, is an evolutionarily 47

Animals 96
Equally sized cohorts of male and female C57BL/6 mice aged 8-12 weeks were used in all experiments. 97 Mice were either bred in house or purchased from The Jackson Laboratory (Bar Harbor, ME). Purchased animals 98 acclimated to the housing facility for >7 days before experimental use. Purchased animals and animals bred in-99 house were not intermixed, but rather used in independent experiments. Animals were randomized to 100 treatment group. All protocols were in accordance with National Institute of Health guidelines and were 101 approved by the Institutional Animal Care and Use Committee at the Medical College of Wisconsin (Milwaukee, 102 WI; protocol #0383). We are not the first to use repeated acetic acid injections as painful stimuli. This 103 procedure was previously used in Stevenson et al. (2006); approximately 4 hr after acetic acid injection, animals 104 no longer exhibited acetic acid-induced feeding suppression in those studies. Similarly, no animal in our study 105 lost more than 15% of their body weight over the course of the experiments. Therefore no animals were 106 excluded from these studies as dictated by approved IACUC endpoints. 107 108 Conditioning paradigms

109
All experiments used two distinct contexts, counterbalanced as Context A and B for all animals. The first 110 context was a 10 x 10 x 15-cm 3 Plexiglas chamber. Two walls of the chamber were solid black Plexiglas, and a 111 horizontal white and black striped pattern covered the remaining two walls. These chambers were cleaned with 112 C-DOX, a chlorine-based disinfectant. The second chamber was a similarly sized Plexiglas chamber, consisting of 113 four black walls, and cleaned with 70% ethanol. Both contexts were placed onto a raised 0.7-cm 2 wire platform 114 to allow for video recordings and mechanical sensitivity testing. Dilute acetic acid (0.1, 0.3, 0.5, or 0.9%) was 115 used as the unconditioned stimulus in all experiments. Animals received an intraperitoneal (IP) injection (10 116 ml/kg body weight) of acetic acid (purchased from Sigma Aldrich; diluted in phosphate buffered saline, PBS, pH 117 7.4) or PBS vehicle according to the experimental details described below. The 0.9% dose of acetic acid was 118 used as the final unconditional stimulus since it has previously been shown to elicit conditioned pain 119 hypersensitivity in male mice (Martin 2019). 120 Within-subject assessment of conditioned pain tolerance. On days 1, 2, and 3, animals received an IP injection of 121 acetic acid immediately before being placed into one of the two contexts. In the escalating dose experiment 122 (Figure 2), the concentrations were 0.1% , 0.3% and 0.5% acetic acid on days 1, 2, and 3 respectively. For the 123 consistent dose experiments (Figure 3), a 0.9% acetic acid injection was given on each of days 1, 2, and 3. 124 Animals received an equivalent volume injection of PBS immediately before being placed into the second of the 125 two contexts. Contexts and time of injection (i.e., morning or afternoon) were counterbalanced across all 126 animals and context pairings were separated by 3 h for each animal. On day 4, animals received an IP injection 127 of 0.9% acetic acid immediately before being placed into Context A. A second IP injection of 0.9% acetic acid 128 was administered 3 h later immediately before placing the animal into Context B. Pain behavior testing was 129 completed after being placed into either context on days 1 and 4, as described below. 130 Between-subject assessment of conditioned pain tolerance. Between-subjects experiments were similar to the 131 within-subject experiments, with the exception that each animal was only exposed to one context. On days 1, 2, 132 and 3, half of all animals received an IP injection of 0.9% acetic acid immediately before being placed into one 133 of the two conditioning chambers described above; chambers were counterbalanced between groups and 134 animals were only exposed to one of the contexts. The remaining half of the animals received an IP injection of 135 PBS before being placed into the chamber. On day 4, all animals received an IP injection of 0.9% acetic acid 136 immediately before being placed into their training context. Pain behavior testing was completed after being 137 placed into the context on days 1 and 4 as described below.

139
Pain behavioral measures 140 Animals were moved to the behavior testing room at approximately 07:00 h each morning. Overhead 141 lights were on throughout the entirety of the habituation and testing period. Animals remained in their home 142 cages in the behavior room for no less than 1 h prior to the first injection each day. Immediately following acetic 143 acid or vehicle injection, animals were placed into Context A or Context B. Mechanical sensitivity was assessed 144 in each animal 45-60 min following injection using von Frey filaments. Calibrated monofilaments were delivered 145 through the wire testing platform and applied to the plantar surface of each hindpaw following the up-down 146 method (Dixon 1965); the 50% withdrawal threshold of each paw was calculated then averaged between paws 147 (Dixon 1980, Chaplan et al., 1994 Corticosterone ELISA

157
Immediately following von Frey mechanical sensitivity testing (i.e., 60 min following acetic acid or 158 vehicle injection on day 4; Figure 3), animals were removed from the testing apparatus, placed into a clean 159 shoebox cage, and transferred (<1 min) to a neighboring necropsy room. Trunk blood was obtained from 160 isoflurane anesthetized mice via cardiac puncture, transferred to a chilled tube containing 3.5% sodium citrate, 161 then centrifuged at 1500 rpm, 4 °C for 15 min. Plasma was collected and stored at -80 °C until analysis. Plasma 162 concentrations of corticosterone were measured using the Corticosterone Enzyme Immunoassay kit (Arbor 163 Assay's DetectX) as previously described (Long et al., 2016).

165
Data reporting and analysis

166
Data presented in this manuscript are those collected during the first performance of each experiment.

167
Results for all animals enrolled in each experiment are reported; no outliers were encountered. All data were 168 analyzed using repeated measures ANOVAs with =0.05 in SPSS v.28. Planned comparisons were used following 169 ANOVAs to determine between-and within-group differences following significant interactions or main effects.

170
Estimates of effect size were calculated using a partial eta-squared. Power analyses were used to determine 171 appropriate sample sizes. Means (μ), standard deviations (σ), and expected effect sizes were obtained from 172 preliminary studies and data of a similar nature previously published by the Stucky Lab. These values were used 173 to calculate the sample sizes needed to achieve a significance level (α) of 0.05 and statistical power (1-β) of 174 >0.8.

177
In chronic pain conditions and models, subjects are repeatedly or continuously presented with a noxious 178 UCS. In these studies, we used associative learning paradigms that repeatedly coupled a painful UCS (i.e., acetic 179 acid) with a unique environment to determine if pain memories affect (1) pain sensitivity (UCR), and (2) pain 180 expectation (CR) (Fig. 1). 181 Figure 1: Using associative learning to investigate contextual control over pain sensitivity and pain expectation. Pavlovian conditioning was used to 183 determine if context could come to exert control over different aspects of pain behaviors. Prior to conditioning, the unconditional stimulus (UCS; 184 0.9% IP acetic acid injection) induced pain behaviors (unconditional response, UCR; hindpaw mechanical hypersensitivity) in mice. During 185 conditioning, the UCS was presented immediately before animals were placed into a unique training environment (conditional stimulus, CS; behavior 186 chambers with unique wall patterns and odors), such that the UCR occurred in the CS. After conditioning, two different experiments were performed 187 to assess how environment influences both pain sensitivity (UCR) and pain expectation (conditional response, CR). In the first experiment, the UCS 188 was presented in both the CS and a novel environment that had different visual, and olfactory cues. The UCR was measured in each context and 189 resulting differences determined if environment affected pain sensitivity. In the second experiment, animals were treated with naloxone and placed 190 into the CS after multiple presentations of the UCS + CS to determine if endogenous opioid signaling influences conditional behavioral responses.

192 193
Female mice develop context-dependent pain tolerance after training with ascending doses of acetic acid.

194
To first determine if context can exert conditional control over pain sensitivity, we designed a within-195 subject, 3-day conditioning paradigm in which animals were trained to associate one unique context with 196 escalating doses of IP acetic acid and a second context with IP vehicle administration ( Fig. 2A). Prior to being 197 placed into either chamber on the fourth day of the paradigm, animals were injected with a dose of acetic acid 198 that was higher than the doses previously administered in the acid-trained context. The primary pain behavior 199 measurement in these experiments was mechanical sensitivity of the hindpaw. Although the hindpaw tissue is contexts on day 1 (p = 0.91) and day 4 (p = 0.29), suggesting that associative learning did not enable 213 environmental control of pain sensitivity in males. Alternatively, hindpaw sensitivity of female mice varied 214 greatly between chambers on day 1 and day 4 of the paradigm (Fig. 2C). On day 1, female mice exhibited 215 mechanical hypersensitivity (i.e., 'anti-DNIC') in the acid-paired context relative to the vehicle chamber (p < 216 0.001). However, on day 4, female hindpaw sensitivity was greater in the vehicle-paired context than the acid-217 paired context (p = 0.001), despite the fact that animals received IP injection of 0.9% acetic acid prior to being 218 placed into either chamber. 219 The individual sexes exhibited similar levels of hindpaw sensitivity in the vehicle-paired chamber on day 220 1 (p = 0.72), but displayed both quantitative and qualitative differences in hindpaw sensitivity following the 221 various acetic acid injections. Following acid injection on day 1, female hindpaw mechanical sensitivity was 222 higher than that observed in males (p < 0.001; Fig, 2D). Similarly, female hindpaw sensitivity was greater than 223 male hindpaw sensitivity following an IP injection of 0.9% acetic acid and placement into the vehicle-paired 224 chamber on day 4 (p = 0.045), despite the fact that both male and female mice exhibited higher mechanical 225 sensitivity in this chamber on day 4 compared to day 1 (males: p = 0.003; females: p < 0.001). However, when 226 injected with acid and placed into the acid-paired context on day 4, females exhibited less hindpaw sensitivity 227 than males (p = 0.011; Fig. 2D); male hindpaw withdrawal thresholds were lower in the acid-paired chamber on 228 day 1 as compared to day 4 (p = 0.002), whereas female hindpaw withdrawal thresholds were higher in the 229 acid-paired context on day 4 relative to day 1 (p = 0.006). This latter effect is especially interesting considering 230 the dose of acetic acid administered in the acid-paired context on day 4 is 9-fold higher than the dose 231 administered in this same context on day 1. Collectively, these data suggest that that prior experience, or pain 232 memories associated with a specific environment, induce a conditioned compensatory response that decreases 233 pain sensitivity -or in other words, increases pain tolerance -only in female mice. 234 235 Figure 2: Female mice develop context-dependent pain tolerance after training with ascending doses of acetic acid. A) Experimental design depicting 236 the within-subject procedure. In daily sessions, mice were given ascending doses of acetic acid in one physical chamber, Context A, and vehicle 237 injections in a separate physical chamber, Context B. On the final day, mice received an injection of 0.9% acetic acid solution in each context. day 4 of the paradigm. Acetic acid injection on day 1 (0.1%) had no effect on hindpaw mechanical sensitivity. Hindpaw mechanical sensitivity was 240 similar in both contexts following 0.9% acetic acid injection on day 4. C) von Frey withdrawal thresholds of female mice (n=8) on day 1 and day 4 of 241 the paradigm. Acetic acid injection on day 1 (0.1%) induced hindpaw mechanical hypersensitivity ('anti-DNIC'). Hindpaw mechanical sensitivity 242 differed between the contexts following 0.9% acetic acid injection on day 4, however; females exhibited contextually mediated pain tolerance (i.e., 243 less pain sensitivity) in the acid-paired chamber relative to the vehicle-paired chamber on day 4. D) von Frey withdrawal thresholds of male and 244 female mice replotted to highlight results in the acetic-acid paired chamber. Male mice exhibit anti-DNIC on day 4 relative to day 1. Female mice 245 exhibit less hindpaw sensitivity following 0.9% acetic acid injection on day 4 than they did following 0.1% acetic acid injection on day 1 suggesting the 246 development of conditional pain tolerance.

248
Female and male mice develop context-dependent pain tolerance after training with high doses of acetic acid.

249
These initial experiments raised the possibility that males were not able to develop conditional pain 250 tolerance. However, unlike females, male mice also failed to develop mechanical hypersensitivity (i.e., 'anti-251 DNIC') after injection of 0.1% acetic acid on day 1 of training (Fig. 2B). Therefore, we repeated this paradigm but 252 administered 0.9% acetic acid on all training and test days (Fig. 3A) to determine if the unconditional stimulus 253 used in the previous experiments was simply not strong enough to support associative learning in males.

254
The significant (largest F = 1.95, p = 0.12). As before, planned comparisons were conducted to assess differences 259 both between-and within-groups. 260 In contrast to the results from the previous experiment, males exhibited hindpaw mechanical 261 hypersensitivity in the acid-paired chamber on day 1 (p = 0.006; Fig. 3B) and conditional pain tolerance in this 262 chamber on day 4 (p = 0.006) relative to the vehicle-paired chamber. A similar pattern was observed in females 263 (Fig. 3C); relative to the vehicle-paired chamber, female mice exhibited hindpaw mechanical hypersensitivity in 264 the acid-paired chamber on day 1 (p = 0.003), and conditional pain tolerance in this chamber on day 4 (p < 265 0.001). Although the magnitude of 'anti-DNIC' exhibited following 0.9% acetic acid injection on day 1 was higher 266 in females than males (p = 0.024; Fig. 3D), the level of conditional pain tolerance exhibited in the acid-paired 267 chamber on day 4 was not significantly different between the sexes (p = 0.078). In planned comparisons that 268 assessed behavior across days, males exhibited 'anti-DNIC' in the vehicle-paired context on day 4 relative to day 269 1 (p < 0.001) but did not exhibit conditional pain tolerance in the acid-paired chamber on day 4 relative to day 1 270 (Fig. 2D; p = 0.42). This was not the case for females as they demonstrated both 'anti-DNIC' in the vehicle-paired 271 context on day 4 relative to day 1 (p = 0.002) and conditional pain tolerance in the acid-paired context on day 4 272 relative to day 1 (p < 0.001). Considering these data, and the fact that male mice had higher withdrawal 273 thresholds in the acid-paired context on day 4 as compared to the vehicle-paired context, we conclude that 274 both sexes are able to develop context-dependent pain tolerance if trained with strong unconditional stimuli, 275 but the magnitude of this phenomenon may differ between the sexes with females being especially sensitive to 276 this type of learning. . Hindpaw mechanical sensitivity differed between the contexts following 0.9% acetic acid injection on day 4, however; females exhibited 287 contextually mediated pain tolerance (i.e., less pain sensitivity) in the acid-paired chamber relative to the vehicle-paired chamber on day 4. D) von 288 Frey withdrawal thresholds of male and female mice replotted to highlight results in the acetic-acid paired chamber. Male mice exhibit similar levels 289 of hindpaw mechanical sensitivity in the acid-paired chamber on days 1 and 4. Female mice exhibit less hindpaw sensitivity following 0.9% acetic acid 290 injection on day 4 than they did following 0.9% acetic acid injection on day 1 suggesting the development of conditional pain tolerance.

292
Conditional pain tolerance is not mediated by changes in circulating corticosterone, but rather increased 293 endogenous opioid signaling. 294 To begin probing the biological basis of this conditional pain tolerance, we next developed a between-295 subject 3-day conditioning paradigm in which animals were only trained in one environment so as to isolate 296 physiological changes occurring in each context (Fig. 4A) paired context in our within-subject experiment, acid-trained animals exhibited referred mechanical 300 hypersensitivity relative to vehicle-trained animals on day 1 (p < 0.001) and conditional pain tolerance relative 301 to vehicle-trained animals on day 4 (p = 0.008; Fig. 4B). To determine if this context-dependent pain tolerance is 302 a stress-mediated phenomenon, circulating corticosterone was measured in animals immediately upon removal 303 from the testing environment. Corticosterone levels were similar between acid-and vehicle-trained mice (Fig.  304 4C; t (14) = 0.56, p = 0.58) suggesting that the observed behaviors were not forms of stress-induced analgesia or 305 hyperalgesia in the acid-and vehicle-trained mice respectively. 306 Another possible biological explanation for conditional pain tolerance is that after repeated exposure to 307 acetic acid, the context may promote the release of endogenous opioids as part of a compensatory response to 308 the upcoming acetic acid treatment. To test this hypothesis, we repeated the between-subject 3-day 309 conditioning paradigm, but injected animals with naloxone, a broad-spectrum opioid receptor antagonist, or 310 vehicle before they were placed into the testing environment on day 4 (Fig. 4D). A 2 (Group: vehicle-trained, 311 acid-trained) x 2 (Drug: vehicle, naloxone) ANOVA found a main effect of group, F (1, 28) = 6.68, MSE = 2.77, p = 312 0.015, ηp 2 = 0.19, a main effect of drug, F (1, 28) = 6.67, MSE = 2.77, p = 0.015, ηp 2 = 0.19, and an interaction 313 between the two, F (1, 28) = 11.57, MSE = 2.77, p = 0.002, ηp 2 = 0.29. Naloxone pre-treatment had no effect on 314 the referred mechanical hypersensitivity exhibited by vehicle-trained animals on day 4 (p = 0.64), but it 315 successfully blocked the conditional pain tolerance exhibited by acid-trained mice on day 4 ( Fig. 4E; p < 0.001); 316 the mechanical withdrawal thresholds of naloxone injected acid-trained mice were similar to those exhibited by 317 vehicle-trained mice after their first exposure to acetic acid. To determine if naloxone treatment blocked 318 conditional pain tolerance to a similar extent in acid-trained males and females, we performed a 2 (Sex: male, 319 female) x 2 (Drug: vehicle, naloxone) ANOVA. This analysis revealed a main effect of drug, F (1, 20) = 31.62, MSE = 320 3.14, p < 0.001, ηp 2 = 0.61, but no main effect of sex nor an interaction between the two (Fs < 1), suggesting 321 that naloxone blocked conditional pain tolerance, in both male (vehicle vs. naloxone; p = 0.005) and female 322 (vehicle vs. naloxone; p < 0.001) acid-trained mice (Fig. 4F). 323 324 Figure 4. Endogenous opioid signaling, and not circulating corticosterone, is associated with contextually mediated conditional pain tolerance. A)

325
Experimental design depicting the between-subjects design. Animals received an IP injection either acetic acid or vehicle for three days. On the receiving vehicle injections before being placed into either chamber. This conditional hypersensitivity was not 357 observed in females. Despite exhibiting the same 'anti-DNIC' effect as males on day 1 (p < 0.001; Fig. 5C), 358 female mice exhibited the same level of mechanical sensitivity in the vehicle-and acid-paired contexts on day 4 359 (Fig. 5C, p = 0.81). Further, in contrast to male mice that exhibited similar withdrawal thresholds in the acid-360 paired chamber on day 1 and day 4, female hindpaw withdrawal thresholds in the acid-paired chamber were 361 significantly higher on day 4 than on day 1 (p = 0.003), suggesting either a lack of any conditional response, or 362 the development of conditional analgesia. 363 To determine if female behavior on day 4 was reflective of context-dependent endogenous opioid 364 release, and thus a form of conditional analgesia, we performed a between-subject, 3-day acetic acid training 365 paradigm and treated animals with naloxone prior to being placed in the testing context on day 4 (Fig. 5D).

366
Mechanical sensitivity was assessed on day 4 (Fig. 5E)  0.43. Naloxone treatment had no effect on the conditional hypersensitivity exhibited by male mice on day 4 (p = 370 0.63), further supporting the notion that this behavior is an example of stress-induced hyperalgesia. When 371 administered in females however, naloxone unmasked an endogenous opioid-mediated conditional 372 compensatory response (p < 0.001). In the absence of naloxone treatment, female behaviors appear to be 373 unaffected by 3 days of acid-training (i.e., females had high withdrawal thresholds in the acid-trained context on 374 day 4 relative to day 1 as they are not receiving an acetic acid injection prior to entering the context; Fig. 5C). 375 However, in the absence of acetic acid injection on day 4, naloxone treatment decreased withdrawal thresholds 376 suggesting that females exhibit context-dependent conditional hypersensitivity that they readily counteract 377 with activation of their endogenous opioid system. 378

394
In the present experiments we asked whether environment can exert control over pain sensitivity and 395 pain expectation through associative learning. We first demonstrated, in a within-subject manner, that a 396 context paired with a painful experience (acetic acid injection; UCS) can control conditional pain tolerance. 397 Although female mice demonstrate this conditional pain tolerance after pairing a relatively weak UCS with the 398 CS, male mice only develop conditional pain tolerance if a stronger UCS is used throughout conditioning trials. 399