Sham surgeries for central and peripheral neural injuries persistently enhance pain-avoidance behavior as revealed by an operant conflict test

Abstract

Studies using rodent models of neuropathic pain employ sham surgery control procedures that cause deep tissue damage. Sham surgeries would thus be expected to induce potentially long-lasting postsurgical pain, but little evidence for such pain has been reported. Operant tests of voluntary behavior can reveal negative motivational and cognitive aspects of pain that may provide sensitive tools for detecting pain-related alterations. In a previously described operant mechanical conflict (MC) test involving lengthy familiarization and training, rodents freely choose to either escape from a brightly lit chamber by crossing sharp probes or refuse to cross. Here, we describe a brief (2-day) MC protocol that exploits rats’ innate exploratory response to a novel environment in order to detect persistently enhanced pain-avoidance behavior after sham surgeries for two neural injury models: thoracic spinal cord injury (SCI) and chronic constriction injury (CCI) of the sciatic nerve. Pitting the combined motivations to avoid the bright light and to explore the novel device against pain from crossing noxious probes disclosed a conflicting, hyperalgesia-related reluctance to repeatedly cross the probes after injury. Rats receiving standard sham surgeries demonstrated enhanced pain-like avoidance behavior compared to naive controls, and this behavior was similar to that of corresponding CCI or SCI rats weeks or months after injury. In the case of sham surgery for SCI, video analysis of voluntary exploratory behavior directed at the probes revealed enhanced forepaw withdrawal responses. These findings have important implications for preclinical investigations into behavioral alterations and physiological mechanisms associated with postsurgical and neuropathic pain.

1. Introduction

Alterations of reflexive responses to hand-delivered test stimuli are the most common measures of pain in animal studies ^{5, 41}. However, reflex assays such as widely used von Frey and Hargreaves tests do not assess motivational and cognitive aspects of pain ^{69, 84, 86, 96}, and they are prone to unconscious tester bias ¹². Better indicators of the motivational-affective properties of pain could increase the predictive value of preclinical models of chronic pain for translational drug discovery ^{10, 95}. In particular, operant behavioral tasks dependent on voluntary behavior, such as conditioned place preference (CPP), ^{4, 57, 94}, place-escape/avoidance paradigm (PEAP) ^{1, 13, 36, 61}, thermal escape ⁸⁵ and passive avoidance ⁹⁰ tests can capture negative motivational components of pain that are particularly important clinically ^{75, 96}.

A mechanical conflict (MC) test developed by Harte and colleagues ⁴⁸ is an operant paradigm in which an animal’s voluntary behavior reveals alterations in the aversiveness of sharp probes the animal must cross to escape bright light. Rats’ latencies to escape from a brightly lit chamber increase as probe height increases, while escape latencies are decreased by treatment with drugs known to reduce pain in humans ⁴⁸. This test assesses motivational and cognitive processing of pain-related information, and it removes the potential for unconscious experimenter bias inherent in reflex tests that utilize hand-delivered stimuli ^{43, 48, 64, 78}. Activation of brain regions associated with motivational-affective components of pain during the MC test ⁷³, as during the PEAP ^{62, 63, 98}, support the view that these operant tests measure pain avoidance.

In principle, an operant test in which an animal’s voluntary behavior discloses the aversiveness of a test stimulus might reveal evoked pain that has not been evident in reflex tests. Humans often experience painful postoperative hypersensitivity long after surgical procedures similar to those used to expose peripheral nerves or the spinal cord in rodent neuropathic pain models ²¹. Therefore, we asked whether an operant MC test could detect postoperative enhancement of evoked pain caused by the sham surgeries typically used as control procedures in rat models of long-lasting neuropathic pain caused by thoracic spinal cord injury (SCI) and chronic constriction injury (CCI) of the sciatic nerve. We modified the MC test to take advantage of rats’ innate drive to explore novel environments ^{11, 32, 52}, and measured pain-related changes in a rat’s motivation to repeatedly cross noxious probes. Prior studies using the same device habituated exploratory behavior prior to testing and permitted only a single crossing of the probes, using increased latency to escape a brightly lit chamber to measure enhancement of evoked pain ^{23, 43, 48, 64, 78}. Here we report that a brief MC protocol, which pits the motivation to continue exploring a novel environment against injury-enhanced pain during repeated crossings of noxious probes, reveals persistent enhancement of pain-avoidance behavior following sham surgeries for central SCI and for peripheral CCI models of neuropathic pain. Some of the work reported here has been described in a published Ph.D. dissertation ⁷¹.

2. Methods

2.1. Animals

All procedures followed the guidelines of the International Association for the Study of Pain and were approved by the McGovern Medical School at UTHealth and University of Texas MD Anderson Cancer Center Animal Care and Use Committees. Male, Sprague-Dawley rats (Envigo, USA) were used at both institutions. At McGovern Medical School, the rats (250–300 g, 2 per cage) were acclimated to a controlled environment (12-hour reverse light/dark cycle, 21 ± 1°C) for ≥4 days before beginning experiments. The corn cob bedding was replaced 2–3 times per week while food and water were provided ad libitum. At MD Anderson Cancer Center, the rats (10 weeks old, 2–3 per cage) were acclimated to the controlled laboratory environment (12-hour light/dark cycle, lights on at 07:00 h, 22 ± 1°C) for at least 7 days before beginning experiments. The corn cob bedding was replaced once per week. Food and water were provided ad libitum.

2.2. Spinal cord injury (SCI) and sham surgeries

Surgeries were performed at McGovern Medical School as previously described ^{7, 8, 72, 92}. Anesthesia in most studies (Fig. 1A–B) was by isoflurane (induction 4–5%; maintenance 1–2%). In a subset (Fig. 1C), intraperitoneal (i.p.) injection of ketamine (60 mg/kg), xylazine (10 mg/kg), and acepromazine (1 mg/kg) was used. Rats were determined to be deeply anesthetized and areflexic before proceeding. Local anesthetic (bupivacaine, 2 mg/kg) was delivered subcutaneously (s.c.) before incising the skin from T8-T12. Laminectomy of the T10 vertebrae was followed by a contusive impact (150 kdyne, 1 s dwell time) using an Infinite Horizon Spinal Cord Impactor (Precision Systems and Instrumentation, LLC, Fairfax Station, VA, USA). Following impact, the paravertebral muscles were closed with vicryl-coated, absorbable suture and the skin incision was closed with 9 mm wound clips. Rats were returned to their home cage and placed on a heating pad maintained at 37°C. The analgesic buprenorphine hydrochloride (0.02 mg/kg; Buprenex, Reckitt Benckiser Healthcare Ltd., Hull, England, UK) was administered in 0.9% saline (2 mL/kg, i.p.) twice, daily up to 5 days post-surgery. The antibiotic enrofloxacin (0.3 mL; Enroflox, Norbrook, Inc., Overland Park, KS, USA), was also administered in 0.9% saline daily up to 10 days post-surgery. Manual bladder evacuations were performed twice daily until rats recovered neurogenic bladder voiding. The day after surgery, hindlimb locomotion was assessed using the Basso, Beattie, and Bresnahan (BBB) Locomotor Rating Scale ⁶. Sham surgery procedures were identical except for the spinal contusion. Only sham-operated rats with BBB scores of 21 for both hindlimbs were accepted. The majority of SCI rats had scores of 0 or 1 for both hindlimbs 1 day after injury. At the time of post-injury testing, SCI rats had BBB scores >9. Naive rats were transported to the surgical suite at the same time as SCI and sham-operated rats but were otherwise left undisturbed in their home cages.

Figure 1.

Timelines of the operant mechanical conflict (MC) test used to measure changes in avoidance of noxious probes. (A) Assessments in the SCI model, providing initial familiarization during a 10-minute session without probes, followed by a test with high probes. (B) Assessments in the SCI model, using a standardized familiarization sequence followed by trials with progressively higher probes. (C) Assessments in the SCI model, with the familiarization sessions and baseline trial followed by trials with high probes. (D) Assessments in the CCI model using the same sequence as in part C. Numbers (in mm) indicate elevation of the sharp probes above the floor of the middle chamber. During the 5-minute familiarization sessions and baseline trial the rat is free to explore the 3-chamber MC device in the absence of elevated probes (0 mm). CCI, chronic constriction injury of the sciatic nerve; SCI, spinal cord injury.

2.3. Chronic constriction injury (CCI) and sham surgeries

Surgeries were performed at MD Anderson Cancer Center. Neuropathic pain from peripheral injury was induced using the CCI model of unilateral sciatic nerve injury ⁹ as previously described ^{39, 40}. Rats receiving CCI or sham surgeries were anesthetized with isoflurane (4% in oxygen for induction; 2–3% maintenance) and placed on an electric heating pad. Skin at the mid-thigh level of the left leg was shaved with an electric razor and cleansed with povidone-iodine and 70% ethanol. An incision of the skin was made with a scalpel blade and the sciatic nerve was exposed through blunt dissection of the biceps femoris muscle. Using glass nerve hooks, a segment of the sciatic nerve was gently liberated from the surrounding connective tissue. For CCI surgeries, 4 ligatures (4–0 chromic gut; Ethicon, USA) were loosely tied around the sciatic nerve approximately 1 mm apart. For sham surgeries, the sciatic nerve was manipulated with nerve hooks and isolated in an identical fashion, but no chromic gut ligations were sutured around the nerve. The muscle layer was closed with non-absorbable sutures (4–0 silk; Ethicon, USA), 9 mm wound clips were applied to close the skin, and rats were then returned to their home cage and monitored post-operatively until fully ambulatory. Naive rats were transported to the surgical suite at the same time as CCI and sham surgeries were performed, but they were otherwise left undisturbed in their home cages.

2.4. Habituation to ambient testing conditions

At McGovern Medical School, rats were acclimated to the behavioral testing room each morning for 1 hour under red light and constant background white noise generated by a TaskMasking speaker (K.R. Moeller Associates Ltd., Burlington, Ontario, Canada). Several days prior to testing the rats were acclimated to the presence of an experimenter and the acrylic chambers (IITC Life Science Inc., Woodland Hills, CA, USA) used to isolate rats for hindpaw reflex tests. During gentling the rats were placed in the chambers on raised wire-mesh platforms for 20 minutes and periodically fed sweetened cereal. Two experimenters were female (<30 years old) and one experimenter was male (>30 years old). Male and female experimenters did not perform tests on the same days in order to limit male-induced stress and analgesia ⁸⁰. At MD Anderson Cancer Center, rats were acclimated to the behavioral testing room for at least 1 hour under red light illumination prior to each behavioral test. Rats were handled by the experimenter in 5 min sessions over 3 days. Habituation to hindpaw reflex testing occurred by placing rats in acrylic chambers on raised wire-mesh platforms for 60 min sessions over 3 days. One female and one male experimenter (>30 years old) performed all experiments, and each rat was always manipulated by the same experimenter.

2.5. Tests of hindpaw reflex sensitivity

Threshold for paw withdrawal to mechanical stimulation

At McGovern Medical School, hindpaw sensitivity to mechanical stimuli was measured by two female experimenters at 1–2 months post-surgery for naive, sham-operated, and SCI rats. Following habituation and gentling procedures, the rats were placed in acrylic chambers and the 50% mechanical withdrawal threshold was assessed using the “up-down” method ^{20, 25, 29, 30} using calibrated von Frey filaments (Stoelting Co., Wood Dale, IL, USA). The set of filaments had logarithmically increasing bending forces (in grams): 0.4, 0.6, 1.0, 1.4, 2.0, 4.0, 6.0, 8.0, 10.0, 15.0, 26.0, 60.0. Filaments (starting with 6.0 g) were presented perpendicularly to the plantar surface of the hindpaw between the footpads at a constant speed until the filament bent. In some SCI experiments filaments were applied for ~1 second, and in others the filaments were applied for ~5 seconds. Each hindpaw was presented with a series of 10 stimuli, spaced 30 s apart to provide consistent testing durations and treatment. A rapid, obvious withdrawal of the hindpaw from the filament was considered a positive response, and care was taken to withhold stimulation during rats’ ambulatory movements. In SCI experiments, the withdrawal thresholds for the left and right hindpaws were calculated separately and then averaged together for a single score per rat. All withdrawal thresholds were log transformed to account for Weber’s Law ⁶⁷. Mills and colleagues demonstrated that the original equation used to calculate the 50% paw withdrawal threshold (in grams) described by Chaplan et al., ²⁰ can be reduced to the following:

$$\mathrm{Log}\left(\mathrm{PWT}\right)=\mathrm{X}\mathrm{f}+\text{Ƙδ}-4$$

where Xf = the final filament used (i.e., the filament handle number), Ƙ = the tabular value for the given sequence of test stimuli ²⁰, and δ = the mean difference between the given sequence of test stimuli (calculated using the filament handle numbers). The filament handle number = Log₁₀ of (10 x filament force in milligrams) (Stoelting Co. Touch Test™ Sensory Evaluators Operation Manual), indicating the handle numbers can be used in the Chaplan or Mills equations without converting gram forces into log units. Also, δ is not a fixed value as not every rat receives the same series of filaments. For CCI experiments at MD Anderson Cancer Center, the naive, sham-operated, and CCI rats were tested 2 weeks post-surgery using the “up-down” method. Rats were placed in acrylic chambers on raised wire-mesh platforms and habituated for at least 30 min before testing. The set of filaments had bending forces (in grams) of 0.4, 0.6, 1.2, 1.4, 2.0, 4.0, 6.0, 8.0, 10.0, 15.0. Notice that when log transformed, forces between 1 and 0 g yield negative numbers (see Results). Filaments were administered to the distal portion of the heel ^{39, 40} and applied for ~5 seconds. The mechanical thresholds for the ipsilateral (injured) and contralateral (uninjured) hindpaws were scored separately.

Latency for paw withdrawal to radiant heat stimulation

At McGovern Medical School, hindpaw sensitivity to radiant heat ⁴⁷ was measured by two female experimenters at 1–2 months post-surgery for naive, sham-operated, and SCI rats ⁸. Once habituated, the rats were placed in acrylic chambers on a glass platform (Plantar Analgesia Meter; IITC Life Science Inc., Woodland Hills, CA, USA) and acclimated to the 30°C temperature-controlled surface ²⁸ for 20 minutes. Settings for the radiant heat stimulus: idle beam intensity = 10%, active beam intensity = 45%, active beam cutoff = 20 seconds. While idle, the light beam was positioned between the footpads of either the left or right hindpaw, and once positioned the active beam was turned on. A rapid, obvious withdrawal of the hindpaw was considered a positive response to the radiant heat stimulus. Rats that did not exhibit a withdrawal by the time of the automatic active beam cutoff were given a score of 20 seconds. Tests continued until the withdrawal latency was recorded 5 times for each hindpaw, switching between the hindpaws every 30 seconds. If ambulatory movements occurred during presentation of the stimulus, the active light beam was turned off, the experimenter waited 30 seconds, and the other hindpaw was tested. The average withdrawal latency for each hindpaw was calculated using the 3 middle latencies; the highest and lowest latencies were omitted. The two hindpaw latencies were then averaged together for a single score per rat.

2.6. Operant mechanical conflict (MC) test

Voluntary, pain-related aversion to a noxious stimulus was assessed using a commercial 3-chambered device, the Mechanical Conflict System (Coy Laboratory Products, Inc., Grass Lake, MI, USA). The basic MC test presents rats with a choice in responding to two aversive stimuli – either to remain exposed to an aversive bright light in one chamber or to escape the light by crossing a middle chamber having a floor covered by a dense array of sharp probes, in order to reach a dark, safe chamber. Longer latencies to leave the light chamber indicate increased motivation to avoid the probes, and this escape latency is currently the most common measure of pain-related behavior in the MC test ^{23, 43, 48, 73, 78}. We found in pilot studies that, when the MC device is still relatively novel, naive and injured rats cross the noxious probes multiple times. These experiments were extended to assess prolonged escape and crossing behavior under conditions of high novelty using two single trial, 10-minute exposures to the MC device, first without and then with noxious probes (Fig. 1A). The rats’ repeated return to the presumably aversive brightly lit chamber across the noxious probes indicates the presence of a second motivation to cross the probes, likely the rats’ exploratory drive, which might be modulated by pain and other emotional states such as anxiety (see Results and Discussion). We took advantage of this second motivation to investigate additional effects of sham surgery, SCI, and CCI. To do this, we shortened the conventional MC paradigm ⁴⁸ to a 2-day protocol so that both motivations to cross -- 1) to escape the light and 2) to explore the MC device -- were in conflict with the aversiveness of the sharp probes.

Familiarization sessions

Harte et al. describe a lengthy procedure for familiarization to the MC device and escape training lasting 4–7 days, with a total of 10 to 19 opportunities (each 5 minutes duration) for the rat to explore the MC device before experiencing the sharp probes ⁴⁸. During this training the rats learn that when the exit door opens they can escape from the brightly lit room to reach the dark room. We shortened the conventional MC test protocol (Fig. 1B) by combining the familiarization and training procedures into three 5-minute familiarization sessions without the probes on a single day, repeated 3 times, spaced 30–60 minutes apart. In each session 1) a rat was placed inside the light chamber with the lid closed, the light off, and the exit door closed; 2) after 20 seconds the light was turned on; 3) after 15 seconds the exit door was opened when (or if) the rat faced the exit; 4) the rat freely explored all 3 chambers in the MC device for 5 minutes, 5) the rat was returned to its home cage, and 6) the device was thoroughly cleaned with 70% ethanol (in distilled water) in preparation for the next session. The rats rapidly learned to escape the light room as soon as the exit door was opened. Indeed, rats sometimes attempted to lift the door on their own by the second or third sessions.

Noxious probe trials

After the familiarization sessions on Day 1, rats underwent one of three different sequences of trials in which they were challenged with sharp probes. In each case, the first trial (baseline) was without probes to reacquaint the rats with the MC device and provide the basis for comparison to noxious probe trials. In the first study (Fig. 1B), probe height was successively increased from 0 to 1, 2, 3, and 4 mm in 3-minute trials spaced ~30 minutes apart. Probe heights were presented in ascending order to minimize possible sensitizing effects from higher probes and to permit testing multiple probe heights per rat on a single day. In subsequent SCI and CCI studies, a 5-minute baseline trial was followed by 2 trials at 4 mm, 5 minutes per trial, spaced ~30 minutes apart (Fig. 1C–D). All SCI and CCI rats were capable of weight-supported plantar stepping (SCI rats had BBB scores ≥10 by the time of testing) and all were observed to readily traverse the probes without bodily harm.

Video scoring and data collection

All sessions and trials were video-recorded in 1080i resolution at 30 frames per second using a Panasonic HC-V750 camcorder (Panasonic Corporation, Osaka, Japan) or 1080p resolution at 30 frames per second using an Apple iPhone (Apple Inc., Cupertino, CA, USA) and scored by a blinded experimenter. The following measures were collected for post-hoc analysis: 1) the escape latency during the first crossing, 2) the number of crossings of the probe chamber, and 3) the total time elapsed to the completion of the second crossing. A subset of videos was scored for behavioral measures that might reveal above-level mechanical hypersensitivity in the forepaws after SCI ⁸. The number of times each rat withdrew a forepaw after voluntary contact with the probes was scored. No formal definitions of a crossing currently exist for the MC test, and measures other than escape latency depend upon this definition. We defined the first crossing as the rat placing all 4 paws inside the dark chamber after leaving the light chamber. Every subsequent crossing was defined as the rat placing its head and two forepaws inside the light or dark chamber.

2.7. Data analysis and experimental design

Statistical analyses were performed using Prism v7.03 (Graphpad Software, Inc., La Jolla, CA, USA). Data are presented as mean ± SD or SEM or as median with interquartile range. The Shapiro-Wilk test was used to assess normality for continuous measures. Planned comparisons were used to test the overriding hypothesis that sham surgeries induce persistent pain-related behavior. Thus, overall differences were assessed using 1-way ANOVA or Kruskal-Wallis tests followed by planned comparisons of sham versus naive groups, as well as sham versus SCI groups using Dunnett’s or Dunn’s post-hoc tests to gauge the similarity of postoperative effects versus effects of the SCI and CCI procedures (see Table 1). Within-group comparisons of crossings measured during probe trials to each groups’ 0-mm baseline trial were performed using repeated measures 1-way ANOVA or Friedman tests followed by Dunnett’s or Dunn’s post-hoc test (see Table 2). Sphericity was not assumed when using the repeated measures 1-way ANOVA and the Geisser-Greenhouse correction was applied. Planned comparisons between groups in Figure 1B were performed for trials with probe heights found to significantly reduce crossings within groups. Subsequent SCI and CCI experiments were performed with just two trials, using the 4 mm probes. Planned comparisons between groups in subsequent SCI and CCI experiments (Fig. 1C–D) were performed for both probe trials. Significance for all statistical tests was set at P < 0.05 and all reported P values are two-tailed.

Table 1.

Summary of planned comparisons

Figure	Graph	Experiment	MC Test Probe Height	Behavioral Measure	Statistical Test, P	Post-hoc Test	Naïve vs Sham, P	Neural Injury vs Sham, P
2	B1	MC 10 minute	0 mm	Crossings	KW, = 0.0011	Dunn’s	= 0.0005	= 0.5238
				Time of final crossing	KW, = 0.0072	Dunn’s	= 0.0051	> 0.9999
	B2		4 mm	Crossings	KW, = 0.4345	Dunn’s	= 0.3963	> 0.9999
				Time of final crossing	KW, = 0.2762	Dunn’s	= 0.4715	= 0.2652
3	A	MC Baseline	Day 2, 0 mm	Escape latency	KW, = 0.4853	Dunn’s	> 0.9999	= 0.5572
	B			Crossings	1-way ANOVA, = 0.3896	Dunnett’s	= 0.6786	= 0.2911
	C			Crossings	KW, = 0.7867	Dunn’s	> 0.9999	> 0.9999
				Time of final crossing	KW, = 0.9234	Dunn’s	> 0.9999	> 0.9999
4	A2	MC 01234	Day 2, 0 mm	Crossings	1-way ANOVA, = 0.4791	Dunnett’s	= 0.5286	= 0.4353
				Time of final crossing	KW, = 0.7001	Dunn’s	> 0.9999	> 0.9999
	B2		3 mm	Crossings	KW, = 0.0417	Dunn’s	= 0.1720	= 0.8256
			4 mm	Crossings	KW, = 0.1530	Dunn’s	= 0.3126	> 0.9999
	C1		3 mm	Escape latency	KW, = 0.2243	Dunn’s	= 0.7884	= 0.7096
			4 mm	Escape latency	KW, = 0.7483	Dunn’s	> 0.9999	> 0.9999
	D1		3 mm	Time to second crossing	KW, = 0.0606	Dunn’s	= 0.2935	= 0.6679
			4 mm	Time to second crossing	KW, = 0.1509	Dunn’s	= 0.2053	> 0.9999
5	A2	MC 044	Day 2, 0 mm	Crossings	KW, = 0.9386	Dunn’s	> 0.9999	> 0.9999
				Time of final crossing	KW, = 0.9246	Dunn’s	> 0.9999	> 0.9999
	B2		First 4 mm	Crossings	KW, = 0.0114	Dunn’s	= 0.0094	> 0.9999
			Second 4 mm	Crossings	KW, = 0.0347	Dunn’s	= 0.1423	= 0.7728
	C1		First 4 mm	Escape latency	KW, = 0.1340	Dunn’s	> 0.9999	= 0.1081
			Second 4 mm	Escape latency	KW, = 0.0435	Dunn’s	> 0.9999	= 0.0924
6	A	von Frey	-	50% PWT – 1 s application	1-way ANOVA, = 0.0792	Dunnett’s	= 0.9689	= 0.0860
	B		-	50% PWT – 5 s application	KW, = 0.0646	Dunn’s	= 0.6118	= 0.0388
	C	Hargreaves	-	Withdrawal latency	1-way ANOVA, < 0.0001	Dunnett’s	= 0.0044	= 0.0587
7	B	MC 01234	1 mm	Forepaw withdrawals	KW, = 0.0368	Dunn’s	= 0.7793	= 0.1670
			2 mm	Forepaw withdrawals	KW, = 0.0113	Dunn’s	= 0.1744	= 0.3461
			3 mm	Forepaw withdrawals	KW, = 0.1987	Dunn’s	= 0.1547	> 0.9999
			4 mm	Forepaw withdrawals	KW, = 0.1243	Dunn’s	= 0.0852	= 0.9160
			Average of 1, 2, 3, 4 mm	Forepaw withdrawals	KW, = 0.0144	Dunn’s	= 0.0986	= 0.6665
	C	MC 044	First 4 mm	Forepaw withdrawals	KW, = 0.0079	Dunn’s	= 0.0252	> 0.9999
8	A2	MC 044	First 4 mm	Crossings	KW, = 0.1167	Dunn’s	= 0.6330	= 0.7078
			Second 4 mm	Crossings	KW, = 0.0035	Dunn’s	= 0.0079	> 0.9999
	B1		First 4 mm	Escape latency	KW, = 0.6266	Dunn’s	> 0.9999	= 0.6671
			Second 4 mm	Escape latency	KW, = 0.0031	Dunn’s	= 0.0178	> 0.9999
	C	von Frey	-	Ipsilateral 50% PWT – 5 s application	KW, = 0.0002	Dunn’s	> 0.9999	= 0.0063
			-	Contralateral 50% PWT – 5 s application	KW, > 0.9999	Dunn’s	> 0.9999	> 0.9999

Summary of planned comparisons between naïve and neural injury groups to sham controls. The experiment column refers to timelines of mechanical conflict tests shown in Figure 1 (corresponding figure label, experiment: 1A, MC 10 minute; 1B-D, MC baseline; 1B, MC 01234; 1C-D, MC 044) or to standard reflex tests (von Frey or Hargreaves). The statistical test and corresponding post-hoc test was chosen based on whether data was normally distributed (Shapiro-Wilk test, results not shown). ANOVA, analysis of variance; KW, Kruskal-Wallis; MC, mechanical conflict; PWT, paw withdrawal threshold.

Table 2.

Summary of results comparing crossings within groups during probe trials

Figure	Graph	Experiment	Probe Height	Group	Statistical Test, P	Post-hoc Test	Probe Height vs 0 mm Baseline	P
4	B1	MC 01234	0, 1, 2, 3, 4 mm	Naïve	RM 1-way ANOVA, = 0.0804	Dunnett’s	1 mm	= 0.9783
							2 mm	= 0.7718
							3 mm	= 0.9998
							4 mm	= 0.1088
				Sham	Friedman, < 0.0001	Dunn’s	1 mm	= 0.4792
							2 mm	= 0.1356
							3 mm	= 0.0002
							4 mm	< 0.0001
				SCI	Friedman, = 0.0009	Dunn’s	1 mm	= 0.8997
							2 mm	= 0.0504
							3 mm	= 0.0185
							4 mm	= 0.0150
5	B1	MC 044	0, first 4, second 4 mm	Naïve	RM 1-way ANOVA, = 0.0043	Dunnett’s	First 4 mm	= 0.0454
							Second 4 mm	= 0.0142
				Sham	Friedman, < 0.0001	Dunn’s	First 4 mm	= 0.0203
							Second 4 mm	= 0.0011
				SCI	Friedman, < 0.0001	Dunn’s	First 4 mm	= 0.0910
							Second 4 mm	= 0.0006
8	A1	MC 044	0, first 4, second 4 mm	Naïve	Friedman, < 0.0001	Dunn’s	First 4 mm	= 0.0003
							Second 4 mm	< 0.0001
				Sham	Friedman, < 0.0001	Dunn’s	First 4 mm	= 0.0203
							Second 4 mm	< 0.0001
				CCI	Friedman, < 0.0001	Dunn’s	First 4 mm	= 0.0203
							Second 4 mm	= 0.0011

3. Results

3.1. Multiple returns to the brightly lit chamber in the absence of probes and repeated crossing of noxious probes by uninjured, sham-operated, and SCI rats indicate that exploratory drive is very strong when the MC device is unfamiliar

Previous studies with the MC device used conditions in which escape from a bright light was assumed to be the primary motivation to cross noxious probes. We explored metrics in addition to the latency to escape the bright chamber to extract more information about probe-avoidance behavior. We first asked whether, upon introduction to the novel MC device, rats voluntarily crossed the middle chamber repeatedly when the bright light was on and when the noxious probes were raised. This study also allowed us to see if SCI rats >1 month post-injury exhibited deficiencies in mobility that might otherwise confound comparisons to uninjured controls. During their first 10-minute exposure to the MC device the probes were not elevated. Naive, sham-operated, and SCI rats made repeated crossings of the middle chamber (Fig. 2A1), indicating that the bright light was not sufficiently aversive to suppress exploratory behavior during the initial exposure to the MC device. All groups also repeatedly crossed the middle chamber when the sharp probes were subsequently raised to a high 4 mm level (Fig. 2A2), indicating that addition of the noxious probes failed to suppress exploratory behavior under these conditions. Summarized data (Figures 2B1, B2) confirm that 1) the bright light did not prevent multiple crossings of the middle chamber in the absence of high probes, 2) the sham group, but not the SCI group, made significantly fewer crossings and stopped crossing sooner than the naive group when no probes were present, 3) exposure to the 4 mm probes greatly increased crossing variability within all groups, and 4) rats in the SCI group made a comparable number of crossings to those in the naive and sham groups, demonstrating that a contusive T10 SCI does not hinder their mobility. These results suggest there is a strong motivation to repeatedly cross the probes, likely from exploratory drive, which can dominate other influences (which may be complex) on crossing behavior when the MC device is novel. This further suggests that, with a limited degree of familiarization (only partial habituation of exploratory drive), counting crossings could serve as a useful metric for assessing pain-related changes in probe-avoidance behavior in conflict with exploratory behavior.

Figure 2.

The first exposures to the mechanical conflict device reveal a strong motivation to repeatedly explore the device despite an aversive bright light and (in the second session) sharp probes. Naive, sham-operated, and SCI rats were exposed to the MC device twice for 10 minutes, without probes on the first day and with 4-mm probes on the second day (see Figure 1A timeline). The bright light was on during both sessions. (A1, A2) The timing of every crossing in each session is shown for each rat. (B1, B2) Summary results for the number of crossings and timing of the final crossing indicate significant differences between sham and naive groups when no probes are present, and large variability among groups when the 4-mm probes are present. Planned comparisons (B1, B2) were performed using the Kruskal-Wallis test followed by Dunn’s post-hoc test. Data in (B1, B2) shown as mean ± SD for the vertical and horizontal axis. *P < 0.05, **P < 0.01, ***P < 0.001. MC, mechanical conflict; SCI, spinal cord injury.

When given three 5-minute familiarization sessions without probes in the MC device, a majority of rats in the naive, sham, and SCI groups quickly learned to exit the light chamber. Escape latencies measured 24 hours later during the single baseline trial without probes were not significantly different among the groups (Fig. 3A), and these values were similar to but slightly longer than the escape latencies without probes reported for rats that underwent more extensive familiarization and escape training ⁴⁸. Many rats voluntarily crossed back into the light chamber as they explored the MC device during each familiarization session. The median number of crossings of the middle chamber measured 24 hours later during the single baseline trial without probes was not significantly different among groups (Fig. 3B), and most rats made repeated crossings for the duration of the 5 min baseline trial (Fig. 3C). This indicates that exploratory behavior of the MC apparatus in the baseline trial on Day 2 is not altered by surgical history, and suggests that, after the 3 standardized familiarization sessions, states such as general anxiety don’t differentially influence the behavior of rats in the naïve, sham, and SCI groups in the absence of elevated probes.

Figure 3.

Exploratory behavior in naive, sham-operated, and spinal cord injury rats in the mechanical conflict device during three familiarization sessions without noxious probes. Each session was 5 minutes long, spaced ~30 minutes apart, with a 5-minute test session 24 hours later serving as baseline for escape latencies and crossings for subsequent trials with elevated probes (see Figures 1B, C timelines). Escape latencies (A) and crossings (B) were not significantly different among naive, sham, and SCI groups. (C) Summary results for the number of crossings and time of the final crossing also do not reveal any significant differences among groups. (D) Two independent measures of SCI severity, the displacement of contused tissue (in μm) measured by the spinal impactor device (x-axis) and the BBB motor score measured 24 h after SCI (y-axis), indicate a majority of SCI rats had received a moderately severe contusive SCI, but this SCI did not prevent injured rats from exploring the MC device 1 month or longer after the SCI. Planned comparisons between groups in (A-C) performed using 1-way ANOVA or Kruskal-Wallis test followed by Dunnett’s or Dunn’s post-hoc test. Data in (A, B) shown as median with interquartile range, and data in (C) shown as mean ± SD. ANOVA, analysis of variance; BBB, Basso, Beatie, and Bresnahan locomotor rating scale; MC, mechanical conflict; SCI, spinal cord injury.

In principle, the lack of differences in crossing behavior found in the SCI group compared to the sham (see subsequent sections) and sometimes the naive groups (Figs. 2–3) might be the result of less spinal injury than produced in previous SCI studies. Two independent measures were used to assess SCI severity ^{19, 25, 58}, tissue displacement during spinal impact recorded by the impactor device (in μm) and the BBB motor score measured 1 day after injury. Contusion displacements were typical for a 150 kdyne impact ¹⁹ and internally consistent with impact data collected by the prior surgeon in the lab (data not shown). The mean displacement of 947 ± 25 μm and the mean BBB score of 1.3 ± 0.3 one day after SCI were also consistent with a moderate SCI ^{19, 58}. All rats in the sham group exhibited BBB scores of 21 for each hindpaw the day after surgery, indicating that unintended damage to the spinal cord during the T10 laminectomy had not occurred. Injury assessments indicate a majority of SCI rats (76%, 37/49 rats with BBB scores of a 0 or 1) used had moderately severe spinal injuries (Fig. 3D). Data shown in Figs. 3A–C are from SCI experiments using the 2-day protocols shown in Figs. 1B and 1C. Multiple returns to the light chamber occurred during familiarization sessions in all experiments (SCI and CCI; see below), suggesting that the motivation to continue exploring the MC device in the absence of noxious probes remained high enough to offset the aversiveness of the bright light, even after 3 exposures to the device, and that this motivation was not diminished by SCI.

3.2. Sham surgery for the SCI model persistently enhances pain-avoidance behavior in the MC test

Having found that 3 familiarization trials without elevated probes elicited similar exploratory behavior in naive, sham-operated, and SCI rats, we then addressed two questions: 1) how would exploratory behavior (as indicated by multiple crossings) be affected by noxious probes in the middle chamber, and 2) would the response to the probes be altered by prior surgical tissue injury or neural injury plus surgical injury? We began by giving rats in naive, sham, and SCI groups a series of trials (3 minutes each) with progressively greater probe heights (0 to 4 mm), spaced 30 minutes apart. No significant differences were found among the groups in the rats’ exploration during the baseline trial (0 mm) prior to elevating the probes (Figs 4A1, A2. However, rats in the SCI and sham, but not the naive, groups crossed the 3 and/or 4 mm probes significantly fewer times than they had crossed the middle chamber during the 0 mm baseline trial (Fig. 4B1). There was a trend for rats in the SCI group to cross the 1- and 2-mm probes fewer times than rats in the naive and sham groups, but post-hoc comparisons did not reveal significant differences (Kruskal-Wallis P = 0.09 for 1-mm and 1-way ANOVA P = 0.07 for 2-mm probes). No significant differences among groups were observed on the 3- and 4-mm probes for number of crossings (Fig. 4B2) or the escape latency during the first crossing (Fig. 4C1). While some SCI and sham-operated rats showed much longer latencies, and a few refused to cross the probes even once, ~60–80% of the sham-operated and SCI rats had latencies that were comparable to naive rats, as shown in a plot of the fraction of rats in each group exhibiting an escape latency equal to or less than the latency indicated on the x-axis (Fig. 4C2). No rats were excluded based on deviant latencies to cross the probes. Rats in the sham and SCI groups tended to cross the probes fewer times than naive rats (Fig. 4D1, D2), with ~40–60% of the SCI and sham-operated rats refusing to cross the 3 and 4 mm probes a second time. Altogether, the results shown in Figure 4 indicate that 1) exploratory behavior (indicated by multiple crossings) in naive rats shows little or no reduction by the noxious probes in the middle chamber and 2) surgical injury used as a sham control for the SCI procedure results in a possible tendency to increase avoidance of the noxious probes (reduced crossings) similar to that apparent in rats with SCI.

Figure 4.

Sham surgery for spinal cord injury enhances pain-avoidance behavior in the MC test. Rats were exposed successively to ascending probe heights from 0 to 4 mm (see Figure 1B timeline). (A1, A2) Times of successive crossings during the 0 mm baseline trial shown for each rat. Summary results show similar numbers and patterns of crossings in naive, sham, and SCI groups. (B1) Crossings on 3- and 4-mm probes were decreased in the sham and SCI groups. Crossings at each probe height were compared to the 0 mm baseline using a repeated measures 1-way ANOVA or Friedman test. Significance levels for sham (stars) and SCI (pound sign) groups shown for Dunn’s post-hoc tests. (B2) Planned comparisons on 3- and 4-mm probe trials revealed no significant reductions in crossings in sham and SCI groups. (C1, C2) Escape latencies were not different among groups. C2 plots the fraction of rats in each group exhibiting an escape latency equal to or less than the latency indicated on the x-axis. (D1, D2) Some sham and SCI rats refused to cross back into the light chamber, as indicated by 180-second second crossing latencies. D2 plots the fraction of rats in each group showing a latency to their second crossing of the probes equal to or less than the time indicated on the x-axis. Planned comparisons in (A2, B2, C1, and D1) were performed using 1-way ANOVA or Kruskal-Wallis followed by Tukey’s or Dunn’s post-hoc test. Data shown as mean ± SD (A2), mean ± SEM (B1) and median with interquartile range (B2, C1, and D1). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ANOVA, analysis of variance; MC, mechanical conflict; SCI, spinal cord injury.

We then modified the 2-day MC protocol to test the effects of two exposures to high probes without prior exposure to less noxious lower probe heights ⁴⁸. There was no difference among the groups on the 0 mm baseline trial on the day of probe exposure (Fig. 5A1, A2). All groups showed decreases in the number of crossings on the probes compared to their baseline crossings without the probes, and this effect was significant for each group during the second 4 mm probe exposure (Fig. 5B1). Planned comparisons showed that rats in the sham group crossed the 4 mm probes fewer times than rats in the naive group during the first exposure to the probes (Fig. 5B2). Although, no statistically significant differences in crossings were observed during the second exposure, comparison of escape latencies (Fig. 5C1, C2) shows that a majority of rats in the sham and SCI groups (sham: 70%, 7/10 rats; SCI: 75%, 6/8 rats) refused to cross the probes even once during the second probe exposure, whereas only 1 of 8 naive rats (13%) refused to cross the probes. This suggests that sham-operated and SCI rats continued to display greater pain-avoidance behavior than naive rats on the second probe exposure, even though the naive rats showed some apparent sensitization from the first noxious probe exposure. The escape latencies on the second 4-mm trials showed a distinct bimodal distribution (Fig. 5C1), with a tendency for SCI latencies to be greater than those in the sham group (Fig. 5C1, C2). These observations indicate that sham surgery for the SCI procedure is sufficient to persistently increase the aversiveness of the probes and enhance pain-avoidance behavior, and that the 2-day MC test reveals similarly increased aversion in the sham-operated and SCI rats.

Figure 5.

Sham surgery for spinal cord injury enhances pain-avoidance behavior compared to naive, uninjured rats in an abbreviated mechanical conflict test. Rats were exposed to the 4-mm probes, twice (see Figure 1C timeline). (A1, A2) Times of successive crossings during the 0 mm baseline trial shown for each rat. Summary results show similar numbers and patterns of crossings in naive, sham, and SCI groups. (B1) Significantly reduced crossings in naive and sham groups during the first probe trial, and all groups during the second probe trial. (B2) Planned comparisons between groups show sham rats made fewer crossings than naive rats on the first probe trial, and SCI rats were not different from shams. A majority of SCI rats (75%, 6 out of 8) refused to cross during the second probe trial, while only 30% of shams (3 out of 6) and 13% of naive rats (1 out of 8) refused. (C1, C2) Escape latencies on the second probe trial reflected crossing results in (B2), but no significant differences were observed. C2 plots the cumulative percentage of rats in each group exhibiting an escape latency equal to or less than the latency indicated on the x-axis. Crossings for groups during both probe trials in (B1) were compared to the 0-mm baseline using a repeated-measures 1-way ANOVA or Friedman test. Significance levels for naive (stars), sham (pound sign), and SCI group (plus sign) shown for Dunnett’s or Dunn’s post-hoc tests. Planned comparisons between groups (A2, B2, and C1) were performed using a 1-way ANOVA or Kruskal-Wallis test followed by Dunnett’s or Dunn’s post-hoc tests. Data shown as mean ± SD (A2), mean ± SEM (B1) and median with interquartile range (B2, C1). *P < 0.05, **P < 0.01. ANOVA, analysis of variance; SCI, spinal cord injury.

3.3. Standard tests of reflex sensitivity indicate that sham surgery for SCI increases hindpaw heat sensitivity without decreasing mechanical threshold

The sham surgery effect in the MC test raised the question of whether commonly used assays of reflex sensitivity, the von Frey mechanical threshold test and the Hargreaves radiant heat test, can also detect persistent differences between naive and sham-operated rats. Hindpaw sensitivity to von Frey filaments was tested using the Chaplan/Dixon up-down method to determine 50% threshold ^{20, 29, 30}, using an extended range of filaments compared to many other studies (0.4–60.0 grams, starting filament = 6 grams, modified from ²⁵) and log transformed for analysis and display (see ⁶⁷). In the first set of experiments von Frey filaments were applied for ~1 second. Unexpectedly, we observed a trend for SCI rats to exhibit reduced mechanical sensitivity of the hindpaws when compared to sham-operated rats (Fig. 6A). Prior studies by several groups, including the Walters group, had reported below-level hypersensitivity after similar mid- to lower thoracic spinal contusion injuries ^{1, 2, 8, 19, 22, 25, 26, 38, 44, 46, 49, 58, 70, 92, 94, 97}. Many groups, including the Grace group (see Section 3.5), use applications of von Frey filaments that are much longer than 1 second. Thus, we performed another study using ~5-second applications of each filament. The longer application method revealed mechanical hypersensitivity in the SCI group compared to the sham group (Fig. 6B), similar to what had been found previously using brief applications after SCI ^{8, 92, 94}. Because filament applications were not timed precisely, it is possible that the tester using ~1-second applications in our recent experiments applied the filament more briefly or with a different velocity than did the testers in our published studies utilizing brief applications. Regardless of the cause, these divergent results show that major differences in test responses can result from what appear to be small differences in von Frey test procedures.

Figure 6.

Standard tests of reflex sensitivity show postsurgical enhancements to heat but not weak mechanical stimuli. (A) Using the Chaplan “up-down” method and applying von Frey filaments for ~1 second does not reveal significant enhancements of the 50% paw withdrawal threshold 1–2 months following spinal cord injury. Thresholds were similar in naive and sham groups, and SCI exhibited a trend for a reduction in mechanical sensitivity. Planned comparisons were made using a 1-way ANOVA. Corresponding gram forces shown on right axis. (B) Increasing the application time of the von Frey filaments to ~5 seconds reveals a significant enhancement of the 50% PWT in SCI rats 1–2 months after injury. Planned comparisons made using a Kruskal-Wallis test followed by Dunn’s post-hoc test. Corresponding gram forces shown on right axis. (C) Withdrawal latency to a heat stimulus was significantly lower in sham rats and there was a trend for SCI rats to exhibit shorter latencies than shams. Groups were compared using a 1-way ANOVA followed by Dunnett’s post-hoc test. Data in (A-C) shown as median with interquartile range. *P < 0.05, **P < 0.01. ANOVA, analysis of variance; PWT, paw withdrawal threshold; SCI, spinal cord injury.

Randomly selected subsets of the naive, sham-operated, and SCI rats were tested using the Hargreaves radiant heat test. The sham group exhibited a significant reduction in withdrawal latency, and there was a trend for the SCI group to exhibit an even greater reduction than found in the sham group (Fig. 6C). Together, these results suggest that sham surgery can induce a modest increase in hindpaw sensitivity to noxious heat without increasing mechanical sensitivity, and that passive reflex tests of mechanical threshold may fail to disclose persistent pain-like alterations after sham surgery that can be revealed by an appropriate operant test.

3.4. Sham surgery for SCI increases forepaw sensitivity as assessed by withdrawals evoked by sharp probes during the MC test

Rats with SCI exhibit at- and above-level mechanical hypersensitivity ^{8, 18} as shown by increased responsiveness to von Frey filaments applied to forepaws and a rostral region of the back. Although one study found no correlation between SCI-induced paw hypersensitivity measured with von Frey filaments and escape latency measured in the MC test ²³, our observations suggested that forepaw hypersensitivity after injury might be expressed during voluntary behavior. Video analysis showed that the rats often pause before crossing and use their forepaws to investigate the probes. Attempts to establish weight support on the probes with their forepaws often produced a rapid withdrawal response (Fig. 7A). To test whether prior injury produced forepaw hypersensitivity expressed during voluntary behavior, the numbers of rapid forepaw withdrawals from the 1, 2, 3, and 4 mm probes made during initial investigation of the probes and immediately after the first crossing were counted. There was a trend for sham and SCI rats to increase the number of forepaw withdrawals as probe height increased, but there were no statistically significant differences between naive vs sham and sham vs SCI rats at any probe height (Fig. 7B, left panel). Averaging across all probe heights to increase statistical power yielded a trend for sham rats to increase the number of forepaw withdrawals compared to the naive group, and there was little difference between the sham and SCI rats (Fig. 7B, right panel).

Figure 7.

Sham surgery for SCI sensitizes forepaw withdrawals evoked during voluntary contact with noxious probes. (A, B) Sham-operated rats and SCI rats exhibited a trend to withdraw their forepaws from the noxious probes more than naive rats when tested successively on 1-, 2-, 3- and 4-mm probes (see Figure 1B timeline). (A) Example sequence of paw movements during probe investigation by a SCI rat. Rapid forepaw withdrawal – middle image, red arrow. (B) Median number withdrawals on each probe (left) and across all 4 probes (right). (C) Forepaw withdrawals were increased in the sham group during the first noxious probe trial, while the SCI group was not significantly different from the sham group (see Figure 1C timeline). A majority of rats in (C) received two 4-mm probe trials after the 0-mm baseline trial, but no comparisons were made during the second probe trial because a majority of injured rats refused to approach the probes. Data in (B, C) shown as median with interquartile range. Planned comparisons were performed using a Kruskal-Wallis test followed by Dunn’s post-hoc test. *P < 0.05. SCI, spinal cord injury.

Persistent sensitizing effects of sham surgery on forepaw responses to the probes were particularly evident in the 2-day protocol that tested sensitivity to high 4-mm probes without prior exposure to less elevated probes. During their first exposure to the probes, the sham group exhibited significantly more forepaw withdrawals than the naive group (Fig. 7C), but there was no difference between sham and SCI groups. Statistical comparisons of forepaw withdrawals during the second exposure to the 4 mm probes were not performed because many of the rats in the SCI group completely avoided the probes, staying on the opposite side of the light chamber and facing away from the probes. These results suggest that, when challenged with a novel, moderately noxious substrate, injured rats investigate the substrate more carefully than uninjured rats do before deciding to cross, and this active, voluntary behavior can disclose heightened sensitivity of the forepaws to noxious stimuli (hyperalgesia) long after surgical injury.

3.5. Probe-avoidance behavior is also enhanced by sham surgery for the CCI peripheral nerve injury model

The studies above show that surgical damage to tissues, including muscle and bone, required to expose the spinal cord for controlled contusive injury is sufficient to persistently enhance probe-avoidance behavior. This hyperalgesic effect of sham surgery was shown in the operant MC test but not the von Frey reflex test, suggesting that the MC test may be a more sensitive test for pain evoked by mechanical stimuli. We asked whether the MC test might also reveal hyperalgesia produced by the sham surgery used for a common peripheral nerve injury model. While one previous study used an MC test to show that rats with a CCI of the sciatic nerve exhibit prolonged escape latencies ⁴⁸, this study did not compare MC tests and von Frey tests, nor did it include sham controls. Mice with apparent allodynia (as assessed with von Frey tests) induced by spared nerve injury (SNI) were also found to exhibit prolonged escape latencies ⁷⁸, but this study did not compare the SNI group to an to a sham group lacking drug treatment in the MC test. Thus, whether peripheral sham surgery is sufficient to enhance pain-avoidance behavior in the MC test and whether the von Frey test is a good predictor of pain-avoidance behavior after hindlimb surgery are unknown. To address these questions we used the sciatic nerve CCI model along with its sham surgery control ^{9, 39, 40}. Rats in naive, sham, and CCI groups were tested 14 days post-surgery, a time when reported mechanical allodynia is well established ^{31, 39, 40}. All groups showed significantly reduced crossings during both 4-mm probe trials compared to their baseline crossings (Fig. 8A1). Note the strong similarity in crossings between groups on the 0 mm baseline trial, indicating the prior familiarization trials had eliminated obvious effects of injury on crossing behavior in the absence of noxious probes. While no significant differences among the groups were observed on the first 4-mm trial, the sham group crossed significantly fewer times than the naive group during the second 4-mm trial (Fig. 8A2). Differences in crossings between sham and CCI groups were statistically indistinguishable. Prolonged escape latencies were often observed in the sham and CCI groups (Fig. 8B1, B2), with a majority of injured rats (60%, 6/10 rats in both sham and CCI groups) refusing to cross even once in the second 4-mm trial. In contrast to the robust hyperalgesic effects found in the sham group in the MC test, no evidence of mechanical hypersensitivity (allodynia) was found in the sham group using the von Frey test (Fig. 8C). In contrast and as expected, CCI produced strong mechanical hypersensitivity in the hindpaw ipsilateral but not contralateral to the injury (Fig. 8C, left panel). In sum, sham-operated rats for the CCI peripheral nerve injury model failed to exhibit mechanical hypersensitivity in von Frey tests that provided evidence for allodynia in CCI rats, yet the sham group showed clear evidence of enhanced pain avoidance in the operant MC test.

Figure 8.

Sham surgery for the chronic constriction injury (CCI) of the sciatic nerve enhances pain-avoidance behavior in a mechanical conflict test. Rats were exposed to the 4-mm probes twice (see Figure 1D timeline). (A1) All groups exhibited a reduction in crossings during both 4-mm probe trials compared to the 0-mm baseline (Friedman test). Significance levels for naive (stars), sham (pound sign), and CCI group (plus sign) shown for Dunn’s post-hoc tests. (A2) No significant differences in crossings were observed on the first probe trial, but sham-operated rats crossed fewer times than naive rats during the second probe trial. A majority of sham and CCI rats (60%, 6 out of 10 in both groups) refused to cross during the second probe trial. (B1, B2) Escape latencies reflected the crossing results and revealed a significant difference between naive and sham groups. B2 plots the cumulative percentage of rats in each group exhibiting an escape latency equal to or less than the latency indicated on the x-axis. (C, left panel) In a subset of rats that were tested using the standard reflex test of mechanical sensitivity, the 50% paw withdrawal threshold of the hindpaw ipsilateral to the side of injury was reduced in rats with CCI, but not on the contralateral (uninjured) hindpaw (C, right panel). Note that 50% PWT measures <1 gram are negative after log transformation. Corresponding gram forces shown on right axis. Planned comparisons between groups in (A2, B1, and C) were performed using a Kruskal-Wallis test followed by Dunn’s post-hoc tests. Data shown as mean ± SEM (A1) and median with interquartile range (A2, B1, and C). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. CCI, chronic constriction injury; PWT, paw withdrawal threshold.

4. Discussion

A novel variant of an operant conflict test ⁴⁸ that pits rats’ exploratory drive plus their aversion to bright light against pain elicited by voluntary crossing of noxious probes has revealed that sham surgeries used as controls for neuropathic pain models (SCI and CCI) can increase pain-avoidance behavior for long periods after surgery. This finding has significant implications for preclinical investigations of postsurgical and neuropathic pain.

4.1. Implications of a sensitive measure of pain-avoidance behavior for preclinical investigations of postsurgical pain

Postsurgical pain is a pervasive clinical problem ^{24, 53, 77, 82, 88}. Patients often develop severe postsurgical pain ^{33, 37} that can become chronic ^{81, 89}. Nonetheless, investigations of postsurgical pain are underrepresented in animal research ^{21, 24, 33, 56}. Both clinical and animal data show that postsurgical pain is greater for deeper, more extensive incisions ^{21, 56}. Sham surgeries for most nerve injury models produce deep tissue damage. Furthermore, relatively mild incision models and more severe postsurgical models in rodents can produce transient to chronic signs of behavioral hypersensitivity, ongoing pain, and nociceptor hyperexcitability ^{3, 14, 16, 17, 40, 55, 59, 65, 66, 93}. Our contusive SCI model involves substantial tissue injury, requiring removal of bone (the T10 dorsal process and lamina) and blunt dissection of paravertebral muscles. A recent study demonstrates that sham surgery for SCI increases immune cell trafficking in surrounding muscle and bone 48 hours post-surgery ¹⁵, suggesting that prolonged immune cell activation may play a role in postsurgical pain near the lesion site. Surgery for the CCI model requires exposure of the sciatic nerve and retraction of neighboring skin, muscles, and nerves ³⁴, which may induce postsurgical pain ³⁵. Considering the degree of tissue damage associated with sham surgery procedures for neuropathic pain models, it is not surprising that persistent pain-like alterations are revealed by interaction with noxious probes in the MC device. Postsurgical alterations may involve direct effects on nociceptive pathways and indirect stimulation from movements of inflamed, previously damaged tissues during behavioral tests. Direct alterations of nociceptive pathways have been shown by increased excitability of primary nociceptors observed months after sham surgery for SCI, and expressed as decreased rheobase, increased membrane resistance, and increased incidence of spontaneous activity ^{8, 72}.

Why has so little evidence been reported previously for long-lasting pain following sham surgeries for neuropathic pain models (for exceptions, see ^{51, 54, 83})? One reason may be that motivational states such as pain compete with many other complicating motivational influences, such as fear of a human tester. Appropriate adjustment of competing motivational influences on voluntary behavior, as we did with exploratory drive in the present study, may permit operant responses to reveal otherwise masked states such as persistent postoperative pain. Another reason may be limitations of the reflex tests investigators largely rely on as indicators of allodynia and hyperalgesia. These tests may only be weakly related to motivational and cognitive dimensions of pain ^{68, 86, 96}, and can be highly sensitive to minor variations in procedure, such as the duration of von Frey filament application (Fig. 6A–B). Moreover, because these tests usually depend on hand-delivered stimuli and subjective assessment of reflex responses, they are prone to investigator-dependent sources of variability including unconscious bias and differential exposure to human pheromones ^{12, 80}. Furthermore, results from naive (uninjured) or sham controls are not always reported, and the sham surgery employed (e.g., skin incision alone) may not match the severity of deep surgical procedures used to produce neural injury (e.g., ^{74, 90}).

Previous studies using the MC test reported no evidence for postoperative pain in sham controls for central or peripheral neural injury. In several of these studies, sham controls were not included ^{23, 48, 64}. A spared nerve injury (SNI) study in mice ⁷⁸ demonstrated that SNI, like CCI in rats (which involves a similar surgery) ⁴⁸, increased escape latency, which was attenuated by an analgesic. However, this study did not include a vehicle-treated sham group or naive group so potential effects on pain-avoidance behavior after sham surgery could not be assessed. Interestingly, this mouse study hinted at a bimodal distribution of escape latencies, raising the possibility that an MC protocol like that described here might reveal increased pain avoidance after sham surgeries in mice.

4.2. Implications of enhanced pain avoidance behavior for preclinical investigations of neuropathic pain

The overriding goal of this study was to test the hypothesis that appropriate operant conditions can reveal a persistent pain-like state induced by sham surgeries in common neuropathic pain models, which was tested statistically by planned comparisons between naive and sham-operated groups. We also used planned comparisons to ask whether these conditions would distinguish pain-like behavior measured in the sham versus the SCI or CCI groups. In general, no significant differences were found in the MC tests between sham and SCI or CCI groups, although a weak trend for greater pain-avoidance behavior in the SCI than sham group was sometimes seen (Fig. 5C1–C2). The similarity in operant behavior between sham and neuropathic models indicates that SCI and CCI procedures, like the corresponding surgery that is part of the SCI and CCI procedures, induce persistent pain-avoidance behavior, as reported previously for SCI ^{23, 64} and CCI ⁴⁸. The surprising similarity between sham groups and both the SCI and CCI groups in the 2-day MC protocol suggests that this test is very sensitive for detecting persistent pain-related states, but in its present form it is not optimal for distinguishing qualities or degrees of different types of pain. In contrast to the similar pain-avoidance behavior of SCI and sham groups described here, many studies have found differences between sham and SCI groups in allodynia-like, hyperalgesia-like, and spontaneous pain-like behaviors ^{27, 38, 45, 60, 79, 87}, and their associated neuronal correlates ^{7, 8, 18, 72, 94}.

The open question of whether neuropathic pain plus postsurgical pain is more severe than postsurgical pain alone in these models is an important one that might be addressed using the 2-day MC protocol by adjusting exploratory drive and/or the aversiveness of the noxious probes to a level that prevents multiple crossings by rats with neural injury but not by rats with sham surgery. Although unlikely in light of clinical observations ⁵⁶, the possibility exists that the motivation to avoid evoked pain might not be substantially greater for these models of neuropathic pain than for the postsurgical pain produced by their sham surgeries. Even if evoked pain turns out to be similar after sham surgery and neural injury, another distinction between sham-operated and neuropathic rats is emerging: SCI and some peripheral neuropathic pain models, but not their sham surgeries, can induce persistent ongoing/spontaneous pain as measured by the CPP test ^{4, 42, 50, 57, 76, 91, 94}. Thus, enhancement of evoked pain sensitivity may be relatively similar in animals with postsurgical pain and animals with neuropathic pain, whereas injury involving damage to large nerve branches may be more likely to induce persistent ongoing pain than is tissue injury that spares nerve branches.

Clues about possible mechanistic differences between postsurgical and neuropathic pain might be found in properties of nociceptors likely to be activated by noxious probes in the MC test. Pain-related behavior after thoracic SCI has been correlated with hyperactivity in widespread primary nociceptors ⁸. Antisense knockdown of a Na⁺ channel (Nav1.8) preferentially expressed in nociceptors and required for SCI-induced nociceptor hyperactivity reverses reflex hypersensitivity and prevents CPP after SCI ⁹⁴. This indicates that nociceptor hyperactivity can play a large role in promoting both pain hypersensitivity and ongoing pain after SCI. While nociceptors from SCI rats show dramatic, persistent increases in excitability and spontaneous activity (SA), similar but much weaker alterations were found in nociceptors isolated from sham-operated rats ^{8, 72}. If the altered nociceptor rheobase and membrane resistance produced by sham surgery also occur in nociceptors’ peripheral terminals, firing evoked by noxious stimuli such as the sharp probes in the MC test should be enhanced, contributing to enhanced postsurgical pain-avoidance behavior. Sham-operated rats pooled across these two published electrophysiological studies also showed an increase in the fraction of isolated nociceptors exhibiting SA (20% of neurons in the sham versus 9% in the naive groups), but this incidence was far lower than the 54% of the pooled SCI group showing SA. Given the evidence that SA in nociceptors can drive ongoing pain ⁹⁴, these results plus indications from various neuropathic pain and incision models ^{21, 33, 75, 81} suggest that postsurgical pain may be manifested primarily as enhanced pain evoked by extrinsic stimuli (hyperalgesia and allodynia) whereas neuropathic pain in these models can be manifested both as hypersensitivity to extrinsic stimuli and as ongoing pain. While nociceptor alterations produced by tissue and nerve injury may be indirectly related to allodynia, they probably contribute directly to hyperalgesia and ongoing pain ⁷². Investigation of the effects on sham-surgery-induced pain-avoidance behavior produced by analgesic drugs selective for nerve injury pain and for inflammatory pain should be useful for distinguishing persistent postsurgical pain from related effects of injury, such as anxiety, and for guiding continuing mechanistic analysis.