Leptin Status Alters Buprenorphine-Induced Antinociception in Obese Mice with Dysfunctional Leptin Receptors

Abstract

Buprenorphine is an opiate used for pain management and to treat opiate addiction. The cytokine leptin can modulate nociception, but the extent to which buprenorphine-induced antinociception varies as a function of leptin signaling has not been characterized. Four congenic mouse lines with phenotypes that include differences in body weight and leptin status were used to test the hypothesis that the antinociceptive effects of buprenorphine vary as function of sex and leptin signaling. Each mouse line was comprised of males (n=12) and females (n=12) for a total of 96 animals. Groups included C57BL/6J (B6) mice (wild type), B6 mice with diet-induced obesity (DIO), obese B6.Cg-Lep^ob/J (ob/ob) mice lacking leptin, and obese B6.BKS(D)-Lepr^db/J (db/db) mice with dysfunctional leptin receptors. The dependent measure was tail flick latency (TFL) in seconds for mouse-initiated tail removal from a warm water bath. Independent variables were intraperitoneal administration of saline (control) or buprenorphine (0.3 mg/kg). Within every mouse line, buprenorphine significantly increased TFL relative to saline. Compared to the other mouse lines, db/db mice with dysfunctional leptin receptors had a significantly longer TFL after saline and after buprenorphine. TFL did not vary significantly by body weight or sex. The results provide novel support for the interpretation that acute thermal nociception is associated with altered leptin signaling.

1. Introduction

Buprenorphine is a mu opiate receptor agonist and kappa opiate receptor antagonist used for pain management [6] and for treating opiate use disorder [13, 14]. Most data regarding buprenorphine have been derived from studies of normal weight males. In contrast, morphine is known to cause sex-specific differences in control of human breathing [10]. Closed claims analyses reveal that being female and obese are the greatest risk factors for opiate-related adverse events [18]. Obesity is co-morbid with increased reports of pain in humans [2, 24] and with altered nociception in mice [29, 34, 37].

Obesity is a proinflammatory disorder with the potential to modulate nociception via multiple tissue systems [9]. Adipocytes secrete leptin, a proinflammatory cytokine that also functions to signal satiety [3]. Nociceptive processing is altered in obese mice [26, 27, 33] and previously we have shown that leptin replacement can restore thermal nociception in obese, leptin-deficient mice [34, 37]. We are aware of no comparable data from mice determining whether the antinociceptive effects of buprenorphine are altered by obesity and/or sex. The current study used four congenic lines of mice with normal or dysfunctional leptinergic transmission. These four lines of mice made it possible to test the hypothesis that the antinociceptive effects of buprenorphine vary as a function of sex and leptin signaling. Preliminary results have been presented [20].

2. Materials and methods

2.1. Animals

All procedures using animals were approved by the University of Tennessee Institutional Animal Care and Use Committee and were conducted in accordance with Guide for the Care and Use of Laboratory Animals (The National Academies Press, 8^th Ed., Washington, D.C., 2011). Adult male and female mice age 6-8 weeks were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). The congenic mouse lines included: C57BL/6J (B6) mice (wild type), B6 mice fed a 60% fat diet (D12492, Research Diets, Inc., New Brunswick, NJ, USA) to create diet-induced obesity (DIO), obese B6Cg-Lep^ob/J (ob/ob) mice that do not produce leptin, and obese B6.BKS(D)-Lepr^db/J (db/db) mice that produce leptin and have dysfunctional leptin receptors. None of these mice are transgenic animals. The ob/ob and db/db mice were discovered as spontaneous mutations in the B6 line. Quality control protocols at The Jackson Laboratory regularly confirm genetics and leptin status of mouse lines [21] via End Point Genotyping to insure the presence of the ob/ob and the db/db mutations. Additionally, genome scans confirm that these congenic lines are on a B6 background.

Each mouse line included 12 males and 12 females for a total sample size of 96. Mice were housed in a temperature-controlled environment and provided with ad libitum access to food and water. B6, ob/ob, and db/db mice were fed Teklad 22/5 rodent chow containing 5.5% fat.

2.2. Nociceptive testing and drug administration

The dependent measure was time in seconds for mouse-initiated tail removal from a water bath maintained between 48° and 50° C. These measures of tail flick latency (TFL) provide an established index of acute thermal nociception [8, 25]. For one week prior to quantifying TFL, mice were conditioned for 2 min to being placed in an acrylic tube with their tails extending outside the tube. During data collection, the distal third of the tail was lowered into the water bath. A timer was started when the tail was immersed in the water. The timer was stopped when the mouse flicked its tail out of the water. Each of three TFL measures was separated by a 10-second interval of no stimulation. A cutoff time of 15 seconds was implemented to prevent tissue damage, and this 15-second interval served as the maximum possible TFL value. Injection volume was 0.3 mL. The 0.3 mg/kg dose of buprenorphine reliably produces antinociception in mice [19, 28] and buprenorphine can have an antinociceptive effect on mouse TFL for as long as 270 min after administration [15]. Based on these previous findings, TFL measures were collected 120 min after intraperitoneal (i.p.) administration of saline (vehicle control) or buprenorphine (0.3 mg/kg). Studies of acute thermal nociception often compare pre-injection (baseline) and post-injection (treatment) measures in order to normalize the data as a percent maximum possible effect. As cautioned previously [1], normalization using percent maximum possible effect is misleading for nociceptive studies involving subjects that are known to differ by strain or line. Therefore, it was not appropriate in the present study to normalize the TFL measures.

2.3. Data analysis

Statistical analyses were performed using Prism 7.0b. To circumvent the potential confound of inflated degrees of freedom, all statistical analyses were performed using TFL values expressed as a mean for each mouse. The data from each of the four lines of mice were confirmed via D'Agostino and Pearson tests to satisfy the assumption of a normal distribution. The hypothesis that the antinociceptive effects of buprenorphine varied as a function of sex and leptin signaling was evaluated using two-way analysis of variance (ANOVA) followed by post hoc multiple comparisons tests. The magnitudes of the treatment effects within each mouse line were quantified with Cohen's d statistic. For each of the four mouse lines and for each drug condition, regression coefficients (r²) quantified the amount of variance in TFL that was accounted for by body weight.

3. Results

Comparisons of nociception after saline or buprenorphine administration revealed no significant differences in TFL between males and females (Table 1). Subsequent analyses thus combined TFL measures from male and female mice within each congenic line (Fig. 1). The increase in TFL caused by buprenorphine for each mouse line was B6=55.7%, DIO=49.5%, ob/ob=38.3%, and db/db=23%. Two-way ANOVA showed statistically significant differences in TFL as a function of mouse line (F=32.09; df=3,184; P<0.0001) and buprenorphine (F=102.3; df=1,184; P<0.0001). Post-hoc analyses indicated that db/db mice had significantly (P<0.0001) longer TFLs than B6, DIO, and ob/ob mice after administration of saline or buprenorphine.

Table 1.

Tail flick latency in seconds (mean ±SD) did not differ significantly between males and females for each of four congenic mouse lines. Values are based on 12 male and 12 female mice per group. B6 = C57BL/6J (wild-type mice); DIO = B6 mice with Diet-Induced Obesity (mice are obese); ob/ob = B6.Cg-Lep ^ob/J (mice are obese and leptin deficient); db/db = B6.BKS(D)-Lepr ^db/J (mice are obese and lack leptin receptors).

Treatment	Sex	B6	DIO	ob/ob	db/db
Saline	Males	3.11 ±0.67	2.94 ±0.59	3.27 ±0.51	5.04 ±0.58
	Females	3.40 ±0.85	3.23 ±0.39	2.99 ±0.62	4.85 ±1.16
Buprenorphine	Males	5.09 ±1.33	4.75 ±0.90	4.48 ±0.71	5.83 ±1.73
	Females	5.05 ±0.72	4.32 ±0.44	4.52 ±1.22	6.33 ±1.97

Fig. 1.

Mean and standard deviation (SD) tail flick latency (TFL) in seconds among four congenic lines of mice. In all four lines of mice buprenorphine significantly (*) increased TFL. Mice (db/db) with dysfunctional leptin receptors displayed a significantly longer TFL than the other three congenic lines after saline (#) and after buprenorphine (+) administration.

Cohen's d statistic provided an additional metric for phenotyping the magnitude of the treatment effect. All Cohen's d values were large, indicating non-overlap in the TFL distributions after buprenorphine relative to saline: DIO (d=2.5), B6 (d=2.0), ob/ob (d=1.6), and db/db (d=0.88). Expressed as percent of non-overlap, the Cohen's d values revealed non-overlapping distributions ranging from >98% percent non-overlap for B6 and DIO mice to 73% for ob/ob, and 52% for db/db mice. Thus, the significant differences in TFL as a function of mouse line (Fig. 1) were not an artifact of inflated degrees of freedom associated with large sample sizes.

ANOVA confirmed that body weight varied significantly (F=142; df=3,92; P<0.0001) as a function of congenic line. Body weights (mean ± standard deviation in g) at the time of TFL testing were B6 female (19.8±1.2), B6 male (27.4±1.9); DIO female (30.7±2.9), DIO male (43.5±5.4); ob/ob female (48.4±2.9), ob/ob male (46.4±2.3); db/db female (49.4±2.0), db/db male (48.0±1.9). Post hoc analyses showed that body weights of DIO, ob/ob, and db/db mice were significantly (P<0.0001) greater than body weight of control (B6) mice. Our assessments of higher-level phenotypes such as body weight, hyperphagia, elevated blood glucose levels, and coat condition were consistent with genotypes comprising the four congenic mouse lines used in the present study https://www.jax.org/jax-mice-and-services/find-and-order-jax-mice/why-jax-mice/genetic-quality-control-program.

Post hoc tests found that body weight of females was significantly (P<0.0001) less than males for B6 and DIO mice, but not for ob/ob and db/db mice. Regression analyses were used to evaluate the percent of variance in TFL that was attributed to body weight. Body weight did not account for a significant amount of variance in TFL in any mouse line after administration of saline or buprenorphine (Fig. 2).

Fig. 2.

Tail flick latency in seconds plotted as a function of body weight for each congenic line of mice after administering saline (A) or buprenorphine (B). Each frame provides the solution to y=mx+b and the amount of variance in TFL accounted for by body weight (r²). The slopes of each of the eight functions did not differ from zero.

4. Discussion

The interacting epidemics of opiate abuse [5] and obesity [30, 38] accentuate the need to better understand the association between opiates, obesity, and nociception. Given that leptin is produced by adipocytes, differentiating the extent to which nociception is modulated by leptin per se versus obesity is not trivial [37]. In an effort to evaluate the contributions of obesity and/or leptin to buprenorphine-induced antinociception, the present study used mice that share the obesity phenotype but differ in leptin signaling. The discussion focuses on the novel finding that obese mice with dysfunctional leptin receptors had the longest latency for antinociception under control conditions and a diminished antinociceptive response to buprenorphine. Differences in nociception identified in the present study also are relevant for mouse models of prevalent human disorders such as diabetes [22] and alterations in morphine pharmacokinetics associated with obesity [11].

Sex-dependent differences in TFL were not observed in any of the four mouse lines (Table 1). The absence of a sexually dimorphic effect of buprenorphine on TFL can be contrasted with results from 12 strains of rats in which opiates had a more potent antinociceptive effect in males than females [32]. In addition to species differences, comparison of the present and previous [32] studies reveals significant differences in methodology. In the rat study TFL was assessed 15 min after opiate administration, whereas in the present study TFL was measured 2 h after buprenorphine administration. We reasoned that testing 2 h after injection would provide a more conservative test of our hypothesis and be consistent with previous data indicating long-duration antinociceptive effects of buprenorphine administered to mice [15].

DIO, ob/ob, and db/db mice share the obese phenotype, but only mice with dysfunctional leptin receptors (db/db) had a diminished antinociceptive response to buprenorphine (Fig. 1). In normal weight B6 mice that produce leptin, buprenorphrine increased TFL by about 56%. In obese db/db mice, buprenorphrine increased TFL only by 23%. The mechanisms causing the diminished antinociceptive effects of buprenorphine in db/db mice are unknown. Many studies have used db/db mice as a model of type-2 diabetes [29]. Previous data, however, have shown that when db/db mice are fed a normal diet, they become euglycemic and do not develop peripheral neuropathy [29]. In the present study only DIO mice received a high fat diet.

Effect size statistics revealed that the well-documented antinociceptive actions of buprenorphine were greatest in lean B6 and obese DIO mice that produce leptin. Although the mechanisms are unknown, these results are consistent with recent evidence that obesity decreases the metabolism of morphine and that leptin normalizes morphine metabolism [11]. The extent to which buprenorphine metabolism is or is not altered in the congenic lines used in the present study is not known.

As noted previously [37] the relationship between leptin and obesity adds to the complexity of determining the degree to which leptin status and/or obesity alter nociception. The Fig. 1 findings prompted analyses quantifying the percent of variance in TFL that could be attributed to body weight (Fig. 2). For no treatment condition and for no congenic mouse line did body weight account for a significant amount of variance in TFL (Fig. 2). Considered together, the results shown in Figs. 1 and 2 support the novel conclusion that differences in leptin signaling among the four lines of mice had a greater impact on buprenorphine modulation of TFL than did body weight.

TFL was altered to a greater extent by leptin status (Fig. 1, ob/ob vs db/db) than body weight (Fig. 2) and this is noteworthy relative to mechanistic data. Leptin inhibits the excitability of neurons in the peribrachial pons [4]. Neurons in the adjacent and overlapping parabrachial pons contribute to ascending and descending nociceptive signaling [16]. Coupled with evidence that inflammatory cytokines promote hyperalgesia [36], these foregoing studies of pontine neurons raise the question of whether the hypoalgesia displayed by db/db mice (Figs. 1 and 2) is related to dysfunctional leptin signaling in peri- and parabrachial neurons. This possibility is open to future investigation and the present results encourage such studies.

In summary, the present use of four congenic mouse lines permits novel inferences regarding the relative influence of sex, leptin status, and body weight on the antinociceptive effects of buprenorphine. The results support the conclusion that leptin signaling among the four lines of mice had a greater impact on buprenorphine modulation of TFL than did body weight. The results should not be misinterpreted, however, as inferring that only leptin modulates TFL. Adipose tissue functions as an endocrine organ secreting more than 600 different adipokines [17]. Thus, leptin signaling is likely only one of many adipocyte-derived molecules that influence nociception. Administering leptin protein to leptin-deficient ob/ob mice [7] or neuron-specific LEPR-B transgenes to db/db mice [12] has long been known to reverse the obesity phenotype and associated dysfunctions. Serum leptin levels are routinely measured during studies designed to test the hypothesis that administering exogenous leptin to leptin-deficient mice modulates breathing [23, 31], nociception [34, 37], or opiate pharmacokinetics [11]. Measuring leptin in the present study was not appropriate, and to do so would have comprised duplicative research for two reasons. First, the experiments were not designed to evaluate whether exogenous leptin administration could rescue nociception. Second, there was good agreement between our assessment of higher-level phenotypes, such as body weight, food consumption, and blood glucose levels, and The Jackson Laboratory's allelic assays confirming the identity of wild type and mutant mice. The present success in differentiating the effects of leptin status versus body weight on thermal nociception is an essential first step encouraging future studies of nociception in relation to functional differences in white, beige, and brown adipose tissue [35].

Highlights.

• Nociception is reduced in mice with dysfunctional leptin receptors

• Mice with dysfunctional leptin receptors have a diminished antinociceptive response to buprenorphine

• Buprenorphine-induced thermal antinociception did not vary by sex or body weight