Disrupted Sleep and Delayed Recovery from Chronic Peripheral Neuropathy Are Distinct Phenotypes in a Rat Model of Metabolic Syndrome

• ❖ Metabolic syndrome is increasingly common in surgical patients and carries increased perioperative risks.

• ❖ A breeding strategy was used to create rats with low intrinsic aerobic capacity that provide a model of metabolic syndrome.

• ❖ These rats have enhanced nociception and disrupted sleep, and show promise for studies of how obesity and diminished aerobic function increase anesthetic and perioperative risks.

METABOLIC syndrome and associated comorbidities have special relevance for anesthesiology.1–3Central obesity is a feature of metabolic syndrome that confers an increased risk of type II diabetes,4and the combination of obesity and diabetes is associated with increased perioperative risks.5–7Obesity increases risk for aspiration during induction,8decreases perioperative tissue oxygenation,9is a predictor for difficult mask ventilation,10and is a risk factor for anesthesia-related maternal mortality.11Metabolic syndrome also is associated with poor sleep,12–20and disrupted sleep can contribute to metabolic syndrome.13,17–19,21,22Metabolic syndrome increases risk for chronic neuropathic pain,23and sleep disruption enhances pain perception.24–28 

The complexity and bidirectional nature of the foregoing pathologies emphasize the need for animal models that can help elucidate mechanisms underlying interactions among metabolic syndrome, sleep, and pain. Two broad observations suggest that aerobic capacity and its attendant subordinate phenotypes are critical in the divide between health and disease, and motivated the development of rat models of disease risk. First, clinical studies reveal a strong statistical relationship between dysfunctional oxygen metabolism and all-cause morbidity and mortality. Indeed, for healthy individuals as well as those with cardiovascular disorders, low exercise capacity is a stronger predictor of decreased survival compared with other conventional risk factors such as body mass index, smoking, hypertension, dyslipidemia, or diabetes.29Second, exercise is an effective prescription for a surprisingly wide range of chronic diseases.30 

These ideas formed the basis for speculating that experimental models that divide for health and disease could be created by selection for low and high intrinsic aerobic endurance running capacity. Starting from a genetically heterogeneous population of rats developed at the National Institutes of Health (Bethesda, Maryland), a divergent (two-way) artificial selection process was used to generate lines of low aerobic capacity runners (LCR) and high aerobic capacity runners (HCR).31At 10 generations of selection, the two lines diverged widely for expected running capacity and hypothesized differences for health risks, including metabolic syndrome.32 

Development of LCR and HCR lines generated a unique model system for dissection of aerobic endurance capacity and its correlated health-related phenotypes. The major hypothesis is that functional alleles at multiple interacting loci that affect intrinsic aerobic capacity have been enriched or fixed differentially between the LCR and HCR lines. It is critical that the models are maintained as genetically heterogeneous lines by using a mating paradigm that minimizes inbreeding.31Compared with inbred strains (in which essentially all loci have been taken to fixation), outbred selected lines maintain genetic complexity that allows combinations of allelic variants at multiple interacting loci to be enriched by selection pressure.33As a result, the LCR-HCR model is better suited to discover epistatic interactions, modifier genes, and synergistic actions.34It is noteworthy that the concurrent breeding of LCR and HCR rats at every generation allows the lines to serve as reciprocal controls for unknown environmental changes.

The present study tested the hypothesis that LCR rats, in addition to demonstrating a predisposition for metabolic syndrome,32,35also have altered nociception and disrupted sleep. The results are consistent with the notion that low intrinsic aerobic capacity and attendant correlated traits such as obesity can at least partly underlie pathology associated with altered nociception and disrupted sleep.

Experiments were conducted in accordance with the Policy on Humane Care and Use of Laboratory Animals established by the National Institutes of Health (publication 80–23). All procedures were reviewed and approved by the University of Michigan Committee on Use and Care of Animals (Ann Arbor). This study used a recently developed rat model selectively bred as LCR and HCR. These rats were derived from a genetically heterogeneous founder population developed at the National Institutes of Health.36As described in detail previously,31a large-scale rotational breeding strategy was started in 1996 to develop lines of rats that differ for intrinsic (i.e. , untrained) aerobic capacity and retain wide heterogeneity. Each rat in the founder population (96 males, 96 females) was tested for maximal endurance running capacity on a speed-ramped treadmill. This test was patterned after clinical treadmill running tests and provided a reliable estimate of aerobic capacity (VO²max) to segregate rats into the lowest and highest capacity runners. Twenty-six mating pairs were selected to create 13 base families of LCR and 13 base families of HCR. Offspring arising from these LCR and HCR breeder animals were tested at 11 weeks of age for intrinsic running capacity. The highest performing male and female in each HCR family and the lowest performing male and female from each LCR family served as parents for each subsequent generation. After 11 generations of selective breeding, LCR and HCR rats differed by almost 350% for running capacity endurance with the LCR rats exhibiting several features consistent with metabolic syndrome.32 

Rats used in the present study were males from generations 20 and greater. Differences in body weight between the two lines are illustrated by figures 1A and 1B. Typical weights ranged from approximately 380 g (HCR) to 550 g (LCR). The difference in aerobic phenotype of distance run-to-exhaustion by HCR and LCR rats has been quantified for generation 26 and is illustrated in figure 1C. Two key points are that (1) these rats were never exercise-trained, and (2) the breeding procedure used made it possible for the present study to evaluate the impact of genetically segregated intrinsic capacity for oxygen metabolism on nociception and sleep. Table 1summarizes previously documented physiologic, behavioral, and metabolic differences between HCR and LCR rats.

There is a convergence of evidence that sleep neurobiology can help elucidate some of the mechanisms by which anesthetics cause a loss of wakefulness.37–41Therefore, the goal of the first experiment was to phenotype states of sleep and wakefulness as a prelude to future studies designed to characterize responses of LCR and HCR rats to intravenous and volatile anesthetics. The procedures used to study sleep in rats have been described in detail.42,43In brief, eight male LCR rats and eight male HCR rats were anesthetized with 3% isoflurane in 100% oxygen (O²), delivered at a flow rate of 1 l/min. Delivered isoflurane concentration was measured by spectrometry (Cardiocap™/5; Datex-Ohmeda, Louisville, CO). Once anesthetized, each rat was placed in a Kopf Model 962 small animal stereotaxic frame with a Model 906 rat anesthesia mask (David Kopf Instruments, Tujunga, CA), and delivered isoflurane concentration was decreased to 1.5%. Core body temperature was maintained at 37°C throughout the surgical procedure with the use of a recirculating heat pump (Gaymar Industries, Inc., Orchard Park, NY). A midline scalp incision was made and electrodes were implanted for electroencephalogram and electromyogram recording. The cortical electroencephalogram electrodes (8IE36320SPCE; Plastics One, Roanoke, VA) were screwed into the cranium at three sites relative to bregma: 2.0 mm posterior, 1.3 mm lateral; 2.0 mm posterior, −1.5 mm lateral; and 1.0 mm anterior, 1.5 mm lateral. Two electromyogram electrodes were constructed from AS632 biomed wire (Cooner Wire, Inc., Chatsworth, CA) and were placed into the dorsal neck muscles. Another electrode was placed subcutaneously near the dorsal aspect of the skull and served as an indifferent electrode. Leads from all six electrodes were placed together in a six-pin multichannel electrode pedestal (MS363, Plastics One) that was secured to the skull with three anchor screws and dental acrylic (Jet Acrylic Self Curing Resin and Liquid; Lang Dental Manufacturing Company, Inc., Wheeling, IL).

After surgery, rats were allowed to recover for 7–10 days while habituating to the sleep recording chambers. Habituation consisted of placing the rats in the recording chambers and connecting the electrode pedestal on the skull to a six-channel cable (363-441/6 80cm 6TCM, Plastics One) that led to amplifiers and a computer. Housing and sleep recordings occurred in a 12:12 light-dark cycle with transitions occurring at 8:00 am and 8:00 pm. States of sleep and wakefulness were recorded across 24 h in two continuous 12-h segments that coincided with light and dark cycle start times. The electrographic signals were amplified, filtered (electroencephalogram, 0.3–30 Hz; electromyogram, 10–100 Hz), and recorded at 128 Hz. After habituation, states of sleep and wakefulness were recorded for 24 h. For each 24-h digital recording, 10-s bins were scored as wakefulness, rapid eye movement (REM) sleep, or non-REM (NREM) sleep. During recording sessions, the rats could move freely and had ad libitum  access to food and water.

A second group of LCR (n = 6) and HCR (n = 6) rats was used to phenotype mechanoreceptor thresholds by measuring paw withdrawal away from 2, 8, and 15 g of pressure caused by von Frey hairs applied to the bottom of a hind paw (Touch Test Sensory Evaluator Kit; Stoelting Co., Wood Dale, IL). As described previously,44,45these nylon hairs were slowly pushed against the skin until the hair bent. Each of the three hairs was applied five times in ascending order. Paw withdrawal was recorded as a positive response and the percentage response to each von Frey hair was tabulated and averaged for each animal.

As described previously,46rats were conditioned daily to being placed in a chamber for nociceptive testing (model 336T; IITC Life Science, Inc., Woodland Hills, CA). The Hargreaves device made it possible to quantify the latency in seconds to paw withdrawal (PWL) away from onset a light beam focused on the hind paw.47On the day of testing, animals were placed in the Hargreaves chamber and allowed to habituate for 60 min before obtaining measures of PWL. The light beam and an electronic timer were activated simultaneously. When the rat lifted its hind paw, the heat source and timer were stopped and PWL was recorded.46,48,49Five measurements were obtained and averaged during 5 min of testing.

A unilateral chronic peripheral mononeuropathy50of the sciatic nerve was created in adult male LCR (n = 7) and HCR (n = 7) rats. Before surgery, each rat was anesthetized with 3% isoflurane until breathing rate decreased to 60 breaths per min. After loss of wakefulness, the rat was fitted with a Kopf Model 906 rat anesthesia mask and isoflurane concentration was lowered to 1.5% for the remainder of the procedure. Delivered isoflurane concentration and core body temperature were measured. The sciatic nerve was accessed via  hind limb incision followed by blunt dissection of the biceps femoris. At a level proximal to its trifurcation, 5–8 mm of nerve were separated from tissue with minimal damage to surrounding muscle and tissue.50Three polyglycolic acid sutures were placed around the sciatic nerve at 1-mm intervals, loosely enough to prevent complete ligation but tightly enough to induce a leg twitch response.50The wound margin was sutured closed and rats were allowed 6 days to recover from surgery.

Thermal nociceptive testing was performed before and after the chronic constriction injury. A beam of light, which served as a noxious thermal stimulus, was directed through a glass floor and onto the plantar aspect of the hind paw. This light remained on until the rat withdrew its paw from the thermal stimulus. PWL was defined as the time between commencement of the light beam and withdrawal of the hind paw.47PWL was measured for all rats the day before sciatic nerve ligation and on days 7, 14, 21, 28, 35, and 42 after nerve ligation. During this 6-week testing period, rats were conditioned in the testing chamber for 2 h daily. At the beginning of each testing day, rats were habituated to the chambers for 1 h before five PWL measurements were obtained with the Hargreaves method applied to the hind paw of the operated and nonoperated limb. To prevent tissue damage during nociceptive testing, the thermal light stimulus was set to terminate automatically (cut-off time) if there was no paw withdrawal after 18 s.

This study tested the one-tailed hypothesis that rats bred for low intrinsic aerobic capacity have enhanced nociception and disordered sleep. The 24-h recordings of sleep and wakefulness were scored for 138,240 epochs, each 10 s in duration. To avoid inflated degrees of freedom, data were averaged for each of the 16 rats. Dependent measures included the percentage of recording time spent in wakefulness, NREM sleep, and REM sleep; episode duration and number of episodes for each of the three states; and number of transitions between states. The data for percent of time spent in each state were normally distributed and fit the assumptions of two-way analysis of variance for repeated measures. The data describing recovery from chronic constriction nerve injury as a function of time were also analyzed using two-way analysis of variance for repeated measures. Post hoc  comparisons were made using Mann-Whitney and Bonferroni statistics. The von Frey data were not normally distributed and were analyzed using the Mann-Whitney statistic. Software used for statistical analyses included SAS (version 9.2; SAS Institute, Inc., Cary, NC), GB-Stat (Dynamic Microsystems, Inc., Silver Spring, MD), and Prism (version 5; GraphPad Software, Inc., La Jolla, CA).

Poincaré analyses are particularly useful for quantitatively evaluating the periodic nature of bistable rhythms such as the cardiac cycle,51breathing,52,53and sleep.54Poincaré analyses were used in the present study to compare LCR and HCR animals for their ability to maintain stable episodes of REM sleep. Poincaré analysis made it possible to quantify two indices of variability. Short-term variability (SD1) expressed the variability in time from one REM sleep episode to the next REM sleep episode. Long-term variability (SD2) quantified the overall variability in duration of REM sleep epochs across the 24-h recording. Calculation of SD1 values for LCR and HCR rats was achieved with a four-step process: (1) duration in seconds of each REM sleep interval was plotted y  relative to x , representing the duration of the previous REM sleep interval. (2) After all REM sleep epochs were plotted, the line of identity (x =y ) was calculated and added to the Poincaré graph. (3) The Pythagorean Theorem was used to calculate the perpendicular distance between each data point and the line of identity. (4) The standard deviation of these raw distances was calculated as the SD1 value for each LCR and HCR rat, and these values were averaged across all LCR and HCR rats, respectively. SD2 values were obtained using the first step noted above and calculating the distance of each point from a line perpendicular to the line y =x  and running through the mean duration of REM sleep intervals (i.e. , centroid of data points). As a final step, individual animal means of SD1 and SD2 for each LCR and HCR rat were used to evaluate rat strain differences by t  test. For all inferential statistics a P  value of less than 0.05 was considered significant.

Figure 2illustrates the temporal distribution of sleep and waking states recorded continuously during a 24-h period for one representative LCR rat and one representative HCR rat. Figure 3summarizes the average of eight 24-h recordings for each strain and quantifies strain-specific differences in sleep architecture. During the 24-h recordings, LCR rats spent significantly (P < 0.009) more time in wakefulness and significantly (P < 0.01) less time in NREM sleep than HCR rats (fig. 3A). These differences were not apparent in the dark portion of the light-dark cycle (fig. 3B) and resulted predominantly from significant differences in wakefulness (P < 0.007) and NREM sleep (P < 0.005) that occurred during the light phase of recording (fig. 3C). LCR and HCR rats did not differ in the percentage of recording time spent in REM sleep (fig. 3).

Figure 4summarizes the variability in duration of REM sleep episodes characteristic of LCR and HCR animals. Comparison of the Poincaré distributions for HCR (fig. 4A) and LCR (fig. 4B) rats indicates differences in the variability of REM sleep duration. Figure 4Cillustrates the significantly (P < 0.05) greater SD2 in REM-sleep episode duration characteristic of LCR rats. Short-term variability (SD1) in a Poincaré analysis quantifies the duration of one REM sleep episode (REMⁱ) relative to the previous REM sleep episode (REMⁱ−¹). The differences in short-term variability between LCR and HCR rats were not significant. Long-term variability (SD2) quantifies the variability in REM-sleep episode duration as a function of rat strain. Figure 4Cshows that SD2 was greater for LCR than HCR rats. Thus, LCR animals could initiate but not maintain stable REM sleep episodes. Comparing REM sleep data in figures 2–4demonstrates how Poincaré analyses made it possible to quantify differences in the temporal organization of REM sleep.

The results of the Poincaré analyses encouraged a more detailed comparison of the temporal organization of sleep-wake states, and figure 5summarizes those results. Averaging across 24 h, LCR animals had significantly (P < 0.001) more transitions from wakefulness to NREM sleep and significantly (P < 0.001) more transitions from NREM sleep to wakefulness than HCR rats (fig. 5A). Consistent with results presented in the figure 1, across the 24-h recording, LCR rats had a significantly (P < 0.04) greater number of wakefulness episodes than HCR rats (fig. 5B). When LCR rats had NREM sleep, the average duration of each NREM sleep episode was significantly (P < 0.002) shorter than that of HCR rats (fig. 5C).

There was no difference between LCR and HCR rats in mechanosensory response as measured by von Frey testing (fig. 6A). Power calculation indicated that 61 rats would be needed to demonstrate a difference in von Frey response at a P  value lower than 0.05. As described previously,46LCR rats revealed a modest but significantly shorter PWL in response to a thermal nociceptive stimulus (fig. 6B). Figure 6Cshows the time course of PWL measured in LCR and HCR rats 1 week before and 42 days after chronic constriction injury of one sciatic nerve.50Each of the seven points in figure 6Csummarizes mean ± SD PWL in response to a thermal stimulus. Before the peripheral neuropathy (preinjury) there was no statistically significant difference in PWL between LCR and HCR animals. For LCR rats (fig. 6C), two-way analysis of variance for repeated measures revealed a significant effect of peripheral neuropathy on PWL. Post hoc  tests revealed significantly (P < 0.05) enhanced nociception in LCR rats on days 7, 14, 21, and 28 after peripheral neuropathy. Measures of PWL in HCR rats (fig. 6C) revealed significantly (P < 0.05) enhanced nociception on days 7 and 14 after chronic constriction injury of the sciatic nerve.

In 2000, the National Task Force on the Prevention and Treatment of Obesity reported delayed recognition of the fact that obesity is a significant health risk.55Although relevance of obesity for anesthesiology was noted in 1975,56only recently has the interaction among obesity, sleep, and pain become apparent.3,5,37,57Obesity has replaced smoking as a major disease burden58and projection data indicate that by 2020, half of all Americans may be obese.59The results are discussed relative to emerging evidence that obesity is associated with disordered sleep, potentially causing hyperalgesia and increase opioid requirement for effective pain management.

Data from LCR rats are consistent with human data documenting a relationship among metabolic syndrome, sleep, and nociception. For example, humans with poor physical fitness and a high body mass index exhibit disrupted and diminished sleep compared with those who maintain a normal body mass index.18,20,60Humans with metabolic syndrome and type II diabetes spend more time in bed and less time asleep.12,13,16,17 

Human data also support the view that reduced sleep leads to increased body mass and decreased fitness by way of increasing cortisol, interleukin-6, and ghrelin, and decreasing leptin.17–19Thus, shortened sleep has been associated with decreased fitness. The present results show for the first time that diminished aerobic fitness in LCR rats is characterized by disordered sleep. In addition, low intrinsic aerobic capacity was associated with decreased NREM sleep time and increased time awake (figs. 2 and 3). Sleep disruption in rats with low aerobic capacity was also characterized by greater variability in REM sleep duration (fig. 4) and more transitions between wakefulness and NREM sleep (fig. 5) compared with rats bred for high intrinsic aerobic capacity. The decreased amount of sleep and inability to maintain sleep characterizing this rodent model of metabolic syndrome31,32,61resemble features of human insomnia. In support of the potential clinical relevance of this rodent model, the disordered sleep and short sleep duration characteristic of human insomnia are risk factors for type II diabetes.62The present results encourage future studies designed to obtain long-term recordings of sleep and wakefulness from LCR and HCR rats.

Obesity and insulin resistance,63–65specific features of metabolic syndrome, are associated with painful, idiopathic neuropathy,23enhanced nociception,66,67and disordered sleep.68–71Although data clearly document an association between metabolic syndrome and pain, no previous studies have characterized the effect of metabolic syndrome on the resolution of experimentally imposed, chronic pain. Compared with HCR rats, the LCR phenotype revealed a delayed recovery from chronic peripheral mononeuropathy (fig. 6C). Results presented in figure 6C, and the fact that LCR and HCR rats have been bred for the intrinsic (i.e. , untrained) component of aerobic capacity, imply that genetic and metabolic factors altered the resolution of the neuropathy. This implication fits with the finding that human obesity is associated with increased hospital admissions and longer hospital stays.72The disordered sleep characteristic of LCR rats may also contribute to hyperalgesia. In normal, pain-free humans, sleep loss causes hyperalgesia25,28and nonrestorative sleep affects pain management.73Opioids disrupt NREM sleep and inhibit the REM phase of sleep.38REM sleep disruption has consistently been shown to diminish the analgesic properties of opioids.74 

There is good precedent for animal models contributing to clinical anesthesiology. Rodent studies have provided the background for spinal drug administration now routine in clinical care,75porcine models helped elucidate the mechanisms of malignant hyperthermia,76and unique features of goat brain vasculature helped differentiate brain versus  spinal cord sites of anesthetic action.77All models have a limited domain of applicability and the strengths and limitations of LCR and HCR rats have been described in detail.61A known limitation of sensory threshold testing is that the data are characterized by a high degree of variability.78,79This variability may contribute to the lack of rat strain–specific differences in response to von Frey testing (fig. 6A), as well as differences between LCR and HCR rats during acute nociceptive testing (fig. 6B), but not during preinjury nociceptive testing (fig. 6C). Likewise, the complexity of animal models of nociception is recognized80and future nociceptive testing should include additional pain models and sensory modalities.

The polygenic nature of complex diseases emphasizes the advantages of models created by the selective breeding for naturally occurring traits over transgenic models that manipulate a single gene or gene product.61This advantage is particularly clear for efforts to model hypertension, metabolic syndrome, and disordered sleep, all of which complicate anesthesia care. The potential for LCR and HCR rats to provide a unique resource for anesthesiology research is supported by evidence emphasizing the biologic centrality of oxygen uptake and delivery61and the special relevance of oxygen for anesthesiology.81 

The present findings suggest homology between obese human and rat for the traits of sleep fragmentation and nociception, and support the interpretation that LCR rats develop some features of metabolic syndrome.32Sleep apnea is the second most common sleep disorder and a potential problem for anesthesia care.82This finding is consistent with evidence that obstructive sleep apnea, similar to metabolic syndrome, is a systemic disease.83The present results encourage future studies using whole body plethysmography53to characterize the actions of intravenous and volatile anesthetics on the respiratory control of LCR and HCR rats. In view of the fact that adenosine is antinociceptive84,85and exercise increases brain adenosine,86future studies also will determine the extent to which adenosinergic neurotransmission87is one mechanism through which intrinsic aerobic capacity modulates sleep and nociception.

For expert assistance, the authors thank Sha Jiang, B.S. (Research Associate), Mary A. Norat, B.S. (Senior Research Associate), Sarah L. Watson, B.S. (Senior Research Associate), and Elizabeth Gauthier, B.S. (Research Assistant), from the Department of Anesthesiology, University of Michigan, Ann Arbor, Michigan, as well as Kathy Welch, M.A., M.P.H. (Statistician Staff Specialist), Center for Statistical Consultation and Research, University of Michigan. They also thank Lori Gilligan, L.V.T. (Animal Technician Senior), and Nathan Kanner, B.A. (Senior Research Laboratory Technician), also from the Department of Anesthesiology, University of Michigan, for their expert care of the rat colony.