Selection of preferred thermal environment and cold-avoidance responses in rats rely on signals transduced by the dorsal portion of the lateral funiculus of the spinal cord

ABSTRACT

Thermoregulatory behaviors are powerful effectors for core body temperature (T_c) regulation. We evaluated the involvement of afferent fibers ascending through the dorsal portion of the lateral funiculus (DLF) of the spinal cord in “spontaneous” thermal preference and thermoregulatory behaviors induced by thermal and pharmacological stimuli in a thermogradient apparatus. In adult Wistar rats, the DLF was surgically severed at the first cervical vertebra bilaterally. The functional effectiveness of funiculotomy was verified by the increased latency of tail-flick responses to noxious cold (−18°C) and heat (50°C). In the thermogradient apparatus, funiculotomized rats showed a higher variability of their preferred ambient temperature (T_pr) and, consequently, increased T_c fluctuations, as compared to sham-operated rats. The cold-avoidance (warmth-seeking) response to moderate cold (whole-body exposure to ~17°C) or epidermal menthol (an agonist of the cold-sensitive TRPM8 channel) was attenuated in funiculotomized rats, as compared to sham-operated rats, and so was the T_c (hyperthermic) response to menthol. In contrast, the warmth-avoidance (cold-seeking) and T_c responses of funiculotomized rats to mild heat (exposure to ~28°C) or intravenous RN-1747 (an agonist of the warmth-sensitive TRPV4; 100 μg/kg) were unaffected. We conclude that DLF-mediated signals contribute to driving spontaneous thermal preference, and that attenuation of these signals is associated with decreased precision of T_c regulation. We further conclude that thermally and pharmacologically induced changes in thermal preference rely on neural, presumably afferent, signals that travel in the spinal cord within the DLF. Signals conveyed by the DLF are important for cold-avoidance behaviors but make little contribution to heat-avoidance responses.

Introduction

Body temperature (T_b) is regulated by autonomic (e.g., nonshivering thermogenesis and skin vasoconstriction), other physiological (shivering), and behavioral effector mechanisms [1,2]. Examples of thermoregulatory behaviors include nest-building, huddling, postural changes, and selection of ambient temperature (T_a). The selected T_a is typically called “preferred” (T_pr), and searching for a preferred thermal environment is a proactive innate behavior, which is ubiquitous in the animal world [2–4]. Furthermore, changing the environment to either warmer or colder can be life-saving, even when the immediate challenge is non-thermal, as in various forms of infection and systemic inflammation [5–7]. In the laboratory, the “spontaneous” selection of T_pr and the closely related induced avoidance behaviors, i.e., warmth-avoidance (cold-seeking) and cold-avoidance (warmth-seeking) responses to various thermal and non-thermal stimuli are often studied in a thermogradient apparatus, a.k.a. thermocline device [1,8–14].

The neural pathways mediating most thermoregulatory behaviors, including the selection of T_pr, are not fully established [2]. Yahiro et al. [15] reported that the inactivation of neurons in the lateral parabrachial nucleus (LPB) suppressed innocuous cold- and warmth-avoidance behaviors in rats. This finding agrees with our results (M. C. Almeida and A. A. Romanovsky, unpublished observation; reviewed in Ref. [2]) showing that bilateral electrolytic lesioning of the LPB attenuated warmth-seeking response to cold exposure in rats in a thermogradient apparatus. It has also been thoroughly documented that the LPB is involved in pathways controlling physiological thermoeffectors, and that it receives thermal information from the periphery, including the skin, through the spinoparabrachial tract [16]. The latter ascends in the spinal cord within the dorsal portion of the lateral funiculus, which we refer to in this paper as the dorsolateral funiculus (DLF) (reviewed in Refs. [17,18]).

In this study, we asked whether bilateral DLF transection would affect the spontaneous selection of T_pr in a thermogradient apparatus and attenuate changes in thermal preference induced by thermal and pharmacological stimuli in rats.

Materials and methods

Animals

Male Wistar rats were purchased from Harlan (Indianapolis, IN, USA). At the time of the surgery, rats weighed 320–370 g. All rats were housed in individual cages placed in temperature-controlled units (T_a of 28°C) and under a 12 h light/dark cycle (lights on at 6:00 AM). Standard laboratory rodent chow and tap water were available ad libitum. All procedures were approved by the St. Joseph’s Hospital and Medical Center Animal Care and Use Committee.

Surgery

General

All surgical procedures were performed under ketamine-xylazine-acepromazine (55.6, 5.5, and 1.1 mg/kg, respectively, intraperitoneally) anesthesia and antibiotic (enrofloxacin, 1.2 mg/kg, subcutaneously) protection. The experiments were performed on days 5–10 post-surgery. Each rat was subjected to the surgical procedures described below.

Bilateral cervical DLF transection (funiculotomy)

The skin over the upper back, neck, and head was shaved and scrubbed. The skin was incised over the vertebral column at the level of the first-second cervical (C) vertebra, and the underlying muscles were retracted. Partial C1 laminectomy was performed, and the dura mater was cut with delicate eye scissors, thus opening the subdural space. Next, a micro knife was used to make a shallow transversal incision aimed at transecting the dorsal aspect of the lateral funiculus at C1. No funiculus transection was performed in sham-operated rats. All the steps listed above were performed bilaterally. A fat pad was inserted in the laminectomy defect, and muscles and skin were sutured in layers. The same funiculotomy procedure was used in another recent study in our laboratory [19].

Implantation of temperature-measuring device

Each rat was implanted with a miniature temperature data logger (SubCue, Calgary, Alberta, Canada) to record abdominal temperature, a measure of core T_b (T_c). The datalogger was implanted into the peritoneal cavity via midline laparotomy and affixed to the abdominal wall with a suture. Abdominal muscles and skin were sutured in layers.

Intravenous catheterization

A small incision was made on the ventral surface of the neck, left of and parallel to the trachea. The left jugular vein was exposed, freed from its surrounding connective tissue, and ligated. A silicone catheter (inner diameter, 0.5 mm; outer diameter, 0.9 mm) filled with heparinized (10 U/ml) saline was passed into the superior vena cava through the jugular vein and secured in place with ligatures. The free end of the catheter was knotted, tunneled under the skin to the nape, and exteriorized. The wound was sutured. The catheter was flushed with heparinized saline (10 U/ml) daily.

Thermogradient apparatus

The present study was conducted in a six-channel thermogradient apparatus, which was described in detail by Almeida et al. [10] and used extensively in our laboratory to study cold- and warmth-avoidance behaviors in rats [14,20–24] and mice [11,25]. Briefly, the apparatus consisted of six 200-cm-long aluminum channels. Each channel had a second, raised stainless-steel grid floor and an acrylic double-wall lid at the top. At each end, all channels shared a common aluminum wall, which separated the channels from a large tank. The tank at the “warm” end of the channels was filled with water heated by two electric units (PolyScience, Niles, IL, USA); the water was heated to maintain the air temperature inside the channels at this end at 30°C. The tank at the “cold” end was constantly perfused with 10% ethylene glycol by an external-circulation cooling pump (Poly-Science) to maintain air temperature inside the channels at this end at 15°C. In this setting, all channels had a common, nearly linear longitudinal temperature gradient of 0.075°C/cm. In each channel, T_a was monitored by five evenly spaced (50 cm apart) thermocouples located under the grid floor, and the position of a rat was monitored with 56 evenly spaced (3.5 cm) infrared emitter-receiver pairs, which formed transversal infrared beams; T_pr was calculated based on the linear relationship between the position in the channel and T_a, with a weighted average calculated for every 5-min interval. The position data were also used to calculate the distance traveled by a rat along its lane in the thermogradient apparatus over periods of time. Briefly, we calculated the distance traveled for every 5 min, based on the recorded IR bean-detected position, and summed these distances up to arrive at the total distance traveled during the selected time period; the latter, total distance was used as an overall measure of gross locomotor activity.

Functional verification of the lesion

The lesion was functionally verified by measuring the pain threshold of rats by thermal tail-flick tests using a cold (−18°C) solution of 75% ethanol in water [26] and hot (50°C) water [27]. A rat was held firmly and had half of its tail submerged in the solution. The latency of the tail-flick response (regardless of whether the flick resulted in the complete removal of the tail from the bath) was measured and used as the nociceptive threshold. For each rat, tail-flick latency was defined as the mean of three determinations performed at 10-min intervals. All rats were tested for both cold and hot, in random order. An interval of at least 2 h between the tests was allowed.

Studying behavioral thermoregulation

Rats were extensively habituated (4 daily training sessions, 6–16 h each) to the thermogradient apparatus. On the day before the experiment, the rats were placed in the apparatus at 6:00 PM and stayed there overnight. The next morning, either the spontaneous selection T_pr or thermoregulatory behavioral responses to thermal or pharmacological stimuli were studied. As in our earlier work [10], thermal stimulation was achieved by confining each rat to a short (22 cm) portion of the channel adjacent to either the cold end (T_a of ~17ºC in the middle of the confinement zone) or warm end (T_a of ~28ºC in the middle of the confinement zone) of the thermogradient apparatus for 2 h. The midpoint of the thermoneutral zone (TNZ) [10] and the T_pr [10,14,20] for rats in this apparatus are both ~24ºC. Confining a rat at 17 or 28ºC results in moderate cold or mild heat exposure, respectively. To induce a warmth-seeking response pharmacologically, rats were shaved (over the entire trunk) the day before the experiment and, during the experiment, treated epidermally (e.d.) with 3 g of a 0.5% menthol ointment (30% polyethylene glycol, 25% glycerin, 25% propylene glycol, and 19.5% ethanol); menthol is a widely used agonist of the cold-sensitive TRPM8 receptor and a ubiquitous “cooling mimetic” [28,29] known to trigger cold-avoidance (warmth-seeking) responses [10,30,31]. To induce a cold-seeking response pharmacologically, rats were injected intravenously (i.v.) with RN-1747 (100 μg/kg), an agonist of the warm-sensitive TRPV4 receptor [32], in saline containing 15% ethanol and 15% propylene glycol.

Histological verification

To determine the exact location and extension of a lesion, each rat was deeply anesthetized and perfused, through the ascending aorta (right atrium cut), with saline (200 ml, 5 min) followed by 10% formalin (200 ml, 5 min). The spinal cord was removed, placed in phosphate-buffered saline (0.1 M, pH 7.4) containing 30% sucrose and 10% formalin, and post- fixed in this solution at 4°C for 48 h. Blocks of fixed spinal cords were frozen in dry ice and sectioned (50-µm thick). Sections were mounted on glass slides, stained with cresyl violet, and examined under a light microscope.

Data analysis

Thermoregulatory behaviors induced by thermal or pharmacological stimuli (T_pr responses) and T_c responses were assessed by a two-way ANOVA for repeated measurements followed by the Holm-Sidak post-hoc test, as appropriate. All other data were analyzed by performing simple pairwise comparisons using unpaired two-tailed Student´s t-test or the Mann- Whitney U-test. Data distribution was tested for normality to decide whether to use parametric or nonparametric tests. All analyses were performed using Sigma Plot 13.0 (Systat Software, Point Richmond, CA, USA). Differences were considered significant at p < 0.05. Data are reported as means with their standard errors.

Results

Cervical DLF transection increases the pain threshold to noxious thermal stimuli

The spinal cords of all rats with DLF transection were examined histologically (Figure 1a), and only rats having a lesion covering > 50% of the targeted area (i.e., the DLF, as outlined by Paxinos and Watson [33]) on each side were included in the analyses. DLF transections were also verified functionally, by the tail-flick responses to noxious cold and heat stimuli (Figure 1b). As compared with sham-operated rats, rats with confirmed successful bilateral DLF transection had longer tail-flick response latencies to both cold and heat (p = 0.029 and p = 0.007, respectively, t-test). Importantly, funiculotomized rats looked healthy and, when observed in their home cages or handled, did not show any behavioral abnormality. At the time of the experiments, their body mass did not differ significantly from that of sham-operated rats (348 ± 12 and 353 ± 10 g, respectively; p = 0.75, t-test). In the thermogradient apparatus (see below), funiculotomized rats did not show any decrease in gross locomotor activity (measured as distance traveled), as compared to sham-operated rats.

Figure 1.

Histological (a) and functional (b) verification of DLF lesions. a. Top panel, left: a schematic of a transversal C1 spinal section from the atlas by Paxinos and Watson [33]. The DLF (targeted area) is shown in orange. VLF: ventrolateral funiculus; the Rexed laminae 1–10 are shown by the corresponding numerals. Top panel, right: a bright-field photomicrograph of a C1 spinal section of a sham-operated rat (50 μm, cresyl violet). Middle panel: a bright-field photomicrograph of a C1 spinal section of a rat with a representative bilateral DLF lesion. Bottom panel: a schematic [33] showing (with different shades of orange) the extent of DLF lesion in all funiculotomized rats included in the analyses (n = 9). Borders of individual lesions were drawn freehand based on photomicrographs. b. The tail-flick latency of the response to noxious cold (−18°) or noxious heat (50°C) in sham-operated and funiculotomized rats. The number of experiments is shown in parentheses. *p <  0.05.

Cervical DLF transection decreases the precision of behavioral thermoregulation

The effects of bilateral DLF transection on the spontaneous innate behavior of selecting T_pr were studied in the thermogradient apparatus over a 24-h period. While the T_pr and T_c dynamics in funiculotomized rats were similar to those in sham-operated rats (Figure 2, Table 1), there was a notable difference. The fluctuations of T_pr were markedly greater in funiculotomized rats than in sham-operated controls, especially during the light (inactive) phase of the diurnal cycle (Figure 2, Table 1). Furthermore, in funiculotomized rats, the changes at the effector level (greater fluctuations in T_pr) translated into the greater amplitude of T_c fluctuations and greater T_c variability (as measured by standard deviation) during the light phase (Table 1). Standard deviation is a measure of data dispersion, or variability, and is often used as such in the thermophysiological literature (see, e.g., Ref. [34]). The wider fluctuations in both T_pr and T_c suggest that funiculotomized rats were less sensitive to environmental thermal changes, and that their behavioral thermoregulation was less precise.

Figure 2.

Effects of cervical DLF transection on the diurnal dynamics of T_pr (a) and T_c (b). To prevent masking the mean T_pr (a) and mean T_c (b) dynamics, error bars are not shown for the mean curves. The histograms in a show the time spent in each of the three T_pr zones (T_a values for each zone are listed) during the light and dark (gray background) phases. The histograms in b show the time during which the T_c was in each of the three T_c ranges (values for each range are listed) during the light and dark (gray background) phases. The number of experiments is shown in the corresponding curve in parentheses. *p < 0.05.

Table 1.

Effect of bilateral DLF transection on T_pr and T_c of rats in a thermogradient apparatus.

	Light phase			Dark phase
	Sham-operated	DLF-lesioned	P value	Sham-operated	DLF-lesioned	P value
Preferred Ta(°C)
Mean	22.9 ± 0.2	22.2 ± 0.2	0.040	22.0 ± 0.3	22.2 ± 0.4	0.626
Minimum	16.5 ± 0.4	14.3 ± 0.5	0.003	15.4 ± 0.3	15.1 ± 0.4	0.563
Maximum	28.4 ± 0.3	29.8 ± 0.4	0.013	27.1 ± 0.4	28.9 ± 0.5	0.012
Fluctuation amplitude	11.9 ± 0.6	15.5 ± 0.7	<0.001	11.7 ± 0.6	13.8 ± 0.6	0.031
Standard deviation	2.1 ± 0.6	3.2 ± 0.9	0.002	2.4 ± 0.5	2.7 ± 0.6	0.113
Deep Tb(°C)
Mean	36.7 ± 0.1	36.5 ± 0.1	0.310	37.2 ± 0.1	37.3 ± 0.1	0.807
Minimum	35.7 ± 0.2	35.3 ± 0.1	0.078	36.3 ± 0.1	36.2 ± 0.1	0.627
Maximum	37.6 ± 0.1	37.8 ± 0.1	0.101	38.2 ± 0.1	38.2 ± 0.1	0.886
Fluctuation amplitude	1.9 ± 0.2	2.6 ± 0.1	0.005	1.9 ± 0.1	2.0 ± 0.1	0.768
Standard deviation	0.4 +/- 0.1	0.5 +/- 0.1	<0.001	0.4 +/- 0.1	0.5 +/- 0.1	0.162

*Statistically significant differences are shown in bold.

A decreased sensitivity to thermal stimuli can also prolong the time that animals spend in the cold or heat, without triggering the escape responses (see Discussion). Indeed, during the light phase, the times spent in the coldest (15–18°C) and warmest (27–30°C) zones of the thermogradient were higher in funiculotomized rats than sham-operated rats (p = 0.021 and p = 0.031, respectively, U-test; Figure 2a). Similarly, during the dark (active) phase, funiculotomized rats had the lowest values of T_c (34.5–35.5°C) for a longer time, as compared to controls, and this difference was statistically significant (p= 0.016, t-test; Figure 2b). In agreement with the funiculotomized rats having a longer cold exposure, almost the entire curve of their mean T_c during the light phase lays below the curve for sham-operated rats – even though the funiculotomized rats could easily move, just for a few centimeters, to expose themselves to a higher T_a and raise their T_b (Figure 2b). Importantly, the lower T_c values were not due to decreased locomotor activity. In fact, the distance traveled by funiculotomized rats during the light phase (22.9 ± 2.6 m over 12 h) was higher than that traveled by sham-operated rats (12.3 ± 1.1 m, 2-way ANOVA; effect of interaction F_1,38 = 8.19, p = 0.006), whereas during the dark phase funiculotomized and sham-operated rats traveled nearly identical distances (21.0 ± 2.3 and 21.4 ± 1.5 m over 12 h, respectively). The reason for the increased gross locomotor activity of funiculotomized rats during the inactive phase is unknown, but it would be in line with an increased anxiety level and decreased sleep.

Cold-avoidance responses are damped in rats with cervical DLF transection

We asked whether cold-avoidance (warmth-seeking) responses were affected by DLF transection. To induce a cold-avoidance response thermally, we confined each rat to the coldest portion of its lane in the thermogradient. When the confining boxes were removed, sham-operated rats immediately moved to a slightly higher T_a (by ~4°C, as compared to the T_pr just before the exposure) – the response that did not occur in funiculotomized rats (F_1,18 = 4.552, p = 0.047, group factor, Figure 3a, F_16,288 = 2.39, p = 0.02, interaction). While funiculotomized rats had a lower T_pr after the cold exposure, their gross locomotor activity did not differ significantly from that of sham-operated rats; the two groups traveled, on average, 7.3 ± 0.7 m and 8.9 ± 0.7, respectively over 2 h (p = 0.153, t-test). The slight but significant difference in T_pr between the two groups did not result in any significant differences in the T_c response (F_1,18 = 0.004, p= 0.952, group factor, Figure 3a, F_16,288 = 1.34, p = 0.159, interaction).

Figure 3.

Effects of cervical DLF transection on T_pr and T_c responses to moderate cold exposure (a) or the e.d. administration of menthol (b). Each rat was confined to the cold end (~17°C) of its lane in the thermogradient apparatus for 120 min (blue bar); a divider was inserted in each lane at time 0 and removed at 120 min. Before and after the cold exposure, each rat moved freely and selected its T_pr. The number of experiments is shown in parentheses at the corresponding T_pr and T_c curves. *p <  0.05.

Next, a cold-avoidance response was induced pharmacologically, by the e.d. application of the TRPM8 agonist menthol. Following menthol administration, sham-operated rats selected a slightly higher T_a (F_21,210 = 7.004, P < 0.001, time factor, F_1,10 = 7.072, p = 0.024, group factor, F_21,210 = 1.189, p= 0.263, interaction; Figure 3b), with no such increase occurred in funiculotomized rats. Menthol also produced a large, almost 2°C, T_c rise in sham-operated rats; the hyperthermic (T_c) response was damped in funiculotomized rats (F_21,210 = 33.825, P < 0.001, time factor, F_21,210 = 1.893, p= 0.013, interaction factor, Figure 3b). No difference in gross locomotor activity was observed between the two groups of rats: the total distance traveled over 3 h was 6.0 ± 0.4 m in funiculotomized rats and 7.5 ± 0.9 m in sham-operated controls (p = 0.200, t-test).

Warmth-avoidance responses are largely unaffected by cervical DLF transection

We asked whether warmth-avoidance (cold-seeking) responses were affected by DLF transection. To induce a warmth-avoidance response thermally, we exposed rats to mild heat by confining them to the warmest portion of the thermogradient, each in its individual lane. When the confining boxes were removed, both sham-operated and funiculotomized rats moved to a slightly lower (by ~4°C, as compared to the time before the exposure) T_a (F_16,304 = 4.204; P < 0.001, time factor); no significant differences between the two groups were observed (Figure 4a). The heat exposure caused marked hyperthermia, with T_c reaching ~39°C, in both sham-operated and funiculotomized groups (F_11,209 = 93.644, P < 0.001, time factor); this T_c response also showed no significant differences between the two groups during the heat exposure (F_1,19 = 0.590, p = 0.452, group factor). After the heat exposure, when rats developed cold-seeking behavior, the T_c of sham-operated animals returned to baseline, whereas the T_c of funiculotomized rats “overshot” by ~0.5°C (F_16,304 = 2.295, p = 0.003, interaction factor). The heat exposure did not result in any significant difference between the groups in the gross locomotor activity: the total distance traveled by sham-operated vs. funiculotomized rats during 2 h was 7.4 ± 0.9 m vs. 6.8 ± 0.7 m, respectively, (p = 0.662, t-test).

Figure 4.

Effect of cervical DLF transection on T_pr and T_c induced by mild heat exposure (a) or the i.v. administration of RN-1747 (dose indicated) (b). Each rat was confined to the warm end (~28°C) of its lane in the thermogradient apparatus for 120 min (pink bar); a divider was inserted in each lane at time 0 and removed at 120 min. Before and after the heat exposure, each rat moved freely and selected its T_pr. The number of experiments is shown in parentheses at the corresponding T_pr and T_c curves. *p < 0.05.

To induce a cold-seeking response pharmacologically, we used the i.v. administration of the TRPV4 agonist RN-1747 (100 μg/kg), which tended to cause a slight (by ~4°C) and short-lived decrease in T_pr in sham-operated, but not funiculotomized, rats; however, the analyses performed were unable to reveal any significant difference in the T_pr response between the two groups (e.g., F_1,18 = 0.685, p = 0.419, group factor). RN-1747 caused a transitory T_c increase in both groups (F_21,378 = 6.842, P < 0.001, time factor), with no significant differences between the groups (F_1,18 = 0.535, p= 0.474, group factor, Figure 4b). The total distance traveled over 3 h after the RN-1747 administration also did not differ significantly between sham-operated and funiculotomized rats (7.4 ± 0.7 m and 9.1 ± 1.7 m, respectively, p = 0.331, t-test).

Even though we did not find any statistically significant differences in the warmth-avoidance (T_pr) responses or the corresponding T_c responses of funiculotomized rats, as compared to sham-operated rats, rats with DLF lesions tended to select a higher T_a immediately after heat exposure or RN-1747 administration, as compared with their sham-operated counterparts. Hence, it is possible that a larger number of experiments would allow for detecting a significant difference. It is clear, however, that cold-avoidance responses in the present study were more readily affected by the funiculotomy than warmth-avoidance responses.

Discussion

Selection of preferred thermal environment

In a lane of a thermogradient apparatus, during the light (inactive) phase, a well-adapted rat usually stays quietly (sleeps) in the middle portion of the lane, at its T_pr, which is usually a neutral T_a [2,10]. From time to time, the rat makes trips toward the cold or warm end of its lane, away from the TNZ. These “spontaneous” trips are probably driven by multiple thermal and non-thermal factors; they may represent the search for a cooler or warmer environment, as well as exploratory or territorial behaviors or attempts to seek food, water, or a mate, or simply an exercise to “stretch legs.” After the departure from the TNZ, the animal may stay at its newly chosen position for some time, while exposing itself to the corresponding subneutral or supraneutral T_a. As the rat cools down or warms up due to such an exposure, its skin temperatures and other T_bs, possibly including T_c, start decreasing or increasing, respectively, which triggers the animal’s return to a neutral T_a. Hence, the spontaneous selection of T_pr is likely the result of cold- and heat-avoidance responses.

Reflecting the described above behavior of a well-adapted rat in a thermogradient apparatus, the T_pr curve typically shows relatively large oscillations, while the T_c fluctuates only slightly. We have not formally described this behavioral pattern previously, but we and others have observed (and to some extent illustrated) it repeatedly in adult rats [10,14,21], adult mice [11], or newborn rabbits [8,9]. The same pattern (relatively large fluctuations in T_pr coupled with very small fluctuations in T_c) was exhibited by sham-operated rats in the present study (Figure 2). Funiculotomized rats, as compared to their sham-operated counterparts, showed much greater oscillations in T_pr and a greater variability of T_c. Similar exaggerations of T_pr fluctuations coupled with a less stable T_c were observed by Szelényi et al. [9] in some (but not all) newborn rabbits in a thermogradient apparatus. Newborn animals are known to have an unstable T_c and a decreased sensitivity to T_a challenges, especially when malnourished or sick [35].

It is understood that different thermoeffectors, including thermoregulatory behaviors, are driven by different combinations of shell and deep T_bs [2,36]. Shell T_bs, especially those of hairy (or furry) skin, are thought to be relatively more important for driving thermoregulatory behaviors [37], whereas deep T_bs are more important for physiological thermoeffectors [38,39]. Such an organization reflects the fact that most behavioral responses are proactive and aimed at escaping forthcoming environmental insults [2]. These escape (rather than active-counteraction) responses are typically triggered when the skin is already exposed to a thermal stimulus, but the T_c is still unaffected. The large oscillations of T_pr coupled with the increased T_c variability, observed in the present study, agree with the notion that thermal signals from the periphery that ascend through the spinal cord within the DLF drive the spontaneous selection of T_pr. Thermosensory-specific spinothalamic fibers have been shown to ascend in the DLF [40,41], and cervical lesions of the DLF were found to cause thermosensory deficiencies in cats, as assessed by a food-reinforced temperature-discrimination behavioral assay [42,43].

As a potential pitfall of our study, it should be noted that the observed large oscillations in T_pr following funiculotomy did not necessarily reflect true changes in the thermopreferendum. As noted above, rats move to different positions in a thermogradient apparatus for various, often non-thermoregulatory, reasons. One common problem with experiments in a linear thermogradient is that the ends of a thermogradient lane have different geometry, as compared to the middle portions: the ends have extra (closing) walls and, therefore, also corners, whereas there are no closing walls or corners in the middle portions. Because rodents tend to avoid open spaces and prefer “hiding” in holes or corners and being in contact with wall surfaces (positive thigmotaxis), the selection of the warm (or cold) end of the lane may be driven by thigmotaxis rather than thermotaxis. Furthermore, the thigmotaxic drive is not constant; e.g., it increases in anxiety [44]. In the past, we were able to exclude the thigmotaxic nature of the cold- and warmth-seeking behaviors in rats treated with bacterial endotoxin by running control experiments in the same apparatus, but without any T_a gradient in it, i.e., when all locations within the apparatus were maintained at the same T_a [10]. Clever circular designs of a thermogradient apparatus, where all positions possess the same geometry, have also been proposed [13,45].

Cold- and warmth-avoidance responses

Cold-avoidance responses elicited by thermal stimuli or the e.d. administration of the TRPM8 agonist menthol were attenuated in funiculotomized rats. Menthol is known [21,30,31,46] to induce hyperthermia (a T_c rise), which was also observed in the present study. Menthol-induced hyperthermia was attenuated in funiculotomized rats – the effect that agrees with the attenuation of cold-avoidance behavior. While we cannot decisively exclude the participation of autonomic effectors, in a parallel study in our laboratory [19], it was shown rats with DLF lesions did not reveal any weakness in the thermogenic (oxygen consumption) and hyperthermic (T_c) responses to CL316243, an agonist of the β3-adrenoreceptor – the “metabolic” receptor in brown adipose tissue [47]. Neither did funiculotomized rats show any decrease in the gross locomotor activity in the present study. As for shivering, its role in thermoregulation in small rodents is typically viewed as less important than that of nonshivering thermogenesis [48]. Furthermore, while TRPM8 has been shown to control brown fat thermogenesis and tail skin vasoconstriction in rats and mice [21,30,49], it is unlikely to control shivering in mice [50]. It is plausible, therefore, that the attenuated hyperthermic response of funiculotomized rats to menthol in this study was due to their sluggish cold-avoidance responses. Hence, our study suggests that cold-avoidance responses critically depend on the spinal afferent signals that travel through the DLF.

We also studied warmth-avoidance responses to mild heat exposure and i.v. administration of RN-1747 (100 μg/kg), the TRPV4 agonist that was previously shown to induce warmth-avoidance behavior [51]. In the present study, in contrast to the cold-avoidance responses, the warmth-avoidance behaviors were not significantly attenuated in funiculotomized rats. While our study might not have the sufficient statistical power to reveal an attenuation of warmth avoidance (see Results), the differential effect of the funiculotomy on the cold- vs. warmth-avoidance behaviors still was pronounced: it was easier to reveal the dampening of cold-avoidance behaviors than any attenuation of warmth-avoidance responses. This differential effect can be attributed to two biological factors.

First, it is likely to reflect the fact that cold-avoidance behaviors were induced in our study without any decrease in T_c, whereas both warmth-avoidance responses (to heat and to RN-1747) were preceded by large increases in T_c. Hence, even if the cutaneous warmth afferent signals that trigger warmth avoidance were blocked by the funiculotomy, the brain could have been a source of thermal signals, and intrabrain pathways for these signals remained intact. The presence vs. absence of concurrent brain warming may also explain the seeming contradiction between the results with the whole-body exposure to innocuous warmth (no clear attenuation of the heat-avoidance response) and the local (tail) noxious heat exposure (a marked delay of the tail-flick response). (In addition, the spinal pathways for innocuous and noxious thermal stimuli may not be identical, and neither may the spinal pathways for thermal signals from different areas of the body.)

Second, the relative importance of central (brain) thermal signals, as compared to peripheral (skin) thermal signals, is greater for driving cold-defense effectors than heat-defense ones [39]. This principle has been documented for autonomic thermoeffectors, but it may also apply to behaviors (reviewed in Refs. [2,36,52]). Some of thermoregulatory behaviors (such as a relaxed postural extension, a.k.a. pronation, that occur in a heat-exposed animal after it becomes hyperthermic) are triggered predominantly or exclusively by thermal signals from the brain [53,54].

Curiously, the significant attenuation of cold-avoidance (but not heat-avoidance) behaviors observed in the present study parallels the observation in human patients with unilateral cervical anterolateral funiculotomy, who exhibited an impairment of skin cooling perception, even though their skin warming perception remained largely unaffected [55].

Implications for future research

While neural pathways that drive physiological thermoeffectors are well-studied, neural substrates of thermoregulatory behaviors are poorly understood, and it is widely known among thermophysiologists that many preliminary studies aimed at inducing deficits in thermoregulatory behaviors by lesioning various brain structures produced negative results and were never published. In the present study, we transected the DLF bilaterally at the highest possible level (C1), thus potentially causing maximal deficits in the thermal somatic and visceral afferentation driving thermoregulatory behaviors. A priori, lesioning the same neural pathway further downstream along the signal flow (i.e., in the brain) cannot be expected to attenuate a targeted behavior to a greater extent – unless the brain lesion also interrupts other signaling pathways, e.g., those that conduct intracranial or visceral T_b signals. The present results suggest that future experiments aimed at identifying neural pathways of thermoregulatory behaviors by inactivating brain structures are more likely to fail if they target warmth-avoidance responses. On the other hand, such experiments may succeed if they target the spontaneous search of T_pr or cold-avoidance behaviors triggered by cold exposure or pharmacological stimulation of TRPM8. Even for cold-avoidance behaviors, the effects are expected to be small and may require large numbers of observations. For studying the selection of T_pr, a detailed analysis of movements of relatively high frequencies (as opposed to mean T_pr for relatively long time periods) will be needed, including a quantitative approach to reveal the effects of T_pr on T_c.

Conclusions

1. Innocuous thermal signals that ascend in the spinal cord through the DLF contribute to driving spontaneous thermal preference; attenuation of these signals is associated with a decreased precision of T_c regulation and T_c stability.

2. Thermally and pharmacologically induced changes in thermal preference rely on neural, presumably afferent, signals that travel in the spinal cord within the DLF. The DLF-conveyed signals are important for innocuous cold-avoidance responses; their importance for warmth-avoidance responses is much lesser and possibly negligible.

Funding Statement

This work was supported in part by the National Institutes of Health (grant R01NS41233 to AAR), St. Joseph’s Foundation (grant to AAR), and São Paulo Research Foundation (grants 2013/25503-5 and 2016/01836-3 to RCLV and 2015/02991-0 to MCA).

List of abbreviations

C

cervical (vertebra)

DLF

dorsal portion of lateral funiculus; dorsolateral funiculus

e.d.

epidermal, epidermally

i.v.

intravenous, intravenously

LPB

lateral parabrachial nucleus

M

melastatin (as in TRPM8)

T_a

ambient temperature

T_b(s)

body temperature(s)

T_c

core body temperature

TNZ

thermoneutral zone

T_pr

preferred ambient temperature

TRP

transient receptor potential (channel)

V

vanilloid (as in TRPV4)

Author contributions

MCA and AAR designed the study; MCA and RNS collected the data; RCLV, MCA, RNS, and AAR analyzed and interpreted data; RCLV, MCA, and AAR prepared figures and wrote the manuscript; all authors read and approved the final version of the manuscript.

Disclosure statement

AAR is an officer and director of Catalina Pharma, Inc. and Zharko Pharma, Inc.; he consulted for the TRPM8 program at Amgen; and his laboratory conducted paid research on TRP antagonists for Amgen, Abbott Laboratories, and AbbVie.