Leptin regulation of core body temperature involves mechanisms independent of the thyroid axis

Abstract

The ability to maintain core temperature within a narrow range despite rapid and dramatic changes in environmental temperature is essential for the survival of free-living mammals, and growing evidence implicates an important role for the hormone leptin. Given that thyroid hormone plays a major role in thermogenesis and that circulating thyroid hormone levels are reduced in leptin-deficient states (an effect partially restored by leptin replacement), we sought to determine the extent to which leptin’s role in thermogenesis is mediated by raising thyroid hormone levels. To this end, we 1) quantified the effect of physiological leptin replacement on circulating levels of thyroid hormone in leptin-deficient ob/ob mice, and 2) determined if the effect of leptin to prevent the fall in core temperature in these animals during cold exposure is mimicked by administration of a physiological replacement dose of triiodothyronine (T₃). We report that, as with leptin, normalization of circulating T₃ levels is sufficient both to increase energy expenditure, respiratory quotient, and ambulatory activity and to reduce torpor in ob/ob mice. Yet, unlike leptin, infusing T₃ at a dose that normalizes plasma T₃ levels fails to prevent the fall of core temperature during mild cold exposure. Because thermal conductance (e.g., heat loss to the environment) was reduced by administration of leptin but not T₃, leptin regulation of heat dissipation is implicated as playing a uniquely important role in thermoregulation. Together, these findings identify a key role in thermoregulation for leptin-mediated suppression of thermal conduction via a mechanism that is independent of the thyroid axis.

INTRODUCTION

Maintenance of core body temperature within a narrow range in the face of rapid and marked changes in environmental temperature is critical for survival (22). During cold exposure, effective thermoregulation requires a substantial increase of energy expenditure in the form of heat, and this response must be offset by comparable increases in energy intake to avert progressive and potentially life-threatening depletion of body fuel stores. This is particularly true in small animals, which are characterized by a high surface area-to-body mass ratio (22) and hence have a greater propensity to lose heat to the environment. Our recent work (29, 30) and that of others (18, 48) suggests that the adipokine leptin plays an important physiological role not only in thermoregulation but also in the coupling of this process to energy homeostasis such that homeostatic defense of core body temperature and body fat stores are seamlessly integrated across a wide range of ambient temperatures.

This integration is evident in the ability of normal mice to adapt to environmental temperatures well below thermoneutrality with little impact on either body fat mass or core body temperature (30), despite pronounced increases of both energy intake and energy expenditure. A key role for leptin in this process is implied by the finding that the core body temperature of leptin-deficient ob/ob mice falls in proportion to declining environmental temperature. In addition, these animals experience a loss of body fat mass during cold exposure (owing to their inability to increase food intake to meet increased energy demands) that is not observed in normal mice (18, 30, 62, 63). A role for leptin deficiency in these outcomes is confirmed by evidence that thermoregulation improves rapidly after administering a physiological replacement dose of leptin (29). Moreover, this effect of leptin to raise core temperature in ob/ob mice cannot be explained by increases of energy expenditure, induction of brown adipose tissue (BAT) thermogenesis, or increased ambulatory activity. Rather, leptin’s effect is associated with an adaptive reduction of thermal conductance, the ease with which heat flows from the animal to the environment (29). The current work focuses on the fact that leptin deficiency is associated not only with hypothermia (62, 63) but also with reduced circulating thyroid hormone levels (1), and both conditions are ameliorated by leptin replacement (1, 24, 29, 45). Moreover, thyroid hormone deficiency, if severe, can cause hypothermia (56). We therefore sought to determine the extent to which the effect of leptin on thermoregulation is secondary to its effect to increase thyroid hormone.

Thyroid hormone is a major determinant of thermogenesis and body temperature (56). Thus, whereas hyperthyroidism is associated with an increased basal metabolic rate and can cause hyperthermia, hypothyroidism reduces both metabolic rate and core body temperature across mammalian species (21), including humans (55). These observations reflect the critical role played by thyroid hormone in “obligatory thermogenesis,” the minimal energy cost required for maintenance of the body in the living state, corresponding to the basal metabolic rate. In addition, however, thyroid hormone can also increase “facultative thermogenesis,” which engages both shivering and non-shivering thermogenic mechanisms to generate heat when exposed to ambient temperatures below thermoneutrality (56). These effects involve direct actions of thyroid hormone on BAT and other thermogenic tissues (6) as well as an action in the central nervous system to increase sympathetic nervous system outflow—the same mechanism implicated in the thermoregulatory action of leptin (9, 25, 31).

The hypothalamic-pituitary-thyroid (HPT) axis is regulated by both nutritional state and by input from other hormones, including leptin. Thyroxine (T₄) secretion and plasma levels are rapidly suppressed during starvation (1), and this effect contributes to the adaptive decrease of metabolic rate that helps to conserve energy and thereby favor survival (7). The mechanism underlying this response appears to involve leptin deficiency, since fasting rapidly and dramatically lowers circulating leptin levels, and systemic administration of leptin during fasting prevents the associated fall of thyroid hormone levels (1). In humans, leptin replacement during weight loss also increases thyroid hormone levels (51, 52), and a similar effect is observed across genetic and acquired models of leptin deficiency, including ob/ob mice, streptozotocin-induced uncontrolled diabetes, and hypothalamic amenorrhea induced by excessive loss of body fat: each is characterized by low thyroid hormone levels that are ameliorated by a physiological replacement dose of leptin (1, 19, 34, 50, 70). Furthermore, the brain is implicated in this process, since administration of leptin directly into the brain at doses that are ineffective when administered systemically raises the low thyroid levels in these conditions (34).

While there can therefore be no question that leptin regulates the HPT axis in a physiologically relevant manner, the extent to which leptin’s role in thermoregulation hinges on this effect remains largely untested. In the current work, we sought to determine 1) whether raising plasma thyroid hormone levels from the low to the normal range is sufficient to ameliorate hypothermia in cold-exposed ob/ob mice (as is observed with leptin replacement), and 2) whether such an effect can account for leptin’s ability to raise core temperature in ob/ob mice during cold exposure.

MATERIALS AND METHODS

Animals.

Adult male C57/Bl6 [wild-type (WT)] and leptin-deficient ob/ob mice bred onto the C57/Bl6 background strain were obtained from Jackson Laboratories (Bar Harbor, ME). All animals were individually housed under specific pathogen-free conditions in a temperature-controlled room with a 12:12 h light-dark cycle and provided with ad libitum access to water and chow unless otherwise stated (PMI Nutrition, St. Louis, MO). All procedures were performed in accordance with National Institutes of Health Guide for the Care and Use of Animals and were approved by the Institutional Animal Care and Use Committee at the University of Washington.

Systemic leptin and T₃ administration.

To examine the effect of leptin deficiency and physiological leptin replacement on thyroid hormone levels, adult male ob/ob and WT littermate controls (n = 7–8 per group) were implanted with an osmotic minipump (Alzet Model 1007D, DURECT Corporation, Cupertino, CA) containing either vehicle (veh; PBS, pH 7.9) or leptin delivered at a rate of 100 ng/hr (made available via Dr. A. F. Parlow, National Hormone & Peptide Program, Torrance, CA) (29, 38). To examine the effect of exogenous thyroid hormone on core temperature during mild exposure, a separate cohort of adult male ob/ob mice (mean body weight: 56.54 ± 0.70 g) was separated into 2 weight-matched groups and underwent subcutaneous implantation of an osmotic minipump containing either saline veh or a physiological dose of triiodothyronine (T₃; 4 μg·kg⁻¹·d⁻¹; Sigma, St. Louis, MO) selected based on previous literature (49). In a subsequent study, ob/ob mice were subjected to the same protocol except that they were implanted with 2 subcutaneous osmotic minipumps, the first containing the same physiological dose of T₃ and the second containing a physiological replacement dose of leptin (100 ng/hr, n = 7–8 per group) (29, 38). Buprenorphine hydrochloride (0.05 mg/kg; Reckitt Benckiser, Richmond, VA) was administered perioperatively.

Core temperature.

Adult male ob/ob mice underwent implantation of a body temperature transponder in the peritoneal cavity (Starr Life Sciences Corp., Oakmont, PA) before study. Animals were allowed at least 1 wk to recover and were then acclimated to metabolic cages enclosed in temperature- and humidity-controlled cabinets (Caron Products and Services, Marietta, OH). Signals emitted by body temperature transponders were sensed by a receiver positioned underneath the cage and analyzed using VitalView software as previously described (29, 30).

Measurements of energy expenditure, energy intake, and ambulatory activity.

Ob/ob mice were acclimated to indirect calorimetry cages before measurement of energy expenditure using a computer-controlled indirect calorimetry system (Promethion, Sable Systems, Las Vegas, NV) with support from the University of Washington Nutrition Obesity Research Center Energy Balance Core as described in detail previously (28–30). This system consists of 16 metabolic cages that contain bedding and are equipped with water bottles and food hoppers connected to load cells for continuous food and water intake monitoring and are housed in a temperature- and humidity-controlled cabinet (Caron Products and Services, Marietta, OH). O₂ consumption and CO₂ production were measured for each animal for 1 min at 10-min intervals, and respiratory quotient (RQ) was calculated as the ratio of CO₂ production to O₂ consumption. Energy expenditure was calculated using the Weir equation (69). Ambulatory activity was determined simultaneously with the collection of calorimetry data. Consecutive adjacent infrared beam breaks in the x-, y- and z-axes were scored as an activity count, and a tally was recorded every 10 min. Data acquisition and instrument control were coordinated by MetaScreen v.1.6.2, and raw data were processed using ExpeData v.1.4.3 (Sable Systems, Las Vegas, NV) using an analysis script documenting all aspects of data transformation.

Thermal conductance.

Whole body thermal conductance was calculated as energy expenditure divided by the difference between core and ambient temperature: C = EE/(T_b − T_a), where C = conductance; EE = energy expenditure; T_b = core temperature; and T_a = ambient temperature (22, 29, 30, 36). This measures the ease with which heat flows from the body core to the environment (via convection, conduction, radiation, and evaporation).

Experimental protocol.

Animals were studied following implantation of body temperature transponders and acclimation to housing within temperature- and humidity-controlled chambers. To allow for pairwise comparisons, measures of energy expenditure, energy intake, ambulatory activity, RQ, and body temperature were recorded continuously in all mice in the absence of treatment for 68 h at a temperature maintained at 22 ± 0.1°C (Baseline). Following a brief washout period, animals were implanted with a subcutaneous osmotic minipump containing either veh or a physiological dose of T₃. After surgery, animals were returned to indirect calorimetry cages on the same day that ambient temperature was lowered from 22°C to 14°C over a 4-h period (1400–1800) before dark cycle onset. Animals remained housed at 14 ± 0.1°C for 112 h for continuous measures of energy expenditure and body temperature (Fig. 1A).

Fig. 1.

Effect of a physiological dose of T₃ on core temperature in ob/ob mice during mild cold exposure. Schematic of study design (A), photoperiod-averaged 24-h core temperature in adult male ob/ob mice housed at under room temperature conditions [22°C; Baseline: Group 1 (blue), Group 2 (red); B] or treated systemically with either veh [Group 1 (blue)] or a physiological dose of T₃ [Group 2 (red)] and subjected to mild cold exposure (14°C; C). The mean change in photoperiod-averaged 24-h core temperature (D) and mean change in core temperature during the dark cycle, light cycle, and 24-h period (E) over all days of treatment in veh-treated mice housed under room temperature conditions (22°C; Baseline) relative to veh- and T₃-treated mice subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. T₃. ns, not significant; T₃, triiodothyronine; veh, vehicle.

To determine if a physiological replacement dose of leptin can normalize core body temperature even in euthyroid ob/ob mice, a separate cohort of ob/ob mice was subjected to a similar protocol. Following indirect calorimetry and core temperature measures at 22 ± 0.1°C (Baseline), mice were implanted with a subcutaneous osmotic minipump containing the same dose of T₃ in combination with a second subcutaneous osmotic minipump containing either veh or leptin to create 2 groups: T₃-veh and T₃-lep. Animals were returned to indirect calorimetry cages on the same day in which ambient temperature was lowered from 22°C to 14 ± 0.1°C using the protocol as described above. Continuous measures of energy expenditure, energy intake, ambulatory activity, and core temperature were recorded for 112 h.

Blood collection and tissue processing.

At the end of each study, whole-blood samples were collected into EDTA-treated tubes and centrifuged, then plasma was removed and stored at −80°C for subsequent assay. All plasma hormone levels were measured using commercially available ELISA kits according to manufacturers’ instructions. Plasma leptin levels (kit no. 90080, Crystal Chem, Chicago, IL) (10) were determined by ELISA, and thyroid stimulating hormone (kit no. LS-F-5125, LifeSpan Biosciences, Seattle, WA), total T₃ (53) (kit no. T-3225-T, Calbiotech, Spring Valley, CA), total T₄ (17) (kit no. T-4044-T-100, Calbiotech), free T₃ (kit no. 1650, Alpha Diagnostic, San Antonio, TX) (5), and free T₄ (32, 54) (kit no. 1110, Alpha Diagnostic) levels were determined by competitive enzyme immunoassay. Previous publications combined with our current work further validate these ELISAs, as the data obtained show the expected biological variation associated with both loss of function (e.g., genetic or acquired models of leptin deficiency and/or hypothyroidism) and gain of function studies (e.g., hormone administration). For example, we show that in leptin-deficient ob/ob mice, plasma leptin levels are undetectable, whereas plasma T₄ levels are reduced relative to WT controls, consistent with the literature (1). Moreover, as expected, systemic T₃ administration raises plasma T₃ levels, an effect associated with the expected compensatory reduction of total T₄ levels.

Statistical analyses.

Results are expressed as means ± SE. Significance was established at P < 0.05, two-tailed. For statistical comparisons involving core temperature, energy expenditure, ambulatory activity, RQ, and energy intake, data obtained during both the 22°C period (Baseline) and the 14°C test periods were analyzed throughout the time course of the study. The data were also reduced into the mean light and dark cycle components and mean 24-h photoperiod for each mouse across all days of drug treatment. The difference in core temperature in mice studied at 14°C relative to the 22°C Baseline period was also calculated as the difference between these averaged 24-h photoperiods. Statistical analyses were performed using Statistica (version 7.1, Statsoft, Inc., Tulsa, OK) and SPSS (SPSS version 23, IBM Corp., Somers, NY). A mixed factorial (group by ambient temperature) ANOVA with a least significance difference pairwise test was used to compare mean values between groups. For changes over time, a repeated measures ANOVA was utilized. A two-sample unpaired Student’s t-test was used for two group comparisons and a paired t-test for within group comparisons.

RESULTS

Effect of ambient temperature, leptin deficiency, and physiological leptin replacement on thyroid hormone levels.

We first examined the effect of leptin deficiency and physiological leptin replacement on plasma levels of thyroid hormone. Here, we found that both plasma total T₄ levels and free T₃ levels were lower in ob/ob mice than WT controls (Table 1), consistent with previous studies (26, 66). In ob/ob mice that received leptin, plasma leptin levels were raised to values similar to those of WT-veh controls (Table 1). Thus, the dose used can be considered a physiological replacement dose. Although the effect was modest, plasma T₄ and free T₃ levels were increased in ob/ob mice receiving a physiological replacement dose of leptin (P < 0.05) but remained below levels seen in WT control mice (Table 1). In addition, although thyroid stimulating hormone levels were elevated in ob/ob relative to WT mice (648 ± 97 vs. 369 ± 53 pg/ml; P < 0.05), they remained in the physiological range and were not suppressed by the leptin-induced increase of circulating thyroid hormone levels (P = not significant).

Table 1.

Effect of leptin treatment on free and total plasma T₄ and T₃ levels in WT and ob/ob mice

Hormones	WT-veh	WT-lep	ob/ob-veh	ob/ob-lep
Leptin, ng/ml	1.73 ± 0.16	3.69 ± 0.35*	0.00 ± 0.00*	2.50 ± 0.25*#
Free T4, pg/ml	4.08 ± 0.34	4.58 ± 0.37	3.94 ± 0.25	5.55 ± 0.57*#
Total T4, µg/dl	5.42 ± 0.14	6.16 ± 0.16*	3.59 ± 0.11*	4.30 ± 0.15*#
Free T3, pg/ml	5.80 ± 0.15	6.56 ± 0.20*	3.86 ± 0.15*	4.46 ± 0.18*#
Total T3, ng/ml	0.96 ± 0.05	1.19 ± 0.05*	0.93 ± 0.03	1.09 ± 0.08#

Means ± SE; n = 7–8 per group. lep, leptin; T₃, triiodothyronine; T₄, thyroxine; veh, vehicle; WT, wild type.

^*P < 0.05 vs. WT-veh;

^#P < 0.05 vs. ob/ob-veh.

Effect of a physiological dose of T₃ on core temperature and energy balance in ob/ob mice during mild cold exposure.

To determine whether raising low thyroid hormone levels characteristic of ob/ob mice into the normal range is sufficient to maintain core temperature when the mice are exposed to sub-thermoneutral temperatures, we studied these animals both at room temperature (22°C) and during mild cold exposure (14°C) while they were receiving continuous subcutaneous infusion of either veh or T₃ (Fig. 1A) at a dose that raised plasma free T₃ levels into the physiological range (49). Specifically, we show that relative to WT mice (Table 1), ob/ob-veh-treated mice exhibit lower free T₃ levels, and in those ob/ob mice receiving T₃, free T₃ levels are raised to that similar of WT controls. (WT-veh: 5.80 ± 0.15 pg/ml vs. ob/ob-veh: 4.53 ± 0.41 pg/ml vs. ob/ob-T₃: 5.31 ± 0.26 pg/ml).

As expected (29, 30), we found that relative to room-temperature housing (22°C), ob/ob mice exhibit a fall in core temperature when subjected to mild cold exposure (14°C) during both the light and dark cycle periods (Fig. 1, B–E). Nevertheless, ob/ob mice exhibited a diurnal temperature rhythm similar to that of normal animals, with a characteristic decline of core body temperature at light cycle onset, a time that coincides with sleep, fasting, reduced activity, and energy conservation. This hypometabolic, hypothermic response is indicative of torpor, an adaptation utilized primarily by small mammals to conserve energy during conditions of low fuel availability and/or decreased ambient temperature (58). Although administration of T₃ failed to prevent the fall in core temperature in ob/ob mice during mild cold exposure, (Fig. 1, B–E), it blunted the characteristic decline in core temperature at light cycle onset, thus attenuating the torpor response.

A more detailed analysis of core temperature throughout the duration of the experiment further reveals that relative to 22°C, core temperature was reduced to a similar extent during the first 24 h of mild cold exposure in both veh- and T₃-treated mice, and these lower core temperatures were comparable between groups throughout the duration of the study (Fig. 2A). Similarly, changes in heat production and ambulatory activity were observed in the first 24 h following T₃ treatment and were maintained throughout the study (Fig. 2, A–D), although the effect of T₃ to attenuate torpor was detected only after the first 24 h of treatment (Fig. 2A). During mild cold exposure, core temperature of ob/ob mice dropped despite a robust increase of energy expenditure during both the dark and light cycle (Fig. 3, A–C). As expected, T₃ treatment elicited a further increase of energy expenditure (Fig. 3, A–C), and although this effect was insufficient to restore normothermia to these animals, it markedly attenuated the incidence of torpor (Fig. 2A).

Fig. 2.

Time-course effect of a physiological dose of T₃ on core body temperature and measures of energy homeostasis during cold exposure. Core body temperature (A), heat production (B), ambulatory activity (C), and energy intake (D) in adult male ob/ob mice treated systemically with either veh or T₃ over time and subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. 14°C veh. T₃, triiodothyronine; veh, vehicle.

Fig. 3.

Effect of a physiological dose of T₃ on energy expenditure, respiratory quotient, and ambulatory activity in ob/ob mice during mild cold exposure. Photoperiod-averaged 24-h profiles and mean dark and light cycle measures of heat production (A–C), respiratory quotient (RQ; D–F), and ambulatory activity (G–I) using indirect calorimetry, respectively, in adult male ob/ob mice housed under room temperature conditions [22°C; Baseline: Group 1 (blue), Group 2 (red)] or treated systemically with either veh [Group 1 (blue)] or a physiological dose of T₃ [Group 2 (red)] and subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. 22°C paired Baseline; #P < 0.05 vs. 14°C veh. T₃, triiodothyronine; veh, vehicle.

Thyroid hormone replacement also did not block the effect of mild cold exposure to cause weight loss in ob/ob mice. Specifically, when ob/ob mice were moved from room temperature (22°C) to a cool environment (14°C), the weight loss effect was exacerbated by administration of a physiological dose of T₃ (−1.20 ± 0.36 g for veh vs. −2.35 ± 0.39 g for T₃; P < 0.05). This effect of T₃ was accompanied by a reduction in RQ (Fig. 3, D–F), indicative of increased fat oxidation, as well as increased ambulatory activity (Fig. 3, G–I), and each of these effects was detected during the first 24 h of treatment and maintained throughout the study (Fig. 2, B and C). Yet these effects of T₃ were insufficient to prevent the fall in core temperature of ob/ob mice subjected to mild cold exposure, suggesting that the effect of leptin to maintain normothermia in ob/ob mice subjected to sub-thermoneutral temperatures is not mediated by increased thyroid hormone levels.

To maintain stable fuel stores during cold exposure, WT mice increase their energy intake in proportion to their increase of energy expenditure (30). Although energy intake was increased during the light cycle in veh-treated ob/ob mice housed at 14°C (Fig. 4C), total daily energy intake did not change significantly, despite a clear increase of energy expenditure (Fig. 4, A–C). The possibility that because the animals are hyperphagic at baseline, ob/ob mice cannot increase food intake further during cold exposure (i.e., a ceiling effect on intake) seems unlikely given that these animals are capable of increasing food intake in other settings, such as fasting-induced refeeding (64). Treatment of ob/ob mice with T₃ did not rescue this energy homeostasis defect, as energy expenditure was again increased during mild cold exposure, and despite increased energy intake during the light cycle, total daily intake did not exceed values obtained when mice were housed at room temperature (22°C) (Fig. 4, A–C). Failure to mount a compensatory hyperphagia during cold exposure, therefore, appears to explain the weight loss phenotype of cold-exposed ob/ob mice, and this defect was, if anything, exacerbated by T₃ treatment.

Fig. 4.

Effect of a physiological dose of T₃ on energy intake and thermal conductance in ob/ob mice during mild cold exposure. Photoperiod-averaged 24-h energy intake (A and B), mean energy intake during the dark cycle, light cycle, and 24-h period (C), and whole-body thermal conductance (D) in adult male ob/ob mice housed under room temperature conditions [22°C; Baseline: Group 1 (blue), Group 2 (red)] or treated systemically with either veh [Group 1 (blue)] or a physiological dose of T₃ [Group 2 (red)] and subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. 22°C paired Baseline; #P < 0.05 vs. 14°C veh. T₃, triiodothyronine; veh, vehicle.

That the increase of energy expenditure induced by T₃ in ob/ob mice failed to prevent the fall of core body temperature during mild cold exposure suggests that the effect must have been offset by increased heat loss. This hypothesis is consistent with evidence that relative to WT mice, in ob/ob mice, hypothermia is linked to increased thermal conductance (i.e., increased heat loss to the environment) and that in these mice, leptin replacement corrects this defect (29, 30). To investigate this possibility, we calculated whole-body thermal conductance of ob/ob mice in the presence or absence of T₃ administration. As expected, relative to baseline conditions at room temperature (22°C), thermal conductance of ob/ob mice was reduced with mild cold exposure (14°C) (Fig. 4D). However, the effect of cold exposure to lower thermal conductance was attenuated in ob/ob mice that received T₃, indicative of increased heat loss to the environment (Fig. 4D). Taken together, these findings suggest that even though a physiological dose of T₃ increases energy expenditure in ob/ob mice during cold exposure, it fails to mimic the effect of leptin to reduce thermal conductance (29) and hence does not ameliorate the fall in core temperature during cold exposure in ob/ob mice.

Effect of co-administration of a physiological dose of T₃ and leptin on core temperature and energy balance in ob/ob mice during mild cold exposure.

To further investigate the extent to which normalization of thyroid hormone levels contributes to the effect of physiological leptin replacement to raise core temperature during mild cold exposure in ob/ob mice, a separate cohort of ob/ob mice received a continuous infusion of T₃ with either veh or leptin (Fig. 5A). By design, this dose of leptin raised plasma leptin levels in ob/ob mice into the physiological range (WT-veh: 1.73 ± 0.16 ng/ml vs. ob/ob T₃-lep: 1.75 ± 0.16 ng/ml) (Tables 1 and 2) (30). This normalization of plasma leptin levels was associated with a modest increase of total plasma T₄ levels in cold-exposed ob/ob mice (relative to veh-T₃-treated controls), and in both ob/ob veh-T₃- and lep-T₃-treated mice, the T₃ administration protocol raised the modestly reduced free T₃ levels characteristic of ob/ob mice (Table 1) into the normal range (Table 2).

Fig. 5.

Effect of co-administration of a physiological dose of T₃ and leptin on core body temperature in ob/ob mice during cold exposure. Schematic of study design (A), photoperiod-averaged 24-h core temperature in adult male ob/ob mice housed under room temperature conditions [22°C; Baseline: Group 1 (blue), Group 2 (red); B] or treated systemically with either a physiological replacement dose of T₃ and leptin [T₃-lep, Group 2 (red)] or T₃ alone [T₃-veh, Group 1 (blue)] and subjected to mild cold exposure (14°C; C). Mean change in photoperiod-averaged 24-h core temperature (D) and mean change in core temperature during the dark cycle, light cycle, and 24-h period (E) over all days of treatment in veh-treated mice housed under room temperature conditions (22°C; Baseline) relative to T₃-lep- and T₃-veh-treated mice subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. T₃-veh. lep, leptin; T₃, triiodothyronine; veh, vehicle.

Table 2.

Effect of co-administration of a physiological dose of T₃ and leptin on plasma T₄ and T₃ levels and leptin levels in ob/ob mice during mild cold exposure

Hormones	ob/ob T3-veh	ob/ob T3-lep
Leptin, ng/ml	0.00 ± 0.00	1.75 ± 0.25*
Total T4, µg/dl	0.60 ± 0.11	0.99 ± 0.17*
Free T3, pg/ml	5.22 ± 0.46	5.41 ± 0.29

Means ± SE; n = 7–8 per group. lep, leptin; T₃, triiodothyronine; T₄, thyroxine; veh, vehicle.

^*P < 0.05 vs. T₃-veh.

Consistent with our earlier observations (Fig. 1), we again found that although T₃ treatment without leptin was insufficient to prevent the fall of core temperature in ob/ob mice during mild cold exposure, this effect was prevented in T₃-treated ob/ob mice that also received a physiological replacement dose of leptin (Fig. 5, A–D). Thus, leptin replacement prevents this fall in core temperature in cold-exposed ob/ob mice irrespective of treatment designed to normalize thyroid status. We also show that in leptin-treated mice, core temperature progressively rises over time (mean core temperature day 1: 34.41 ± 0.35°C vs. day 2: 34.72 ± 0.39°C vs. day 3: 34.73 ± 0.38°C vs. day 4: 34.90 ± 0.34°C vs. day 5: 35.04 ± 0.36°C (Fig. 6A), consistent with our previous observations (29). Taken together, these findings show that leptin’s thermoregulatory actions involve mechanisms additional to increased thyroid hormone secretion.

Fig. 6.

Time-course effect during co-administration of a physiological dose of T₃ and leptin on core body temperature and measures of energy homeostasis during cold exposure. Core body temperature (A), heat production (B), ambulatory activity (C), and energy intake (D) in adult male ob/ob mice treated systemically with either a physiological dose of T₃ and leptin administered systemically (T₃-lep) or T₃ alone (T₃-veh) and subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. 14°C T₃-veh. lep, leptin; T₃, triiodothyronine; veh, vehicle.

Effects of co-administration of a physiological dose of T₃ and leptin on thermoeffector mechanisms in ob/ob mice during mild cold exposure.

Again, we found that during mild cold exposure (14°C), ob/ob mice lost weight when treated with T₃, and that this weight loss was further accentuated in mice that also received physiological leptin replacement (−3.77 ± 0.41 g for T₃-veh vs. −6.46 ± 0.41 g for T₃-leptin; P < 0.05). As expected, relative to room temperature conditions, mild cold exposure induced a marked increase of energy expenditure both in ob/ob mice that received T₃ alone and in those treated with T₃ and leptin (Fig. 7, A–C). Although this increase of energy expenditure was comparable between the two groups, RQ was lower among animals that received leptin as well as T₃, suggesting a further increase of fat oxidation (Fig. 7, D–F). Ambulatory activity levels were also increased in the T₃-leptin group compared with T₃ alone (Fig. 7, G–I). Although these data confirm the ability of leptin replacement to increase ambulatory activity of ob/ob mice, based on our previous findings (29), this effect is unlikely to play a major role in the adaptive increase of energy expenditure exhibited by these animals when subjected to sub-thermoneutral environments.

Fig. 7.

Effect of co-administration of a physiological dose of T₃ and leptin on energy expenditure, respiratory quotient, and ambulatory activity in ob/ob mice during mild cold exposure. Photoperiod-averaged 24-h profiles and mean heat production (A–C), respiratory quotient (RQ; D–F), and ambulatory activity (G–I) during the dark cycle and light cycle, respectively, in adult male ob/ob mice housed under room temperature conditions [22°C; Baseline: Group 1 (blue), Group 2 (red)] or treated systemically with either a physiological dose of T₃ and leptin [T₃-lep, Group 2 (red)] or T₃ alone [T₃-veh, Group 1 (blue)] and subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. 22°C paired Baseline; #P < 0.05 vs. 14°C T₃-veh. lep, leptin; T₃, triiodothyronine; veh, vehicle.

Our findings once again showed that despite a marked increase of energy expenditure, ob/ob mice treated with T₃ failed to increase energy intake during mild cold exposure (Fig. 8, A–C). Moreover, as expected, in those mice that also received leptin, energy intake was also reduced (Fig. 8, A–C). Thus, whereas energy expenditure was increased to a similar extent in T₃-veh- relative to T₃-leptin-treated ob/ob mice, energy intake was inhibited to a further extent in the latter group, resulting in greater weight loss. In addition, T₃ treatment failed to lower thermal conductance during cold exposure, whereas thermal conductance was reduced (i.e., reduced heat loss) in animals that also received leptin (Fig. 8D). Taken together, these findings demonstrate that the effect of leptin to defend core body temperature in ob/ob mice is mediated via mechanisms additional to activation of the HPT axis.

Fig. 8.

Effect of co-administration of a physiological dose of T₃ and leptin on energy intake and thermal conductance in ob/ob mice during mild cold exposure. Photoperiod-averaged 24-h energy intake (A and B), mean energy intake during the dark cycle, light cycle, and 24-h period (C), and whole-body thermal conductance (D) in adult male ob/ob mice housed under room temperature conditions [22°C; Baseline: Group 1 (blue), Group 2 (red)] or treated systemically with either a physiological dose of T₃ and leptin [T₃-lep, Group 2 (red)] or T₃ alone [T₃-veh, Group 1 (blue)] and subjected to mild cold exposure (14°C). n = 7–8 per group. Means ± SE. *P < 0.05 vs. 22°C paired Baseline; #P < 0.05 vs. 14°C T₃-veh. lep, leptin; T₃, triiodothyronine; veh, vehicle.

DISCUSSION

To better understand leptin’s role in thermoregulation and the coupling of this process to energy homeostasis, we investigated the contribution made by the action of leptin on the HPT axis. Thyroid hormone is a major determinant of both body temperature and thermogenesis, and leptin-deficient states are associated with subnormal thyroid hormone levels, an effect partially corrected with leptin treatment (1, 26, 34, 37, 41). Based on these observations, we sought to determine whether exogenous administration of a physiological replacement dose of thyroid hormone is sufficient to mimic the effect of leptin to raise the lower defended level of core temperature characteristic of ob/ob mice when exposed to a sub-thermoneutral environment. Our results extend previous evidence that circulating total T₄ and free T₃ hormone levels are lower in leptin-deficient ob/ob mice than WT controls and that this effect is partially attenuated by ameliorating leptin deficiency. We also show that energy expenditure, fat oxidation, and ambulatory activity are each increased when plasma T₃ levels are raised into the normal range in ob/ob mice, and evidence of torpor at light cycle onset is also diminished by thyroid hormone replacement. Nevertheless, this intervention failed to prevent the fall in core temperature during mild cold exposure characteristic of ob/ob mice, whereas restoring normal plasma leptin levels to these animals did prevent this hypothermic response (29), irrespective of whether T₃ was also administered.

A key difference in the response of ob/ob mice to administration of leptin versus thyroid hormone is that leptin reduces thermal conductance during mild exposure, whereas thyroid hormone does not. Together, these findings suggest that leptin’s ability to maintain core body temperature during cold exposure depends on both increased thermogenesis and reduced thermal conductance and that this combination of effects is not elicited by treatment with thyroid hormone alone. It therefore follows that the fall in core temperature in leptin-deficient mice when housed in a sub-thermoneutral environment is not due to suppression of the HPT axis and that the thermoregulatory actions of leptin are mediated by mechanisms additional to activation of the HPT axis.

Work from us and others demonstrates that in WT mice, both core body temperature and energy homeostasis are preserved across a variety of different environmental temperatures (22, 30, 47). During cold exposure, for example, increases of both energy expenditure and heat conservation are accompanied by a compensatory increase of energy intake that precisely offsets increased energy demand without changing overall body fuel stores (22). In ob/ob mice, however, direct exposure to the cold causes core temperature to decline in a manner that parallels the drop in environmental temperature (30, 62, 63). Since this effect is ameliorated by a physiological replacement dose of leptin (29), these data point to a crucial but poorly understood requirement for leptin in thermoregulation. Among potential mechanisms underlying the defense of a lower core temperature of ob/ob mice when housed at sub-thermoneutral temperatures, suppression of the HPT axis stands out, as hypothyroidism can cause hypothermia (8), and thyroid hormone levels are reduced in leptin deficient states (1). Based on these considerations, we first sought to better understand the effects of leptin deficiency and cold exposure on thyroid hormone levels.

Consistent with previous studies (26, 41), we found that plasma total T₄ levels and free T₃ levels were lower in ob/ob relative to WT controls and that this effect was partially corrected with leptin treatment. Humans with congenital leptin deficiency also exhibit abnormalities in the thyroid axis, and leptin administration can at least partially attenuate them (12, 15, 16, 37). However, the thyroid phenotype of leptin-deficient humans is heterogeneous, and some congenitally leptin-deficient individuals have relatively normal thyroid function, regardless of whether or not they undergo leptin treatment (44). Moreover, in humans with acquired leptin deficiency such as hypothalamic amenorrhea (a condition common among elite female athletes with markedly reduced body fat mass), thyroid hormone levels can be either low or in the normal range, and leptin administration is sufficient to raise plasma free T₃ levels (11, 70).

Our finding that the effect of leptin to prevent the fall in core temperature in ob/ob mice when subjected to mild cold exposure is accompanied by an increase in thyroid hormone levels is consistent with a role for the latter in leptin’s thermoregulatory effects. Consistent with this hypothesis, the ability of leptin administration to support survival of ob/ob mice during cold exposure, even if they are deficient in uncoupling protein-1 (the mitochondrial uncoupling protein that drives BAT thermogenesis), is accompanied by increased circulating thyroid hormone concentrations (65) that result from leptin stimulation of hypothalamic thyrotropin-releasing hormone neurons (23).

To test this hypothesis directly, we administered a physiological dose of thyroid hormone to ob/ob mice and subjected them to mild cold exposure (14°C). The effect of T₃ administration to increase energy expenditure is well established in both rodents and humans (39, 56), and treatment of ob/ob mice with a pharmacological dose of T₃ (>100 μg/kg) housed at room temperature increased both energy expenditure and core temperature to control levels (41). The physiological relevance of this effect remains uncertain, however, since the study design resulted in supraphysiological T₃ levels in plasma, and the effect in cold-exposed animals was not examined. To address these limitations, we investigated whether the fall in core temperature exhibited by ob/ob mice subjected to mild cold exposure is prevented by administration of a physiological dose of T₃ (4 μg/kg) for 5 days. Our finding that a physiological dose of thyroid hormone did not prevent the fall in core temperature during cold exposure is consistent with previous work showing that treatment of ob/ob mice with a similar dose of T₃ (5 μg/kg) failed to restore normal body temperature to ob/ob mice exposed to a transient but more severe cold stress (4°C for 45 min), despite having other unambiguous physiological effects (42). Moreover, even very large doses of T₃ (2 mg/kg) prolong survival of ob/ob mice during acute cold exposure (4°C) only for a few hours (35). Combined with evidence that 1) the fall in core temperature during cold exposure is greater in ob/ob mice than in WT mice rendered hypothyroid (by methimazole) (42), and 2) our current finding that treatment of ob/ob mice with T₃ at physiological doses does not prevent the fall in core temperature during mild cold exposure, we conclude that the effect of leptin to prevent hypothermia in ob/ob mice exposed to a sub-thermoneutral environment cannot be explained by its effect on thyroid hormone levels.

The failure of T₃ treatment to prevent the fall in core temperature in ob/ob mice during mild cold exposure occurs despite appropriate thermogenic responses, including increased energy expenditure, a shift in substrate utilization from carbohydrate to fat, and an increase of ambulatory activity. Although thyroid hormone also regulates heat dissipation (67), we found that in ob/ob mice, T₃ administration was associated with increased thermal conductance. This tendency of T₃ to promote heat loss may explain its inability prevent the fall in core temperature of ob/ob mice during mild cold exposure despite its effect to increase energy expenditure. Since leptin replacement in these animals increases thermogenesis while also reducing thermal conductance (29), we infer that the combination of both effects is indispensable for maintenance of normal body temperature during cold exposure, and that leptin, but not thyroid hormone, plays a physiological role to engage both heat production and heat conservation. To our knowledge, leptin is unique in this capacity, and these effects involve leptin targets additional to the HPT axis.

Daily torpor episodes are a state characterized by an acute fall in both core temperature and energy expenditure during the resting phase of the day and are one mechanism that conserves body fuel stores in conditions of reduced food availability and/or decreased ambient temperature (46, 58). A low leptin level is implicated as a physiological signal for entering torpor, since even ad libitum fed ob/ob mice housed at room temperature often enter torpor at light cycle onset (58, 68) and since leptin replacement blunts torpor bouts in both food-restricted WT (13) and ob/ob mice (20). Our finding that T₃ administration blunts torpor in cold-exposed ob/ob mice at light cycle onset (even in the absence of leptin replacement) supports previous evidence that reduced thyroid hormone signaling can predispose to torpor (4). While the mechanisms whereby T₃ prevents torpor remain to be elucidated, thyroid hormone can modulate the effect whereby the sympathetic nervous system activates BAT thermogenesis (56). Given the importance of BAT thermogenesis in small mammals, such an effect may well contribute to protection against torpor observed in T₃-treated ob/ob mice. A potential mechanism may involve a direct action of thyroid hormone in the hypothalamus, which has been implicated in seasonal regulation of torpor in hibernating mammals (14, 40). However, torpor onset appears to be controlled by mechanisms additional to reduced leptin or thyroid hormone signaling since neither hormone prevents torpor in a transgenic (A-ZIP/F-1) mouse model of white adipose tissue depletion (20). Moreover, although T₃ treatment attenuated daily episodes of torpor, it was not sufficient to prevent the fall in core temperature in ob/ob mice housed in a sub-thermoneutral environment.

To meet the increased thermogenic needs associated with cold exposure, energy intake of normal mice can increase dramatically. In normal animals, food intake increases in direct proportion to increased energy expenditure such that body fat mass does not change, and the underlying mechanisms appear to be leptin-dependent since ob/ob mice fail to show this response (3, 30, 57, 60). These findings contrast with a previous report suggesting that both WT and ob/ob mice increase their food intake under a cold challenge (27). However, the discrepancy between these findings and ours is likely explained by the fact that these latter studies were conducted using a paradigm in which ambient temperature was progressively and modestly reduced over a significantly greater time than ours, whereas the current studies involved acute and sustained changes in ambient temperature. Since mice deficient in leptin signaling are cold intolerant when acutely exposure to a cold challenge (62, 63) yet are able to adapt to a lower temperature if the exposure is gradual (18, 61, 65), it suggests that although leptin-deficient animals are less metabolically flexible, they have the capacity to engage leptin-independent mechanisms to regulate feeding and core temperature over time. However, our finding that T₃ treatment fails to restore the normal relationship between energy intake and energy expenditure in ob/ob mice subjected to mild cold exposure implies that leptin-sensitive coupling of thermoregulation to energy homeostasis also involves mechanisms additional to activation of the thyroid axis. Consistent with this view, T₃-treated ob/ob mice lose more body weight than veh-treated controls because they fail to stimulate energy intake appropriately. We acknowledge, however, that this observation may be related to the duration of treatment, since T₃ treatment increases energy expenditure within 24 h, yet the effect to induce hyperphagia may require several days (43).

In contrast to T₃ treatment alone, physiological leptin replacement prevented the fall in core temperature during mild cold exposure in ob/ob mice regardless of whether it was combined with T₃ treatment. Consistent with previous observations (18, 29), these effects were not explained by effects of leptin on energy expenditure, substrate oxidation, or ambulatory activity; rather, the effect was accompanied by a reduction in thermal conductance (i.e., reduced heat loss) that did not occur with T₃ alone. Interestingly, as was observed with leptin replacement alone (29), core body temperature increased progressively over time when leptin was given in combination with T₃. Whether this time-dependent increase of temperature involves heat-conserving or heat-producing mechanisms (or both) and whether it requires BAT activation are questions that warrant further study.

While the thermogenic effects of thyroid hormone clearly involve actions on BAT and other peripheral tissues, growing evidence suggests that the brain is a target for these effects as well. Specifically, thyroid hormone action in the brain reduces phosphorylation of AMP-activated protein kinase activity in the ventromedial nucleus of the hypothalamus, an effect that increases sympathetic tone to BAT (31, 33). Such a mechanism is consistent with evidence that central administration of T₃ increases thermogenesis and raises core temperature in WT mice housed under thermoneutral conditions via a mechanism that requires uncoupling protein-1 (2). Our data raise the possibility that intact leptin action may be required for thermoregulatory effects of T₃, and additional studies are warranted to investigate this hypothesis.

Together, our findings support an overarching model whereby in response to a fall of ambient temperature, a series of physiological and behavioral responses are engaged to maintain core temperature within a narrow range (59). While these responses are influenced by external factors, including the availability of food, severity of cold exposure, access to shelter, or warmer environments, key features of the physiological responses to cold exposure include activation of BAT thermogenesis and skeletal muscle shivering to increase heat production and vasoconstriction of cutaneous blood vessels to prevent heat loss. In addition to increased food intake to meet increased fuel needs, associated behavioral responses include warm seeking (e.g., nesting/burrowing, etc.) (Fig. 9A). However, when leptin-deficient ob/ob mice are subjected to sub-thermoneutral environments, they fail to maintain core temperature even though they retain the capacity to increase energy expenditure (18, 30, 62, 63). Their defect appears to lie in their ability both to conserve heat and to increase food intake, with the result that they exhibit not only hypothermia punctuated by episodes of torpor but also weight loss (20, 30) (Fig. 9B). A causal role for leptin deficiency in these defects is suggested by our evidence that physiological leptin replacement corrects both the fall in core temperature and episodes of torpor during mild cold exposure in these mice, effects mediated in part by reduced thermal conductance (i.e., increased heat conservation) (29) (Fig. 9C). While these leptin effects are associated with increased circulating thyroid hormone levels, our finding that a physiological dose of T₃ (in the absence of leptin) does not protect these mice from hypothermia during mild cold exposure (despite the expected thermogenic response) suggests that leptin’s role in thermoregulation involves mechanisms independent of the action of thyroid hormone.

Fig. 9.

Regulation of body temperature. Conceptual model whereby physiological and behavioral effectors are engaged for controlling body temperature during mild exposure in wild-type mice (A), ob/ob mice (B), and ob/ob mice treated with a physiological dose of leptin or T₃ (C). BAT, brown adipose tissue; T₃, triiodothyronine; WT, wild type.

GRANTS

This work was supported by National Institutes of Health Grant DK-089056 (G. J. Morton) and the National Institute of Diabetes and Digestive and Kidney Diseases-funded Nutrition Obesity Research Center Grant DK-035816, and the University of Washington Nutrition, Obesity and Atherosclerosis Training Grant T32 HL007028 and Diabetes, Obesity and Metabolism Training Grant T32 DK007247.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.D.D., K.J.K., and G.J.M. conceived and designed research; J.D.D., K.M., K.O., J.T.N., K.R.V., and G.J.M. performed experiments; J.D.D., K.M., K.O., J.T.N., K.R.V., K.J.K., and G.J.M. analyzed data; J.D.D., K.M., K.O., K.J.K., and G.J.M. interpreted results of experiments; J.D.D. and G.J.M. prepared figures; J.D.D. and G.J.M. drafted manuscript; J.D.D., K.M., K.O., J.T.N., K.R.V., K.J.K., and G.J.M. edited and revised manuscript; J.D.D., K.M., K.O., J.T.N., K.R.V., K.J.K., and G.J.M. approved final version of manuscript.