Role for kisspeptin/neurokinin B/dynorphin (KNDy) neurons in cutaneous vasodilatation and the estrogen modulation of body temperature

Abstract

Estrogen withdrawal in menopausal women leads to hot flushes, a syndrome characterized by the episodic activation of heat dissipation effectors. Despite the extraordinary number of individuals affected, the etiology of flushes remains an enigma. Because menopause is accompanied by marked alterations in hypothalamic kisspeptin/neurokinin B/dynorphin (KNDy) neurons, we hypothesized that these neurons could contribute to the generation of flushes. To determine if KNDy neurons participate in the regulation of body temperature, we evaluated the thermoregulatory effects of ablating KNDy neurons by injecting a selective toxin for neurokinin-3 expressing neurons [NK₃-saporin (SAP)] into the rat arcuate nucleus. Remarkably, KNDy neuron ablation consistently reduced tail-skin temperature (T_SKIN), indicating that KNDy neurons facilitate cutaneous vasodilatation, an important heat dissipation effector. Moreover, KNDy ablation blocked the reduction of T_SKIN by 17β-estradiol (E₂), which occurred in the environmental chamber during the light phase, but did not affect the E₂ suppression of T_SKIN during the dark phase. At the high ambient temperature of 33 °C, the average core temperature (T_CORE) of ovariectomized (OVX) control rats was significantly elevated, and this value was reduced by E₂ replacement. In contrast, the average T_CORE of OVX, KNDy-ablated rats was lower than OVX control rats at 33 °C, and not altered by E₂ replacement. These data provide unique evidence that KNDy neurons promote cutaneous vasodilatation and participate in the E₂ modulation of body temperature. Because cutaneous vasodilatation is a cardinal sign of a hot flush, these results support the hypothesis that KNDy neurons could play a role in the generation of flushes.

Estrogen withdrawal leads to hot flushes in the majority of menopausal women (1). Hot flushes are also experienced by men and women treated with tamoxifen for breast cancer, men undergoing androgen-ablation therapy for prostate cancer, young oophorectomized women, and hypogonadal men (2, 3). A hot flush is characterized by episodic activation of heat dissipation effectors, including cutaneous vasodilatation, sweating, and behavioral thermoregulation. After a flush is initiated, the activation of heat dissipation mechanisms is so effective that core temperature frequently drops (4). Despite the vast numbers of individuals affected, the etiology of flushes remains an enigma.

Hot flushes are closely timed with luteinizing hormone (LH) pulses, providing a clue that the generation of flushes is linked to the hypothalamic neural circuitry controlling pulsatile gonadotropin-releasing hormone (GnRH) secretion (5, 6). Current evidence suggests that pulsatile GnRH secretion is modulated by a subpopulation of neurons in the arcuate (infundibular) nucleus that express estrogen receptor α (ERα), neurokinin 3 receptor (NK₃R), kisspeptin, neurokinin B (NKB), and dynorphin (7–11). In the hypothalamus of postmenopausal women, these kisspeptin/NKB/dynorphin (KNDy) neurons undergo an unusual somatic hypertrophy and express increased kisspeptin and NKB gene transcripts (12–15). Studies of cynomolgus monkeys indicate that the changes in KNDy neurons in postmenopausal women are secondary to withdrawal of ovarian estrogens and not due to aging per se (13, 15–17). Mutations in the genes encoding kisspeptin, NKB, or their receptors result in hypogonadotropic hypogonadism, a syndrome characterized by lack of pubertal development, impaired gonadotropin secretion, absence of secondary sex characteristics, and infertility (18–21). Thus, the hypertrophied neurons in the hypothalamus of postmenopausal women express two peptides, kisspeptin and NKB, that are essential for human reproduction.

Because ERα-expressing KNDy neurons are markedly altered in the hypothalamus of postmenopausal women, we hypothesized that these neurons could be involved in the generation of hot flushes (12). Consistent with this hypothesis, tract-tracing studies in the rodent have shown that KNDy neurons project to structures critical for the regulation of body temperature (22, 23) including the median preoptic nucleus (MnPO), an important component of a thermosensory heat-defense pathway (24). Moreover, pharmacological activation of NK₃R-expressing neurons in the MnPO elicits a robust decrease in T_CORE and alters tail skin vasomotion (25). Whereas these data indicate that NK₃R signaling may influence body temperature at the level of the MnPO, there is no information on whether KNDy neurons modulate body temperature or mediate the effects of estrogen on the thermoregulatory system.

Studies of laboratory rats have provided useful information on the effects of estrogens on the activation of heat-dissipation effectors. Cutaneous vasodilatation of the rat's tail is a major heat-defense effector that can be evaluated indirectly by measuring changes in tail skin temperature (T_SKIN) (26, 27). The vasomotor state of the tail varies across the estrous cycle, with the lowest levels of vasodilatation on proestrous night, when estrogen levels are high (28). Consistent with these findings, tail skin vasodilatation is increased by ovariectomy and decreased by treatment with estrogens (28–31). Although the increase in the average T_SKIN after ovariectomy does not reflect an episodic event, like a flush, the reduction of T_SKIN by a pharmacological agent has been widely used to evaluate its efficacy for reducing flushes in women (29, 32–35). Moreover, 17β-estradiol (E₂) replacement in ovariectomized (OVX) rats shifts the thermoneutral zone (36), reduces Fos protein expression in the MnPO (37), and alters heat-escape behavior (38). Finally, at high ambient temperatures (T_AMBIENT), OVX rats exhibit higher core temperatures (T_CORE) relative to OVX rats treated with E₂ (36, 39). Whereas these studies show pronounced effects of E₂ on thermoregulation in the rat, the underlying neural circuitry mediating these changes is largely unknown.

In the present study, we evaluate the thermoregulatory effects of ablating KNDy neurons using stereotaxic injections of NK₃-saporin (SAP), a selective neurotoxin for NK₃R-expressing cells. The utility of NK₃-SAP for selective ablation of KNDy neurons was recently documented (11). Circadian rhythms of T_SKIN, T_CORE, and activity were evaluated in KNDy-ablated and control rats that were OVX and then replaced with E₂. Rats were also exposed to subneutral, neutral, and supraneutral T_AMBIENT in an environmental chamber to determine if KNDy neuron ablation altered the ability to defend T_CORE in response to environmental temperature challenges. Our previous studies have shown that KNDy neurons are required for tonic gonadotropin secretion, the rise in LH secretion after ovariectomy, and the E₂ modulation of body weight (11). Here we determine if KNDy neurons also participate in the E₂ modulation of thermoregulation.

Results

Experiment 1: KNDy Neuron Ablation Decreased Tail Skin Vasodilatation.

The experimental protocol is illustrated in Fig. 1. Prominent circadian rhythms of T_SKIN and T_CORE were observed in both NK₃-SAP and Blank-SAP controls. T_CORE and T_SKIN exhibited a circadian periodicity of 24 h, with no significant differences in circadian rhythms between groups. However, inspection of the average circadian waveforms revealed consistently lower T_SKIN in NK₃-SAP rats compared with controls (Fig. 2). Statistical analysis confirmed that the average T_SKIN of NK₃-SAP rats was significantly lower than Blank-SAP controls in both phases of the light/dark cycle and in OVX and OVX + E₂ conditions (Figs. 2 and 3B). During the dark phase of the circadian rhythm recordings, E₂ treatment of OVX Blank-SAP and NK₃-SAP rats reduced T_SKIN (Figs. 2 and 3B). During the light phase, E₂ treatment did not lower T_SKIN in either Blank-SAP or NK₃-SAP rats (Fig. 3B). The heat loss index (HLI), an index of tail skin vasomotion that controls for ambient and core temperatures (26), was also consistently lower in NK₃-SAP rats than Blank-SAP controls (Table 1). These results indicate that KNDy-ablated rats have lower levels of tail skin vasodilatation than Blank-SAP controls.

Fig. 1.

Experimental protocol. On day 0, the rats were ovariectomized, injected with Blank-SAP or NK₃-SAP in the arcuate nucleus, implanted with telemetry devices and returned to their home cages. After a 12- to 15-d recovery period (shown as day 14 in this example), T_CORE, T_SKIN, and activity were recorded every 10 min over 5 d to evaluate changes over the light/dark cycle (circadian). The next three mornings, T_CORE and T_SKIN were recorded in rats exposed to T_AMBIENT of 26 °C, 11 °C, and 33 °C (in that order) in an environmental chamber. One to 3 d later (shown as day 22 in this example), rats received s.c. E₂ capsules. The circadian rhythm recordings were repeated, followed by the T_AMBIENT exposures. The rats were killed 11 d after E₂ treatment. E₂,17β-estradiol; OVX, ovariectomized; SAP, saporin; T_AMBIENT, ambient temperature.

Fig. 2.

Effects of KNDy neuron ablation on average T_SKIN in OVX (A) and OVX + E₂ rats (B). (A) Circadian rhythms of T_SKIN were not different between NK₃-SAP and Blank-SAP rats. However, OVX NK₃-SAP rats exhibited consistently lower T_SKIN than OVX Blank-SAP rats. (B) After E₂ treatment, T_SKIN was still consistently lower in NK₃-SAP rats than Blank-SAP rats. E₂ reduced T_SKIN during the dark phase in both Blank-SAP and NK₃-SAP rats (compare with A). These waveforms represent the average T_SKIN for each group (6–11 rats per group) and were generated using a moving average of five data points. Dark bars represent the dark phase of the 24-h cycle, and numbers refer to circadian recording days 3, 4, and 5. Statistical analyses of these data are shown in Fig. 3B.

Fig. 3.

Effects of KNDy neuron ablation and E₂ treatment on T_CORE (A), T_SKIN (B), and activity (C) during the light (Left) and dark (Right) phases. (A) Average T_CORE was increased in the dark phase (compared with light) in all groups with no significant effect of NK₃-SAP or E₂ treatment. (B) Average T_SKIN was consistently lower in NK₃-SAP rats compared with Blank-SAP rats, indicative of lower levels of cutaneous vasodilatation. In the dark phase, T_SKIN was decreased by E₂ treatment in both Blank-SAP and NK₃-SAP rats. (C) Activity (detected by the DSI telemetry probe) was markedly increased in the dark (active) phase, with no significant differences between Blank-SAP– and NK₃-SAP–treated groups. There was a trend (P = 0.06) for E₂ to increase activity in the dark phase in both Blank-SAP and NK₃-SAP rats. Values represent mean ± SEM, n = 6–11 rats per group. *Significantly different (NK₃-SAP vs. Blank-SAP, within OVX or OVX + E₂). ⁺Significantly different (OVX vs. OVX + E₂, within Blank-SAP or NK₃-SAP).

Table 1.

Heat loss index in control (Blank-SAP) and KNDy-ablated rats (NK₃-SAP)

	Blank-SAP		NK3-SAP
	OVX	OVX + E2	OVX	OVX + E2
Light phase	0.49 ± 0.01	0.47 ± 0.02	0.39 ± 0.01*	0.42 ± 0.01*
Dark phase	0.39 ± 0.01	0.24 ± 0.02+	0.29 ± 0.01*	0.15 ± 0.02*,+

Values represent mean ± SEM calculated from 6-h time blocks in the middle of each phase, n = 8–11 rats per group.

*Significantly different (NK₃-SAP vs. Blank-SAP, within OVX or OVX + E₂).

⁺Significantly different (OVX + E₂ vs. OVX, within Blank-SAP or NK₃-SAP).

T_CORE was ∼0.5 °C higher during the dark phase than in the light phase, but within each phase, there was no effect of NK₃-SAP or E₂ treatment on the average T_CORE (Fig. 3A). The average activity of all groups of rats was markedly increased in the dark phase but there was no significant effect of NK₃-SAP on activity within either the dark or the light phase (Fig. 3C). There was a trend for E₂ to increase activity in the dark phase in both Blank-SAP and NK₃-SAP rats (P = 0.06).

Experiment 2: At the High T_AMBIENT of 33 °C, the T_CORE of OVX KNDy-Ablated Rats Was Lower than in OVX Blank-SAP Rats and Unresponsive to E₂ Replacement.

At the supraneutral T_AMBIENT of 33 °C, the average T_CORE of all groups was elevated, compared with rats exposed to the neutral T_AMBIENT of 26 °C. However, the highest body temperature was observed in OVX Blank-SAP controls at the T_AMBIENT of 33 °C (Fig. 4A). Consistent with previous studies (36, 37, 39), this hyperthermic T_CORE was significantly reduced by E₂ treatment (Fig.4A). At the T_AMBIENT of 33 °C, the average T_CORE of OVX NK₃-SAP rats was significantly lower than that of OVX Blank-SAP rats. Moreover, unlike OVX controls, the average T_CORE of OVX NK₃-SAP rats was not significantly reduced by E₂ treatment. At the lower T_AMBIENT of 11 °C, T_CORE was not altered by E₂ or NK₃-SAP, except for a mild increase in T_CORE in OVX E₂-treated NK₃-SAP rats, compared with OVX E₂-treated Blank-SAP controls.

Fig. 4.

Average T_CORE (A), T_SKIN (B), and heat loss index (HLI) (C) in Blank-SAP and NK₃-SAP rats exposed to T_AMBIENT of 11 °C, 26 °C, or 33 °C (Left to Right). (A) At the T_AMBIENT of 33 °C the T_CORE was the highest in OVX Blank-SAP rats and reduced in these rats by E₂ treatment. At the T_AMBIENT of 33 °C, the T_CORE of NK₃-SAP rats was significantly lower than Blank-SAP animals and was not affected by E₂. (B) At the T_AMBIENT of 11 °C and 26 °C, the T_SKIN of NK₃-SAP animals was significantly lower than in Blank-SAP animals. The T_SKIN of the Blank-SAP rats was reduced by E₂ at all T_AMBIENT, but T_SKIN was not lowered by E₂ treatment in NK₃-SAP rats at any T_AMBIENT. (C) At the T_AMBIENT of 11 °C and 26 °C, the average HLI of NK₃-SAP rats was significantly lower than in Blank-SAP rats, indicating lower levels of vasodilatation. HLI was reduced by E₂ in Blank-SAP rats but not in NK₃-SAP rats. Values represent mean ± SEM, 6–11 rats per group. *Significantly different (NK₃-SAP vs. Blank-SAP, within OVX or OVX + E₂). ⁺Significantly different (OVX + E₂ vs. OVX, within Blank-SAP or NK₃-SAP).

KNDy Neuron Ablation Decreases Tail Skin Vasodilatation in Rats Exposed to Subneutral and Neutral T_AMBIENT and Prevents the Effects of E₂ on T_SKIN.

At all three T_AMBIENT, OVX Blank-SAP rats exhibited the highest T_SKIN (Fig. 4B). In agreement with previous studies (e.g., ref. 37), E₂ treatment significantly decreased T_SKIN and HLI in OVX Blank-SAP controls (Fig. 4 B and C). In contrast, E₂ had no effect on T_SKIN or HLI in NK₃-SAP rats at any ambient temperature. Similar to the findings observed during the circadian recordings, at the T_AMBIENT of 11 °C and 26 °C, the T_SKIN and HLI were significantly lower in NK₃-SAP rats than Blank-SAP rats, in both OVX and OVX + E₂ conditions, indicative of lower levels of tail skin vasodilatation. NK₃-SAP rats exposed to the T_AMBIENT of 33 °C exhibited a higher HLI than NK₃-SAP rats at the T_AMBIENT of 26 °C (Fig. 4C). Thus, despite the consistently lower levels of tail vasodilatation in NK₃-SAP rats at lower T_AMBIENT, these rats were still capable of tail skin vasodilatation when exposed to a high T_AMBIENT. At the supraneutral T_AMBIENT of 33 °C, OVX NK₃-SAP rats exhibited a lower T_SKIN than OVX Blank-SAP controls but the HLI was not significantly different between these two groups.

Discussion

Estrogen withdrawal alters thermoregulation in rats and causes hot flushes in humans, but the neural circuits underlying these effects are unknown. In the present study, we determined the thermoregulatory effects of ablating KNDy neurons using stereotaxic injections of NK₃-SAP. Remarkably, KNDy-ablated rats exhibited robust decreases in T_SKIN and HLI, indicating decreased tail skin vasodilatation. This effect occurred independent of E₂ treatment, throughout the light/dark cycle, and in the environmental chamber at the T_AMBIENT of 11 °C and 26 °C. Although tail skin vasoconstriction is a physiological mechanism that conserves heat, KNDy-ablated rats were able to regulate T_CORE. In addition, KNDy-ablated rats were still capable of tail skin vasodilatation when exposed to the high T_AMBIENT of 33 °C. Under all other conditions, the HLIs of KNDy-ablated animals were lower than controls, indicating that KNDy neurons facilitate cutaneous vasodilatation.

KNDy neurons could facilitate tail skin vasodilatation via projections to rostral hypothalamic structures that control thermoregulatory effectors, such as the medial preoptic area and MnPO (22). Notably, using single-cell transcriptomics, tachykinin receptor 3 (Tacr3) gene expression (encoding NK₃R) was detected in warm-sensitive, GABAergic neurons in the medial preoptic area (40). Excitation of warm-sensitive, GABAergic neurons results in tail skin vasodilatation through projections that inhibit tonic activation of sympathetic (vasoconstrictor) premotor neurons in the ventromedial medulla (41, 42). Therefore, in KNDy-ablated rats, removal of an excitatory input to NK₃R-expressing warm-sensitive medial preoptic neurons could explain the lower levels of tail skin vasodilatation. KNDy neurons could also influence tail vasomotion via projections to the MnPO, a major component of a thermosensory pathway regulating heat-defense effectors (24). MnPO Fos expression is modulated by changes in T_AMBIENT or chronic E₂ treatment, implicating this nucleus as a site for integrating thermoregulatory and reproductive functions (37). MnPO neurons express the NK₃R protein (25) and Tacr3 mRNA (43). Moreover, focal microinfusion of an NK₃R agonist activates MnPO neurons (as measured by Fos), elicits a robust drop in T_CORE and inhibits sympathetic vasoconstriction (25). MnPO neurons project to other thermoregulatory regions of the hypothalamus (24, 44, 45) as well as neurons in the ventromedial medulla controlling tail vasomotion (42, 46–48). Thus, removal of excitatory KNDy neuron projections to the MnPO or medial preoptic area (or both) could provide a mechanism for the lower levels of tail skin vasodilatation in KNDy-ablated rats.

Consistent with numerous studies, tail skin vasodilatation was decreased by E₂ in OVX control rats (28–31, 36, 38). During the dark phase of the circadian recordings, E₂ treatment of OVX controls markedly reduced T_SKIN and HLI. E₂ did not alter T_SKIN or HLI during the light phase of the circadian recordings (3–5 d after E₂ capsules), but did decrease T_SKIN and HLI during the light phase in the T_AMBIENT experiments (6–9 d after E₂ capsules). This discrepancy is consistent with previous studies that have reported that E₂ effects on T_SKIN during the light phase do not appear until 7 d after E₂ replacement (28). Interestingly, ablation of KNDy neurons blocked the effect of E₂ on T_SKIN in the light phase (during the T_AMBEINT challenges), although it did not prevent the effect of E₂ on T_SKIN during the dark phase. Thus, KNDy neurons are essential for E₂ to reduce tail skin vasodilatation during the light phase, but not during the dark phase.

At the supraneutral T_AMBIENT of 33 °C, the average T_CORE of OVX control rats increased to 39.0 °C and this value was reduced by E₂ treatment. These findings are similar to other studies showing that E₂ treatment improved the ability to maintain core temperature when exposed to high ambient temperatures (36, 39). Because the primary function of hypothalamic thermoregulation is to defend T_CORE across a wide range of T_AMBIENT (49), these data indicate that ovariectomy induces thermoregulatory dysfunction, which is corrected by E₂ treatment. Interestingly, OVX KNDy-ablated rats exposed to the high T_AMBIENT were able to maintain their body temperature closer to normal than OVX control rats. Moreover, unlike controls, E₂ did not reduce T_CORE in rats with KNDy neuron ablation. These results indicate that KNDy neurons mediate the E₂ reduction of body temperature in OVX rats exposed to a high T_AMBIENT.

We do not understand how KNDy neuron ablation improved T_CORE regulation at the supraneutral T_AMBIENT, just as we do not understand how E₂ treatment protects against hyperthermia at this T_AMBIENT. A change in metabolic activity does not explain the lower T_CORE in OVX E₂-treated controls because E₂ treatment of OVX rodents either has no effect (38), or increases metabolic activity (50), which would result in a higher T_CORE, not lower. The E₂ effect on T_CORE in control rats was also not secondary to increased vasodilatation, because the HLI was reduced by E₂, reflecting decreased vasodilatation. However, in rats exposed to a supraneutral T_AMBIENT, vasodilatation becomes maximal and insufficient to further reduce T_CORE. Other heat-defense effectors are activated, including evaporative cooling from saliva spreading, postural changes, and heat escape behavior (26, 51). T_CORE is also influenced by numerous other factors, such as respiratory rate, the surface to mass ratio, insulation, and integration with homeostatic systems controlling water balance and food intake (51). Therefore, many factors could potentially play a role in the E₂ reduction of T_CORE in a supraneutral T_AMBIENT. Although the mechanism remains to be determined, these findings are relevant to understanding thermoregulatory dysfunction in menopausal women, who tolerate heat stress better after estrogen replacement therapy (52).

We have previously shown that KNDy neurons are required for the rise in LH secretion after the removal of E₂ negative feedback (11). Fig. 5 illustrates a hypothetical circuit for KNDy neurons to mediate E₂ effects on LH secretion and cutaneous vasodilatation in the rat. We hypothesized that estrogen withdrawal in humans leads to menopausal flushes by increasing the activity of KNDy neurons. This hypothesis is supported by the increased NKB and kisspeptin gene expression in postmenopausal women (12, 13), the close timing of LH pulses and flushes (5), the putative role of KNDy neurons in stimulating pulsatile LH secretion (7–10), the expression of ERα in KNDy neurons (12), and the demonstration that estrogen suppresses hot flushes (3) as well as NKB and kisspeptin gene expression in young ovariectomized monkeys (13, 16). In the present study, ablation of KNDy neurons consistently decreased tail skin vasodilatation in OVX rats, mimicking the thermoregulatory effects of E₂ treatment. In addition, KNDy-neuron ablation blocked the effects of E₂ on T_CORE in a warm environment and T_SKIN during the light phase. The demonstration that KNDy neurons facilitate tail skin vasodilatation is particularly relevant, because cutaneous vasodilatation is a cardinal feature of the menopausal hot flush. These findings provide strong support for the hypothesis that KNDy neurons could play a role in the generation of flushes.

Fig. 5.

(A) Relationship between KNDy neurons, GnRH neurons, and the heat-defense pathway in the rat. KNDy neurons branch within the arcuate nucleus and project to GnRH terminals in the median eminence (22, 56) and preoptic areas that regulate body temperature, including the MnPO (22–24, 42, 45, 48). Secretion of GnRH into portal capillaries stimulates LH secretion from the anterior pituitary gland, which stimulates the secretion of E₂ from the ovaries. E₂ negative feedback reduces serum LH and decreases NKB and kisspeptin mRNA in KNDy neurons (57, 58). ERα, the isoform required for estrogen negative feedback (59), is expressed in arcuate KNDy neurons (60) but not GnRH neurons (61). NK₃R is expressed on arcuate KNDy neurons (60) and GnRH terminals in the median eminence (56). GnRH neurons express kisspeptin receptor mRNA (62), but the distribution of the kisspeptin receptor protein on GnRH neurons has not been described. MnPO neurons express NK₃R and pharmacological activation of these neurons reduces body temperature (25). The MnPO receives information from warm-sensitive, cutaneous thermoreceptors and projects to CNS centers to modulate heat-dissipation effectors (24, 42). (B) Effects of ovariectomy on serum LH and tail skin vasodilatation: Loss of E₂ after ovariectomy markedly increases NKB and kisspeptin gene expression in KNDy neurons (57, 58), increases GnRH secretion into the portal capillaries (63) and LH secretion from the anterior pituitary gland. In the MnPO, at neutral T_AMBIENT, Fos is increased in OVX rats, compared with OVX + E₂ rats (37). Tail skin vasodilatation is increased by ovariectomy and decreased in OVX rats by E₂ replacement (28–31). (C) Effects of KNDy neuron ablation in OVX rats: KNDy neuron ablation lowers serum LH and prevents the rise in LH secretion after ovariectomy (11). Because KNDy neurons do not project to the portal capillary system to directly influence pituitary gonadotrophs (56), this effect is likely mediated via lower levels of GnRH secretion. Ablation of KNDy neurons (with loss of their projections to the rostral hypothalamus) also reduces tail skin vasodilatation. These findings suggest that withdrawal of E₂ in OVX rats increases LH secretion and tail skin vasodilatation by increasing the activity of KNDy neurons. AP, anterior pituitary gland; ERα, estrogen receptor α; E₂, estradiol-17β; GnRH, gonadotropin-releasing hormone; LH, luteinizing hormone; Kiss, kisspeptin; KNDy, kisspeptin, neurokinin B, and dynorphin-expressing neurons; MnPO, median preoptic nucleus; NKB, neurokinin B; NK₃R, neurokinin 3 receptors; oc, optic chiasm.

Methods

Young adult, female Sprague-Dawley rats (approximately 12 wk old, 200–250 g; Harlan Laboratories) were individually housed in a quiet, temperature- and humidity-controlled room (ambient temperature of 21.1–22.5 °C, humidity set at 50%) in the University of Arizona Animal Care Facility with a 12:12 h light:dark cycle (lights on at 0700 h). Rats had ad libitum access to water and a low-phytoestrogen diet (Harlan Teklad 2014; Harlan Laboratories). All protocols were approved by the University of Arizona Animal Care and Use Committee and conformed to National Institutes of Health (Bethesda, MD) guidelines. The serum hormone levels, body weights, and the brain histology from rats used in the present study have been previously described (11). Please refer to ref. 11 for detailed methods and documentation of the selectivity of KNDy neuron ablation.

Rats (n = 24) were OVX under general anesthesia using a mixture (1.0 mL/kg i.m.) containing ketamine (33.3 mg/mL), xylazine (10.7 mg/mL), and acepromazine (1.3 mg/mL). A PhysioTel transmitter (TA10-F40; Data Sciences International, DSI) was inserted into the peritoneal cavity for measurements of T_CORE and activity. Stereotaxic surgery was used to make bilateral injections of 10 ng NK₃-SAP (Advanced Targeting Systems; n = 14) in 100 nL PBS into the arcuate nucleus (two on each side). Control rats (n = 10) received 4 arcuate injections of 10 ng of Blank-SAP (11-aa peptide conjugated to saporin, Advanced Targeting Systems) in 100 nl PBS. Twenty to 23 d after the initial surgery, under isoflurane anesthesia, rats were implanted with two s.c. capsules (each 20 mm effective length, 1.57 mm inner diameter, 3.18 mm outer diameter; Dow Corning) containing 360 μg/mL 17β-estradiol dissolved in sesame oil. RIA showed that this regimen produced physiological levels of serum E₂ that were similar in Blank-SAP– and NK₃-SAP–treated animals (11).

Temperature Recordings.

T_CORE and gross motor activity were measured via telemetry using the implanted DSI transmitter. A previous study showed that activity detected by telemetry is comparable to the activity detected from infrared motion sensors (53). Individual cages were placed on an RPC-1 Physiotel receiver that was connected by a Data Exchange Matrix (DSI) to a computer equipped with Dataquest A.R.T. software (DSI). T_SKIN was recorded with a SubCue Mini datalogger (SubCue Dataloggers) as previously described (28). The dataloggers were housed in a protective nylon casing taped to the lateral surface of the tail (4.0 cm from the base) under brief (<5 min) isoflurane anesthesia. The magnitude of E₂ effects on T_SKIN recorded by these probes is similar to that obtained by surgically implanted telemetry devises (28, 32, 34). T_AMBIENT was recorded with an IT-18 thermocouple (Physitemp) inserted into a QuadTemp datalogger (Madgetech). Temperature measurement devices were calibrated according to manufacturer specifications and validated against a National Institute of Standards and Technology certified TC4000 thermocouple datalogger (Madgetech).

Experiment 1: Effects of KNDy Neuron Ablation on Circadian Rhythms of T_CORE, T_SKIN, and Activity in OVX and OVX + E₂ Rats.

Twelve to 15 d after the initial surgery, T_CORE, T_SKIN, activity, and T_AMBIENT were recorded every 10 min for five consecutive 24-h light/dark cycles (Fig. 1). One day after E₂ capsule implantation, these recordings were repeated for another five 24-h cycles. For these 5-d recording sessions, the rats were housed in their home cages in a dedicated room in the animal facility that was relatively isolated from noise and had access allowed only to the investigators. Cage cleaning was conducted before and after the 5-d recordings. The rats were placed in transparent, plastic shoebox-style cages containing wood-shaving bedding and the cages were placed on individual telemetry receivers. Although individual cages were needed for telemetry, the rats maintained visual, auditory, and olfactory contact with other animals.

Experiment 2: Effects of KNDy Neuron Ablation on the E₂ Modulation of T_CORE, T_SKIN, and HLI in Rats Exposed to T_AMBIENT of 26 °C, 11 °C, or 33 °C.

To determine if KNDy neuron ablation altered the ability of the rats to defend T_CORE in response to environmental temperature challenges, rats were exposed to temperatures that were within the thermoneutral zone (26 °C), subneutral (11 °C), or supraneutral (33 °C) T_AMBIENT. The thermoneutral zone is defined as the range of T_AMBIENT in which thermoregulation is achieved only by sensible (dry) heat loss, without regulatory changes in metabolic heat production or evaporative cooling (54). Within the thermoneutral zone, T_CORE is regulated primarily by skin vasomotion. The selection of T_AMBIENT was based on our previous analysis of the thermoneutral zone in OVX and OVX + E₂ rats (36).

After the completion of the circadian rhythm recordings, animals were brought to the laboratory for three consecutive mornings and exposed to one of three T_AMBIENT in an environmental chamber. The experiments were conducted in the morning to avoid the confounding influence of E₂ positive feedback. Animals were habituated to the experimental procedure on three occasions within the first 14 d after surgery. The environmental chamber (Forma model 3940; Thermo Scientific) was equilibrated to the required temperature with humidity set at 50%. Each rat was placed in a 6 inch × 6 inch × 4 inch plastic grid cage, which allowed free movement and ad libitum access to food and water as described previously (36). The cages were placed on telemetry receivers in the environmental chamber and T_CORE, T_SKIN, and T_AMBIENT were recorded every 10 min for 3 h. This procedure was repeated after the OVX rats were implanted with E₂ and the second set of circadian rhythm recordings were obtained (Fig. 1).

For sacrifice, animals were injected with a lethal dose of sodium pentobarbital (100 mg/kg i.p.) and perfused through the ascending aorta with 200 mL of 0.1 M phosphate buffered heparinized saline followed by 400 mL of 4% (wt/vol) paraformaldehyde in 0.1 M PBS, pH 7.4. Brains were extracted and immunohistochemical methods were used to characterize the effects of NK₃-SAP injections. Three NK₃-SAP rats with preservation of KNDy neurons were considered to be “missed” injections and excluded from further analysis. As reported previously, the rats injected with NK₃-SAP in the arcuate nucleus exhibited near-complete degeneration of KNDy neurons with variable, incomplete loss of lateral hypothalamic NK₃R neurons adjacent to the injection site (11). Selectivity was demonstrated by intact Nissl architecture and preservation of proopiomelanocortin, neuropeptide Y, and GnRH immunoreactive elements in the arcuate nucleus and adjacent median eminence (11).

Data Analysis.

A temperature probe on the tail skin surface will reflect not only active changes in vasomotor tone, but also passive changes in T_AMBIENT and T_CORE. To provide a more accurate assessment of tail skin vasomotion, we calculated the heat loss index [HLI = (T_SKIN − T_AMBIENT)/(T_CORE − T_AMBIENT)], which removes the passive influences of ambient and core temperatures and is strongly correlated with blood flow (26). The theoretical range of the HLI is between 0 (corresponding to maximal skin vasoconstriction) and 1 (maximal skin vasodilatation).

Circadian rhythms were evaluated over the 5-d period using circadian physiology software (55). Averages of T_SKIN, T_CORE, HLI, and activity were calculated for each animal during light (inactive) or dark (active) phases. Six-hour time blocks were used from the middle of each phase (1000 h to 1600 h and 2200 h to 0400 h for light and dark, respectively) to avoid the lights on/off transition. Because E₂ effects on T_SKIN during the dark phase are not significant until 3 d after capsule implantation (28), days 3–5 of the circadian recordings were used for statistical analysis. Group averages of T_SKIN, T_CORE, activity (counts/6 h), and HLI were generated from the means of individual rats. Group means were compared using a two-way ANOVA with Tukey’s post hoc analysis (α = 0.05).

For the temperature challenges, the mean T_CORE and T_SKIN for each animal was calculated from the second and third hour of recording in the environmental chamber. The averages from each animal were used to calculate group averages. Data were analyzed using a two-way ANOVA and Tukey’s post hoc tests (α = 0.05). Normothermia (37.2 ± 0.3 °C) was considered to be the average T_CORE (±1 SD) during the light phase of all rats exposed to a T_AMBIENT of 26 °C, which is within the thermoneutral zone of both OVX and OVX + E₂ rats (36).