A divergent astrocytic response to stress alters activity patterns via distinct mechanisms in male and female mice

Abstract

The lateral hypothalamus is a brain region that regulates activity levels, circadian, and motivated behaviour. While disruption of these behaviours forms a hallmark of stress-related neuropsychiatric disorders, the underlying cellular mechanisms of how stress affects this brain region remain poorly understood. Here, we report that the effects of stress on behavioural activity levels correlate with spontaneous firing of orexin neurons, inducing hyperactivity in males and hypoactivity in female mice. These neuronal changes are accompanied by astrocyte remodelling, with causal manipulations identifying lateral hypothalamic astrocytes as key regulators of neuronal firing and physical activity patterns. In the context of stress, sex-specific changes in orexin neuron firing were driven by distinct astrocytic mechanisms with elevated purinergic signaling in male mice and reduced extracellular L-lactate in female mice. Finally, we show that genetic deletion of glucocorticoid receptors in lateral hypothalamic astrocytes restores key aspects of astrocyte morphology, rescues the effects of stress on orexin neuron firing, and recovers activity levels in both males and females. Overall, these data causally implicate astrocytes in the regulation of orexin neuron firing, behavioural activity patterns, and reveal that astrocytes are primary drivers of stress-induced behavioural change.

Fig. 1. ELS induces elevations in blood corticosterone and differentially alters diurnal activity rhythms in male and female mice.

A Experimental timeline with ELS occurring between P10-P17. B ELS paradigm; maternal separation 4 h/day and 70% reduced bedding in home cage and separation cages. C Serum corticosterone measurements significantly elevated at ZT2 (two-way ANOVA; F(3,24) = 0.0020. Tukey’s multiple comparisons; Naïve ZT2 = 48.04 ng/ml, ±4.76 (N = 11 mice), ELS ZT2 = 139.51 ng/ml, ±22.5 (N = 12 mice), p = 0.0011. Naïve ZT14 = 71.21 ng/ml, ±6.36 (N = 10 mice) vs ELS ZT14 = 93.75 ng/ml, ±16.44 (N = 6 mice), p = 0.876). D Adrenal weight from Naïve and ELS mice (unpaired t test; Naïve = 5.91 mg, ±0.287 (N = 19 mice), ELS = 6.95 mg, ±0.349 (N = 31 mice), p = 0.0427). E Running wheel protocol; 2-week single housing with ad libitum access to horizontal running wheels. F Distance ran per day of Naïve (N = 13) and ELS (N = 12) mice (two-way ANOVA; F(1,23) = 8.591, p = 0.0075). G Representative 24-h trace of running wheel activity from Naïve and ELS mice. H Analyses of 24-hour running wheel activity curves from Naïve and ELS mice. I FWHM (Mann–Whitney test: Naïve = 32.44, ±5.78 (N = 14), ELS = 30.41, ±6.45 (N = 12), p = 0.849). J AUC (Mann–Whitney test: naïve = 31987, ±3950 (N = 14), ELS = 19649, ±2635 (N = 12), p = 0.031). K Peak wheel activity (unpaired t test: Naïve = 595.8 revolutions, ±34.6 (N = 14), ELS revolutions = 580.6, ±43.9 (N = 12), p = 0.773). L 24-h traces from Naïve and ELS male mice. M 24-h traces from Naïve and ELS female mice. N Distance ran per day (two-way ANOVA; F(1,9) = 0.046, p = 0.834) Naïve N = 5, ELS N = 6. O Total distance ran (Mann–Whitney test: Naïve = 82.40, ±13.3 (N = 5), ELS = 78.42, ±12.7 (N = 6), p = 0.9307), and (P) light-period activity counts (Mann–Whitney test: Naïve = 23.83, ±82.1 (N = 5), ELS = 898.7, ±360 (N = 6), p = 0.0087) for Naïve and ELS male mice. Q Distance ran per day (two-way ANOVA; F(1,13) = 15.63, p = 0.0016), Naïve N = 9, ELS N = 6. R Total distance ran (Unpaired t-test: Naïve = 123.3, ±8.48 (N = 9), ELS = 70.64, ±10.2 (N = 6), p = 0.0016), and (S) light-period activity counts (Mann–Whitney test: naïve = 50.17, ±59.7 (N = 9), ELS = 108.8, ±38.5 (N = 6), p = 0.282) for Naïve and ELS female mice. N = mice. Bar charts = mean ± S.E.M. Panels P and S = Median and I.Q.R = not significant, *P < 0.05, **P < 0.01.

Fig. 2. ELS modifies intrinsic firing rates of orexin neurons in a sex specific manner.

A Representative viral targeting of orexin neurons in the LH, scale bar = 1 mm & 25 µm. B Validation of AAV-2/8-miniHCRT-tdTomato construct, scale bar = 50 µm. C Representative neurons from acute slice preparation under brightfield and Cy3 filter cube fluorescence with representative spontaneous firing rates of 3-4 Hz. D, E Representative traces of spontaneous excitatory post-synaptic currents (sEPSC) and spontaneous inhibitory post-synaptic currents (IPSCs) from orexin neurons in voltage-clamp (−70mV) configuration from Naïve and ELS mice. Male mice orexin neuron. F sEPSC peak amplitude (unpaired t test: Naïve = −38.14pA, ±3.81 (N = 3, n = 8 cells), ELS = −36.35pA, ±1.39 (N = 2, n = 5 cells), p = 0.728). G Event frequency (unpaired t test: Naïve =1.022 Hz, ±1.44 (N = 3, n = 10 cells), ELS = 1.215 Hz, ±2.30 (N = 2, n = 5 cells), p = 0.616). H Inter-event interval (unpaired t test: Naïve=1665 ms, ±340 (N = 3, n = 10 cells), ELS = 1478 ms, ±537 (N = 2, n = 5 cells), p = 0.765). I Female mice orexin neuron sEPSC peak amplitude (unpaired t test: Naïve = −37.57pA, ±3.32 (N = 3, n = 7 cells), ELS = −36.31 pA, ±1.50 (N = 3, n = 10 cells), p = 0.705). J Event frequency (Mann–Whitney test: Naïve = 0.836 Hz, ±0.183 (N = 3, n = 7 cells), ELS = 0.853 Hz, ±0.485 (N = 3, n = 10 cells), p = 0.720). K Inter-event interval (unpaired t test: Naïve = 1306 ms, ±257 (N = 3, n = 7 cells), ELS = 2049 ms, ±550 (N = 3, n = 10 cells), p = 0.303). Male mice orexin neuron. L sIPSC peak amplitude (unpaired t test: Naïve =−106.8 pA, ±15.1 (N = 2, n = 7 cells), ELS = −103.8 pA, ±10.1 (N = 2, n = 7 cells), p = 0.871). M Event frequency (unpaired t test: Naïve =1.022 Hz, ±0.135 (N = 2, n = 7 cells), ELS = 1.215 Hz, ±0.246 (N = 2, n = 8 cells), p = 0.230). N Inter-event interval (unpaired t test: Naïve=980.1 ms, ±169 (N = 2, n = 7 cells), ELS = 1489 ms, ±555 (N = 2, n = 7 cells), p = 0.398). Female mice orexin neuron. O sIPSC peak amplitude (unpaired t test: Naïve = − 116.5 pA, ±8.77 (N = 2, n = 9 cells), ELS = −76.9 pA, ±8.34 (N = 2, n = 9 cells), p = 0.0064). P Event frequency (unpaired t test: Naïve = 1.012 Hz, ±0.203 (N = 2, n = 7 cells), ELS = 1.230 Hz, ±0.203 (N = 2, n = 9 cells), p = 0.468). Q Inter-event interval (Mann–Whitney test: Naïve = 1431 ms, ±373 (N = 2, n = 7 cells), ELS = 824.5 ms, ±194 (N = 2, n = 9 cells), p = 0.252). R Representative traces of spontaneous firing rates of orexin neurons from Naïve and ELS male mice. S Representative traces of spontaneous firing rates of orexin neurons from Naïve and ELS female mice. T Resting membrane potentials (RMP) of orexin neurons from Naïve and ELS male mice (unpaired t-test: Naïve =−56.65 mV, ±1.04 (N = 3, n = 12 cells), ELS = −48.36 mV, ±1.78 (N = 4, n = 11 cells), p = 0.0005). U Spontaneous firing rates of orexin neurons from Naïve and ELS male mice (unpaired t test: Naïve = 3.85 Hz, ±0.459 (N = 4, n = 12 cells), ELS = 6.789 Hz, ±1.13 (N = 4, n = 12 cells), p = 0.025. V Instantaneous AP frequency of orexin neurons from Naïve (N = 5, n = 14) and ELS (N = 5, n = 11) male mice in response to 5 pA incremental current injections (two-way repeated measures ANOVA; F(1,23) = 40.60, p = < 0.001). Naive N = 4, n = 14 cells, ELS N = 4, n=11cells. W Resting membrane potentials (RMP) of orexin neurons from Naïve and ELS female mice (unpaired t test: Naïve =−54.19 mV, ±1.53 (N = 4, n = 16 cells), ELS = −49.66 mV, ±0.930 (N = 4, n = 16 cells), p = 0.0169). X Spontaneous firing rates of orexin neurons from Naïve and ELS female mice (unpaired t test: Naïve = 6.91 Hz, ±0.807 (N = 4, n = 15 cells), ELS = 3.648 Hz, ±0.618 (N = 4, n = 14 cells), p = 0.0037). Y Instantaneous AP frequency of orexin neurons from Naïve (N = 3, n = 9) and ELS (N = 3, n = 10) female mice in response to 5 pA incremental current injections (two-way repeated measures ANOVA; F(1,17) = 15.67, p = < 0.01). N = mice, n = cells. Bar charts = mean ± S.E.M. ns = not significant, *P < 0.05, **P < 0.01.

Fig. 3. ELS modifies astrocyte morphology in the LH.

A Representative IHC of GFAP, Orexin- A, and Cx43, In the LH of Naïve and ELS mice, scale bars = 25 µm. B, C Normalised fluorescence intensity (Arb.U.) measures of GFAP and Cx43 in the LH of Naïve and ELS male mice. B GFAP - unpaired t test: Naïve = 80.33Arb.U., ±8.64 (N = 4, n = 20 cells), ELS = 55.91Arb.U., ±3.01 (N = 4, n = 20 cells), p = 0.011. C Cx43 - unpaired t test: Naïve= 93.74Arb.U., ±6.79 (N = 4, n = 20 cells), ELS = 72.68Arb.U., ±4.02 (N = 4, n = 20 cells), p = 0.011). D, E Normalised fluorescence intensity (Arb.U.) measures of GFAP and Cx43 in the LH of Naïve and ELS female mice. D GFAP - unpaired t test: Naïve =119.2Arb.U., ±12.7 (N = 4, n = 19 cells), ELS = 58.87Arb.U., ±3.91 (N = 4, n = 19 cells), p < 0.0001. E Cx43 - unpaired t test: Naïve = 100.8Arb.U., ±6.02 (N = 4, n = 20 cells), ELS = 79.27Arb.U., ±4.98 (N = 4, n = 20 cells), p = 0.009). F Quantification of Orexin-A+ cells between Naïve and ELS male mice (unpaired t test: Naïve = 17.33 cells, ±1.09 (N = 6), ELS = 16.83 cells, ±0.872 (N = 6), p = 0.727) and (G) Naïve and ELS female mice (unpaired t test: Naïve =15.60 cells, ±1.66 (N = 5), ELS = 15.71 cells, ±0.747 (N = 7), p = 0.933). H Quantification of S100β + astrocytes in the lateral hypothalamus (unpaired t test: Naïve = 14.50 (N = 8), ELS = 12.00 (N = 8), p = 0.367) scale bars = 25 µm. I Experimental timeline and viral constructs for sparse labelling of astrocytes in the LH. Scale bars = 50μm. J Representative AAV-GfaABC1D-eGFP expressing astrocytes in the LH from Naïve and ELS mice, scale bars = 5 μm. Morphological reconstruction of LH astrocytes from Naïve and ELS male mice. K Volume (unpaired t test: Naïve = 5916 μm³, ±2.797 (N = 4, n = 18 cells), ELS = 4430 μm³, ±308 (N = 3, n = 13 cells), p = 0.0004). L Total process length (unpaired t test: Naïve =3365 μm ±293 (N = 3, n = 12 cells), ELS = 3168μm, ±471(N = 3, n = 10 cells), p = 0.717). M No. of branch points (Mann–Whitney test: Naïve = 1256 branch points ±264 (N = 3, n = 12 cells), ELS = 943 branch points, ±263 (N = 3, n = 10 cells) p = 0.314). N Sholl analysis (two-way ANOVA; F(1,25) = 0.09, p = 0.764). Morphological reconstruction of LH astrocytes from Naïve and ELS female mice. O Volume (unpaired t test: Naïve = 7740 μm³, ±1041 (N = 4, n = 16 cells) ELS = 3659μm³, ±257 (N = 4, n = 19 cells), p = 0.0006). P Total process length (unpaired t test: Naïve =4396 μm, ±627 (N = 3, n = 10 cells), ELS = 2701 μm, ±278 (N = 4, n = 17 cells), p = 0.009). Q No. of branch points (Mann–Whitney test: Naïve = 1209 branch points, ±211 (N = 3, n = 11 cells), ELS = 720.5 branch points, ±278 (N = 4, n = 16 cells) p = 0.488). R Sholl analysis (two-way ANOVA; F(1,23 = 4.31, p = 0.049). N = mice, n = cells. Bar charts = mean ± S.E.M. ns = not significant, *P < 0.05, **P < 0.01.

Fig. 4. Gq-coupled calcium fluctuations in LH astrocytes excites orexin neurons to perturb diurnal activity rhythms in naïve mice.

A Chemogenetic construct & timeline. B, C Expression in LH astrocytes and chronic delivery of clozapine-N-oxide (CNO) in drinking water (1.25 mg/kg), Scale bar = 100 μm. D Representative traces of running wheel activity from hM3Dq-injected mice (N = 7). E Distance ran per day, (N = 7). F Quantification of light period activity pre and during delivery of CNO (ratio paired t test: t6 = 3.34, p = 0.016, N = 7). G Quantification of dark period activity pre and during delivery of CNO (paired t test: t5 = 3.96, p = 0.0074, N = 7). H Experimental timeline and representative AAV-GfaABC₁D-eGFP expressing astrocytes in the LH from hM3Dq- or mCherry mice, scale bars = 20 µm. Morphological reconstruction of LH astrocytes from or mCherry and hM3Dq injected male mice: (I) Volume (unpaired t test: mCherry = 5389 μm³, ±708 (N = 3, n = 13 cells), hM3Dq = 5015 μm³, ±670 (N = 3, n = 15 cells), p = 0.707), (J) Total process length (unpaired t test: mCherry = 5813 μm, ±257 (N = 3, n = 13 cells), hM3Dq = 5593 μm, ±441 (N = 3, n = 17 cells), p = 0.793), (K) No. of branch points (Mann–Whitney test: mCherry = 1632 branch points, ±207 (N = 3, n = 13 cells), hM3Dq = 1445 branch points, ±137 (N = 3, n = 17 cells) p = 0.893), (L) Sholl analysis (two-way ANOVA; F(1,27) = 0.0233, p = 0.880). Morphological reconstruction of LH astrocytes from or mCherry and hM3Dq injected female mice: (M) Volume (unpaired t test: mCherry =6418 μm³,±527 (N = 3, n = 14 cells), hM3Dq = 5132 μm³, ±434 (N = 3, n = 16 cells), p = 0.0678), (N) Total process length (unpaired t test: mCherry = 5798 μm, ±534 (N = 3, n = 14 cells), hM3Dq = 4083 μm ±330 (N = 3, n = 16 cells), p = 0.009), (O) No. of branch points (unpaired t test: mCherry = 1754 branch points ±156 (N = 3, n = 13 cells), hM3Dq = 1139 branch points, ±109 (N = 3, n = 17 cells) p = 0.0027), (P) Sholl analysis (two-way ANOVA; F(1,28) = 8.76, p = 0.0062). Q Co-injection of hM3Dq and miniHCRT-tdTomato virus into the LH. R Representative trace of orexin neuron spontaneous AP firing in current clamp with 3 min bath application of CNO (10 µM. S Change orexin neuron spontaneous firing rate in male mice (N = 3, n = 6 cells) after bath application of CNO (paired t test: t5 = 2.94, p = 0.032). T Change orexin neuron spontaneous firing rate in female mice (N = 3, n = 8 cells) after bath application of CNO (paired t test: t7 = 4.94, p = 0.0017). U Representative firing rate of orexin neuron firing rates at baseline (1) and with CNO (2). N = mice, n = cells. Bar charts = mean ± S.E.M. ns = not significant, *P < 0.05, **P < 0.01.

Fig. 5. Elevated P2X receptor signalling drives increased firing rates in male ELS mice.

A Schematic of orexin neuron tripartite synapse and overview of mechanism of action of pharmacological agents. B Percentage change in firing after bath application of modulators of the purinergic system. two-way ANOVA; Uncorrected Fisher’s LSD: 8-cyclopentyltheophylline (CPT) (p = 0.0786, Naïve N = 2, n = 6 cells, ELS N = 2, n = 7 cells), ZM-241385 (p = 0.973, Naïve N = 2, n = 6 cells, ELS N = 3, n = 5 cells), PSB12379 (p < 0.001, Naïve N = 2, n = 6 cells, ELS N = 3, n = 6 cells), PPADS (p = 0.0102, Naïve N = 2, n = 6 cells, ELS N = 3, n = 7 cells), iso-PPADS (p < 0.001, Naïve N = 3, n = 5 cells, ELS N = 3, n = 7 cells). C Representative traces of orexin neurons from Naïve and ELS male mice before and after bath application of adenosine A1 receptor antagonist, CPT (200 nM). D Change in firing rate in Naïve mice with application of adenosine A1 receptor antagonist, CPT (Paired t test: baseline= 3.035, ±0.822, CPT (200 nM) = 1.56 ±0.564, (N = 2, n = 6 cells) p = 0.0178. E Change in firing rate in ELS mice with application of CPT (Paired t test: baseline = 5.60 Hz ±1.47, CPT = 4.44 Hz ±1.69 (N = 2, n = 7 cells) p = 0.068). F Representative traces of orexin neurons from Naïve and ELS male mice before and after bath application of adenosine A2 receptor antagonist, ZM241385 (50 nM). G Change in firing rate in Naïve mice with ZM241385 (Paired t test: baseline= 2.671 Hz ±0.461, ZM241385 = 2.422 Hz ±0.468 (N = 2, n = 6 cells) p = 0.448). H Change in firing rate in ELS mice with application of adenosine A2 receptor antagonist, ZM241385 (Paired t test: baseline = 5.756 Hz ±0.938, ZM241385 = 5.421 Hz ±1.15 (N = 3, n = 5 cells) p = 0.528). I Representative traces of orexin neurons from Naïve and ELS male mice before and after bath application of Ecto-5’-Nucleotidase (CD73) inhibitor, PSB12379 (10 μM). J Change in firing rate in Naïve mice with PSB12379 (Paired t test: baseline = 2.78 Hz ±0.826, PSB12379 = 5.29 Hz ±1.39 (N = 2, n = 6 cells) p = 0.0086). K Change in firing rate in ELS mice with PSB12379 (Paired t test: baseline = 5.23 Hz ±0.720, PSB12379 = 6.105 Hz ±6.11 (N = 3, n = 6 cells) p = 0.029). L Representative traces of orexin neurons from Naïve and ELS male mice before and after bath application of PPADS tetrasodium salt (50 μM). M Change in firing rate in Naïve mice with PPADS (Paired t test: baseline= 2.13 Hz ±0.440, PPADS = 1.88 Hz ±0.361 (N = 2, n = 6 cells) p = 0.111). N Change in firing rate in ELS mice with PPADS (Paired t test: baseline= 6.52 Hz ±1.34, PPADS = 3.32 Hz ±0.737 (N = 3, n = 7 cells) p = 0.025). O Representative trace of orexin neurons from ELS male mice before and after bath application of P2X purinoceptor antagonist, iso-PPADS tetrasodium salt (50 μM). P Change in firing rate in Naïve mice with iso-PPADS (Paired t test: baseline = 4.13 Hz ±0.368, iso-PPADS = 4.09 Hz ±0.377 (N = 3, n = 5 cells) p = 0.237). Q Change in firing rate in ELS mice with iso-PPADS (Paired t test: baseline= 4.96 Hz ±0.759, iso-PPADS = 1.77 Hz ±0.448 (N = 3, n = 7 cells) p = 0.006). R Zoom in representative trace after application of iso-PPADS. S Schematic highlighting shift in extracellular adenosine to ATP in ELS male mice. Extracellular ATP signalling through P2X receptors on the post-synaptic membrane to induce elevated spontaneous firing rates in orexin neurons. N = mice, n = cells. Bar charts = mean ± S.E.M. ns = not significant, *P < 0.05, **P < 0.01.

Fig. 6. ELS disrupts extracellular Lactate availability in female mice.

A–D Representative trace of orexin neuron firing from (A) Naïve female, (B) Naïve male, (C) ELS female, (D) ELS male before and during bath application of 5 mM L-Lactate. E Absence of change in firing rate in Naïve female mice with 5 mM L-Lactate (Paired t test: baseline= 7.36 Hz ±1.18, 5mM L-Lactate= 7.10 Hz ±1.27 (N = 3, n = 8 cells) p = 0.311). F Increase in firing rate in ELS female mice with 5 mM L-Lactate (Paired t test: baseline= 3.08 Hz ±0.898, 5 mM L-Lactate= 5.83 Hz ±0.911 (N = 3, n = 8 cells) p = 0.0014). G Absence of change in firing rate in Naïve male mice with 5 mM L-Lactate (Paired t test: baseline= 2.61 Hz ±0.583, 5 mM L-Lactate = 2.50 Hz ±0.517 (N = 3, n = 7 cells) p = 0.354). H Absence of change in firing rate in ELS male mice with 5 mM L-Lactate (Paired t test: baseline = 5.70 Hz ±1.00, 5 mM L-Lactate = 5.10 Hz ± 1.14 (N = 3, n = 8 cells) p = 0.113). I Injection of fluorescent intracellular L-Lactate sensor into neurons in the LH of Naïve and ELS female mice. J Representative images of iLACCO2.0 expressing neurons in acute slices of the LH, scale bars = 25 µm. K Change in normalised fluorescence intensity (df/f) of iLACCO2.0 before and during 5 min of bath application of 5 mM L-Lactate. L Change in df/f of in Naïve female neurons expressing iLACCO2.0 (paired t test: baseline = −0.0017df/f ±0.00191, 5 mM L-Lactate= 0.0075 df/f ±0.0331 (N = 3, n = 12 cells) p = 0.778). M Change in df/f of in ELS female neurons expressing iLACCO2.0 (paired t test: baseline = 0.0040df/f ±0.00157, 5 mM L-Lactate = 0.1495df/f ±0.0323 (N = 3, n = 10 cells) p = 0.0012). N = mice, n = cells. Bar charts = mean ± S.E.M. ns = not significant, *P < 0.05, **P < 0.01.

Fig. 7. Astrocyte specific deletion of glucocorticoid receptors (GR) in the LH corrects cellular and behavioural impairments induced by ELS.

A Experimental timeline for combined ELS and genetic deletion of astrocyte GRs in the LH. Separate cohorts of mice subjected to behavioural protocols or IHC. B Viral construct used to achieve astrocyte specific deletion of GR receptors in Nr3c1fl/fl mouse line, scale bar = 25 µm. C–E Validation of AAV2/5-GfaABC₁D-Cre-eGFP expression in LH astrocytes. C quantification of orexin neurons (Mann–Whitney test: Naïve = 16 cells ±1.77, (N = 4), Cre-eGFP injected = 19 cells ±1.66 (N = 5), p = 0.381), (D) Quantification of S100β astrocytes (Mann–Whitney test Naïve = 17 cells ±1.20 (N = 6), Cre-eGFP injected = 18.40 cells ±0.759 (N = 5), p = 0.495), (E) Cellular identity of Cre-eGFP and expression specificity of 94.5% in S100β + astrocytes (N = 7). F Representative LH astrocytes expressing control virus (AAV2/5-GfaABC₁D-eGFP) or AAV2/5-GfaABC₁D-Cre-eGFP in 3D view and 2D slice view of branching structures (branch points & branching processes), 3D volume scalebars = 5 µm, 2D slice scale bar = 2 µm. G Volume of -eGFP control or cre- expressing LH astrocytes from ELS male mice (unpaired t test: eGFP = 4803 μm³ ±342 (N = 3, n = 19 cells), Cre = 4365 μm³, ±529 (N = 3, n = 19 cells), p = 0.491). H Total branch points/volume (unpaired t test: eGFP = 18.18 ±2.08 (N = 3, n = 19 cells), Cre = 31.91 ±4.30 (N = 3, n = 19 cells), p = 0.007). I Normalised Cx43 fluorescence intensity (a.u) in LH of -eGFP control or cre- astrocytes (unpaired t test: eGFP = 71.37Arb.U. ±8.56 (N = 3, n = 14 cells), Cre = 102.6Arb.U. ±11.1 (N = 3, n = 15 cells), p = 0.036), scale bars = 25 µm. J Volume of eGFP control or cre- expressing LH astrocytes from ELS female mice (Mann-–Whitney test: eGFP = 4622 μm³, ±518 (N = 3, n = 21 cells), Cre = 3545 μm³, ±484 (N = 3, n = 16 cells), p = 0.130). K Total branch points/volume (unpaired t test: eGFP = 15.15 ±0.994 (N = 3, n = 21 cells), Cre = 31.06 ±5.27 (N = 3, n = 16 cells), p = 0.0018). L Normalised Cx43 fluorescence intensity (a.u) in LH of -eGFP control or cre-astrocytes (unpaired t test: eGFP = 91.32Arb.U. ±7.46 (N = 3, n = 15 cells), Cre = 158.5Arb.U. ±13.1 (N = 3, n = 15 cells), p < 0.001), scale bars = 25 µm. M Experimental timeline for combined expression of AAV2/5-GfaABC₁D-Cre-eGFP and AAV2/8-miniHCRT-tdTomato in Nr3c1-ELS mice, scale bars = 25 µm. N Representative firing rates of orexin neurons in Naive eGFP (N = 3, n = 7), ELS eGFP (N = 3, n = 14) and ELS Cre (N = 3, n = 14) injected male mice. O Restoration of RMP (one-way ANOVA: F = 10.32, p < 0.001) and (P) firing rates (one-way ANOVA: F = 5.78, p = 0.0072) in Naive eGFP (N = 3, n = 7), ELS eGFP (N = 3, n = 14) and ELS Cre (N = 3, n = 17) injected male mice. Q Representative firing rates of orexin neurons in Naïve eGFP (N = 3, n = 12) ELS eGFP (N = 3, n = 11) and ELS Cre (N = 3, n = 18) injected female mice (one-way ANOVA: F = 7.90, p = 0.0014). R Restoration of RMP (one-way ANOVA: F = 7.90, p = 0.0014) and (S) firing rate (one-way ANOVA: F = 8.32, p = 0.001) in Naïve eGFP (N = 3, n = 11), ELS eGFP (N = 3, n = 12), and ELS Cre (N = 4, =16) injected female mice. T, V Running wheel behaviour of eGFP and Cre injected male mice: (T) Distance ran per day, (U) Total distance (one-way ANOVA: F = 1.96, p = 0.173), and (V)Llight cycle activity (one-way ANOVA: F = 3.23, p = 0.001). W–Y Running wheel behaviour of eGFP and Cre injected female mice: (W) Distance ran per day, (X) Total distance (one-way ANOVA: F = 6.02, p = 0.011), and (Y) Light cycle activity (one-way ANOVA: F = 1.24, p = 0.315) N = mice, n = cells. Bar charts = mean ± S.E.M. ns = not significant, *P < 0.05, **P < 0.01, ***P < 0.001.