Astrocyte-Specific Knockout of a Mitochondrial Protein in Mice Increases Neural Inertia Renjini Ramadasan-Nair PhD, Jessica Hui, Philip G Morgan MD, Margaret M Sedensky MD Center for Integrative Brain Research, Seattle Children’s Research Institute, Seattle, Washington, United States Anesthesiology & Pain Medicine, University of Washington, Seattle, Washington, United States Background Characterization of astrocytic Cre expression Astrocytic Ndufs4(KO) Viral injections to the Locus coeruleus B WT - Cre A A B B The cell types in the CNS which contribute to the anesthetic response are unknown. Animals with defects in mitochondrial complex I have been shown to be hypersensitive to volatile anesthetics (VAs)1, 2. Restriction of the knock out of the mitochondrial complex I protein, NDUFS4, to glutamatergic neurons confers the VA hypersensitivity3. While the importance of neuronal circuitry in mediating the anesthetic state is undisputed, recent research shows that astrocytes may also play a role by altering Ca2+ transients independent of neuronal activity4. In addition, norepinephrine inputs from the locus coeruleus (LC) have been implicated in astrocyte-mediated cortical state switching5. We analyzed the effects of knocking out Ndufs4 in astrocytes and in the LC on anesthetic sensitivity. EC50s for each VA defined by the first loss of tail flick or righting reflex were not different between the KO and control animals. The dose-response curve for ISO and HAL for the astrocytic KO (gfap-KO) mice indicated hysteresis between induction and emergence at 1 and 2 months post tamoxifen injection (Table 1, Figure 3). The values of hysteresis (DAD) were not altered by prolonged times (30 minutes) at each concentration of anesthetic to eliminate a pharmacokinetic effect. a-NDUFS4 DAPI C C D - Cre a-GFAP a-GFAP/GFP Ai6/+ Pgfap-CreERT2/+; Ai6/+ a-NDUFS4 DAPI Figure 1: Pgfap driven Cre expression in Pgfap-CreERT2/+; Ai6/+ reporter mice result in co-expression of functional GFP in the cells. Immunohistochemistry using anti-GFAP antibody reveals the co-localization in astrocytes (B). Nuclei were stained with DAPI. Sibling Ai6/+ mice were used as controls (A). Tail Clamp-ISO Emerge/Induce Tail Clamp-HAL LORR - ISO LORR – HAL Control Gfap-KO 1 month 0.93 1.18/1.27 0.56* 0.69/1.22 0.84 1.08/1.28 0.49* 0.59/1.20 0.98 0.98/1.01 0.69* 0.68/0.99 0.97 1.0/1.03 0.66* 0.66/1.0 2 months 1.20/1.24 0.57* 0.67/1.18 0.96 1.15/1.19 0.54* 0.66/1.22 0.89/0.92 0.68* 0.67/0.98 0.94/0.98 0.59* 0.55/0.93 D E *** *** ** Loss of NDUFS4 in astrocytes * p<0.01 ** p<0.01, *** p<0.001 A B Table 1: EC50s in ISO and HALO for TC and LORR, expressed as ratios between emergence and induction doses (* p-value < 0.005). N=6 for controls, N≥6 for knock-outs. Figure 4: A, Viral infected cells in the bilateral LC which co-express GFP. WT-Cre virus caused loss of red NDUFS4 immunofluorescence (B), while D-Cre virus infected cells retained it (C). DAD, the change in the anesthetic dose between induction and emergence required to demonstrate righting reflex (D) or response to tail clamp (E) upon exposure to ISO or HAL (p-values: ** < 0.01; *** < 0.001). Methods *** *** *** *** a-NDUFS4 a-GFAP DAPI a-NDUFS4 a-GFAP DAPI DISCUSSION Control Gfap-KO The SCRI IACUC approved all studies. We constructed a transgenic mouse line Pgfap-CreERT2;Ndufs4D/lox, which conditionally expresses Cre-recombinase in GFAP-expressing cells (astrocytes). These mice were injected with 4-hydroxytamoxifen (50 mg/g, once a day, PND33-40) to generate animals lacking NDUFS4 in astrocytes (GFAP-KO) and were subjected to behavioral tests three weeks post injections. To assess the Cre expression, the Pgfap-CreERT2 line was crossed with the Ai6 Cre reporter line and characterized. Successful astrocytic KO of Ndufs4 was confirmed by immunohistochemistry. Cre-positive sibling mice (Ndufs4 /+ or Ndufs4 lox/+) were used as controls. For the LC-specific knockout (KO), we injected adeno-associated viruses (109 pfu) encoding Cre-recombinase (WT-Cre) or inactive virus (-Cre) bilaterally into the LC of Ndufs4 floxed mice, and allowed them to recover for 1-2 months. The injected mice were exposed to isoflurane or halothane and tested for the loss of righting reflex or response to a tail clamp both during induction of and emergence from anesthesia. C Our conditional GFAP-KO uncovered a surprising role of astrocytes in response to VAs: recovery from the anesthetized state. Although the astrocytic KO animal has normal anesthetic sensitivity during induction, it emerges from general anesthesia at a much lower concentration than controls. The hysteresis between the induction of and emergence from anesthesia, termed neural inertia by Kelz6, is not explained by pharmacokinetics. Our study shows that astrocytes play a role in generating this hysteresis (glial inertia), likely by supporting metabolism in glutamatergic neurons. We hypothesize that energy deficits in astrocytes disrupt Ca2+ homeostasis, thereby delaying glutamate cycling leading to inability to regain synaptic function in the presence of low concentrations of VA. Preliminary results from KO of Ndufs4 in the LC lend support to the hypothesis that the LC mediates arousal from anesthesia through modulation of astrocytes. *** *** *** *** Figure 2: Characterization of NDUFS4 expression in Pgfap-CreERT2/+; Ndufs4D/lox mice. Immunohistochemistry using anti-NDUFS4 and anti-GFAP antibodies show the co-localization in control astrocytes (A) and loss of red NDUFS4 immunofluorescence in the Gfap-KO astrocytes (B) (quantitated in C). Nuclei were stained with DAPI. Sibling Pgfap-CreERT2/+; Ndufs4+/lox mice were used as controls. *** p<0.001 Figure 3: DAD, the change in the anesthetic dose between induction and emergence required to demonstrate righting reflex or response to tail clamp upon exposure to ISO or HAL (*** p-value < 0.001). References 1. Kayser, E.B. et al. Anesthesiology 90, 545-554 (1999). 2. Morgan, P.G. et al. Anesth 96, 1268-1270 (2002). 3. Quintana, A. et al. PLoS One 7, e42904 (2012). 4. Thrane, A.S. et al. Proc Natl Acad Sci U S A 109, 18974-18979 (2012). 5. Bekar, L.K.et al. Cereb Cortex 18, 2789-2795 (2008). 6. Friedman, E.B., et al. PLoS One 5, e11903 (2010). Inject with 4-hydroxy tamoxifen Cells with active Cre recombinase (Astrocytes) Cells without active Cre recombinase (Neurons) Pgfap-CreERT2; Ndufs4 D/lox (Gfap KO) Acknowledgements NDUFS4 NDUFS4 R01 GM105696 and the Northwest Mitochondrial Research Guild.