A novel anesthesia regime enables neurofunctional studies and imaging genetics across mouse strains

Abstract

Functional magnetic resonance imaging (fMRI) has revolutionized neuroscience by opening a unique window that allows neurocircuitry function and pathological alterations to be probed non-invasively across brain disorders. Here we report a novel sustainable anesthesia procedure for small animal neuroimaging that overcomes shortcomings of anesthetics commonly used in rodent fMRI. The significantly improved preservation of cerebrovascular dynamics enhances sensitivity to neural activity changes for which it serves as a proxy in fMRI readouts. Excellent cross-species/strain applicability provides coherence among preclinical findings and is expected to improve translation to clinical fMRI investigations. The novel anesthesia procedure based on the GABAergic anesthetic etomidate was extensively validated in fMRI studies conducted in a range of genetically engineered rodent models of autism and strains commonly used for transgenic manipulations. Etomidate proved effective, yielded long-term stable physiology with basal cerebral blood flow of ~0.5 ml/g/min and full recovery. Cerebrovascular responsiveness of up to 180% was maintained as demonstrated with perfusion- and BOLD-based fMRI upon hypercapnic, pharmacological and sensory stimulation. Hence, etomidate lends itself as an anesthetic-of-choice for translational neuroimaging studies across rodent models of brain disorders.

Figure 1. Comparison of cerebral perfusion at rest and upon injection of acetazolamide in spontaneously breathing SD rats, C57BL/6J, BTBR T+tf/J and CD1 mice under different anesthetic conditions.

Left panels: mean cerebral perfusion at rest under (a) isoflurane, (b) medetomidine or (c) etomidate anesthesia. Numbers in the bars reflect sample sizes, error bars indicate standard deviations. Right panels: time course of perfusion response in whole brain and the PFC before and after injection of acetazolamide as a proxy for cerebrovascular reserve capacity. Responses are presented as percentage change relative to baseline (black dashed line at 100%) to allow comparison of different regions of interest. The left border of the gray-shaded area designates the time-point of acetazolamide application. Acetazolamide (30 mg/kg) was injected intravenously in isoflurane- and medetomidine-anesthetized animals and intraperitoneally under etomidate, thus explaining the slightly delayed response onset for the latter. Data are shown as sample means, and error bars represent 95% confidence intervals to visualize significant deviations from baseline. Sample sizes are provided below each graph. Color-coded magnetic resonance images represent the percentage change of perfusion (according to the color bar along the y axis) as an average over three consecutive time-points for a representative coronal plane (rats: +1.00 mm; mice: +1.34 mm relative to bregma). PFC, prefrontal cortex; SD, Sprague Dawley.

Figure 2. Comparison of basal blood perfusion in the brain and PFC of six mouse models of autism spectrum disorder under etomidate anesthesia.

Perfusion in the (a) brain and (b) PFC was assessed in spontaneously breathing Tsc2^+/−, Cntnap2^−/−, Shank3^−/−, NL3^R451C KI, BTBR T+tf/J, VPA-exposed mice, and corresponding wild-type controls. Sample sizes are provided in each bar. Error bars indicate standard deviations to visualize the spread of data for individuals. All the models were on a C57BL/6J background, except VPA (CD1) and Tsc2^+/− (B6.129S4 mixed background). Wild-type littermates were used as controls, except for BTBR T+tf/J mice where age-matched C57BL/6J mice served as controls. KI, knock-in; PFC, prefrontal cortex; VPA, valproic acid; WT, wild-type.

Figure 3. Pharmacological fMRI upon acute challenge with the antipsychotic olanzapine in C57BL/6J mice under different anesthetics.

(a) Mouse brain atlas superimposed on T₂-weighted anatomical images with outlined regions of interest and indicated distance to bregma. (b–d) phMRI activation maps showing differences in normalized perfusion between olanzapine- (3 mg/kg) and vehicle-treated C57BL/6J mice anesthetized with either (b) isoflurane, (c) medetomidine or (d) etomidate. Differences are given in percent, according to the color bar. Sample sizes for vehicle- and olanzapine-treated animals, respectively, were n = 8 and 9 under isoflurane, n = 8 and 8 under medetomidine, and n = 6 and 8 under etomidate. Cg, cingulate cortex; CPu, striatum; Thal, thalamus.

Figure 4. BOLD fMRI of neural activation elicited by unilateral electrical hindpaw stimulation in etomidate-anesthetized C57BL/6J mice.

(a) Mouse brain atlas superimposed on T₂-weighted anatomical images with distances to bregma indicated. Regions of interest are outlined in the hemisphere contralateral to the stimulated hindpaw. (b) Statistical parametric maps of two brain sections, located at −0.34 mm and −1.22 mm relative to bregma, showing neural activation pattern upon electrical stimulation (2 mA, 5 Hz). The strongest stimulus-induced activation was observed both in ipsi- and contralateral hemisphere S1HL and S2 cortices, and Thal. The color bar represents F-values (F = 13 corresponds to p < 0.05, family-wise error rate (FWER) corrected with minimum cluster size ≥20 voxels). (c) Time course of BOLD signal change (∆BOLD; in percent of baseline values) in contralateral S1HL during the four stimulation periods of 20 s duration each (gray bars). (d,e) ∆BOLD (shown for the first stimulus period only) in contralateral (d) S2 and (e) Thal illustrating region-specific activation strength. Data are depicted as mean ± standard deviation for n = 5 animals. BOLD, blood oxygenation level dependent; S1HL, primary somatosensory hindlimb cortex; S2, secondary somatosensory cortex; Thal, thalamus.