Isoflurane disrupts excitatory neurotransmitter dynamics via inhibition of mitochondrial complex I

Abstract

Background: The mechanisms of action of volatile anaesthetics are unclear. Volatile anaesthetics selectively inhibit complex I in the mitochondrial respiratory chain. Mice in which the mitochondrial complex I subunit NDUFS4 is knocked out [Ndufs4(KO)] either globally or in glutamatergic neurons are hypersensitive to volatile anaesthetics. The volatile anaesthetic isoflurane selectively decreases the frequency of spontaneous excitatory events in hippocampal slices from Ndufs4(KO) mice.

Methods: Complex I inhibition by isoflurane was assessed with a Clark electrode. Synaptic function was measured by stimulating Schaffer collateral fibres and recording field potentials in the hippocampus CA1 region.

Results: Isoflurane specifically inhibits complex I dependent respiration at lower concentrations in mitochondria from Ndufs4(KO) than from wild-type mice. In hippocampal slices, after high frequency stimulation to increase energetic demand, short-term synaptic potentiation is less in KO compared with wild-type mice. After high frequency stimulation, both Ndufs4(KO) and wild-type hippocampal slices exhibit striking synaptic depression in isoflurane at twice the 50% effective concentrations (EC₅₀). The pattern of synaptic depression by isoflurane indicates a failure in synaptic vesicle recycling. Application of a selective A₁ adenosine receptor antagonist partially eliminates isoflurane-induced short-term depression in both wild-type and Ndufs4(KO) slices, implicating an additional mitochondria-dependent effect on exocytosis. When mitochondria are the sole energy source, isoflurane completely eliminates synaptic output in both mutant and wild-type mice at twice the (EC₅₀) for anaesthesia.

Conclusions: Volatile anaesthetics directly inhibit mitochondrial complex I as a primary target, limiting synaptic ATP production, and excitatory vesicle endocytosis and exocytosis.

Keywords: adenosine; anaesthesia; hypersensitivity.

Fig. 1

Isoflurane reversibly inhibits complex I dependent oxidative phosphorylation. ADP-stimulated respiration (state 3) of mitochondria from whole mouse brain was measured with a Clark electrode. (a) Complex I dependent state 3 respiration powered by pyruvate and malate was inhibited by isoflurane in a dose dependent fashion. The knockout (KO) was hypersensitive to isoflurane compared with wild-type with estimated IC₅₀ values of 0.16 mM for the KO and 0.31 mM for control [* denotes significant difference (P=0.0013) between the genotypes at 0.34 mM isoflurane]. The absolute initial rates for KO and wild-type were significantly different [KO: 0.9(0.3), control: 1.7(0.6) nmol_O2 s⁻¹ mg_protein⁻¹, P=0.035]. Inhibition was fully reversible after 10 min of removal from isoflurane. (b) Complex II dependent state 3 respiration was minimally affected for either genotype. In wild-type mitochondria inhibition caused by the highest isoflurane concentration, although small, was not reversible. (c) Diagram of the electron transport chain to illustrate electron flow (arrows) and site of isoflurane inhibition. Data presented as mean (standard deviation), n>4.

Fig. 2

Ndufs4(KO) inhibits short-term potentiation (STP) after high frequency stimulation and leads to short-term depression in the presence of isoflurane. (a) The depiction of the experimental flow. Axons from CA3 are stimulated and postsynaptic potentials in CA1 are recorded. (b) Representative traces recorded during baseline, 2 min after high-frequency stimulation (HFS) and 20 min after HFS in wild-type and Ndufs4(KO) hippocampal slices in the absence of isoflurane. (c) In the absence of isoflurane, compared with wild-type, there is less STP after HFS in Ndufs4(KO) slices [control: n=10 slices, Ndufs4(KO): n=9 slices]. (d) There are small, but statistically significant, differences in field excitatory postsynaptic potentials (fEPSPs) during HFS trains in the absence of isoflurane. We interpret those differences as not biologically substantial. (e) Representative traces recorded during baseline, 2 min after HFS and 20 min after HFS in wild-type and Ndufs4(KO) slices in 0.25 mM isoflurane. (f) Ndufs4(KO) slices pre-exposed to 0.25 mM isoflurane show pronounced short-term depression after HFS [control: n=8 slices, Ndufs4(KO): n=9 slices]. (G) There are no differences between genotypes in fEPSPs during the first HFS train. fEPSPs in Ndufs4(KO) slices do not recover after the first and second HFS train as evidenced by the smaller than baseline initial fEPSPs of the second and third trains. Here and in subsequent figures, the black downward arrow represents the first HFS train, and error bars represent standard error of the mean, unless stated differently. Red dots show statistically significant differences between wild-type and Ndufs4(KO) slices.

Fig. 3

Control and Ndufs4(KO) slices exhibit similar high-frequency stimulation (HFS)-induced short-term depression when exposed to behaviourally equipotent isoflurane concentrations. (a) Short-term and long-term changes after HFS are similar in wild-type and Ndufs4(KO) slices in presence of ∼1.4 EC₅₀ isoflurane for each genotype. (b) No pronounced differences in field excitatory postsynaptic potentials (fEPSPs) during HFS trains between wild-type and Ndufs4(KO) slices at 1.4 EC₅₀ isoflurane [wild-type: n=6 slices, Ndufs4(KO): n=5 slices]. (c) Short-term changes are similar in wild-type and Ndufs4(KO) slices in the presence of 2 EC₅₀ isoflurane for each genotype, however long-term fEPSP responses are diminished in the wild-type slices [wild-type: n=7 slices, Ndufs4(KO): n=9 slices]. (D) Ndufs4(KO) slices show diminished facilitation, but similar rates of depression in the three trains of HFS. Red dots show statistically significant differences between wild-type and Ndufs4(KO) slices.

Fig. 4

Paired-pulse ratios, first field excitatory postsynaptic potential (fEPSP) and second fEPSP plots of wild-type and Ndufs4(KO) synapses after high frequency stimulation. (a) In the absence of isoflurane, control and Ndufs4(KO) paired-pulse ratios decrease from corresponding baseline values between 2 and 3 min after high-frequency stimulation [HFS; wild-type: n=7 slices, Ndufs4(KO): n=7 slices]. (b) In the presence of isoflurane, wild-type paired-pulse ratios decrease between 2 and 28 min after HFS, while the KO paired-pulse ratios do not change from baseline values between 2 and 60 min after HFS [wild-type: n=7 slices, Ndufs4(KO): n=6 slices]. Red dots show statistically significant changes from baseline values.

Fig. 5

Role of A₁ adenosine receptor signalling in short-term depression after high frequency stimulation. (a) In the absence of isoflurane, DPCPX increases short-term potential after high-frequency stimulation (HFS) in the Ndufs4(KO) without a pronounced effect on wild-type [DPCPX wild-type: n=5 slices, DPCPX Ndufs4(KO): n=5 slices]. (b) While statistically significant changes were detected (red dots and crosses), DPCPX exposure did not substantially affect field excitatory postsynaptic potentials (fEPSP) responses during the three trains of HFS in either the wild-type or Ndufs4(KO). (c) DPCPX partially alleviated HFS-induced short-term depression in the Ndufs4(KO) in 0.25 mM isoflurane, but did not affect wild-type [DPCPX wild-type: n=5 slices, DPCPX Ndufs4(KO): n=5 slices]. (d) In Ndufs4(KO) slices, DPCPX decreased responses in the first and second train of HFS and eliminated depression of the first fEPSP of the third train of HFS. (e) DPCPX lessened HFS-induced short-term depression in wild-type slices in 0.74 mM isoflurane (DPCPX wild-type: n=7 slices, wild-type without DPCPX: n=7 slices). (f) No profound changes were observed in fEPSP responses during HFS trains between DPCPX-treated and DPCPX-untreated wild-type slices in the presence of 0.74 mM isoflurane. Red dots represent statistically significant differences between wild-type vehicle and wild-type DPCPX slices. Red crosses represent statistically significant differences between Ndufs4(KO) vehicle and Ndufs4(KO) DPCPX slices. Error bars in the anaesthetic behaviour experiments represent standard deviations.

Fig. 6

Ndufs4(KO) synaptic function is completely depressed by isoflurane in conditions favouring oxidative phosphorylation. Replacement of glucose with pyruvate (blue bar) leads to similar synaptic depression in Ndufs4(KO) and wild-type synapses. Addition of 0.25 mM isoflurane (green bar) leads to complete loss of synaptic function in Ndufs4(KO) synapses (closed circles), while causing ∼30% depression in wild-type synapses (open circles). Addition of 0.74 mM isoflurane (green bar) with pyruvate leads to complete loss of synaptic function in wild-type slices (open triangles). Prolonged incubation of slices in pyruvate solution without isoflurane did not lead to additional depression of field excitatory postsynaptic potentials (fEPSPs) in Ndufs4(KO) synapses (closed triangles). Changing solution to glucose solution with isoflurane (purple bar) led to partial recovery of fEPSPs.

Fig. 7

Proposed model for volatile anaesthetic mechanism of action at excitatory synapses. (a) Volatile anaesthetics (VAs) directly inhibit complex I function to reduce ATP in the presynapse which, in turn, inhibits synaptic vesicle endocytosis. A decrease in the ATP/ADP ratio leads to an increase in adenosine, which, through A₁ adenosine receptors, inhibits Ca²⁺ influx. The decreased Ca²⁺ influx, in turn, inhibits synaptic vesicle exocytosis. (b) The left lower panels [wild-type and Ndufs4(KO)] represent the mechanisms underlying the findings depicted immediately above the panels. In the presence of 0.25 mM isoflurane, mitochondrial function is inhibited in both genotypes, more so in the KO. This inhibition leads to more adenosine release (green dots) in the KO. Before high-frequency stimulation (HFS), control synapses balance ATP synthesis with ATP use, while Ndufs4(KO) synapses show larger isoflurane-induced depression because of increased adenosine signalling. The right lower panels represent the mechanisms underlying the findings depicted immediately above those panels. After HFS, ATP concentrations decrease (red dots) more so in the KO than wild-type synapses. This leads to decreased synaptic vesicle endocytosis in the KO and increased release of adenosine. Immediately after HFS in Ndufs4(KO), ATP synthesis fails to keep up with demand and the readily releasable pool of synaptic vesicles is depleted because of impaired endocytosis. This is depicted by black arrows in extracellular space and by less release of glutamate. The small black arrows in the intracellular spaces depict the recycling of synaptic vesicles into the readily releasable pools. At 0.75 mM isoflurane, synaptic dynamics would be the same in the wild-type as in the KO at 0.25 mM isoflurane. CI, complex I; iso, isoflurane.