Long-Term GABAergic Modulation of Brain Tissue Oxygenation and Electrophysiological Activity Following Neonatal Anesthesia Exposure

Abstract

Millions of children are exposed to general anesthesia annually, raising concerns about its impact on neuronal development, subsequent behavioral impairments, and the conditions under which these functional deficiencies can be detected instrumentally. In this study, we investigated the delayed consequences of administering sevoflurane, a widely used anesthetic, to newborn rabbits to mimic repeated clinical anesthesia exposure. When these rabbits reached adolescence/adulthood, we implanted electrodes into their cerebral cortex to record electrophysiology and tissue oxygen. Tissue oxygen recording was chosen to show translational changes in hemodynamics, which can be detected by functional MRI. Despite observing no differences in baseline oxygen between the group exposed to neonatal anesthesia and the control group, subsequent administration of GABA antagonists in small concentrations near the oxygen electrodes revealed significant findings. While the local administration of the GABA antagonist picrotoxin resulted in a noticeable shift in oxygen oscillations to lower frequencies in both groups, the rabbits exposed to neonatal anesthesia also generated large and sustained response curves associated with epileptic activity. This study suggests the specific effects of neonatal anesthesia in inducing a state of susceptibility to epileptic activity, which can be triggered by specific and mild challenges (e.g., the application of picrotoxin in small concentrations). We generalize these findings to suggest that this mechanism can explain the inconsistency of clinical manifestations in populations exposed to neonatal anesthesia, as the anesthesia alone may not be sufficient for clinical symptoms to appear; an additional trigger or challenge is required.

23.1. Introduction

Concerns about the long-term behavioral effects of neonatal anesthesia are growing (Aksenov et al. 2023a; Borzage and Peterson 2024; Robinson et al. 2022). Research has shown that neonatal anesthesia can impact learning in later age, with effects varying depending on the cumulative dose of general anesthetics such as sevoflurane, isoflurane, and propofol. Human studies present a variable picture, with some showing behavioral deficiency and others not (Aksenov 2024). Moreover, it remains unclear how neuronal or neurovascular markers of these functional deficiencies can be detected instrumentally. Previous studies (Aksenov et al. 2020; Taghon et al. 2015; Neudecker et al. 2023) have shown that functional MRI (fMRI) has inconsistently detected these effects, and the reasons for this variability remain unclear.

We propose a plausible explanation for this inconsistency. Numerous studies have demonstrated that neonatal anesthesia induces apoptosis, which selectively affects certain neurons, particularly subpopulations of GABAergic interneurons (Gascoigne et al. 2021; Zhou et al. 2021). As the number of these interneurons decreases, those remaining attempt to compensate by forming additional synapses. This compensation is effective only up to a certain limit. Consequently, children exposed to neonatal anesthesia may appear functionally identical to non-exposed children under normal conditions. However, under stress, overstimulation, excessive workload, and other challenges, this compensation fails, and the differences between exposed and non-exposed children become apparent.

We hypothesize that if interneurons are affected by neonatal anesthesia, and if functional deficiency of interneurons is a cause of epileptic activity (Mueller-Buehl et al. 2023), then neonatal anesthesia may result later in a state of “hidden” (clinically and electrophysiologically silent) epilepsy. We term it “hidden” because it can only be revealed by specific challenges. To expose the functional deficiency of the inhibitory system, we employed local injections of the GABA-antagonist picrotoxin (PTX) at a concentration that does not induce epileptic activity in the control group (Aksenov et al. 2023b).

Since it is known that fMRI provides better volume resolution compared to EEG, and we have shown that brain tissue oxygen (PO₂) and fMRI follow each other (Aksenov et al. 2023c), we simultaneously recorded local field potentials as a marker of neuronal postsynaptic activity and PO₂ fluctuations as a marker of hemodynamics.

23.2. Methods

Animal preparation. Dutch-belted rabbits (N = 10) were used in accordance with the National Institutes of Health guidelines and NorthShore University HealthSystem Research Institute Institutional Animal Care and Use Committee-approved protocol. Infant rabbits underwent sevoflurane exposures (N = 5) on days 8, 11, and 14 for 2 h each following a previously described approach (Aksenov et al. 2020). At the age of 3 months, animals were implanted with restraining headbolts and a microdrive containing electrodes and an injection cannula (Aksenov et al. 2019). After 1 week of recovery from surgery, each subject was habituated for 3–5 days. The electrophysiological signals from the microwires were fed through a miniature preamplifier to a multi-channel differential amplifier system (Neuralynx, Inc., Bozeman, MT, USA). The signals were amplified, band-pass filtered (0.3–3 kHz for single units and 1–150 Hz for local field potentials (LFP)), and digitized (32 kHz/channel) using a Neuralynx data acquisition system. Single units were monitored to ensure that the volume effect was minimized (Aksenov et al. 2014).

For oxygen recording, microwires were polarized to −0.7 V with respect to a reference electrode (located between the skull and dura), and the current was measured with a Keithley model 614 electrometer (Keithley Instruments, Cleveland, OH, USA). The output voltage from the electrometer was low-pass filtered at 30 Hz, amplified, and digitized at 2 Hz. LFP were recorded simultaneously with PO₂ recordings. Epileptic events were recorded for rabbits and were determined from LFP measurements. Picrotoxin (50 μM) was dissolved in artificial cerebrospinal fluid (ACSF) for injection. The concentration of picrotoxin (PTX) was based on our previous studies which showed this concentration does not cause epileptic activity (Aksenov et al. 2023b).

23.3. Results

For rabbits exposed to anesthesia as neonates, a small concentration of picrotoxin (PTX) induces focal seizures, as highlighted in Fig. 23.1. Figure 23.1a and b shows the drastic change in the LFP trace of a neonatal anesthesia-exposed rabbit after the injection of PTX. Figure 23.1a displays a representative 2-min LFP trace before injection, and Fig. 23.1b displays a 5-min LFP trace after injection, containing two major epileptic events. These major epileptic events are characterized by highly frequent large deviations in LFP from the baseline, followed by a refractory period with minimal LFP fluctuations. The increased magnitude of LFP fluctuations is evident in the LFP spectrograms (Fig. 23.1c and d). Figure 23.1c shows the LFP spectrogram for a rabbit not exposed to neonatal anesthesia. After PTX injection at 15 min, the effect on the power of LFP frequencies is minimal. In contrast, Fig. 23.1d shows the LFP spectrogram for a rabbit exposed to neonatal anesthesia, revealing a substantial increase in the power of all frequencies below 40 Hz, consistent with an epileptic state.

Fig. 23.1.

Injections of picrotoxin induce epileptic activity in rabbits exposed to neonatal anesthesia. From a rabbit that was exposed to anesthesia as a neonate, (a) shows an illustrative 2-min time window of normalized LFP prior to the injection of PTX and (b) shows a 5-min window of normalized LFP after injection of PTX, which contains two epileptic events. LFP spectrograms for neonates (c) not exposed to neonatal anesthesia and (d) exposed to neonatal anesthesia suggest the substantial change in the power of the lower frequency bands shortly after the timing of the PTX injection. (e) A PO₂ spectrogram of a rabbit not exposed to neonatal anesthesia demonstrates the abolition of the resting state higher frequency component after the injection of PTX. (f) A PO₂ spectrogram of a rabbit exposed to neonatal anesthesia shows an initial decrease in the power of the higher frequency component after PTX injection; however, during major epileptic events, there is a sudden increase in the power across most frequency bands, with a subsequent abolition across all frequencies above 1 cpm. (g) shows a significant increase (p < 0.05) in the occurrence of epilepsy between the rabbits not exposed to neonatal anesthesia and the rabbits exposed to neonatal anesthesia. Blue arrows in (c–f) denote the time of PTX injection. Examples are representative of the group means reported above

Figure 23.1e and f compare the PO₂ spectrograms between a rabbit not exposed to neonatal anesthesia and a rabbit exposed to neonatal anesthesia, respectively. In both rabbits, the higher frequency (approximately 10 cpm) of PO₂ oscillations is substantially reduced after the injection of PTX. However, during epileptic events, the rabbits exposed to neonatal anesthesia (Fig. 23.1f) exhibit increases in the power of all PO₂ frequencies up to 15 cpm, corresponding to major epileptic events. The exposed group showed 3.5 ± 2.38 events per 30-min recording (3/5), compared with 0 events in controls (unpaired t-test, p < 0.05). Focal seizures were observed in 0/5 control rabbits and 3/5 anesthesia-exposed rabbits. All anesthesia-exposed rabbits had epileptiform discharges across the recording period.

To illustrate how hemodynamics can be used to gauge various levels of functional deficiency in the inhibitory system, Fig. 23.2 displays illustrative PO₂ traces from our experiments. Figure 23.2a shows the typical PO₂ fluctuations during the resting state across 10 min (prior to the injection of PTX). Notably, there is a prominent high-frequency component at around 10 cpm. Figure 23.2b shows a 10-min PO₂ trace after PTX injection from the same rabbit in Fig. 23.2a, which was not exposed to neonatal anesthesia. The higher frequency fluctuations of PO₂ are smaller compared to the lower frequency fluctuations, suggesting a neurovascular decoupling resulting in reduced frequencies. Figure 23.2c shows a 10-min PO₂ trace after PTX injection from a different rabbit that was exposed to neonatal anesthesia. Major epileptic events are reflected by the extended and significant fluctuations in PO₂, along with a preceding increase in higher frequency oscillations.

Fig. 23.2.

PO₂ reveals different levels of functional deficiency in the inhibitory system. (a) displays 10 min of PO₂ prior to the injection of PTX, highlighting the dominant 10 cpm higher frequency component typical for the resting state. (b) displays 10 min of PO₂ after PTX injection for a rabbit not exposed to anesthesia as a neonate, where slower frequencies have higher power over higher frequencies. (c) displays 10 min of PO₂ after PTX injection for a rabbit exposed to neonatal anesthesia, which has large fluctuations in PO₂ corresponding to epileptic events

23.4. Discussion

Our results indicate the presence of “hidden” epileptic activity in adults previously exposed to neonatal anesthesia. We used the GABA antagonist picrotoxin at a small concentration (negligible for healthy subjects at the level of single neurons (Aksenov et al. 2023b)) to mimic a minor functional deficiency of the inhibitory system. The rabbits exposed to neonatal anesthesia experienced focal seizures and epileptic activity, which could be visualized using recordings of both electrophysiology and brain tissue oxygen. This suggests that any conditions, diseases, or states that induce a minor functional deficiency of the inhibitory system can trigger epileptic-like activity in subjects exposed to neonatal anesthesia. All animals underwent identical implantation surgery, so the differential response cannot be attributed to acute wound-healing effects.

Neonatal anesthesia produces initial damage, but the brain can compensate for it under normal conditions—that is why we do not see epileptic activity in the subjects exposed to neonatal anesthesia without picrotoxin. However, when a challenge (small concentration of picrotoxin in rabbits or, presumably, stress, overstimulation, or additional damage in humans) occurs, this compensation can fail, leading to the manifestation of epileptic activity or increased excitation. This can explain the inconsistency of clinical manifestations in populations exposed to neonatal anesthesia, as the anesthesia alone may not be sufficient for clinical symptoms to appear; an additional trigger or challenge is required. Moreover, different clinical manifestations can appear based on specific brain regions involved, depending on the nature of secondary damage or challenge.

This phenomenon concerns not only clinical epilepsy but also other conditions related to increased excitability in specific brain regions, such as autism and attention-deficit/hyperactivity disorder (ADHD) (Rubenstein and Merzenich 2003; Bogdanska-Chomczyk et al. 2024). These conditions have also been inconsistently reported in patients previously exposed to neonatal anesthesia. Additionally, post-traumatic conditions can follow the same logic.

Recordings of brain tissue oxygen illustrate the utility of hemodynamic signals in diagnosing this epileptic-like activity, even if it is highly localized, clinically silent, and located in the deep structures of the brain.