Volatile Anesthetics Rapidly Increase Dendritic Spine Density in the Rat Medial Prefrontal Cortex during Synaptogenesis

• Long periods of exposure to anesthetics can accelerate and enhance cell death in neonatal animals, but their effect on synaptic formation later in early life is unknown

• Volatile anesthetic exposure for up to 2 h did not cause cell death, but altered dendritic spine architecture in 16-day-old rat pups

• The behavioral relevance in rats of this second type of effect on brain development requires further study

THE central nervous system requires the proper formation of exquisitely precise neuronal circuits to function correctly. In humans, the brain growth spurt period, including dendritic development and synaptogenesis, starts at the beginning of the third trimester of pregnancy and is thought to end up only several years after birth.1–3In the rodent cerebral cortex, the most intense phase of synaptogenesis takes place between the second and fourth postnatal weeks.4–6Extensive experimental work demonstrates that, during this period, interference with physiologic neuronal activity can alter wiring patterns and, thereby, alter information processing in the central nervous system.7–10General anesthetics are powerful modulators of neuronal activity, raising the possibility that these drugs might significantly affect activity-dependent neuronal network development during critical periods of brain maturation. This issue is of potential clinical interest, because millions of human newborns undergo general anesthesia worldwide every year, and a growing body of clinical observations suggests a casual relationship between anesthesia/surgery and adverse neurocognitive outcome in these young populations.11–13 

Although there is compelling evidence that anesthetics can induce neuroapoptosis at distinct stages of development,14–16less is known about the impact of these drugs on developing dendritic arbor and synaptogenesis. Recently, exposure of newborn rodent pups to anesthetics has been shown to decrease dendritic spine and synapse number in the hippocampus,17,18and this decrease in synaptic connectivity was proposed to mediate anesthesia-related neuroapoptosis.17Because anesthesia-triggered extensive neuroapoptosis seems to be confined to a critical period terminating around the tenth day of postnatal life in the rodent cerebral cortex,14and that rodent cortical synaptogenesis is actively ongoing until the end of the first postnatal month,4–6an important remaining question is to explore the effect of anesthetics on neuronal cytoarchitecture during these later but still very intense stages of synaptic network development.

To elucidate this issue, in a recent series of experiments, we have demonstrated that exposure of mice pups to sedative concentrations of midazolam, propofol, or ketamine for 5 h at postnatal day (PND) 15 rapidly induced a significant and persistent increase in dendritic spine and synapse density in several brain areas.19This study was aimed to pursue these investigations by assessing the exposure time-dependent effects of three volatile anesthetics (isoflurane, sevoflurane, and desflurane) on neuronal cytoarchitecture in the medial prefrontal cortex (mPFC), an area of utmost importance in higher order cognitive functions.20,21In these experiments, we specifically focused on the acute effects of volatile anesthetics on layer 5 pyramidal neurons at PND 16, a developmental stage when these drugs do not trigger significant neuroapoptosis, but there is still actively ongoing synaptogenesis in the cerebral cortex.

The experimental protocol was reviewed and approved by the Ethics Committee of the University Medical Center of Geneva and by the Cantonal Veterinary Office, Geneva, Switzerland. Animals were group housed and bred in the animal facilities of the University of Geneva Medical School under light- (12-h light/dark cycle) and temperature-controlled (22°± 2°C) conditions. Food and water were available ad libitum . Every effort was made to minimize the number of animals used and their suffering. Sixteen-day-old (PND 16) Whistar rats were used for all experiments.

Awake animals were placed in an induction chamber and received isoflurane (1.5%, isoflurane vaporizer; Ohmeda, Steeton, West Yorkshire, United Kingdom), sevoflurane (2.5%, sevoflurane vapore; Abbott, Chicago, IL), or desflurane (7%, desflurane vaporizer; Ohmeda), administered through a continuous oxygen flow of 1 l/min for 5 min. Animals were then withdrawn from the induction chamber and positioned in dorsal recumbency, while volatile anesthetics at the same concentration were further administrated on spontaneous respiration via  a facemask. At these concentrations, these volatile drugs induced deep sedation (absence of the righting reflex > 10 s) without significant compromise of hemodynamic parameters and blood gas values (table 1). Body temperature was monitored and maintained between 37° and 38°C by means of a heating pad (Harvard Apparatus, Holliston, MA). At the end of the anesthesia (30, 60, or 120 min after induction), the left femoral vein was punctured for blood gas analysis to determine adequacy of spontaneous respiration. Blood gas analysis was performed with the handheld i-STAT analyzer (Abbott) and blood glucose with Bayer Ascensia Contour (Bayer Health Care, Tarrytown, NY). Control animals underwent the same maternal separation and handling as anesthetized animals and were kept in individual cages for the duration of the experimental procedure.

Six hours after anesthesia induction, animals were killed by an overdose of pentobarbital (100 mg/kg intraperitoneally) and perfused transcardially with a 4% paraformaldehyde and 0.125% glutaraldehyde (pH 7.4) solution. Brains were then removed and postfixed for 2 h in 4% paraformaldehyde. Coronal sections (300-μm thick) of the left hemisphere were then cut on a Vibratome in ice-cold phosphate-based saline solution (pH 7.4). Coronal sections were prestained for 30 min with methylene blue that enabled the visualization of neuronal somata, mounted in an injection chamber, and placed on a fixed-stage Zeiss microscope equipped with a micromanipulator. Layer 5 pyramidal neurons were loaded iontophoretically with a 5% Lucifer Yellow solution (Sigma-Aldrich, St. Louis, MO), using sharp micropipettes with a negative current of 70 nA. Loading time per cell was 4 min. Six to nine cells were loaded per slice and animal. The prefrontal cortex and layer 5 were visually identified according to cytoarchitectural features.

For immunolabeling to detect Lucifer Yellow, coronal sections were incubated with an anti-Lucifer-Yellow antibody (Sigma-Aldrich; 1:5000 dilution) for 48 h at ambient temperature in a phosphate-based saline solution containing sucrose (5%), bovine serum albumine (2%), triton (1%), and azide (0.1%). Slices were then rinsed in phosphate-based saline solution and incubated for an additional 48 h with an Alexa-conjugated secondary antibody (Molecular Probes, Carlsbad, CA; 1:1000). Slices were mounted and coverslipped using immunomount (Thermo Scientific, Pittsburgh, PA).

Twelve-micrometer-thick sections were cut, collected on gelatin-coated slides, and dried overnight. Potassium permanganate (0.06%) was applied for 10 min to ameliorate signal-to-noise ratio, and slides were then incubated with a 0.004% Fluoro-Jade B solution (Histo-Chem, Inc., Jefferson, AR) for 20 min. Slides were then rinsed through three changes of distilled water for 1 min per change and air dried. Sections were then costained with 4′,6′-diamidino-2-phenyllindole (Sigma-Aldrich), a fluorescent nuclear stain, and coverslipped with DPX nonfluorescent mounting media (Sigma-Aldrich).

Samples were rinsed 2 × 5 min with phosphate-based saline and incubated for 15 min with the terminal deoxynucleotidyl transferase deoxyuridine triphosphate nick-end labeling buffer (30 mm Tris, 140 mm sodium cacodylate, and 1 mm cobalt chloride). Then, 0.3 U/μl of terminal transferase (Roche, Mannheim, Germany) and 6 μm 7′-fluorescein-labeled deoxyuridine triphosphate (Roche) were applied for 90 min at room temperature. The reaction was stopped with 2× sodium citrate buffer and the samples were washed again with phosphate-based saline and mounted with immunomount (Thermo Scientific).

Only pyramidal neurons lying within layer 5 of the mPFC with proper filling of distal dendritic tips were included in the analysis. Reconstruction of 3D dendritic structure was performed on a computer-based Neurolucida system (Microbrightfield, Williston, VT) with a 40× objective on a Nikon microscope (Nikon Corporation, Tokyo, Japan). Total dendritic length, number of branching points, dendritic ends, and surface area for both basal and apical dendritic arbors were quantified.

An LSM 510 metaconfocal microscope (Carl-Zeiss, Göttingen, Germany) equipped with a 63× oil-immersion objective was used for dendritic spine analysis. Only second-order dendritic shafts, situated in a distance between 100 and 200 μm from the soma, were considered. Spine density analysis was performed on the LSM image viewer software (Carl-Zeiss). We counted protrusions located behind each other on z stacks whenever distinction was possible. Protrusions were considered as spines whenever they displayed a clearly distinguishable head under high magnification. Using this criterion, filopodia (i.e. , protrusion without head) constituted less than 1% of total protrusions under each experimental condition and were not further considered in this study. Spine length and head diameter analyses were carried out by scrolling across single z stacks of raw images using a plug-in specifically developed for OsiriX software.**

Note that for illustration purposes, images presented in figures 1–6are maximum intensity projections of z stacks with volume rendering, further treated with a Gaussian blur filter.

All statistics are given with SD. Normality was tested for each distribution (D'Agostino and Pearson test), and α was set to 5% for all tests. For each volatile anesthetic tested, we compared the mean spine density obtained from individual cells with the control condition, distinguishing either between basal and apical dendrites or between spine head diameter-defined groups. For this, we performed for each group one-way ANOVA with Bonferroni post hoc  tests using the Prism Software (Version 5.0a; GraphPad Software, Inc., La Jolla, CA). Note that for graphical convenience, spine density was further normalized and represented as percentage of control for figures 2E, 3E, and 4E.

To explore the effects of volatile anesthetics on the developing cerebral cortex during the intense phase of cortical synaptogenesis,4–6we exposed PND 16 rat pups to isoflurane (1.5%), sevoflurane (2.5%), and desflurane (7%) for a period extending from 30 min up to 2 h. As given in table 1, neither of these volatile drugs induced important respiratory or metabolic derangements at these above-mentioned concentrations up to a 2-h-long exposure.

To study the effect of volatile anesthetics on cortical synaptogenesis, we performed intracellular injections of Lucifer Yellow into layer 5 pyramidal neurons of the mPFC 6 h after initiation of anesthesia in PND 16 animals. As seen in figure 1, this technique allowed complete filling of apical and basal dendrites with dendritic spines, representing primary postsynaptic sites of excitatory synaptic contacts in neurons, throughout the entire extent of the dendritic tree. Using this technique, we first analyzed the effects of volatile anesthetics on dendritic spines on second-order dendritic segments of layer 5 pyramidal neurons in the mPFC with confocal microscopy. In these experiments, anesthesia was maintained for 30, 60, and 120 min, and the effect of these treatment paradigms on dendritic spine density and head diameter was assessed 6 h after the initiation of anesthesia. Because apical and basal dendrites develop differently,22,23we discriminated between spines that were located on apical and those on basal dendrites (see Materials and Methods section for selection criteria).

In animals exposed to 1.5% isoflurane, we observed a gradual, exposure time-dependent increase in dendritic spine density on both apical and basal dendrites (figs. 2A–E). In control, sham-treated PND 16 pups, dendritic spine density was 0.91 ± 0.11 μm⁻¹on apical and 0.81 ± 0.12 μm⁻¹on basal second-order dendrites of layer 5 pyramidal neurons in the mPFC. As short as a 30-min-long exposure resulted in a small, nonsignificant increase in spine density on both apical (7 ± 13%, 0.98 ± 0.12 spines μm⁻¹, P = 0.21) and basal (14 ± 15%, 0.92 ± 0.13 spines μm⁻¹, P = 0.066) dendrites. The increase in spine density reached significance after a 60-min-long exposure (26 ± 11%, 1.14 ± 0.1 spines μm⁻¹, P = 0.0003 on apical; and 17 ± 8%, 0.95 ± 0.07, P = 0.01 on basal dendrites), and then further accentuated after 120 min (37 ± 29%,1.25 ± 0.27 spines μm⁻¹, P = 0.002 on apical; and 31 ± 24%, 1.1 ± 0.19 spines μm⁻¹, P = 0.004 on basal dendrites). The higher spine density observed after isoflurane exposure was primarily due to a significant increase in the number of spines with smaller head diameters, representing, most probably, newly formed spine populations (fig. 2F).

As short as a 30-min-long anesthesia with sevoflurane at 2.5% was sufficient to induce a highly significant increase in spine density on both apical (25 ± 14%, 1.14 ± 0.13 spines μm⁻¹, P = 0.001) and basal dendrites (48 ± 21%, 1.19 ± 0.17 spines μm⁻¹, P < 0.0001; fig. 3). Similar to that observed in the isoflurane group, extending exposure time also resulted in a gradual time-dependent increase in spine density on apical dendrites of sevoflurane-anesthetized animals (at 60 min: 32 ± 11%, 1.2 ± 0.1 spines μm⁻¹, P < 0.0001; at 120 min: 38 ± 19%, 1.25 ± 0.17 spines μm⁻¹, P < 0.0001). In contrast, the effect of sevoflurane on spine density of basal dendrites plateaued at 30 min, and no further increase could be observed after exposure to this drug up to 2 h (at 60 min: 40 ± 16%, 1.13 ± 0.13 spines μm⁻¹, P < 0.0001; at 120 min: 38 ± 24%, 1.11 ± 0.2 spines μm⁻¹, P = 0.001). In line with the observations made in isoflurane-exposed animals, the increased spine density after sevoflurane treatment was also mainly due to an important increase in the number of spines with smaller head diameters (fig. 3F).

In contrast to isoflurane and sevoflurane, exposure to desflurane (7%) up to 60 min induced no increase in spine density either on apical (at 30 min: 3 ± 9%, 0.94 ± 0.09 spines μm⁻¹; at 60 min: 2 ± 9%; 0.93 ± 0.03 spines μm⁻¹) or on basal dendrites (at 30 min: 2 ± 5%, 0.82 ± 0.02 spines μm⁻¹; at 60 min: 98 ± 7%, 0.80 ± 0.09 spines μm⁻¹; fig. 4). However, a significant increase in spine density in this group was found after a 2-h-long exposure (20 ± 16%, 1.09 ± 0.15 spines μm⁻¹, P = 0.01 on apical; 35 ± 8%, 1.09 ± 0.07 spines μm⁻¹, P < 0.0001 on basal dendrites). Analysis of spine head diameters revealed a significant increase in the number of spines with large spine head diameter (>0.5 μm) after 30 min of exposure, indicating that desflurane might rapidly induce alterations in dendritic spine morphology without affecting spine density at short-term exposure (fig. 4F). The desflurane-induced increase in spine density, after a 2-h-long exposure to this drug, was also associated with a significant increase in the number of spines with smaller head diameters (fig. 4F).

To further evaluate the effect of volatile anesthetics on neuronal cytoarchitecture, we also performed semiautomatic quantitative analysis of dendritic arbor architecture in control and volatile-exposed animals 6 h after the initiation of anesthesia (fig. 5A). This analysis revealed no significant differences in total dendritic length or in the number of branching points after isoflurane, sevoflurane, or desflurane up to 2-h-long exposure times (figs. 5B–E). This indicated that overall dendritic outgrowth/retraction or the degree of dendritic branching was not changed by these volatiles. In addition, no differences were observed in Sholl's distribution of dendritic arbor (figs. 5F–G), suggesting that the three-dimensional organization of the dendritic tree was not affected either.

Finally, we also assessed potential anesthesia-related cell loss associated with volatile drug exposure at this developmental stage. To this end, the necrotic and the apoptotic type of cell death were evaluated in PND 16 animals 6 h after the initiation of a 2-h-long anesthesia, maintained by isoflurane (1.5%), sevoflurane (2.5%), or desflurane (7%). As seen in figures 6A–C, neither Fluoro-Jade B staining nor the terminal deoxynucleotidyl transferase deoxyuridine triphosphate nick-end labeling assay revealed cellular degeneration or apoptosis after a 2-h-long isoflurane treatment, and similar results were obtained in the sevoflurane and desflurane group (data not shown). As a positive control for detecting anesthesia-induced cell death, we also performed a 6-h-long anesthesia in PND 7 rat pups using intraperitoneally administered ketamine (40 mg/kg). In line with previous observations, this treatment readily induced necrotic and apoptotic cell death in the mPFC (figs. 6D–E). Altogether, these data strongly suggest that exposure to volatile anesthetics up to 2 h does not induce cell death in the developing cerebral cortex of PND 16 rat pups.

In this study, we provide in vivo  evidence that exposure to clinically relevant concentrations of volatile anesthetics up to 2 h can rapidly impair neuronal cytoarchitecture without inducing cell death in rats at PND 16, a developmental period characterized by active synaptogenesis in the cerebral cortex.4–6By focusing on dendritic arbor architecture of layer 5 pyramidal neurons in the mPFC, we demonstrate that while isoflurane, sevoflurane, or desflurane did not alter important morphofunctional aspects of gross dendritic arbor, sevoflurane or desflurane significantly increased dendritic spine density on apical and basal dendritic shafts of these cells. Importantly, our data also reveal considerable differences between these three volatile agents in terms of exposure time-dependent effects on dendritic spine density. Given that dendritic spines represent primary sites of synaptic contacts in neurons, these new results suggest that volatile anesthetics, with different potencies and without inducing cell death, could rapidly interfere with physiologic patterns of synaptogenesis and thus might impair appropriate circuit assembly in the developing cerebral cortex.

To our knowledge, this is the first study designed to evaluate the effect of volatile anesthetics on precise neuronal cytoarchitecture at a developmental stage (PND 16), when cortical synaptogenesis is still very intense,4–6but the critical period for naturally occurring neuroapoptosis is already terminated.14,24In fact, in the majority of previous studies, aimed to investigate the effects of general anesthetics on the developing brain during the brain growth spurt period, anesthetics-induced cell death in the developing brain was reported when rodents were exposed to these drugs at some point between developmental stages of PND 5 and 10 (for review see Loepke and Soriano16). In line, recent in vitro  and in vivo  studies reported that the cell death-inducing effects of volatile anesthetics are also age dependent and peak around PND 5–7.25,17Our results, showing that exposure to volatile drugs does not induce cell death in the brain of PND 16 animals, are in agreement with these observations. The fact that we found no significant changes in gross dendritic arbor architecture in volatile-treated animals compared with control littermates further indicates that the observed effects of these drugs are specifically restricted to dendritic spines at this developmental stage. However, it is important to note that we only evaluated the possibility of volatile-induced cell death at 6 h postexfosure. Thus, we cannot formally exclude the possibility that these drugs induce delayed cell death several days after exposure and this should be considered as a potential limitation of the study.

Although the important role of neural activity to drive synaptogenesis during critical periods of neural network formation is well established, little is known about how exposure to anesthetics affects synapse formation. A powerful approach to address this question is to focus on dendritic spines, because these structures are the recipients of most excitatory inputs to pyramidal cells26,27and play a key role in the expression of synaptic plasticity.28Importantly, several potential links between spine morphology and function have been reported.27Among them, spine head volume has been shown to be directly proportional to the size of postsynaptic density, the number of postsynaptic receptors, and to the presynaptic number of docked vesicles and thus the readily releasable pool of neurotransmitters.29–32These observations suggest that the analysis of the morphology of a spine could give important indication as to its function, notably synaptic strength.27In line with experimental studies, dendritic spine abnormalities are the most consistent anatomical correlates of altered cognitive function associated with multiple human neurologic disorders.33Imaging dendritic spines thus provide a powerful tool to obtain insights into neural activity-driven morphofunctional plasticity in the cerebral cortex.34 

Recently, we demonstrated that a 5-h-long exposure of PND 15 mice pups to general anesthetics, such as midazolam, propofol, and ketamine, induces a significant increase in the dendritic spine density of the cerebral cortex.19More importantly, we provided evidence that these effects persist for several days to weeks and that the majority of newly formed spines form functional synapses.19The effect of shorter, and thus clinically more relevant, exposure times on dendritic spines remained, however, an open question. Another intriguing issue is to elucidate whether and how volatile anesthetics impair dendritic spine development. Hence, an important new observation of this work is that volatile anesthetics can rapidly induce significant increases in dendritic spine density without affecting gross dendritic arbor morphology. Given that none of the volatiles, up to a 2-h-long exposure, used in this study induced significant metabolic or respiratory perturbations at the systemic level, these results strongly suggest a specific and selective action of these drugs on dendritic spines and thus at the synaptic level.

Our observations are seemingly in contradiction with recent results demonstrating an anesthesia-induced decrease in dendritic spine density.17,18The most plausible explanation for this discrepancy is related to the fact that these studies were performed at rather different developmental stages of neural circuitry assembly. In fact, in those aforementioned studies, anesthesia was provided at earlier developmental stages (PND 0–10 pups), when signaling through the γ-aminobutyric acid receptor type A complex has excitatory modalities.9On the other hand, our study was performed in PND 16 animals, when γ-aminobutyric acid has inhibitory actions.9,35The developmental switch from excitatory toward inhibitory nature of γ-aminobutyric acid-ergic neurotransmission occurs during the second postnatal week in the rodent cerebral cortex and is primarily linked to the developmental expression of the of the K⁺–Cl⁻cotransporter (KCC2), extruding actively intracellular Cl⁻from neurons.9,35,36Thus, although exposure to anesthetics will enhance γ-aminobutyric acid receptor type A receptor-mediated neuronal excitation at earlier development, these same drugs will have an opposite, inhibitory action at the later stages of synaptogenesis. In line with this assumption, we have previously provided evidence that a potential mechanism, underlying anesthesia-induced rapid synaptogenesis, is the modulation of the excitatory versus  inhibitory balance in favor of inhibition.19 

Assessing the effects of different anesthetics in the same experimental model provides useful comparison between the effects of these drugs on particular aspects of brain development. Here, we thus compared the exposure time-dependent effects of three widely used volatile drugs at mean alveolar concentrations that are also relevant to human practice. Importantly, in an attempt to better fit into potential clinical relevance, we focused on short exposure times, ranging from 30 min to 2 h. Our data revealed that the extent of increase in spine density was dependent on the exposure time, with longest administration exhibiting the most robust effects in the case of each volatile anesthetic. These data thus strongly suggest an exposure time-dependent effect. However, as evaluation of spine density systematically occurred 6 h after the initiation of anesthesia, recuperation times also differed between experimental groups. Thus, while we previously demonstrated that the effects of anesthetics on spine density persists for days to weeks,19we cannot formally exclude the possibility that the observed exposure time-dependent increases in spine density are not fully independent from these differences in recuperation time.

Our results revealed important differences between volatile drugs as to the time course of their effects on spine density, with sevoflurane being the fastest and desflurane being the slowest to induce spinogenesis. The reason for these differences remains unclear. Despite the fact that all animals in each experimental group displayed comparable sedation profile, we can of course formally not exclude the possibility that the depth of anesthesia differed between volatile drugs at some level. Alternatively, the observed differences in the time course of the effects might not primarily be related to the depth of anesthesia but would rather be attributable to different pharmacologic effects exerted by these drugs at the cellular and network level. In fact, equivalent minimum alveolar anesthetic concentrations values of volatile anesthetics have been shown to have fairly different effect on bispectral index values, and thus on neuronal activity patterns, in children.37In line with these clinical observations, an increasing number of experimental studies also indicate quite different potencies of volatile drugs to induce apoptosis both in neurons and in other cell types.38–42It has recently been suggested that this differential effect is related to the different potencies of these drugs to induce calcium release from the endoplasmic reticulum via  activation of the inositol-1,4,5-trisphosphate receptors.38Further studies are needed to determine whether such mechanisms are also involved in volatile anesthetics-induced spinogenesis during development.

The functional relevance of our results remains to be explored. In this study, we focused on and compared the effect of volatile anesthetics on neuronal cytoarchitecture in the rat mPFC. In rodents, this brain region is considered as the functional equivalent of the human prefrontal cortex,21a cortical structure that is implicated in higher order cognitive and emotional functions.20To our knowledge, there are currently no studies available that explores the long-term cognitive/behavioral effects of general anesthetics when these drugs were administered at the developmental stage targeted by our current work.16Because rodent models of prefrontal cortical function are now accessible,43performing such experiments could give further insights into this issue. We believe that the volatile anesthetics-induced rapid changes in dendritic spine density and shape in this cortical region could constitute the morphological substrate of functional alterations. In fact, dendritic spines are the primary sites of excitatory synaptic contacts in neurons, and increasing evidence suggest that dendritic spine density, shape, and distribution play fundamental roles as to the excitable properties of a neuron.27,44Because this study focused on a developmental period when dendritic spine turnover is especially rapid and that immature spines may be transient without synapse formation,34anesthesia-induced increase in spine density does not necessarily imply functional interference with circuitry development. However, we have previously demonstrated that the majority of newly formed spines after anesthesia exposure integrate into the circuitry by forming functional synapses, suggesting that anesthetics might indeed modify network characteristics.19In line with this assumption, the majority of even small, immature-like spines have been demonstrated to present postsynaptic density,45and it has recently been demonstrated that newly formed spines rapidly acquire postsynaptic density and exhibit functional maturity.46 

Finally, the extrapolation of the developmental stage targeted in this rodent study to human development is a delicate question. It has been long time considered that the brain developmental stage of PND 7 in mice and rats, with the majority of experimental works has been performed so far, corresponds approximately to the 32–36 weeks of gestation in humans.47,48However, recent work suggests that the PND 7 rodent brain is rather equivalent to 17–20 weeks of human gestation.49,50In fact, with regard to neuronal circuit formation, the peak synaptogenic period in humans takes place between the third trimester of pregnancy and the first few years of postnatal life,1–3whereas, in rodents, this period is situated between the second and fourth postnatal weeks.4–6Thus, we believe that conducting experiments in PND 16 animals might reflect a developmental stage somewhere during the first few years at the human scale.

In conclusion, this study demonstrates that volatile anesthetics can rapidly increase dendritic spine density during the peak synaptogenic period in the rodent cerebral cortex. These results further strengthen our recent data, demonstrating that nonvolatile general anesthetics induce synaptogenesis via  the modulation of the excitatory/inhibitory activity balance at this same stage of central nervous system development.19Most importantly, the ensemble of these observations raise the intriguing possibility that general anesthetics might interfere with physiologic patterns of neural network formation at a developmental stage when exposure to anesthetics is not associated with significant neuroapoptosis, but when there is still actively ongoing synaptogenesis.

The authors thank Beatrice King (Technician, Department of Fundamental Neuroscience, University of Geneva Medical School, Geneva, Switzerland) and Manuel de Costa (Technician, Department of Anesthesiology, Pharmacology and Intensive Care, University Hospital of Geneva, Geneva, Switzerland) for technical support.