Isoflurane-induced neuroapoptosis in the neonatal rhesus macaque brain

Abstract

Background: Brief isoflurane anesthesia induces neuroapoptosis in the developing rodent brain, but susceptibility of non-human primates to the apoptogenic action of isoflurane has not been studied. Therefore, we exposed postnatal day 6 (P6) rhesus macaques to a surgical plane of isoflurane anesthesia for 5 h, and studied the brains 3 h later for histopathologic changes.

Method: With the same intensity of physiologic monitoring typical for human neonatal anesthesia, five P6 rhesus macaques were exposed for 5 h to isoflurane maintained between 0.7 and 1.5 end-tidal Vol% (endotracheally intubated and mechanically ventilated) and five controls were exposed for 5 h to room air without further intervention. Three hours later, the brains were harvested and serially sectioned across the entire forebrain and midbrain, and stained immunohistochemically with antibodies to activated caspase-3 for detection and quantification of apoptotic neurons.

Results: Quantitative evaluation of brain sections revealed a median of 32.5 (range, 18.0-48.2) apoptotic cells/mm of brain tissue in the isoflurane group and only 2.5 (range, 1.1-5.2) in the control group (difference significant at P = 0.008). Apoptotic neuronal profiles were largely confined to the cerebral cortex. In the control brains, they were sparse and randomly distributed, whereas in the isoflurane brains they were abundant and preferentially concentrated in specific cortical layers and regions.

Conclusion: The developing non-human primate brain is sensitive to the apoptogenic action of isoflurane and displays a 13-fold increase in neuroapoptosis after 5 h exposure to a surgical plane of isoflurane anesthesia.

Figures

Figure 1

Density of activated caspase 3-stained neurons (profiles/mm3) in isoflurane-exposed versus control brains. Transverse sections (approximately 800) were cut serially across the entire forebrain and midbrain and counts were performed on sections selected at 2.24 mm intervals (every 32nd section). An outline of each section was entered into the computer and all stained neuronal profiles within the outlined space were counted and marked for location. The total number of stained neurons divided by the total tissue space (determined by the applied software) yielded a neuronal density count (number of activated caspase 3-stained profiles per mm3) for each brain. The median density (range) of activated caspase 3-positive neurons per mm3 for the isoflurane-exposed brains was 32.5 (18.0 to 48.2), and for the control brains was 2.5 (1.1 to 5.2). This amounts to a 13-fold increase in density of apoptotic neurons in the isoflurane-exposed brains. The difference in median density of apoptotic neurons between the isoflurane and control groups was 30 (95% CI, 15.5 to 45.7, p = 0.008).

Figure 2

Computer plots based on the number and location of each activated caspase 3-stained neuronal profile in homologous sections from the primary visual cortex of a control and isoflurane brain. The sections shown are from the animal in each treatment group (isoflurane and control) that had the highest mean activated caspase 3 density count. Note the striking laminar pattern of distribution of stained neuronal profiles in the isoflurane brain and the randomly scattered pattern in the control brain. The layers primarily affected in the isoflurane brain are layers II and V, but there are also many stained profiles in between these two layers. The primary cell types affected are depicted in figure 3.

Figure 3

The appearance of activated caspase 3-stained neurons in layer II (panel A & B) and layer V (panel C) of the primary visual cortex of an isoflurane-exposed brain. The layer II profiles have the shape and arborization pattern characteristic of γ-aminobutyric acid-ergic inhibitory interneurons. The layer V profiles are predominantly small pyramidal neurons (presumably glutamatergic) of a type that are thought to project to visual neurons in the contralateral hemisphere. The large activated caspase 3-positive multipolar neuron in panel C was seen occasionally in layer V and more frequently in the superficial portion of layer VI. This cell has the appearance of a Martinotti cell, which has been described as γ-aminobutyric acid-ergic and is thought to mediate inhibition.

Figure 4

Computer plots based on the number and location of each activated caspase 3-stained profile in homologous sections from three different divisions of the neocortex (temporal, somatosensory, frontal) of control versus isoflurane-exposed brains. In all of the control sections the activated caspase 3-stained profiles are sparse and randomly distributed in no particular relationship to the laminar organization of the cortical tissue. In the temporal cortex of the isoflurane-exposed brain a laminar pattern is evident, with layers II and IV being the most severely affected. The density count in these particular temporal cortical sections is 17-fold higher in the isoflurane compared to the control section. In the somatosensory cortex, a laminar pattern is also evident, but not quite as pronounced, and the overall count is 11.5 fold higher in the isoflurane than control section. In the frontal cortex, the laminar pattern is only faintly evident and the overall count is only 3-fold greater in the isoflurane than control section.

Figure 5

The appearance of activated caspase 3-stained (AC3) neurons in layer II (panel A) and deeper layers (panel B) of the temporal cortex of an isoflurane-exposed brain. The layer II AC3-positive profiles are interneurons, presumably γ-aminobutyric acid-ergic, and are identical in appearance to layer II neurons that frequently undergo apoptosis in various divisions of the infant mouse neocortex following exposure to either anesthetics or ethanol. Panel B is a scene from deeper layers of the isoflurane-exposed temporal cortex depicting the appearance of several types of pyramidal neurons in an early stage of degeneration. In layer IV, many AC3-positive profiles are large pyramidal neurons having a typical triangular cell body and long apical dendrites. These are accompanied in the same layer by very small AC3-positive neurons and in layer III by occasional AC3-positive neurons that are relatively large. The latter two cell types would probably be classified as pyramidal neurons, although their cell bodies are shaped more like a tear- drop than a pyramid. These several types of pyramidal neurons are identical in appearance to those that are preferentially affected in the same cortical layers of infant rodent brain following exposure to either anesthetics or ethanol.

Figure 6

Comparison of pyramidal neurons in early versus late stages of degeneration following exposure to isoflurane. Panel A shows the typical appearance of a relatively well-preserved activated caspase 3 positive pyramidal neuron in an early stage of degeneration. Note that the shaft of the apical dendrite is straight and does not appear dysmorphic. Panels B, C, & D are from layer V of the primary visual cortex and illustrate a later stage of degeneration when the apical dendrite has begun to shrivel and assume a corkscrew-like configuration. Panels E, F, & G are from the deep layers of the temporal cortex and show neurons in a late stage of degeneration with shriveled apical dendrites that are beginning to undergo fragmentation. Compare these dysmorphic neuronal profiles with those in figure 5B that are in an earlier stage of degeneration.

Figure 7

Timing of the brain growth spurt period in human versus nonhuman primates. This graph was adapted from a publication by Dobbing and Sands providing comparative brain growth spurt data for numerous mammalian species. The curves represent the changing rates of brain growth for humans and rhesus monkeys in the perinatal period. In both species the rate of brain growth increases rapidly to a peak level then gradually descends to a lower level sustained into adolescence. Note that the timing of the growth spurt in relation to birth occurs earlier in monkeys than in humans, and it tapers down to a low level at an earlier age. Thus, in the monkey infant at postnatal day 6, the brain growth rate has decreased to about 25% of its peak value, whereas in the human infant at postnatal day 6 the brain growth rate is just reaching its peak and does not taper down to the 25% level until the age of 6 months (see dashed line). Therefore, in terms of comparative brain growth spurt timing, the 6-day-old monkey infant is the equivalent of a 6-month-old human infant. Another relevant comparison is that the brain weight of the infant rhesus monkey at birth is 76% of its adult brain weight, whereas the brain weight of the newborn human is only 27% of its adult brain weight.