General anesthesia alters CNS and astrocyte expression of activity-dependent and activity-independent genes

Abstract

General anesthesia represents a common clinical intervention and yet can result in long-term adverse CNS effects particularly in the elderly or dementia patients. Suppression of cortical activity is a key feature of the anesthetic-induced unconscious state, with activity being a well-described regulator of pathways important for brain health. However, the extent to which the effects of anesthesia go beyond simple suppression of neuronal activity is incompletely understood. We found that general anesthesia lowered cortical expression of genes induced by physiological activity in vivo, and recapitulated additional patterns of gene regulation induced by total blockade of firing activity in vitro, including repression of neuroprotective genes and induction of pro-apoptotic genes. However, the influence of anesthesia extended beyond that which could be accounted for by activity modulation, including the induction of non activity-regulated genes associated with inflammation and cell death. We next focused on astrocytes, important integrators of both neuronal activity and inflammatory signaling. General anesthesia triggered gene expression changes consistent with astrocytes being in a low-activity environment, but additionally caused induction of a reactive profile, with transcriptional changes enriched in those triggered by stroke, neuroinflammation, and Aß/tau pathology. Thus, while the effects of general anesthesia on cortical gene expression are consistent with the strong repression of brain activity, further deleterious effects are apparent including a reactive astrocyte profile.

Keywords: astrocytes; general anesthesia; neurodegenerating diseases; synaptic activity; transcriptional regulation.

FIGURE 1

Consequences of anesthesia on cortical transcription. (A) RNAseq used to determine cortical gene expression changes in mice exposed to 6 h isoflurane anesthesia. Genes significantly induced (red) and repressed (blue) are highlighted (expression cut-off > 1 FPKM, > 1.3 FC up or down, p values adjusted for multiple testing by the Benjamini–Hochberg procedure to give a false discovery rate of 5% (p_adj < 0.05) here and in all RNA-seq analyses; n = 5 animals per condition). (B) Ontological analysis of genes induced and repressed by anesthesia. Top ten most significantly enriched pathways are shown, with pathways with less than four significant genes or not relevant to tissue type omitted. (C) Fold change expression of immediate early and activity-regulated neuroprotective genes in mice undergoing 6 h isoflurane anesthesia v controls (Mean ± SEM fold change anesthesia v control, N = 5 animals per condition, *p_adj < 0.05, overall p-value < 0.001, 1-way ANOVA, all gene expression data shown here and elsewhere, unless otherwise stated, determined using RNA-sequencing).

FIGURE 2

Visual sensory deprivation and light-re-exposure determines activity-mediated CNS gene expression. (A) Visual cortex innervation from the retina allows manipulation of neuronal activity by exposure to conditions of light v darkness. (B) Schematic of experimental protocol. Mice were kept in normal light conditions, total darkness for 7 days, or total darkness for 6 days followed by 24 h of light-re-exposure. (C) Fold change immediate early gene expression in mice exposed to normal light conditions, 7 days of darkness or 6 days of darkness followed by 24 h light re-exposure v mean expression in mice kept in darkness (mean ± SEM, ***p-value < 0.001, 1-way ANOVA, N = 4 mice per condition here and in all experiments involving visual sensory deprivation and light-stimulation). (D) RNAseq analysis reveals the consequences of visual sensory deprivation on visual cortex gene expression. Genes significantly induced by normal light conditions v darkness (red) and repressed (blue) are highlighted (expression cut-off > 1 FPKM, FC up or down > 1.3, p_adj < 0.05). (E) 24 h light re-exposure following sustained darkness alters CNS gene expression. Genes significantly induced or repressed by 24 h light re-exposure v darkness (red) and repressed (blue) are highlighted (expression cut-off > 1 FPKM, FC up or down > 1.3, p_adj < 0.05). (F) Gene expression changes after 24 h light re-exposure v darkness in mice correlated with changes identified in mice kept in normal light conditions v darkness. Significantly altered genes in either paradigm are plotted to show log2 fold change expression in normal light v darkness and 24 h light-exposure v darkness (R2 and p-value from Pearson correlation).

FIGURE 3

Consequences of anesthesia on genes regulated by visual sensory stimulation. (A) Anesthesia overall suppresses genes enhanced in normal light conditions v darkness (expression cut-off > 1 FPKM, p_adj < 0.05, FC > 2), or enhanced by light-re-exposure v darkness (expression cut-off > 1 FPKM, p_adj < 0.05, FC > 1.5). Fold change 6 h anesthesia v control (expression cut-off > 1 FPKM, p-value < 0.001, ratio paired 2-tailed t-test). (B) Visual activity-regulated genes are significantly enriched for genes suppressed by anesthesia. Genes induced > 2 fold by normal light conditions or > 1.5 fold in light re-exposure (expression cut-off > 1 FPKM, p_adj < 0.05) taken and enrichment analysis performed using genes found to be downregulated (expression cut-off > 1 FPKM, p_adj < 0.05, > 2 FC) by anesthesia. Fold enrichment is shown, and 95% confidence interval (CI) of the fold enrichment depicted by the error bar. (*p < 0.05 two-sided Fisher’s exact test).

FIGURE 4

Consequences of anesthesia on genes altered by neuronal-activity blockade in vitro. (A) Anesthesia overall upregulates genes enhanced by neuronal activity blockade (genes upregulated following 24 h 100 nM TTX in neuronal culture, expression cut-off > 1 FPKM, FC > 2, p_adj < 0.05). Fold change anesthesia v control shown (expression cut-off > 1 FPKM, ratio paired 2-tailed t-test). (B) Anesthesia overall downregulates genes enhanced by neuronal activity blockade. Fold change anesthesia v control shown (expression cut-off > 1 FPKM, ratio paired 2-tailed t-test). (C) Genes altered up or down by neuronal activity blockade in vitro are significantly enriched for genes respectively enhanced or suppressed by anesthesia. Enrichment analysis performed on genes induced or suppressed by TTX using genes found to be upregulated and downregulated (expression cut-off > 1 FPKM, p_adj < 0.05, > 1.5 FC) by anesthesia. Fold enrichment is shown, and 95% confidence interval (CI) of the fold enrichment depicted by the error bar. (*p-value < 0.05 two-sided Fisher’s exact test). (D) Activity-regulated pro-death genes found to be upregulated by neuronal activity blockade in vitro, are also upregulated by anesthesia. Expression of pro-death genes upregulated by in vitro neuronal activity blockade using TTX (expression cut-off 1 FPKM, p_adj < 0.05) following 6 h isoflurane anesthesia (Mean ± SEM fold change expression anesthesia v control, N = 5 animals per condition, *p_adj < 0.05, overall p-value < 0.001, 1-way ANOVA).

FIGURE 5

Anesthesia alters genes not regulated by neuronal activity. (A) Anesthesia-regulated genes (FPKM cut-off > 1, p_adj < 0.05, FC > 1.3) not regulated by neuronal activity identified by removing possible activity-regulated genes. These are genes found to be changed in normal light v dark conditions, 24 h light-re-exposure or with neuronal-activity blockade in vitro (expression cut-off > 1 FPKM, unadjusted p-value < 0.1). (B) Confirmation that identified non-activity-mediated anesthesia-regulated genes are not altered in vivo by visual sensory stimulation or in vitro by neuronal-activity blockade. Log2 fold change expression in normal light conditions v darkness (top) or in TTX v control (bottom) shown for activity-independent genes identified as shown in (A). (C) Ontological analysis of non-activity-regulated genes induced or repressed by anesthesia. Top ten most significantly enriched pathways are shown, with pathways with less than 4 significant genes or not relevant to tissue type omitted.

FIGURE 6

TRAPseq allows determination of astrocyte-specific gene expression. (A) Example images from cortical slices from ALDH1l1_eGFP-RPL10a mice with immunofluorescence with anti-EGFP plus an astrocyte-specific marker (Aldh1l1), neuron-specific marker (Neurochrom) and microglial specific marker (Iba1); scale bar 50 uM. (B) Schematic of TRAP protocol, allowing isolation of astrocyte-specific mRNA via immunoprecipitation of EGFP-tagged astrocyte ribosomes. (C) TRAPseq successfully enriches for astrocyte-specific markers and depletes for marker for other CNS cell-types. Mean fold change enrichment or depletion of cell-type specific gene markers following TRAP v pre-TRAP (input) sample (Mean ± SEM fold change TRAP v input, ***p < 0.001, 1 way ANOVA).

FIGURE 7

(A) Anesthesia alters astrocyte gene expression and upregulates genes associated with both acute and chronic reactive states. Genes significantly induced (red) and suppressed (blue) are highlighted (expression cut-off 1 FPKM, FC > 1.3, p_adj < 0.05) (N = 4 animals per condition). (B) Ontological analysis of astrocyte genes induced and repressed by anesthesia. Top ten most significantly enriched pathways are shown, with pathways with less than 4 significant genes or not relevant to tissue type omitted. (C) Anesthesia overall upregulates pan-reactive astrocyte genes enhanced by both acute LPS and middle cerebral artery occlusion (MCAO). Gene lists derived from Zamanian et al. (2012), re-derived as described in Jiwaji and Hardingham (2022). Fold change astrocyte genes, anesthesia v control (expression cut-off > 1 FPKM, ratio paired 2-tailed t-test). (D) Anesthesia overall upregulates reactive astrocyte genes found to be commonly upregulated in end-stage amyloidopathy (APP/PS1 model) and end-stage tauopathy (MAPT-P301S model). Gene-list from Jiwaji and Hardingham (2022). Fold change astrocyte genes, anesthesia v control (expression cut-off > 1 FPKM, ratio paired 2-tailed t-test).

FIGURE 8

Consequences of visual sensory deprivation and light stimulation on astrocyte gene expression, and comparison with astrocyte genes changed by alternative stimulation paradigms and by ageing and disease. (A) TRAPseq identifies astrocyte genes altered by altered visual sensory experience. Genes significantly induced (red) and repressed (blue) are highlighted (expression cut-off > 1 FPKM, FC > 1.3, p_adj < 0.05) for normal light conditions v darkness (left) and with 24 h light-exposure (right). (B) Astrocyte genes upregulated by light re exposure overlap with astrocyte genes significantly enhanced by neuronal activity in vitro (gene-set from Hasel et al., 2017). Fold change light re-exposure v continuous darkness (expression cut-off > 1 FPKM, ratio paired 2-tailed t-test). (C) (left). Activity-dependent astrocyte genes upregulated by visual sensory stimulation (Normal light conditions v continuous darkness, expression cut-off > 1 FPKM, FC > 2, p_adj < 0.05) are significantly enriched in sets of astrocyte genes upregulated after 4 h light exposure in single-cell analysis of visual cortex; drug-induced seizures; and in mice experiencing sleep-deprivation v sleep. (*p-value < 0.05, Fisher’s exact test). (C) (right). Activity-dependent astrocyte genes are enriched in sets of astrocyte genes reduced by ageing and in a mouse model of tauopathy. (*p-value < 0.05, Fisher’s exact test).

FIGURE 9

Consequences of anesthesia on astrocyte genes and pathways altered by neuronal activity (A) Anesthesia overall downregulates (left) astrocyte genes enhanced by light re-exposure v darkness (expression cut-off > 1 FPKM, FC > 1.5, p_adj < 0.05) and (right) upregulates astrocyte genes suppressed by light re-exposure. Fold change anesthesia v control (expression cut-off > 1 FPKM, ratio paired two-tailed t-test). (B) Anesthesia downregulates activity-regulated glycolytic genes. Fold change expression of astrocyte glycolytic genes shown for anesthesia v control; in vitro expression following suppression neuronal activity with TTX [gene-list from Hasel et al. (2017)]; and visual cortex expression following 6 days of darkness v normal light conditions (expression cut-off > 1 FPKM, *p_adj < 0.05 for each individual gene, ***p-value < 0.001 for overall downregulation of glycolysis gene-set, 2-way ANOVA).

FIGURE 10

Consequences of anesthesia on astrocyte genes not regulated by neuronal activity. (A) Astrocyte anesthesia-regulated genes (expression cut-off > 1 FPKM, p_adj < 0.05, FC > 1.3) not regulated by activity identified by removing possible activity-regulated genes as those changed in normal light v dark conditions, 24 h light-re-exposure or with neuronal-activity blockade in vitro (expression cut-off > 1 FPKM, unadjusted p-value < 0.1). (B) Identified astrocyte non activity-mediated anesthesia-regulated genes are not altered in vivo by visual sensory stimulation or in vitro by neuronal-activity blockade. Log2 FC expression of astrocyte genes normal light conditions v darkness (top) or in TTX v control (bottom) shown of genes identified in (A). (C) Ontological analysis of astrocyte non-activity-regulated genes induced or repressed by anesthesia. Top ten pathways are shown, with pathways with less than 3 significant genes or fold enrichment < 2 omitted.