Electroacupuncture attenuates surgical pain-induced delirium-like behavior in mice via remodeling gut microbiota and dendritic spine

Abstract

Surgical pain is associated with delirium in patients, and acupuncture can treat pain. However, whether electroacupuncture can attenuate the surgical pain-associated delirium via the gut-brain axis remains unknown. Leveraging a mouse model of foot incision-induced surgical pain and delirium-like behavior, we found that electroacupuncture stimulation at specific acupoints (e.g., DU20+KI1) attenuated both surgical pain and delirium-like behavior in mice. Mechanistically, mice with incision-induced surgical pain and delirium-like behavior showed gut microbiota imbalance, microglia activation in the spinal cord, somatosensory cortex, and hippocampus, as well as an enhanced dendritic spine elimination in cortex revealed by two-photon imaging. The electroacupuncture regimen that alleviated surgical pain and delirium-like behavior in mice also effectively restored the gut microbiota balance, prevented the microglia activation, and reversed the dendritic spine elimination. These data demonstrated a potentially important gut-brain interactive mechanism underlying the surgical pain-induced delirium in mice. Pending further studies, these findings revealed a possible therapeutic approach in preventing and/or treating postoperative delirium by using perioperative electroacupuncture stimulation in patients.

Keywords: delirium; dendritic spine; electroacupuncture; gut microbiota; mice; microglia; surgical pain.

Figure 1

EA attenuates surgical pain (SP) in mice. (A) Diagram of procedures and behavioral tests. (B) Representative photos of paw deformation induced by SP in the foot. The red arrow indicates paw deformation in surgery mice; green arrows indicate normal paw in sham and SP plus EA mice. (C) Paw deformation counted at 12 h after the surgical incision. Effects of EA on paw deformation counted in mice that underwent foot incision. p-values indicate the difference in the counts between SP vs. sham; SP vs. SP plus EA (DU20+KI1); SP vs. EA (n = 12, Fisher’s exact test). (D) Effects of EA on mechanical withdrawal threshold. p-values indicate the difference in the threshold between SP vs. sham; SP vs. SP plus EA (DU20+KI1); SP plus EA (DU20+DU16) vs. SP plus EA (DU20+KI1) (n = 12, two-way ANOVA followed by Bonferroni post-hoc analyses). (E) Effects of EA on thermal paw withdrawal latency. p-values indicate the difference in the latency between SP vs. sham; SP vs. SP plus EA (DU20+KI1) (n = 12, two-way ANOVA followed by Bonferroni post-hoc analyses). (F) Effects of EA on mechanical withdrawal threshold. p-values indicate the difference in the threshold between SP vs. sham; SP vs. SP plus EA (GB30+ST36); SP plus EA (GB30+ST36) vs. sham (n = 12, two-way ANOVA followed by Bonferroni post-hoc analyses). (G) Effects of EA on thermal paw withdrawal latency. p-values indicate the difference in the latency between SP vs. sham; SP vs. EA (GB30+ST36); EA (GB30+ST36) vs. sham (n = 12, two-way ANOVA followed by Bonferroni post-hoc analyses). Data are represented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. SP, surgical pain; EA, electroacupuncture; SEM, standard error of mean; BL60, DU20, DU16, KI1, GB30, and ST36: acupoint.

Figure 2

EA attenuates the SP-induced delirium-like behavior in mice. (A) Diagram of procedures and behavioral tests. (B) Effects of EA and SP on latency to the center (left panel), freezing time (middle panel), and the time spent in the center (right panel) of the open field test. p-values indicate the difference in these changes between SP vs. sham; SP vs. SP plus EA (DU20+KI1); SP vs. EA (DU20+KI1) (n = 12, one-way ANOVA followed by Bonferroni post-hoc analyses). (C) The carton of the Y maze test (left panel). The effects of EA and SP on entries in the novel arm (middle panel) and duration in the novel arm (right panel) of Y maze test in mice at 9 h after the surgery. p-values indicate the difference in these changes between SP vs. sham; SP vs. SP plus EA (DU20+KI1); SP vs. EA (DU20+KI1) (n = 12, one-way ANOVA followed by Bonferroni post-hoc analyses). (D) The picture of the buried food test (left panel). The effects of EA and SP on the latency to eat food in the mice at 9 h after the surgery (right panel). The p-values indicate the difference in the latency between SP vs. sham, SP vs. SP plus EA (DU20+KI1), SP vs. EA (DU20+KI1) (n = 12, one-way ANOVA followed by Bonferroni post-hoc analyses). (E) The effects of EA and SP on composite Z score in mice at 9 h after the surgery. EA (DU20+KI1) attenuated the SP-induced delirium-like behavior in the mice. p-values indicate the difference in the composite Z score between SP vs. sham; SP vs. SP plus EA (DU20+KI1); SP vs. EA (DU20+KI1) (n = 12, one-way ANOVA followed by Bonferroni post-hoc analyses). Data are represented as mean ± SEM. *p < 0.05, **p < 0.01. SP, surgical pain; EA, electroacupuncture; SEM, standard error of mean; BL60, DU20, DU16, KI1: acupoint.

Figure 3

EA (DU20+KI1) attenuates the SP-induced changes in gut microbiota in mice. (A) Log-scaled percentage heatmap of phylum level. (B-G) SP increased the abundance of Verrucomicrobia at the phylum, family, and class levels. EA mitigated such increases. The p-values indicate the difference in the abundance between SP and SP plus EA (n = 3, one-way ANOVA followed by Bonferroni post-hoc analyses). (H) SP increased the abundance of Cyanobacteria at the phylum level indicated in (E). EA mitigated such increases. The p-values indicate the difference in the abundance between SP and SP plus EA (n = 3, one-way ANOVA followed by Bonferroni post-hoc analyses). (I–L) SP increased the abundance of Lactobacillales at the order level. EA mitigated such increases. The p-values indicate the difference in the abundance between SP and SP plus EA (n = 3, one-way ANOVA followed by Bonferroni post-hoc analyses). (J, K, M, N) The taxonomic composition distribution in samples at genus and species level. SP increased the abundance of Akkermansia at the genus level. EA mitigated such increases. The p-values indicate the difference in the abundance between SP and SP plus EA (n = 3, one-way ANOVA followed by Bonferroni post-hoc analyses). Data are represented as mean ± SEM. *p < 0.05. SP, surgical pain; EA, electroacupuncture; SEM, standard error of mean.

Figure 4

EA (DU20+KI1) attenuates the SP-induced increases in microglia density in the spinal dorsal horn, hippocampus, and somatosensory cortex in mice. (A–C) Representative Iba1 staining of sham, SP, SP plus EA, and EA mice were taken at 4× (scale bar represents 500 µm); the boxed regions were zoomed-in (20×) with scale bar representing 50 µm. (D) Spinal cord microglia density significantly increased in mice that underwent plantar surgical incision. p-values indicate the difference in the microglia density between SP vs. sham, SP vs. SP plus EA, and SP vs. EA (n = 4, one-way ANOVA followed by Bonferroni post-hoc analyses). Data are represented as mean ± SEM. *p < 0.05, **p < 0.01. (E) Microglia density upregulated in the hippocampus of mice that underwent plantar surgical incision. p-values indicate the difference in the microglia density between SP vs. sham, SP vs. SP plus EA, and SP vs. EA (n = 4, one-way ANOVA followed by Bonferroni post-hoc analyses). (F) Somatosensory cortex microglia density significantly increased in mice that underwent plantar surgical incision. The p-values indicate the difference in the microglia density between SP vs. sham, SP vs. SP plus EA, and SP vs. EA (n = 4, one-way ANOVA followed by Bonferroni post-hoc analyses). SP, surgical pain; EA, electroacupuncture; SEM, standard error of mean.

Figure 5

EA (DU20+KI1) attenuates the SP-induced changes in the formation and elimination of the cortical dendritic spine in mice 9 h after surgical incision. (A) Diagram of transcranial two-photon imaging workflow. (B) Left two panels, representative two-photon images of same pyramidal dendrites acquired at baseline and 9 h after surgical incision individually. Right three panels, repeated imaging of the boxed dendritic segment after surgical incision and EA stimulation with a higher-magnification view of the boxed dendritic segment at baseline, 6 h, and 9 h after the surgical incision. Arrowheads indicate dendritic spines dynamics at the indicated time points. Yellow arrowheads represent spines that remained in situ; red arrowheads represent spine elimination. (C) SP did not cause significant differences in the dendritic spines formation rate in the somatosensory cortex of mice compared to sham condition at 6 or 9 h after the procedure (n = 4, two-way ANOVA followed by Bonferroni post-hoc analyses). (D) SP significantly increased the dendritic spines elimination rate in the somatosensory cortex of mice compared to sham condition at 6 or 9 h after the procedure (n = 4, two-way ANOVA followed by Bonferroni post-hoc analyses). p-values indicate the difference in spine elimination rate between SP vs. sham, SP vs. SP plus EA, and SP vs. EA (n = 4, two-way ANOVA followed by Bonferroni post-hoc analyses. (E) Lower panel, mouse head restrained for transcranial two-photon imaging; upper panel, representative images of cortical dendrites and pyramidal neurons at different depths of the S1HL cortex in YFP mouse. Scale bars represent 50 µm. (F) Schematic of the mouse brain and transcranial window for two-photon imaging. (G) Representative YFP mouse brain slice taken at 4×. Scale bar represents 500 µm. All images were scanned with a green filter. Data are represented as mean ± SEM. *p < 0.05. SP, surgical pain; EA, electroacupuncture; SEM, standard error of mean.

Figure 6

Hypothesized pathway of surgical pain-induced delirium-like behavior. Surgical incision in the foot causes pain and delirium-like behavior in mice; changes in gut microbiota; microglia activation at spinal dorsal horn, hippocampus, and somatosensory cortex in mice; and remodeling of the cortical dendritic spine in mice. EA stimulation at specific acupoints DU20+KI1 ameliorated these changes. Mechanistically, we hypothesize that surgical pain increased microglia activation and dendritic spine remodeling, which are mediated by gut microbiota dysbiosis, leading to delirium-like behavior in mice.