Igfbp2 Downregulation in PVT-CeA Glutamatergic Circuits Drives Neonatal Anesthesia-Induced Fear Memory Deficits

Abstract

Repeated neonatal general anesthesia results in long-term cognitive dysfunction; however, the underlying mechanisms remain unclear. This study finds that repeated neonatal anesthesia impaired fear memory in adolescent mice of both sexes, along with hypoactivated glutamatergic neurons in the paraventricular thalamus (PVT). Optogenetic activation of PVT glutamatergic neurons rescued fear memory deficits in anesthesia-treated mice, whereas optogenetic inhibition of these neurons recapitulated memory deficits in control mice. Specifically, repeated neonatal anesthesia reduced insulin-like growth factor-binding protein 2 (Igbp2) expression and dendritic spine density in PVT glutamatergic neurons in both males and females. Selectively manipulating PVT glutamatergic Igfbp2 mediated anesthesia-induced fear memory deficits through modulating neuron excitability and spine density. Notably, optogenetic activation or restoring Igfbp2 expression in glutamatergic projections from the PVT to the central amygdala (CeA) blocked anesthesia-induced memory impairment, whereas optogenetic inhibition or knocking down of Igfbp2 expression in these projections is sufficient to engender similar memory impairment in control mice. The findings demonstrate that Igfbp2 in glutamatergic neurons in the PVT afferents to the CeA mediates fear memory deficits caused by repeated neonatal anesthesia in mice of both sexes, highlighting Igfbp2 as a potential therapeutic target for repeated anesthesia-induced cognitive impairment.

Keywords: Insulin‐like growth factor‐binding protein 2; anesthesia; central amygdala; fear memory; paraventricular thalamus.

Figure 1

Repeated neonatal anesthesia induced fear memory deficits in both male and female adolescent mice. A) Schematic of mice undergoing repeated neonatal anesthesia and subsequent behavioral tests. B,D) No significant difference in training performance between control and anesthesia groups in males (B) and females (D). C,E) Anesthesia group showed impaired fear memory retrieval compared to the control group in males (C) and females (E), while baseline freezing levels showed no between‐group difference in either sex. F,G) Representative trajectory plots of males (F) and females (G) in open field test. H,J) Total distance traveled in the open field arena showed no significant difference between control and anesthesia groups in males (H) and females (J). I,K) Time spent in the center region of the open field arena also showed no between‐group difference in males (I) and females (K). L,M) No significant difference in foot shock sensitivity between control and anesthesia groups in males (L) and females (M). Analyzed by two‐way repeated‐measure ANOVA in B and D, and unpaired t test in C, E, H, I, J, K, L and M; n = 8 mice per group; **P < 0.01, ns: not significance. Data are presented as means ± SEM.

Figure 2

Repeated neonatal anesthesia impaired PVT activity following fear memory retrieval in both male and female mice. A) Schematic of labeling thalamic neurons related to fear memory retrieval with tdTomato in TRAP2/Ai9 mice. B,C) Histograms summarize the tdTomato⁺ expression levels in thalamus following fear memory retrieval of males (B) and females (C); n = 4 brain sections from 4 mice per group. D,E) Representative images showing tdTomato⁺ expression in PVT of males (D) and females (E); scale bar: 100 µm. Analyzed by unpaired t test; *P < 0.05, **P < 0.01. Data are presented as means ± SEM.

Figure 3

Ex vivo and in vivo recordings of PVT glutamatergic neurons activity. A) Schematic of labeling PVT glutamatergic neurons with virus and ex vivo whole‐cell recordings in PVT slices. B) Representative images of real‐time recordings of PVT glutamatergic neurons; scale bar: 10 µm. C‐F) Representative traces of AP elicited by 200 pA current injections (C, E) and AP frequency in response to a range of current injections (D, F) in males (C, D) and females (E, F); n = 10‐12 neurons from 4 mice per group. G) Schematic of experimental design and timeline of fiber photometry. H,I) Diagram of virus injection and fiber implantation into the PVT (H) and representative image showing virus expression and fiber site (I); scale bar: 100 µm. J,M) Average calcium‐dependent GCaMP fluorescent signal (top) of PVT glutamatergic neurons and the corresponding heatmaps (down) in males (J) and females (M); duration of the tone is indicated by the red box. K,N) Anesthesia group showed significantly reduced peak ΔF/F of calcium activity in the PVT glutamatergic neurons compared to control group in males (K) and females (N); n = 8 mice per group. L,O) Anesthesia group showed significantly decreased the AUC of calcium activity in the PVT glutamatergic neurons compared to control group in males (L) and females (O); n = 8 mice per group. Analyzed by two‐way repeated‐measure ANOVA in D and F, and unpaired t test in J, K, M and N; **P < 0.01, ***P < 0.001, **** P < 0.0001. Data are presented as means ± SEM.

Figure 4

PVT glutamatergic neuron activity drove fear memory impairment induced by repeated neonatal anesthesia. A) Schematic of experimental design and timeline of optogenetic modulation. B,C) Diagram of virus injection and fiber implantation into the PVT of anesthesia‐treated mice (B) and representative image showing virus expression and fiber site (C); scale bar: 100 µm. D) Representative trace of neuronal firing in ChR2‐expressing neurons evoked by 473‐nm photostimulation at 10 Hz. E,G) No significant difference in training performance between eYFP and ChR2 groups in both males (E) and females (G). F,H) Optogenetic activation of PVT glutamatergic neurons ameliorated fear memory impairment in both males (F) and females (H), while baseline freezing levels showed no significant differences in either sex. I,J) Diagram of virus injection and fiber implantation into the PVT of control mice (I) and representative image showing virus expression and fiber site (J); scale bar: 100 µm. K) Representative trace of neuronal firing in NpHR‐expressing neurons suppressed by 590‐nm laser. L,N) No significant difference in training performance between eYFP and NpHR groups in both males (L) and females (N). M,O) Optogenetic inhibition of PVT glutamatergic neurons induced fear memory impairment in both males (M) and females (O), while baseline freezing levels showed no significant differences in either sex. Analyzed by two‐way repeated‐measure ANOVA in E, G, L and N and unpaired t test in F, H, M and O; n = 8 mice per group; **P < 0.01, ns: not significance. Data are presented as means ± SEM.

Figure 5

Repeated neonatal anesthesia reduced Igfbp2 expression in PVT glutamatergic neurons and decreased spine density in PVT. A,C) Heatmaps showing DEGs between control and anesthesia groups identified by RNA‐seq of PVT tissues from males (A) and females (C). Normalized Z score values (high, red; low, blue) were calculated for each DEG (row). B,D) Volcano plots illustrating the distribution of upregulated (red), downregulated (blue), and unchanged (gray) DEGs in PVT tissues in males (B) and females (D); each dot represents one gene. E,G) RT‐qPCR of up‐regulated DEGs in males (E) and females (G); n = 4 mice per group. F,H) RT‐qPCR of down‐regulated DEGs in males (F) and females (H); n = 4 mice per group. I‐L) Representative western blot bands (I, K) and quantification analysis of Lcn2, Vwf, Igfbp2 and Hspa1b expression in PVT tissues (J, L) between control and anesthesia groups in males (I, J) and females (K, L); n = 4 mice per group. M‐P) Representative images (M, O) and quantification analysis of the percentage of Igfbp2‐expressing glutamatergic neurons in PVT (N, P) between control and anesthesia groups in males (M, N) and females (O, P); scale bar: 100 µm; n = 4 brain sections from 4 mice per group. Q‐T) Representative Golgi‐Cox staining images (Q, S) and quantification analysis of dendritic spine numbers per 10 µm in PVT (R, T) between control and anesthesia groups in males (Q, R) and females (S, T); scale bar: 5 µm; n=12 brain sections from 4 mice per group. Analyzed by unpaired t test; *P < 0.05, **P < 0.01, ***P < 0.01, ns: not significance. Data are presented as means ± SEM.

Figure 6

Overexpression of Igfbp2 in PVT glutamatergic neurons rescued fear memory deficits, neuronal excitability impairments, and spine loss induced by repeated neonatal anesthesia. A) Schematic of experimental design and timeline of Igfbp2 overexpression in PVT glutamatergic neurons. B,C) Diagram of virus injection into PVT of anesthesia‐treated mice (B) and representative image showing virus expression (C); scale bar: 100 µm. D,E) Representative western blot bands (D) and quantification analysis of Igfbp2 in PVT (E) between AAV‐eYFP and AAV‐Igfbp2 groups; n = 4 mice per group. F,H) No significant difference in training performance between AAV‐eYFP and AAV‐Igfbp2 groups in males (F) and females (H); n = 8 mice per group. G,I) Restoration of Igfbp2 in PVT glutamatergic neurons rescued fear memory impairment in males (G) and females (I), while baseline freezing levels showed no significant differences in either sex; n = 8 mice per group. J‐M) Representative traces of AP elicited by 200 pA current injections (J, L), and AP frequency in response to a range of current injections (K, M) between AAV‐eYFP and AAV‐Igfbp2 groups in males (J, K) and females (L, M); n = 10 neurons from 4 mice per group. N‐Q) Representative Golgi‐Cox staining images (N, P) and quantification analysis of dendritic spine numbers per 10 µm in PVT (O, Q) between AAV‐eYFP and AAV‐Igfbp2 groups in males (N, O) and females (P, Q); scale bar: 5 µm; n = 12 brain sections from 4 mice per group. Analyzed by two‐way repeated‐measure ANOVA in F and H, and unpaired t test in G, I, K, M, O and Q; *P < 0.05, **P < 0.01, ns: not significance. Data are presented as means ± SEM.

Figure 7

Knockdown of Igfbp2 in PVT glutamatergic neurons impaired fear memory retrieval, neuronal excitability and spine density. A) Schematic of experimental design and timeline of Igfbp2 knockdown in PVT glutamatergic neurons. B,C) Diagram of virus injection into PVT of control mice (B) and representative image showing virus expression (C); scale bar: 100 µm. D,E) Representative western blot bands (D) and quantification analysis of Igfbp2 in PVT (E) between AAV‐scramble and AAV‐shRNA groups; n = 4 mice per group. F,H) No significant difference in training performance between AAV‐scramble and AAV‐shRNA groups in males (F) and females (H); n = 8 mice per group. G,I) Knockdown of Igfbp2 in PVT glutamatergic neurons induced fear memory impairment in males (G) and females (I), while baseline freezing levels showed no significant differences in either sex; n = 8 mice per group. J‐M) Representative traces of AP elicited by 200 pA current injections (J, L), and AP frequency in response to a range of current injections (K, M) between AAV‐scramble and AAV‐shRNA groups in males (J, K) and females (L, M); n =10 neurons from 4 mice per group. N‐Q) Representative Golgi‐Cox staining images (N, P) and quantification of dendritic spine numbers per 10 mm in PVT (O, Q) from AAV‐scramble and AAV‐shRNA groups in males (N, O) and females (P, Q); scale bar: 5 µm; n = 12 brain sections from 4 mice per group. Analyzed by two‐way repeated‐measure ANOVA in F and H, and unpaired t test in G, I, K, M, O and Q; *P < 0.05, **P < 0.01, ns: not significance. Data are presented as means ± SEM.

Figure 8

PVT‐CeA glutamatergic projections contributed to fear memory impairment induced by repeated neonatal anesthesia. A,B) Schematic of anterograde virus tracing (A) and representative images showing virus injection site within PVT (left) and virus expression in CeA (right) (B); scale bars: 100 µm. C,D) Schematic of retrograde virus tracing (C) and representative images of virus injection site within CeA (left) and virus expression in PVT (right) (D); scale bars: 100 µm. E) Schematic of experimental design and timeline of optogenetic modulation. F,G) Diagram of virus injection and fiber implantation (F) and representative image showing virus expression and fiber site in CeA (G); scale bars: 100 µm. H,J) No significant difference in training performance between eYFP and ChR2 groups in anesthesia‐treated males (H) and females (J). I,K) Optogenetic activation of PVT‐CeA glutamatergic projections ameliorated fear memory impairment in anesthesia‐treated males (I) and females (K), while baseline freezing levels showed no significant differences in either sex. L,N) No significant difference in training performance between eYFP and NpHR groups in control males (L) and females (N). M,O) Optogenetic inhibition of PVT‐CeA glutamatergic projections induced fear memory deficits in control males (M) and females (O), while baseline freezing levels showed no significant differences in either sex. Analyzed by two‐way repeated‐measure ANOVA in H, I, L and N, and unpaired t test in I, K, M and O; n = 8 mice per group; **P < 0.01, ***P < 0.001, ns: not significance. Data are presented as means ± SEM.

Figure 9

Igfbp2 in glutamatergic neurons of PVT projecting to CeA modulated fear memory impairment induced by repeated neonatal anesthesia. A) Schematic of experimental design and timeline of specific modulation of Igfbp2 in PVT glutamatergic neurons projecting to CeA. B,C) Diagram of virus injection into PVT (B) and representative image showing virus expression (C); scale bars: 100 µm. D,F) No significant difference in training performance between AAV‐eYFP and AAV‐Igfbp2 groups in anesthesia‐treated males (D) and females (F). E,G) Restoration of Igfbp2 expression in PVT glutamatergic neurons projecting to CeA rescued fear memory impairment in anesthesia‐treated males (E) and females (G), while baseline freezing levels showed no significant difference in either sex. H,J) No significant difference in training performance between AAV‐scramble and AAV‐shRNA groups in control males (H) and females (J). I,K) Knockdown of Igfbp2 in PVT glutamatergic neurons projecting to CeA induced fear memory deficits in control males (I) and females (K), while baseline freezing levels showed no significant differences in either sex. Analyzed by two‐way repeated‐measure ANOVA in D, F, H and J, and unpaired t test in E, G, I and K; n = 8 mice per group; **P < 0.01, ***P < 0.001, ns: not significance. Data are presented as means ± SEM.

Figure 10

Schematic summary of this study. Repeated neonatal anesthesia impaired fear memory in adolescent mice of both sexes, accompanied by reduced PVT glutamatergic neuron activity and downregulated Igfbp2 expression. Manipulating Igfbp2 expression in PVT glutamatergic neurons mediated anesthesia‐induced fear memory deficits by modulating neuronal excitability and dendritic spine density. Selectively optogenetic activation or restoration of Igfbp2 expression in PVT‐CeA glutamatergic projections ameliorated anesthesia‐induced memory deficits, while optogenetic inhibition or Igfbp2 knockdown in these projections recapitulated memory deficits in control mice. PVT paraventricular thalamus, CeA central amygdala, Igfbp2 Insulin‐like growth factor‐binding protein 2.