Fig 2: Recombinant Cbln1 in CeA increases mechanical hypersensitivity in normal animals. (A) Local deletion of GluD1 in the CeA using AAV constructs. Conditional deletion of GluD1 from CeA was achieved using Cre-lox strategy. Mice were injected with with AAV9-hsyn-eGFP or AAV9-hsyn-eGFP-Cre, bilaterally into CeA. Deletion of GluD1 from CeA was confirmed by immunohistochemistry. (B) Deletion of GluD1 from the CeA does not lead to significant change in mechanical hypersensitivity. (C) Impaired averse fear learning in mouse with GluD1 ablation in the CeA. Significant deficits in fear acquisition at 4th CS US (Two-way repeated measures ANOVA, F (1, 10) = 6.31, p = 0.031; Bonferroni’s post hoc test, 4th CS-US: AAV-control (n = 7 mice) vs. AAV-Cre (n = 5 mice): 54.764 ± 4.765 vs. 26.66 ± 6.665, *** p = 0.0008) as well as in retrieval 24 h later (AAV-control vs. AAV-Cre: 40.47 ± 5.282 vs. 11.46 ± 4.619, ** p < 0.005, Unpaired t-test with Welch correction). (D) Mice underwent cannulation surgery for the implantation of bilateral CeA cannulas. Effect of recombinant Cbln1 (250 ng in 0.5 µL) injected into CeA of naïve mice. After measuring basal mechanical hypersensitivity, Cbln1 was administered intracranially into the CeA and mechanical sensitivity (von Frey test) was measured in the right hind paw at 3 h, 6 h, 24 h, and 1 week. Significant increase in mechanical sensitivity is observed from 3 h to 24 h. (Two-way repeated measures ANOVA, treatment F (3, 15) = 8.77, p = 0.0009; Bonferroni’s post hoc test, WT-vehicle (n = 4 mice) vs. WT- Cbln1 (n = 5 mice): 3 h, 5.371 ± 0.223 vs. 1.553 ± 0.110, **** p < 0.0001; 6 h, 5.549 ± 0.253 vs. 1.418 ± 0.114, **** p < 0.0001; 24 h, 6.347 ± 0.341 vs. 2.088 ± 0.217, **** p < 0.0001). No change was seen in GluD1 KO mice after Cbln1 administration (n = vehicle (5 mice), Cbln1 (5 mice)). (E) Recombinant Cbln1 led to a significant reduction of perisomatic GluD1 at 6 h timepoint predominantly in right CeA compared to vehicle (PBS) injected mice. (Two-way ANOVA treatment F (1, 10) = 19.45, p = 0.0013; Bonferroni’s post hoc test, WT-vehicle (3 mice) vs. WT-Cbln1 (4 mice): Right CeA: 29.74 ± 3.08 vs. 8.72 ± 1.57, ** p = 0.0048; Left CeA: 24.31 ± 3.91 vs. 12.82 ± 4.89, p = 0.10).

Fig 3: Recombinant Cbln1 rescues hypersensitivity and averse-affective behaviors in a neuropathic pain model. (A) Schematic representation of SNL model. The left L5 spinal nerve was ligated. Cbln1 (500 ng in 1 µL) or PBS was injected into right CeA of SNL rat. (B) Intra-CeA Cbln1 showed reduction in noxious stimuli induced audible vocalization in SNL rats; n = Sham-vehicle (6 mice), Sham-Cbln1 (5 mice), SNL-vehicle (7 mice), SNL-Cbln1 (7 mice) (Two-way ANOVA treatment F (3, 103) = 120, p < 0.0001; Bonferroni’s multiple comparison test, SNL-vehicle vs. SNL-Cbln1: Day 1: 6.948 ± 0.793 vs. 4.273 ± 0.410, ** p = 0.0012; Day 2: 6.711 ± 0.796 vs. 4.076 ± 0.299, ** p = 0.0015; Day 3: 6.853 ± 0.606 vs. 4.226 ± 0.251, ** p = 0.0016; Day 7: 6.973 ± 0.712 vs. 4.604 ± 0.202, * p = 0.0126). (B’) Injection of recombinant Cbln1 into the right CeA reduced audible vocalization to innocuous stimuli in SNL rats (Two-way ANOVA treatment F (3, 103) = 204.5, p < 0.0001; Bonferroni’s multiple comparison test, SNL-vehicle vs. SNL-Cbln1: Day 2: 2.130 ± 0.171 vs. 1.497 ± 0.068, * p = 0.0284; Day 7: 2.258 ± 0.191 vs. 1.343 ± 0.145, ** p = 0.0015; n = 5–7 rats per group). (C) Intra-CeA Cbln1 reduced ultrasonic vocalization in SNL rats (Two-way ANOVA treatment F (3, 103) = 119.3, p < 0.0001; Bonferroni’s multiple comparisons test, SNL-vehicle vs. SNL-Cbln1: Day 2: 5.845 ± 0.765 vs. 3.974 ± 0.351, * p = 0.0323; Day 3: 6.213 ± 0.868 vs. 3.907 ± 0.199, ** p = 0.0041; Day 7: 6.221 ± 0.709 vs. 4.211 ± 0.212, * p = 0.0340; n = 5–7 rats per group). (C’) Ultrasonic vocalization to innocuous stimuli (Two-way ANOVA treatment F (3, 103) = 172.1, p < 0.0001; Bonferroni’s multiple comparison test, SNL-vehicle vs. SNL-Cbln1: Day 3: 2.070 ± 0.226 vs. 1.515 ± 0.073, * p = 0.0408; Day 7: 1.968 ± 0.143 vs. 1.333 ± 0.115, * p = 0.0256; n = 5–7 rats per group). (D) Reduction in mechanical thresholds in the von Frey test were mitigated by intra-CeA Cbln1 (Two-way ANOVA treatment F (3, 103) = 147.1, p < 0.0001; Bonferroni’s multiple comparison test, SNL-vehicle vs. SNL-Cbln1: Day 2: 3.781 ± 0.615 vs. 7.001 ± 1.111, * p = 0.0498; n = 5–7 rats per group). (E) Open arm entries in the elevated plus maze test were also increased by Cbln1 in SNL rats suggesting anxiolytic effect in neuropathic pain (Two-way ANOVA treatment F (3, 40) = 18.36, p < 0.0001; Bonferroni’s multiple comparison test, SNL-vehicle vs. SNL-Cbln1: Day 7: 14.078 ± 4.842 vs. 31.997 ± 6.443, * p = 0.0399; n = 5–7 rats per group). (F) Recombinant Cbln1 restored downregulated GluD1 in CeLC in SNL rats. Immunohistochemical analysis of perisomatic GluD1 levels in right CeA was performed in tissue from SNL rats injected with PBS or recombinant Cbln1. Recombinant Cbln1 restored GluD1 levels in SNL rats to the levels of sham rats. (One-way ANOVA F (2, 9) = 7.49, p = 0.012; Tukey’s post hoc test, Sham-vehicle (n = 4 rats): 9.54 ± 1.57 vs. SNL-vehicle (n = 4 rats): 4.53 ± 1.08, * p = 0.035; SNL-Vehicle vs. SNL-Cbln1 (n = 4 rats): 10.51 ± 0.69; * p = 0.014). (G) Recombinant Cbln1 normalized surface GluA1 levels in SNL rats. Immunohistochemical analysis was performed in SNL rats injected with either PBS or recombinant Cbln1 for AMPA levels. Cbln1 injection reduced the upregulated GluA1 expression in SNL rats. SNL (One-way ANOVA F (2, 8) = 27.17, p = 0.0003; Sham-vehicle (n = 4 rats): 58.35 ± 7.51 vs. SNL-vehicle (n = 4 rats): 110.92 ± 3.70, **** p < 0.0001; SNL-vehicle vs. SNL-Cbln1 (n = 4 rats): 91.48 ± 2.40, * p = 0.049).

Fig 4: Recombinant Cbln1 administration normalizes synaptic dysfunction in the CeA in inflammatory pain. (A) Intracerebroventricular administration of Cbln1 (1.5 µg in 1.5 µL PBS) 48 h after CFA also attenuated mechanical hypersensitivity for at least one week. Two-way repeated measures ANOVA, F (3, 14) = 12.81, p = 0.0001, Bonferroni’s post hoc test, WT CFA-vehicle (n = 4 mice) vs. WT CFA-Cbln1 (n = 6 mice), ** p = 0.0051, **** p < 0.0001. No effect of Cbln1 was observed in GluD1 KO (n = 4 mice (vehicle), 4 mice (Cbln1)). (B) No change in mechanical threshold in von Frey analysis in control (non-CFA injected) paw. (C) Immunohistochemical analysis of right CeLC for the effect of ICV administration of recombinant Cbln1 on perisomatic GluD1 volume in the CFA pain model. Recombinant Cbln1 restored GluD1 levels in CFA mice compared to mice injected with PBS (One-way ANOVA, treatment F (2, 9) = 14.38, p = 0.009; Bonferroni’s multiple comparison; Saline-vehicle (n = 4 mice): 27.3 ± 1.53, vs. CFA-vehicle (n = 4 mice): 10.87 ± 1.94, ** p = 0.068; CFA-vehicle vs. CFA-Cbln1 (n = 4 mice): 30.3 ± 4.08, ** p = 0.0023). (D) Recombinant Cbln1 normalized surface GluA1 upregulation in the right CeLC in CFA pain model (One-way ANOVA treatment F (2, 9) = 9.23 p = 0.006; Bonferroni’s multiple comparison; Saline-vehicle (n = 4 mice) vs. CFA-vehicle (n = 4 mice): 112.9 ± 7.07 vs. 169.2 ± 6.83, * p = 0.013 and vs. CFA-Cbln1 (n = 4 mice): 114.2 ± 15.4, * p = 0.015).

Fig 5: GluD1 is critical for regulating excitatory neurotransmission at parabrachio-amygdala synapses. (A) Whole-cell mEPSC recordings at -70mV (in the presence of picrotoxin and tetrodotoxin) from CeC neurons demonstrate significantly lower mEPSC frequency and amplitude in GluD1 KO (WT (n = 8) vs. GluD1 KO (n = 9): Frequency: 6.17 ± 0.83 Hz vs. 2.91 ± 0.96 Hz, * p = 0.022, two-tailed unpaired t-test with Welch’s correction; Amplitude: 16.78 ± 0.869 pA vs. 13.66 ± 1.124 pA, * p = 0.045, unpaired t-test). (B) No change in mIPSC frequency and amplitude in CeC due to deletion of GluD1 (Frequency: p = 0.74, Amplitude: p = 0.77; two-tailed unpaired t-test, n = 8). (C) Whole-cell evoked EPSCs from CeC neurons at -70 mV in the presence of picrotoxin with recording electrode in CeC (yellow asterisk) and stimulating electrode on PB fibers (red asterisk). The input–output curve shows significant reduction in evoked EPSC amplitude in GluD1 KO mice. (Two-way repeated measures ANOVA, genotype F (1, 28) = 23.03, p < 0.0001; Bonferroni’s multiple comparisons test: WT (n = 15) vs. GluD1 KO (n = 25); 800 µA current injection: 48.97 ± 7.39 vs. 17.16 ± 1.43, **** p < 0.0001; 1 mA current injection: 47.00 ± 6.40 vs. 24.46 ± 3.56, **** p < 0.0001).