A bipolar disorder-associated missense variant alters adenylyl cyclase 2 activity and promotes mania-like behavior

Abstract

The single nucleotide polymorphism rs13166360, causing a substitution of valine (Val) 147 to leucine (Leu) in the adenylyl cyclase 2 (ADCY2), has previously been associated with bipolar disorder (BD). Here we show that the disease-associated ADCY2 missense mutation diminishes the enzyme´s capacity to generate the second messenger 3',5'-cylic adenosine monophosphate (cAMP) by altering its subcellular localization. We established mice specifically carrying the Val to Leu substitution using CRISPR/Cas9-based gene editing. Mice homozygous for the Leu variant display symptoms of a mania-like state accompanied by cognitive impairments. Mutant animals show additional characteristic signs of rodent mania models, i.e., they are hypersensitive to amphetamine, the observed mania-like behaviors are responsive to lithium treatment and the Val to Leu substitution results in a shifted excitatory/inhibitory synaptic balance towards more excitation. Exposure to chronic social defeat stress switches homozygous Leu variant carriers from a mania- to a depressive-like state, a transition which is reminiscent of the alternations characterizing the symptomatology in BD patients. Single-cell RNA-seq (scRNA-seq) revealed widespread Adcy2 mRNA expression in numerous hippocampal cell types. Differentially expressed genes particularly identified from glutamatergic CA1 neurons point towards ADCY2 variant-dependent alterations in multiple biological processes including cAMP-related signaling pathways. These results validate ADCY2 as a BD risk gene, provide insights into underlying disease mechanisms, and potentially open novel avenues for therapeutic intervention strategies.

Fig. 1. Consequences of Val147Leu substitution on ADCY2 activity and subcellular localization.

a Organization of human ADCY2 gene and localization of BD-associated SNPs rs17826816 and rs13166360. b rs13166360 causes a Val147 Leu substitution in the 4^th transmembrane helix (TM4) of the first transmembrane domain (M1). c Alignment of TM4 and adjacent regions of human and murine ADCY2 with human membrane bound ADCY paralogues. d Quantification and representative Western blot showing expression levels of ADCY2-V151, ADCY2-L151 and ADCY2-N1030A/R1034S in transiently transfected HEK293 cells (n = 3). e Forskolin (FSK) concentration-dependent stimulation of cAMP production by ADCY2 variants compared to the catalytically inactive variant ADCY2-N1030A/R1034S (n = 3; two-way ANOVA, ADCY2 variants: F_(1,14) = 127.6, ^****p < 0.0001, concentration: F_(6,14) = 207.9, p < 0.0001, variants × concentration: F_(6,14) = 20.18, p < 0.0001). f Time course of cAMP production of ADCY2 variants at baseline and following repeated FSK stimulation (n = 3; two-way ANOVA repeated measures, ADCY2 variants: F_(2,3) = 98.37, ^**p = 0.0018, time: F_(2,7) = 301^, p < 0.0001, variants × time: F_(12,18) = 22.17, p < 0.0001). Quantification of co-localization of HA-tagged ADCY2-V151 and HA-tagged ADCY2-L151 with: (g) FLAG-tagged ADCY2-L151, (h) RFP-tagged Rab5, GFP-tagged Rab7, Rab9, RAB11, (i) endogenous Rab5 and (j) phalloidin using Mander’s co-localization co-efficient. In all comparisons, unpaired t test, ^****p < 0.0001. Each dot represents an individual cell. All data are presented as mean ± s.e.m..

Fig. 2. Generation of Adcy2^V151L mice.

a Organization of the murine Adcy2 gene. b CRISPR/Cas9-based strategy targeting exon 3 using sgRNAs sgAdcy2-a and sgAdcy2-b in combination with a ssODN to replace Val by Leu at position 151. c Representative electropherograms depicting possible genotypes of offspring from heterozygous breedings. d Quantification of Adcy2 mRNA expression in the cortex of WT and L151 mice (n = 5 each) by RT-qPCR. e Representative electropherograms of Adcy2 cDNA from WT and L151 mice covering the transition from exon 2 (E2) to exon 3 (E3). Highlighted are the AluI restriction site and the Val151Leu substitution present in L151 mice.

Fig. 3. Behavioral assessment of L151 mice reveals signs of mania-like behavior and cognitive deficits.

a Locomotor activity during adaptation to a novel home cage (WT n = 8, L151 n = 9; two-way ANOVA repeated measures, 1^st dark phase: genotype: F_(1,15) = 4.748, ^*p = 0.0457, time: F_{(2, 32)} = 5.468, p = 0.006, genotype × time: F_(9,120) = 0.8536, p = 0.5687; 2nd dark phase: genotype: F_(1,14) = 5.215, ^*p = 0.0385, time: F_(5,75) = 12.54, p < 0.0001, genotype × time: F_(11,154) = 0.9820, p = 0.4652). b Distance traveled, number of entries into (unpaired t test, ^*p = 0.0199⁾ and time spent in the inner zone (unpaired t test, ^*p = 0.035) of the open field test. c Time spent in and number of entries into the lit zone of the dark-light box (unpaired t test, ^*p = 0.0344⁾. d Exploration of the object (unpaired t test, ^*p = 0.0227) and object zone entries (unpaired t test, ^Tp = 0.087) in the novel object exploration test (b–d: WT n = 16, L151 n = 13). e Time spent floating and swimming in the forced swim test (WT n = 16, L151 = 12; unpaired t test, floating and swimming time: p < 0.0001 for both). f Latency to find the platform in the Morris water maze test (WT n = 10, L151 n = 9; two-way ANOVA repeated measures, genotype: ^***p = 0.0003, F_(1,17) ⁼ 20.33, Bonferroni multiple comparisons test: day 2: ^*p = 0.0478, day 3: ^*p = 0.0187) and time spent in the target quadrant during memory retrieval (unpaired t test, ^*p = 0.0325). g LTP at CA3-CA1 synapses in dorsal hippocampal slices from WT and L151 mice (n = 17 slices from 5 WT mice and 18 slices from 5 L151 mice; unpaired t test, ^*p = 0.0298). Representative recording traces depict fEPSPs before and 60 min after LTP induction. All data are presented as mean ± s.e.m..

Fig. 4. L151 mice show amphetamine hypersensitivity, lithium chloride responsiveness and a shift towards more synaptic excitation.

a Locomotor activity following amphetamine treatment (WT n = 11, L151 n = 9; one-way ANOVA repeated measures, genotype: F_(3,7) = 8.1386, ^**p = 0.0014). b Assessment of dopamine content in the prefrontal cortex (PFC), striatum (STR), nucleus accumbens (NAc) and hippocampus (HIP) following amphetamine treatment (WT n = 3; L151 n = 3; two-way ANOVA; PFC: genotype: F_(1,21) = 6.167, p = 0.0091, treatment: F _(1,21) = 21.63, p < 0.001, genotype × treatment: p = 23.514; STR: genotype: F_(1,21) = 8.914, p = 0.0141, treatment: F_(1,21) = 13.615, p = 0.010, genotype × treatment: F_(1,21) = 21.631, p = 0.017; NAc: genotype: F_(1,22) = 9.762, p = 0.610, treatment: F_(1,22) = 12.218, p = 0.012, genotype × treatment: F_(1,22) = 21.426, p = 0.047; HIP: genotype: F_(1,21) = 10.21, p = 0.0493, treatment: F_(1,21) = 6.514, p = 0.0121, genotype × treatment: F_(1,21) = 11.426, p = 0.0521). c, d Behavioral assessment of WT and L151 mice following 2 weeks of LiCl treatment (WT control n = 8, L151 control n = 10, WT LiCl n = 9, L151 LiCl n = 10). c Time spent in the inner zone of the open field test (two-way ANOVA, genotype: F_(1,31) = 8.641, p = 0.0062, treatment: F_(1,31) = 6.763, p = 0 .0141, genotype × treatment: F_(1,31) = 13.29, p = 0.001). d Time spent floating (two-way ANOVA, genotype: F_(1,33) = 4.530, p = 0.0409, treatment: F_(1,33) = 0.1745, p = 0.6788, genotype × treatment: F_(1,33) = 324.1, p < 0.0001) and swimming (two-way ANOVA, genotype: F_(1,33) = 169.4, p = 0.0409, treatment: F_(1,33) = 0.1745, p = 0.6788, genotype × treatment: F_(1,33) = 169.4, p < 0.000) in the forced swim test. e Representative traces, (f) amplitude and (g) frequency (unpaired t test, p = 0.0108) of patch-clamp recordings in the ventral hippocampus of miniature excitatory postsynaptic currents (mEPSCs; WT: n = 5 mice, 19 cells; L151: n = 4 mice, 22 cells). h Representative traces, (i) amplitude and (j) frequency (unpaired t test, p = 0.0005) of patch-clamp recordings in the ventral hippocampus of miniature inhibitory postsynaptic currents (mIPSCs; WT: n = 4 mice, 37 cells; L151: n = 4 mice, 32 cells). Bonferroni post hoc tests in (b–d): ^Tp = 0.0531, ^*significantly different from WT of the same condition, p < 0.05; ^#significantly different from the basal condition of the same genotype, p < 0.05. All data are presented as mean ± s.e.m..

Fig. 5. L151 mice show a switch in their behavioral state in response to chronic social defeat stress.

a–c Validation of efficiency of chronic social defeat stress (CSDS) paradigm. Adrenal weight (a) (two-way ANOVA, stress: F_(1,45) = 99.06, p < 0.001), thymus weight (b) (two-way ANOVA, stress: F_(1,45) = 76.27, p < 0.001) and social avoidance ratio (c) (two-way ANOVA, stress: F_(1,47) = 43.88, p < 0.0001) of Adcy^V151L mice subjected to the CSDS. Distance traveled (d) (two-way ANOVA, stress: F_(1,49) = 24.33, p < 0.001), time spent in the inner zone (e) (two-way ANOVA, genotype: F_(1,48) = 32.97, p < 0.001, stress: F_(1,48) = 23.88, p < 0.001, genotype × stress: F_(1,48) = 19.02, p < 0.001) and entries to the inner zone (f) (two-way ANOVA, genotype: F_(1,50) = 11.56, p = 0.0013, stress: F_(1,50) = 19.72, p < 0.001, genotype × stress: F_(1,50) = 10.25, p = 0.0024) of the open field. g Time spent in the lit zone of the dark-light box (two-way ANOVA, genotype: F_(1,47) = 0.6375, p = 0.4286, stress: F_(1,47) = 1.142, p = 0.2906, genotype × stress: F_(1,47) = 10.73, p < 0.002). Time spent floating (h) (two-way ANOVA, genotype: F_(1,48) = 2.437, p = 0.1250, stress: F_(1,48) = 39.05, p < 0.001, genotype × stress: F_(1,48) = 34.55, p < 0.001) and swimming (i) in the forced swim test (two-way ANOVA, genotype: F_(1,48) = 2.972, p = 0.0912, stress: F_(1,48) = 33.15, p < 0.001, genotype × stress: F_(1,48) = 40.31, p < 0.0001). Groups: WT basal n = 16-17, L151 basal n = 10–14, WT stressed n = 10–15, L151 stressed n = 9–11. Bonferroni post hoc tests in (a-i): p < 0.05. ^*significantly different from WT of the same condition, p < 0.05; ^#significantly different from the basal condition of the same genotype. All data are presented as mean ± s.e.m..

Fig. 6. Single cell RNA sequencing of the ventral hippocampus of Adcy2^V151L mice.

a Dimensionality reduction Uniform Manifold Approximation and Projection (UMAP) plot showing 33 clusters belonging to 17 main cell types. b Expression of ADCY2 in various cell types of the ventral hippocampus. c Violin plots comparing ADCY2 expression in different cell types of the ventral hippocampus of WT and L151 mice. d Number of differentially expressed genes between L151 and WT cells in various cell types. e Gene ontology functional enrichment analysis using Kyoto Encyclopedia of Gene and Genome (KEGG) biological pathways. f List of significantly enriched KEGG terms.