Altered transcriptomes, cell type proportions, and dendritic spine morphology in hippocampus of suicide deaths

Abstract

Suicide is a condition resulting from complex environmental and genetic risks that affect millions of people globally. Both structural and functional studies identified the hippocampus as one of the vulnerable brain regions contributing to suicide risk. Here, we have identified the hippocampal transcriptomes, gene ontology, cell type proportions, dendritic spine morphology, and transcriptomic signature in iPSC-derived neuronal precursor cells (NPCs) and neurons in postmortem brain tissue from suicide deaths. The hippocampal tissue transcriptomic data revealed that NPAS4 gene expression was downregulated while ALDH1A2, NAAA, and MLXIPL gene expressions were upregulated in tissue from suicide deaths. The gene ontology identified 29 significant pathways including NPAS4-associated gene ontology terms "excitatory post-synaptic potential", "regulation of postsynaptic membrane potential" and "long-term memory" indicating alteration of glutamatergic synapses in the hippocampus of suicide deaths. The cell type deconvolution identified decreased excitatory neuron proportion and an increased inhibitory neuron proportion providing evidence of excitation/inhibition imbalance in the hippocampus of suicide deaths. In addition, suicide deaths had increased dendric spine density, due to an increase of thin (relatively unstable) dendritic spines, compared to controls. The transcriptomes of iPSC-derived hippocampal-like NPCs and neurons revealed 31 and 33 differentially expressed genes in NPC and neurons, respectively, of suicide deaths. The suicide-associated differentially expressed genes in NPCs were RELN, CRH, EMX2, OXTR, PARM1 and IFITM2 which overlapped with previously published results. The previously-known suicide-associated differentially expressed genes in differentiated neurons were COL1A1, THBS1, IFITM2, AQP1, and NLRP2. Together, these findings would help better understand the hippocampal neurobiology of suicide for identifying therapeutic targets to prevent suicide.

Figure 1:

Differentially expressed genes (DEGs) in hippocampal CA1 region of suicide deaths. The DEGs were analyzed through DESeq2. The volcano-plot shows NPAS4, ALDH1A2, NAAA and MLXIPL as DEGs in suicide deaths (CON vs Sui, FDR≤0.05, p≤0.05) [A]. The NPAS4 gene expression was significantly downregulated in suicide deaths (FC=5.81, p=8.98E-06) [B] while ALDH1A2 (FC=4.16, p=5.59E-07) [C], MLXIPL (FC=1.59, p=8.39E-06) [D] and NAAA (FC=1.32, p=4.72E-06) [E] gene expression was significantly upregulated in suicide group. qPCR validated the decreased relative gene expression of NPAS4 in the suicide deaths [F]. CON=control group; Sui=suicide group; FC=fold change; FDR=false discovery rate; **, p<0.01.

Figure 2:

Estimated cell type proportions in hippocampal CA1 of suicide deaths. The cell type deconvolution was performed using Bisque with the raw bulk RNA-Seq data with published single-nucleus RNA-Seq data from human hippocampus as a reference. Linear regression analysis revealed that the estimated excitatory neuron proportion [A] was significantly higher while inhibitory neuron proportion [B] was significantly lower in suicide deaths (Sui) compared to controls (CON). The estimated cell type proportion for astrocytes [C], oligodendrocytes [D], OPC [E], mural cells [F], and microglia/T-cell [G] was not significantly different between controls and suicide deaths. OPC=Oligodendrocyte Progenitor Cells, CON=Controls, Sui=Suicide deaths. n=28 for CON, n=22 for Sui. *, p<0.05; **, p<0.01.

Figure 3:

Dendritic spine morphology of hippocampal CA1 pyramidal neurons of suicide deaths. The 3D deconvolved apical (upper panel) and basal (lower panel) dendritic segments from control and suicide deaths were modeled and analyzed using Neurolucida 360 [A]. Both apical [B] and basal [C] dendritic spine density were significantly higher in suicide deaths compared to controls. The increased apical and basal spine densities were associated with decreased apical spine head width [D] and unaltered basal spine head width [E] in suicide deaths. The increased apical and basal dendritic spine density was due to an increase of thin dendritic spines in both apical [F] and basal [G] dendrites. The stubby and mushroom spines were not significantly different in suicide deaths compared to controls. Horizontal lines=Least Square Mean (LSM), CON=control group; Sui=suicide group; n=number of neurons (2–3 dendritic segments/neuron). Scale bar = 4 μm. *, p<0.05; **, p<0.01.

Figure 4:

Characterization of iPSCs and iPSC-derived neurons. The skin fibroblast-derived iPSCs were positive for pluripotency markers OCT4, SOX2, SSEA4 and TRA-1–60 [A]. The iPSC-derived neurons were positive for neuronal marker MAP2 [B] and beta-Tubulin-III [E]. The majority (~78%) of cells were positive for glutamatergic neuron marker vGluT1 [C]. The differentiated neurons also expressed synaptic marker PSD95 [D]. About 7–10% cells were positive for GABA [E]. The CA3 marker GRIK4 was positive for ~67% cells [F] whereas ~35% cells also expressed DG marker PROX1 [G].