Spatial transcriptomic analysis of adult hippocampal neurogenesis in the human brain

Abstract

Background: Adult hippocampal neurogenesis has been extensively characterized in rodent models, but its existence in humans remains controversial. We sought to assess the phenomenon in postmortem human hippocampal samples by combining spatial transcriptomics and multiplexed fluorescent in situ hybridization.

Methods: We computationally examined the spatial expression of various canonical neurogenesis markers in postmortem dentate gyrus (DG) sections from young and middle-aged sudden-death males. We conducted in situ assessment of markers expressed in neural stem cells, proliferative cells, and immature granule neurons in postmortem DG sections from infant, adolescent, and middle-aged males.

Results: We examined frozen DG tissue from infant (n = 1, age 2 yr), adolescent (n = 1, age 16 yr), young adult (n = 2, mean age 23.5 yr), and middle-aged (n = 2, mean age 42.5 yr) males, and frozen-fixed DG tissue from middle-aged males (n = 6, mean age 43.5 yr). We detected very few cells expressing neural stem cell and proliferative markers in the human DG from childhood to middle age. However, at all ages, we observed a substantial number of DG cells expressing the immature neuronal marker DCX. Most DCX ⁺ cells displayed an inhibitory phenotype, while the remainder were non-committed or excitatory in nature.

Limitations: The study was limited by small sample sizes and included samples only from males.

Conclusion: Our findings indicate very low levels of hippocampal neurogenesis throughout life and the existence of a local reserve of plasticity in the adult human hippocampus. Overall, our study provides important insight into the distribution and phenotype of cells expressing neurogenesis markers in the adult human hippocampus.

Figure 1

Schematic overview of the experimental workflow used for the spatial transcriptomic analysis of 4 hippocampal samples with 10× Genomic Visium and of 8 hippocampal samples with multiplexed fluorescent in situ hybridization (RNAscope). Figure created with BioRender (www.biorender.com). See Related Content tab for accessible version. DAPI = 4′,6-diamidino-2-phenylindole; H&E = hematoxylin and eosin.

Figure 2

Detection of neurogenesis markers in Visium Spatial Gene Expression data of the adult human hippocampus. (A) Plots showing the 7 clusters (white matter, blood vessel, CA1 and subiculum, polyform cell layer and subgranular zone [SGZ], granule cell layer [GCL], CA3 and CA4, and molecular layer [ML]) generated with the BayesSpace t-distributed error model algorithm. Sections on the left are from middle-aged males (n = 2, mean age 42.5 yr), and sections on the right are from young adult males (n = 2, mean age 23.5 yr). (B) Percentage of spots in each of the 7 BayesSpace clusters in each section. Supporting data are presented in Appendix 1, Table 3. (C) Validation of dentate gyrus (DG) morphology with hematoxylin and eosin (H&E) staining and RNAscope using a probe directed against the PROX1 (in green) dentate-lineage marker, with 4′,6-diamidino-2-phenylindole (DAPI) nuclear staining in blue (scale = 800 μm). The lines are labelling the SGZ, GCL, and ML of the DG. (D) Dotplot visualization of the scaled average expression of canonical cell-type marker genes in each BayesSpace cluster in all sections, including SATB2, SLC17A7, and SLC17A6 (excitatory neurons); GAD1 and GAD2 (inhibitory neurons); TMEM119, CX3CR1, and MRC1 (microglia); CLDN5 and VTN (endothelial cells); GLUL, SOX9, AQP4, GFAP, ALDH1L1, and VIM (astrocytes); PDGFRA, OLIG1, and OLIG2 (oligodendrocyte precursor cells); PLP1, MOG, MOBP, and MBP (oligodendrocytes); and SNAP25, STMN2, and RBFOX3 (neurons). The size of the dot corresponds to the percentage of spots within a cluster enriched for the marker. The colour of the dot represents the scaled average expression levels of the marker across spots within a cluster. Supporting data are presented in Appendix 1, Table 4.

Figure 3

Detection of neurogenesis markers in Visium Spatial Gene Expression data of the adult human hippocampus, showing log-normalized expression of neurogenesis markers spatially plotted onto each section at the subspot level using BayesSpace, including NES, SOX2, and ASCL1 (neural stem cells); PCNA and MCM2 (proliferative cells); NEUROD1 (neuroblasts); DCX and CALB2 (immature granule neurons); CALB1 (mature granule neurons); PROX1 (dentate lineage cells); SLC17A7 (excitatory neurons); and GAD1 (inhibitory neurons). Higher values in the scale correspond to higher expression levels, whereas lower values reflect lower expression levels. Supporting data can be accessed using our data set, deposited in Gene Expression Omnibus and following our analysis steps.

Figure 4

Qualitative comparison of gene expression between Visium and multiplexed fluorescent in situ hybridization (RNAscope) within and outside the dentate gyrus (DG). Enlarged sub-spot level plots showing NES, PCNA, MCM2, and DCX log-normalized expression on each section, with representative RNAscope images of NES, PCNA, MCM2, and DCX expression within and outside the DG from 2 samples. The delineated area in red corresponds to the subgranular zone (SGZ) and granule cell layer (GCL) on both sections (scale bars of whole DG sections = 1 mm; scale bars of expanded insets = 50 μm). Supporting data can be accessed using our data set, deposited in Gene Expression Omnibus and following our analysis steps.

Figure 5

Spatial reconstruction of a human hippocampal single-nucleus RNA sequencing data set with spot deconvolution, including the dentate gyrus (DG) granule neuron (exDG) and the neural stem cell (NSC) clusters from Habib and colleagues. The coloured scale corresponds to the proportions of cell type markers from the single-nucleus clusters represented in each spot. DAPI = 4′,6-diamidino-2-phenylindole. Supporting data presented in Appendix 4, Tables 1 and 2, available at https://www.jpn.ca/lookup/doi/10.1503/jpn.240026/tab-related-content.

Figure 6

Distribution of neural stem cell and proliferation marker genes, and density of DCX⁺ cells in the subgranular zone (SGZ) and granule cell layer (GCL) of the human dentate gyrus (DG). (A) Percentage of NES⁺SOX2⁺ cells with or without ALDH1L1 expression in the SGZ and GCL of the whole DG in samples from an infant, an adolescent, and 6 adults. The lower panel shows a NES⁺, SOX2⁺, and ALDH1L1⁺ cell in the GCL of the infant DG (age 2 yr). Scale bars indicate 20 μm (scale bars for expanded insets = 10 μm). Supporting data presented in Appendix 3, Table 1. (B) Expression of PCNA and MCM2 in the infant (age 2 yr) and adult DG (age 55 yr). Scale bars indicate 20 μm. (C) DCX expression detected in PROX1⁺ and PROX1⁻ cells in the DG of different age groups. Scale bars indicate 1 mm (scale bars for expanded insets = 20 μm). Dotted lines delineate the SGZ and GCL of the DG. (D) Number of DCX⁺ cells/mm² in the SGZ and GCL of the whole DG in infant (n = 1, age 2 yr), adolescent (n = 1, age 16 yr), and adult (n = 6, mean age 43.5 yr) samples. The graph shows values of 4 staining replicates per participant. Supporting data presented in Appendix 3, Table 2. DAPI = 4′,6-diamidino-2-phenylindole.

Figure 7

Distribution of DCX expression and different cell-type marker and immature neuronal genes in the subgranular zone (SGZ) and granule cell layer (GCL) of the human dentate gyrus (DG). (A) DCX expression detected in cells expressing PROX1 (dentate lineage), SLC17A7 (excitatory neurons), and GAD1 (inhibitory neurons) in the SGZ and GCL of the DG from a 31-year-old adult (scale bars = 20 μm). (B) Percentage of DCX⁺ cells expressing PROX1, NEUROD1, SLC17A7, and GAD1 in the SGZ and GCL of the whole DG by participant age. Supporting data are presented in Appendix 3, Table 3. (C) DCX⁺ and GAD1⁺ cells located in the inner GCL (closer to the SGZ) and outer GCL (closer to the molecular layer) of a 55-year-old adult (scale bars = 20 μm). White arrows point to double-positive cells. (D) DCX⁺ cells expressing PROX1 and CALB2 in the SGZ and GCL of the DG from a 37-year-old adult (scale bars = 20 μm; scale bars for expanded insets = 5 μm). (E) Percentage of PROX1⁺ and DCX⁺ cells expressing CALB2 in the SGZ and GCL of the whole DG by participant age. Supporting data are presented in Appendix 3, Table 4. DAPI = 4′,6-diamidino-2-phenylindole.

Figure 8

Specificity and distribution of DCX in the human brain. (A) Left panel shows manual annotations of samples 151671, 151672, 151673 and 151674 from Maynard and colleagues (n = 4, 2 males and 2 females, mean age 38.41 yr), including the 6 cortical layers (layers 1–6) and a white-matter region. Middle and right panels show the log-normalized expression of DCX and PROX1, spatially plotted onto each section at the sub-spot level. Higher values in the scale correspond to higher expression levels, whereas lower values reflect lower expression levels. (B) DCX expression with 4′,6-diamidino-2-phenylindole (DAPI) staining in a dorsolateral prefrontal cortex (DLPFC) section of a 26-year-old adult (scale bars = 50 μm). White arrows point to DCX⁺ cells. (C) DCX expression detected in the molecular layer (ML) of the dentate gyrus (DG) of a 37-year-old adult (scale bars = 50 μm; scale bars of expanded inset = 10 μm). (D) DCX expression in cells located in the CA3 region of the hippocampus in a 51-year-old adult (whole DG slide scanner images with scale bar = 500 μm; scale bars; higher magnification confocal images scale bars = 20 μm). (E) DCX expression in PROX1⁺ cells located in highly MBP-expressing regions in the adult human hippocampus (whole DG slide scanner images with scale bars = 800 μm; higher magnification confocal images with scale bars = 20 μm [5 μm for expanded insets]). GCL = granule cell layer; NA = not available.

Figure 9

Graphical representation of the the main findings from Visium and RNAscope. See Related Content tab for accessible version. DEG = differentially expressed gene; DG = dentate gyrus; GCL = granule cell layer; ML = molecular layer; OPC = oligodendrocyte precursor cell; SGZ = subgranular zone. Figure created with BioRender (www.biorender.com).