Transcriptome dynamics of hippocampal neurogenesis in macaques across the lifespan and aged humans

Whether adult hippocampal neurogenesis (AHN) persists in adult and aged humans continues to be extensively debated. A major question is whether the markers identified in rodents are reliable enough to reveal new neurons and the neurogenic trajectory in primates. Here, to provide a better understanding of AHN in primates and to reveal more novel markers for distinct cell types, droplet-based single-nucleus RNA sequencing (snRNA-seq) is used to investigate the cellular heterogeneity and molecular characteristics of the hippocampi in macaques across the lifespan and in aged humans. All of the major cell types in the hippocampus and their expression profiles were identified. The dynamics of the neurogenic lineage was revealed and the diversity of astrocytes and microglia was delineated. In the neurogenic lineage, the regulatory continuum from adult neural stem cells (NSCs) to immature and mature granule cells was investigated. A group of primate-specific markers were identified. We validated ETNPPL as a primate-specific NSC marker and verified STMN1 and STMN2 as immature neuron markers in primates. Furthermore, we illustrate a cluster of active astrocytes and microglia exhibiting proinflammatory responses in aged samples. The interaction analysis and the comparative investigation on published datasets and ours imply that astrocytes provide signals inducing the proliferation, quiescence and inflammation of adult NSCs at different stages and that the proinflammatory status of astrocytes probably contributes to the decrease and variability of AHN in adults and elderly individuals.

The hippocampal formation is one of the main brain regions affected in neurological diseases, 48 such as Alzheimer's disease, stress, and depression, and has attracted tremendous attention due 49 to its physiological and clinical significance 9,10 . Extensive evidence demonstrates that 50 hippocampal neurogenesis is involved in memory processing, cognitive function, and mood 51 regulation in animal models 11,12 . However, controversy still exists regarding whether adult 52 hippocampal neurogenesis persists in aged humans, although remarkable efforts have been 53 attempted, including using different antibodies, shortening the intervals of postmortem 54 sampling, and improving the fixation conditions and procedures for immunostaining 2,5,7,8,[13][14][15] . 55 Recently, another round of debate was ignited between two labs 3,6 . Despite the conflicting 56 results, they both reached a consensus that new technologies, such as single-cell RNA-seq, will 57 provide new insights in this field, and reveal novel markers for immature neurons and adult 58 NSCs.

74
Some DCX + EdU + and NeuN + EdU + cells were also identified, which strongly indicated that 75 these cells were newly generated by mitosis (Fig. 1g, h, Extended Data Fig. 1f, h). Overall, 76 NSCs in aged macaques proliferate continuously and produce immature granular cells, 77 although the neurogenesis rate is decreased compared with samples obtained at younger stages.

117
We next analyzed the NSC/Astrocyte residing in the DG. Three subclusters with distinct 118 gene expression profiles were identified (Fig. 2c,  to the well-studied rodent models. We compared our dataset with the dataset containing 153 information on postnatal stages in mice 17 . Cells in the Astrocyte and NSC1 clusters matched 154 perfectly with mouse astrocytes and RGL, respectively. Additionally, NSC2 partially merged 155 with nIPC in mice, which strengthened that NSC2 was more active (Fig. 2f, Extended Data Fig.   156 4a-c). Then, the expressional features of macaque NSCs and mouse RGLs were compared.  With the identified NSCs, we next investigated the developmental process and regulatory 170 continuum of postnatal and adult neurogenesis in macaques. The neurogenic lineage-related 171 clusters (NSC1, NSC2, ImmN, and GC) were integrated for further analysis. Representative 172 markers for each cell type were highlighted (Fig. 2j, k). The trajectory inference with monocle3 173 showed that NSC1 transited to NSC2, followed by ImmN, and then differentiated into mature GC (Fig. 2j). The expression of DEGs along the pseudotime plot exactly recapitulated the 175 regulatory continuum of the process of adult neurogenesis (Fig. 2l). An analysis of enriched 176 GO terms validated the biological process and function of each cell type, such as 'cell 177 proliferation' for NSC, 'cell migration' and 'axon extension' for ImmN, and 'neurotransmitter 178 secretion' and 'synapse assembly' for mature GC (Fig. 2m). and Micro4 were two small population (Fig. 3i). The analysis of enriched GO terms showed 254 that Micro2 was associated with the phagocytosis pathway, inflammatory response and 255 lysosome pathways (Fig. 3j). Moreover, over two-thirds of the Micro2 cells were from the 256 aging samples (Fig. 3k). Active microglia exert beneficial functions via phagocytosis or 257 detrimental functions by secreting cytotoxic cytokines 28-32 . To detail the nature of Micro2, it 258 was divided into two subclusters, Micro2-1 and Micro2-2 (Fig. 3l). The Micro2-1 subcluster 259 mainly containing young samples exhibited anti-inflammatory signatures, such as complement 260 activation and apoptosis, which might be associated with synaptic pruning and phagocytosis of 261 debris from apoptotic cells during early postnatal stages. Micro2-2 subcluster populated with 262 aging samples were involved in the inflammatory response and PI3K/AKT/mTOR pathway 263 activation, which is reportedly related to aging-associated microglial activation (Fig. 3l-n). the results showed that astrocytes and NSCs had a stronger interaction (Fig. 3o). Thus, we 273 further delineated the potential interactions between astrocytes and NSCs (Fig. 3o). The Eph-Ephrin signaling, and cytokine-receptor interaction were enriched in the old samples (Y4-277 Y23) (Fig. 3p, q). Our results were consistent with previous findings that VEGF and FGF are 278 involved in the mechanism by which astrocytes promote adult NSC proliferation 21,34 ; the 279 Integrin pathway, BMP signaling and Eph-Ephrin signaling have consistently been shown to 280 be involved in regulating NSC quiescence by niche cells 21,22 . Therefore, at a young stage, 281 astrocytes may increase NSC proliferation, and at an old stage, astrocytes may promote NSC 282 quiescence and inflammation. 283 We next investigated the expression levels of the ligands in the subclusters of astrocytes. Astro2. In Astro3, the cytokine TNFSF4 might lead to an inflammatory effect on NSCs.

287
Interestingly, we also found that extracellular matrix-related genes, which play an important 288 role in modulating the local inflammatory milieu, as well as homeostasis of NSCs, were 289 enriched in Astro3 (Fig. 3r). Overall, we provided evidence that distinct astrocyte subtypes  (Fig. 4c, d). We also observed the specific expression of ETNPPL in human NSCs 322 (Fig. 4e). MKI67 + ETNPPL + double-positive cells were detected in the DG of the hippocampus 323 (Fig. 4f), which strongly indicated that ETNPPL was a primate-specific NSC marker. The 324 expression of STMN1 and STMN2 was also verified in aged human samples (Fig. 4g, h). The 325 colocalization of STMN1 and STMN2 with CALB2 but not with NEUROD1 strongly indicated 326 that both markers label immature neurons, which confirmed that adult neurogenesis exists in 327 aged human hippocampi (Extended Data Fig. 9f-i). 328 Last, we examined the NSC niche in aged humans. The astrocytes in humans were 329 subgrouped into two populations, H-Astro1 and H-Astro2 (Fig. 4i). H-Astro1 was enriched in 330 genes related to the inflammatory response, similar to those in macaques (e.g., JAK/STAT 331 pathway, NFKB1, and TNFSF4), which strengthened the hypothesis that a group of astrocytes 332 was activated during brain aging (Fig. 4j). H-Astro2 expressed genes related to regulating NSC 333 quiescence (e.g., PDGFA/B/C, MFGE8, and Ephrin family). We integrated the two datasets to 334 investigate whether the regulation of active astrocytes was conserved between humans and 335 macaques. River-plot showed that active astrocytes (M-Astro3 and H-Astro1) were conserved 336 between macaques and humans (Extended Data Fig.10a, b). Likewise, the analysis of microglia 337 in aged humans also revealed a cluster of active microglia that correlated strongly with the 338 active microglia in macaques and expressed CD74, IGNGR1, TLR2, and TNFRSF1B (Fig. 4k,   339 l, Extended Data Fig. 10c, d). The interaction of astrocytes and adult NSCs revealed conserved 340 pathways between humans and macaques, such as the EPH-EPHRIN pathway, BMP pathway, 341 and inflammatory signaling pathway. However, in the Eph-Ephrin pathway, EFNA5 was the 342 only ligand enriched in macaques, while EFNB1, EFNB2, and EFNA1 were also significantly 343 expressed in humans, indicating conserved pathways but distinct preferred ligand-receptor 344 counterparts between humans and macaques (Fig. 4m).

346
In this study, we systematically survey the molecular and cellular dynamics of the 347 hippocampi in macaques across the lifespan and in aged humans using high-throughput    BrdU (Thermo Fisher Scientific, B23151) labeling was conducted. Briefly, EdU labeling (5 mg/kg body 463 weight) was administered to macaques intravenously (i.p.) twice a week for 10 weeks. With a two-month-464 interval, BrdU was intravenously administered (10 mg/kg) twice a week for 10 weeks. Then, the animals 465 were sacrificed 2 months of post-administration. 466 For macaque brain sample preparation, adult and senile macaque monkeys were anesthetized 467 with Katamine (10 mg/kg body weight) and perfused with cold artificial cortico-spinal fluid 468 (ACSF). The whole brains were rapidly dissected on ice. For single nucleus RNA-seq, the 469 hippocampal formations and other brain regions in the right hemisphere were isolated and 470 quick frozen in liquid nitrogen. For the immunostaining, the left part of each brain was cut 471 coronally into around 1cm-thickn slabs and fixed in 4% PFA for 12 h, following cryoprotected 472 in a sucrose gradient from 10%, 20% to 30%, and then embed the tissue with OCT.

474
Frozen sections were equilibrated to room temperature (RT) and rinsed with PBST buffer (0.3%

475
Triton-X in PBS) for 10 min. To reduce the remaining aldehyde groups, the sections were 476 incubated with a 0.5% sodium triacetoxyborohydride (NaBH(OAc)3; Merck, 316393) solution 477 for 30 min at room temperature. Next, the samples were boiled in a sodium citrate buffer (10 478 mM sodium citrate tribasic dihydrate, pH 6.0) for 20 minutes, and cooled down to room 479 temperature. For PSA-NCAM antibody staining, we conducted conduct antigen retrieval in Fluoromount-G (SouthernBiotech, 0100-01) and stored at 4℃ in the dark.

489
To visualize EdU-labelled cells, a Click-iT EdU Imaging Kit (Invitrogen, C10639) was used according to 490 manufacturer's guidelines. For BrdU staining, the tissue was incubated with 2 M HCl at RT for 20 minutes, 491 followed by sequential incubation with 5% NDS, the anti-BrdU antibody and secondary antibody, 492 respectively. 493

Imaging 494
Serial coronal sections of the hippocampus were collected from macaques and humans. Images were 495 acquired using an Olympus FV3000 confocal microscope with a 30x oil immersion objective (NA 1.30). 496 Tiled images were acquired and stitched using Olympus FluoView software. Image processing and analysis 497 were conducted using FLUOVIEW FV3000 and Fiji/ImageJ software and compiled with Adobe Photoshop 498

499
Single nuclei isolation, 10 X Genomics Chromium, and sequencing 500 Nuclei were isolated from fresh-frozen tissue as previously described with some modifications and 501 improvements. Briefly, tissue samples were minced into pieces <5 mm and then homogenized using a glass 502 dounce tissue grinder (Sigma, Cat# D8938) in 2 ml of Nuclei EZ lysis buffer (Sigma, Cat# NUC-101) on 503 ice. After two incubations on ice (with 4 mL of lysis buffer and 5 min each time), the homogenate was 504 filtered through a 70 µm strainer, and then we used Debris Removal Solution (Miltenyi Biotech, #130-109-505 398) to perform density gradient centrifugation to clean the nuclear suspension according to the 506 manufacturer's protocol. Isolated nuclei were resuspended and washed with nuclei suspension buffer (NSB, consisting of 1× PBS, 0.1% BSA and 0.4 U/µl Ambion™ RNase inhibitor (Thermo Fisher, Cat# AM2684)) 508 and filtered through a 35 µm cell strainer. Nuclei were counted using a hemocytometer and diluted to 1000 509 nuclei/μL for optimal 10X loading. Approximately 8000 nuclei were targeted and captured for each reaction. 510 Steps from Chip B loading to cDNA library construction was carried out with Chromium Single Cell 3' 511 Reagent Kits v3 according to the 10X official instructions. 512 Processing of macaque snRNA-seq data 513 A cell-by-gene count matrix was generated following sequence alignment with Cell Ranger software 514 (https://support.10xgenomics.com/). The raw count matrix from each sample was loaded to the scrublet 515 pipeline individually, with the parameter expected_doublet_ratio set to 0.06 to identify possible doublets 35 . 516 A total of 32049/182036 cells with a doublet score greater than the doublet score threshold was considered 517 as doublets and thus discarded from subsequent analysis. The filtered cell-by-gene count matrix was then 518 loaded into the Seurat pipeline 36 for downstream analysis. Cells that did not meet the following criteria were 519 omitted: 1) cells with a number of detected genes ranging from 800 to 7500; 2) cells with nUMI greater than 520 1000; and 3) cells with a percentage of mitochondrial counts less than 1%. Overall, 132524 cells were 521 retained for downstream analysis. Next, the filtered cell-by-gene matrix was imported to function 522 CreateSeuratObject to create a Seurat object followed by log-normalization of the gene expression matrix 523 with function NormalizeData. Then, we identified variable genes that exhibited high cell-to-cell variations 524 Distinguishing DG and CA cells based on regional transcriptional identities 530 We first computed differentially expressed genes between cells obtained from individual dissection of CA 531 and DG for the each of the following individual cell types to distinguish DG and CA cells from whole 532 hippocampal samples: astrocytes, microglia, OPCs, oligodendrocytes and GABAergic neurons. Then, we 533 performed principal component analysis of each cell type in whole hippocampal samples based on the 534 identified regional differentially expressed genes. Next, we fitted the resulting principal components to the 535 k-means algorithm for clustering analysis. Subsequently, we determined the regional characteristics of the 536 identified clusters by calculating the enrichment score for genes enriched in CA and DG across all clusters 537 using the AUCell package 37 . A cluster with a higher enrichment score for DG-enriched genes was considered 538 as a DG-derived group, and a cluster with a higher enrichment score for CA-enriched genes was identified 539 as a CA-derived group. 540 Construction of the macaque adult neurogenic trajectory 541 We first extracted NSCs, ImmNs and granule from the whole dataset and performed the dimension reduction 542 analysis using those cells to infer the adult neurogenic trajectory in macaque hippocampi. Then, for trajectory 543 construction, we converted the Seurat object to a SingleCellExperiment object and performed trajectory 544 graph learning analysis and pseudotime measurement with the functions learn_graph and order_cells from 545 monocle3 38-40 , respectively. 546 Comparison of transcriptional profiles between macaque and mouse NSCs. 547 We first assembled cells from our dataset with those from a published scRNA-seq dataset of the mouse 548 dentate gyrus 41 based on orthologues identified between the macaque and mouse genomes with the R 549 package biomaRt 42 . Next, we subdivided NSCs from the merged dataset and computed differentially 550 expressed genes between macaque and mouse NSCs with the Seurat function FindMarkers. 551 Identification of cell-cell communication 552 We studied cell-cell communication using iTALK (https://github.com/Coolgenome/iTALK), a package 553 designed to study cell interactions. Briefly, the rawParse function was applied to retain the top 50% highly 554 expressed genes as input for the FindLR function to identify ligand-receptor pairs between cell types. We 555 mainly focused on the ligand-receptor pairs where ligands were obtained from astrocytes and receptors were 556 expressed in NSCs. The identified interactions were visualized using the NetView and LRPlot functions. 557 Processing of human snRNA-seq data 558 Reads were aligned to the human reference genome hg19 with Cell Ranger software 559 (https://support.10xgenomics.com/), and a cell-by-gene count matrix was then generated. Then, candidate 560 doublets were identified by individually importing the raw count matrix from each human sample into the 561 scrublet software by setting the parameter expected_doublet_ratio to 0.06 35 . A total of 2287/26865 cells 562 with a doublet score exceeding the setting threshold were then omitted from the subsequent analysis. The 563 filtered cell-by-gene count matrix was then loaded into the Seurat pipeline. Then, cells that did not meet the 564 following criteria were discarded: 1) cells with more than 200 detected genes and 2) cells with fewer than 565 10000 detected genes. After removing low-quality cells, 22119 cells were retained for the subsequent 566 analysis. The filtered count matrix was then log-normalized with the NormalizeData function. Next, variable 567 features were computed with the function FindVariableFeatures, and dimension reduction analysis was 568 performed on these variable features with the functions RunPCA and RunUMAP. Batch effects derived from 569 nonbiological variations were removed by conducting a canonical correction analysis among individual 570 samples with the functions FindIntegrationAnchors and IntegrateData. The clustering analysis was carried 571 out with the functions FindNeighbors and FindClusters. 572

Construction of the human adult neurogenic trajectory 573
Similar to the trajectory inference analysis conducted for the macaque dataset, we first extracted NSCs, 574 ImmNs and granule cells from the whole human dataset and then performed the dimension reduction analysis 575 on those cells. Next, we used the UMAP embedding and clusters generated by Seurat as input for the 576 trajectory learning analysis and pseudotime measurement analysis with the functions learn_graph and 577 order_cells from monocle3. 578 Quantification and statistical analysis 579 All data obtained from macaques and humans at each age were collected from at least three The single-nucleus RNA-seq data used in this study has been deposited in the Gene Expression Omnibus 586