Expression of Exosome Biogenesis Genes Is Differentially Altered by Aging in the Mouse and in the Human Brain During Alzheimer’s Disease

Abstract

Extracellular vesicles like exosomes are secreted by numerous cell types in a variety of tissues. Exosomes have been implicated in both aging and age-related disorders like Alzheimer’s disease (AD). However, how aging and AD affect exosome biogenesis within and across cell types is poorly understood. Moreover, cells acquire characteristics based on tissue niche, but the impact of tissue residence on cell type exosome biogenesis is unknown. We explored the Tabula Muris Senis, Mayo RNA-seq and Rush Religious Order Study/Memory and Aging Project data sets to characterize the cell and tissue-specific effects of aging and AD on genes involved in exosome biogenesis. Specifically, we examined the age-dependent expression (age coefficient) of genes involved in exosome biogenesis (22 genes), exosome cargo (3 genes), and senescence (5 genes). Of the 131 cell populations (cell type × tissue) studied, 95 had at least 1 exosome biogenesis gene affected by age. The most common gene/transcript increased by age was charged multivesicular body protein 2A (CHMP2A) (54 cell populations). The most common gene/transcript decreased by age was syndecan-binding protein (SDCBP) (58 cell populations). The senescence-associated genes cyclin-dependent kinase 1A (CDKN1A) and CDKN2A were not related to changes in CHMP2A and SDCBP and were altered by age in fewer cell populations. Finally, individuals with AD had decreased CHMP2A and increased SDCBP expression, opposite of what is observed during mouse aging in the absence of disease. These findings indicate that exosome biogenesis gene expression is modified by age in many cell populations mostly independent of senescence, and may be further altered in AD.

Advancing age is a primary predictor for the development of chronic disease (1). The mechanisms linking aging to chronic disease are incompletely defined, but recent work indicates an important role for extracellular vesicles like exosomes. Exosomes are lipid-encapsulated particles ranging from 50 to 200 nm in diameter (2). Exosome biogenesis occurs through an endosomal process coordinated by a series of endosomal sorting complexes required for transport (ESCRTs) (3). To date, very little is known regarding how the ESCRT pathway changes during aging. In fact, to our knowledge, the only evidence linking the ESCRT pathway to aging comes from recent work demonstrating that the ESCRT-II (4) and ESCRT-III (5) complexes are important for replicative lifespan in yeast.

Recent work showing that exosomes from young donors (6) or young mesenchymal stem cells (7) extend lifespan in rodents demonstrates the potential for using exosomes as an antiaging therapy. However, somewhat in contrast, there also is growing evidence that senescence leads to an increase in total exosome secretion (8–10), although the mechanism(s) involved and the function of these exosomes have not yet been described. While senescence is a hallmark of aging in some cell types, senescence and aging are not mutually inclusive (11). It is not known if, and to what extent, aging itself affects the biogenesis and secretion of exosomes (ie, in the absence of senescence). Additionally, it is not currently known whether exosome biogenesis is related to classic age-related disorders like Alzheimer’s disease (AD). Elucidating the impact of aging on exosome biogenesis will not only improve our understanding of the basic biology of aging, but may also improve the design and implementation of exosome-based therapeutics by identifying the cell types, and molecular mechanisms, most affected by aging.

Recent advances in single-cell RNA sequencing (scRNA-seq) have provided new opportunities to study the heterogeneity of aging within and between tissues and cell types that would otherwise be impossible or impractical. For example, the Tabula Muris Senis is a scRNA-seq data set comprising 131 different cell populations (type × tissue) from 22 different tissues across the life span of the C57BL/6JN mouse (12). Investigators from the Tabula Muris consortium generated age coefficients to describe the relationship between age and gene expression across the lifespan. Here, we analyzed age coefficients from the Tabula Muris Senis to define how aging affects exosome biogenesis pathways/gene expression, and the association of these changes with senescence and exosome cargo. We then compared our findings to publicly available RNA-seq data on human AD. Our analysis: (i) illustrates that aging alters exosome biogenesis gene expression in the mouse, (ii) identifies 2 novel aging-associated genes/transcripts involved in exosome biogenesis via the ESCRT pathway, and (iii) implicates exosome biogenesis as an affected process in the brain of human AD patients.

Materials and Methods

The methods used to obtain and process data for the Tabula Muris Senis are described here (12). In brief, tissues were harvested from male and/or female mice at 1 (n = 2 male mice), 3 (n = 7 male, n = 4 female), 18 (n = 2 male, n = 4 female), 21 (n = 3 female), 24 (n = 4 male), and 30 (n = 4 male) months of age. Single cells were isolated and scRNA-seq was performed. The number of cells from each population analyzed is available in Supplementary Materials. Age coefficients were generated in the Tabula Muris Senis using a linear model (13) where age was a numerical value and coefficients were adjusted for age and technology. Genes that were not expressed in at least 3 cells and cells that did not have at least 250 detected genes were not used in these analyses. We obtained differential gene expression data described in the Tabula Muris Senis from the Gene Expression Omnibus website (accession number: GSE149590). Data on human AD were obtained from the Mayo RNA-seq study and the Rush Religious Order Study/Memory and Aging Project (ROSMAP) scRNA-seq study, both available on the Synapse database (accession numbers: syn3163039 and syn18485175). Log2 fold gene expression differences were extracted from differential expression tables for the Mayo RNA-seq study provided on the Synapse server (syn6090802), and from supplementary tables in a published study of ROSMAP scRNA-seq data (https://pubmed.ncbi.nlm.nih.gov/31042697/).

The primary gene set of interest was the Gene Ontology category “extracellular exosome biogenesis”, GO: 0097734. In addition, CD9, CD63, and CD81 were chosen because these genes encode established exosome cargo proteins (2). Finally, cyclin-dependent kinase 1A (CDKN1A), CDKN2A, interleukin-6 (IL-6), interleukin-1β (IL-1β), CCL2, and tumor necrosis factor (TNF) were chosen based on their established roles in senescence (14). Alpha for age coefficient data was set as an adjusted p value of <.01. An alpha of p < .001 was also tested and did not affect the conclusions herein. All figures were generated using GraphPad Prism software.

Results and Discussion

We analyzed data from the Tabula Muris Senis to identify exosome biogenesis genes/transcripts altered by aging in mice. First, we calculated the number of cell populations (population = cell type × tissue; n = 131) across all tissues with significantly altered expression of exosome biogenesis genes with age (Figure 1A). The 10 most commonly decreased (Figure 1B; top) or increased (Figure 1B; bottom) genes/transcripts with age are shown. Age coefficients for every cell population are provided in Supplementary Materials. From this analysis, we found that 95 different cell populations have at least 1 exosome biogenesis gene altered by aging. In general, genes more frequently decreased in expression with age. Syndecan-binding protein (SDCBP; also known as Syntenin-1) was the exosome biogenesis gene most commonly decreased with age. Charged multivesicular body protein 2A (CHMP2A) was the exosome biogenesis gene most commonly increased with age. Importantly, recent reports suggest that SDCBP may be a “universal biomarker” of exosomes (15), and CHMP2A has been shown to decrease with senescence (16) (although its exact function in this context is unknown). These results demonstrate that aging affects the expression of exosome biogenesis genes across many cell types and tissues.

Figure 1.

Age-related changes in exosome biogenesis gene expression from the Tabula Muris Senis. (A) Dot plot showing the number of cell populations and age-related expression changes (age coefficients) for exosome biogenesis genes across all 23 tissues in the Tabula Muris Senis data set. N = 131 cell populations. (B) Exosome biogenesis genes ranked based on the number of cell populations displaying decreased (top) or increased (bottom) expression with age. (C) Proportion of cell populations with decreased SDCBP and/or increased CHMP2A (key exosome biogenesis genes) expression among the most common cell types in the Tabula Muris Senis.

We next asked whether decreased expression of SDCBP and increased expression of CHMP2A with aging are commonly observed in the 5 most abundant cell types in the Tabula Muris Senis (endothelial cells, T cells, B cells, macrophages, and mesenchymal stem cells). As shown in Figure 1C, the majority (5/6) of mesenchymal stem cell populations displayed a concurrent increase in CHMP2A and decrease in SDCBP with age. Multiple (6/14) endothelial cell populations also displayed both an increase in CHMP2A and a decrease in SDCBP. By comparison, macrophages, B cells and T cells showed some, but infrequent, co-occurrence of decreased SDCBP and increased CHMP2A expression. These findings indicate that aging preferentially affects gene expression of CHMP2A and SDCBP in mesenchymal stem cells and endothelial cells, and less commonly in immune cell populations.

Senescence is a hallmark of aging and is known to promote the release of exosomes (8–10). Others have shown that SDCBP gene expression is increased in senescent cells compared to proliferating cells in culture (17), and that senescence decreases CHMP2A expression (16). However, our analysis generally suggests an age-related decrease in SDCBP expression and increase in CHMP2A expression. Thus, we sought to better understand the cell population-specific relationship between changes in expression of CHMP2A and SDCBP and senescence-associated genes. Expression of the senescence-associated genes CDKN1A, CDKN2A, IL-1β, IL-6, CCL2, and TNF was only increased in 36, 9, 17, 20, 35, and 6 cell populations (out of 131), respectively (Figure 2A; right). Fewer cell populations displayed a decrease in these same genes (Figure 2A; left). Therefore, at a glance, the exosome biogenesis genes SDCBP, CHMP2A, syndecan-4 (SDC4), programmed cell death 6-interacting protein (PDCD6IP aka ALIX), vacuolar sorting 4 homolog B (VPS4B), Rab7, and CD34 were all more commonly altered by age than any of these senescence markers.

Figure 2.

Extracellular vesicle biogenesis genes are more responsive to aging than senescence-associated genes. (A) Number of cell populations from the Tabula Muris Senis data set with decreased (left) and increased (right) expression of senescence genes with age. N = 131 cell populations. (B) Proportion of cell populations with concurrent age-related changes in the expression of CHMP2A (left) or SDCBP (right) and the senescence genes CDKN1A and CDKN2A. Black numerals represent the total number of cell populations with the specific gene expression pattern with aging. White numerals represent the number of cell populations that overlap with other gene expression patterns.

Co-occurrence analysis (Figure 2B) indicated that only 48% of cell populations with decreased SDCBP expression (28 out of 58) or increased CHMP2A expression (26 out of 54) also displayed an increase in CDKN1A. Furthermore, only 16% of cell populations with decreased SDCBP expression (9 out of 58) and 17% of cell populations with increased CHMP2A expression (9 out of 54) also displayed an increase in CDKN2A. Collectively, our analysis strongly suggests that aging alters exosome biogenesis genes in a wide variety of cell populations, and that this can occur independent of classic markers of senescence (ie, CDKN1A and CDKN2A). Further, this finding suggests that distinct cellular alterations in exosome biogenesis occur during healthy aging (prior to the onset of pathological/senescence-associated aging).

Could the age-related changes in exosome biogenesis gene expression we observe in mice be relevant to human aging/disease? To investigate this possibility, we examined exosome biogenesis gene expression in brain aging and AD, a quintessential disease of aging (18). Exosomes have been implicated in the etiology of AD via trafficking of tau (19,20), and silencing of CHMP2A causes the propagation of tau seeds (21), which lead to tau tangles that are associated with AD pathology (22)—but a link to aging has not been established. Because tau accumulates even during normal aging (23), we first examined whether the aging brain is characterized by altered exosome biogenesis gene expression in the Tabula Muris Senis. We found that 7 of the 9 cell types in the brain displayed age-related changes in CHMP2A expression (Figure 3A). CHMP2A expression was increased in 6 of these cell types, consistent with the general response of CHMP2A to aging, but CHMP2A expression was decreased with age in neurons. In contrast, consistent with its general response to aging, SDCBP expression was decreased in 5 of 9 brain cell types but not increased in any brain cell type. To determine if these changes in exosome-associated genes/transcripts are relevant to AD, we next examined temporal cortex tissue gene expression data from the Mayo RNA-seq study and scRNA-seq data from the ROSMAP study. In postmortem cortex samples from the Mayo RNA-seq study, individuals with diagnosed AD displayed numerous alterations in exosome biogenesis genes/transcripts (Figure 3B) compared to neurotypical individuals. Many of these gene expression patterns were opposite of the age-related changes observed in the Tabula Muris Senis (eg, decreased CHMP2A and increased SDCBP expression). Interestingly, scRNA-seq data from the ROSMAP study confirmed these AD-related changes and suggested they may be: (i) especially pronounced in neurons and microglia; and (ii) progressively increased with pathology (Figure 3C). These data collectively demonstrate that AD is associated with gene expression changes in CHMP2A and SDCBP that oppose those observed during aging in mice. Future studies could explore whether age-related changes in CHMP2A/SDCBP expression in neurons contribute to the accumulation of tau and the glial cell activation that are thought to contribute to AD pathology.

Figure 3.

Mouse brain aging and human Alzheimer’s disease (AD)-associated changes in exosome biogenesis and cargo gene expression. (A) Age-related changes (age coefficients) in expression of the key exosome biogenesis genes SDCBP and CHMP2A in brain cell types in the Tabula Muris Senis mouse data set. (B) Differential expression of SDCBP and CHMP2A in bulk brain samples of AD patients compared to neurotypical controls from Mayo RNA-seq data set. (C) Differential expression of SDCBP and CHMP2A in different brain cell types with increasing AD pathology from the Rush Religious Order Study/Memory and Aging Project (ROSMAP) scRNA-seq data set. (D) Age-related changes in expression of key exosome cargo genes in brain cells in the Tabula Muris Senis mouse data set. (E) Differential expression of exosome cargo genes in bulk brain samples from AD patients compared to neurotypical controls in the Mayo RNA-seq data set. (F) Differential expression of exosome cargo genes in different brain cell types with increasing AD pathology from the ROSMAP scRNA-seq data set.

Recent work has demonstrated that exosome cargo is also altered in patients with AD (24). In the Tabula Muris Senis, we found that microglia have increased expression of CD9, CD63, and CD81 with age (Figure 3D), whereas neurons have increased expression of CD9 but decreased expression of CD63 and CD81 with age. In the Mayo RNA-seq study (bulk brain RNA-seq), patients with AD displayed increased expression of CD9, CD63, and CD81 compared to neurotypical controls (Figure 3E), and ROSMAP scRNA-seq data indicated that this could be explained by pathology-associated expression increases in microglia (Figure 3F), which are highly abundant in the brain. This differential expression of CD63 in microglia (increased) and neurons (decreased) is interesting since CD63 identifies a unique population of neuroprotective microglia in the 5xFAD mouse model of AD (25). Additionally, circulating blood amyloid-beta bound to CD63⁺ exosomes has been recently shown to better predict AD than unbound circulating amyloid-beta (26). These results further support the need for future mechanistic studies to determine if gene expression changes in exosome biogenesis in the aging brain contribute to the development and progression of AD.

There are a few limitations and considerations in the current report worth noting. First, the number of mice studied at some ages in the Tabula Muris Senis is small. Therefore, these gene expression data are worthy of replication in a larger cohort of aged mice and/or human participants. Furthermore, the original Tabula Muris Senis consortium reported that the number of cells expressing CDKN2A was more than doubled in aged cells compared to young cells (12). Nonetheless, within cell populations, we found that exosome biogenesis genes like CHMP2A and SDCBP are more commonly altered than CDKN2A and other senescence-associated genes (Figure 2). Whether this finding means that changes in exosome biogenesis gene expression precede, protect against, or occur independently of senescence remains to be determined. Furthermore, since only gene expression was examined in this report, understanding the functional impact of these observed changes will require the study of protein expression and activity/function in aged cells. Finally, it should be noted that while age-related changes in exosome biogenesis gene expression were observed in mice, it is not currently known to what extent these changes also occur in human aging, especially in organs like the brain that may age differently in mice versus humans. The recently published Tabula Sapiens data set does not include data on aging, but it may be a platform for initial comparisons of gene expression of exosome genes in mice versus humans at the single-cell level.

Overall, our findings demonstrate that advancing age alters the expression of exosome biogenesis genes in mice. The 2 most commonly altered exosome biogenesis genes are CHMP2A and SDCBP, which have not been previously implicated in aging. We also show that age-related changes in exosome biogenesis can occur independent of classical markers of senescence. We further found that AD affects exosome biogenesis gene expression in the human brain, including CHMP2A, which has been recently implicated in the regulation of tau trafficking. These findings offer new insight into the relationship between aging and exosome biogenesis within and across cell types and the potential relevance of these events to neurodegenerative diseases like AD.

Funding

This work was supported by the National Institutes of Health – National Institute on Aging (AG060302 to T.J.L.) and the American Heart Association – Innovative Project Award (18IPA34110052 to D.S.L.).

Conflict of Interest

None declared.