Mitochondrial Small RNA Alterations Associated with Increased Lysosome Activity in an Alzheimer's Disease Mouse Model Uncovered by PANDORA-seq

Abstract

Emerging small non-coding RNAs (sncRNAs), including tRNA-derived small RNAs (tsRNAs) and rRNA-derived small RNAs (rsRNAs), are critical in various biological processes, such as neurological diseases. Traditional sncRNA-sequencing (seq) protocols often miss these sncRNAs due to their modifications, such as internal and terminal modifications, that can interfere with sequencing. We recently developed panoramic RNA display by overcoming RNA modification aborted sequencing (PANDORA-seq), a method enabling comprehensive detection of modified sncRNAs by overcoming the RNA modifications. Using PANDORA-seq, we revealed a previously unrecognized sncRNA profile enriched by tsRNAs/rsRNAs in the mouse prefrontal cortex and found a significant downregulation of mitochondrial tsRNAs and rsRNAs in an Alzheimer's disease (AD) mouse model compared to wild-type controls, while this pattern is not present in the genomic tsRNAs and rsRNAs. Moreover, our integrated analysis of gene expression and sncRNA profiles reveals that those downregulated mitochondrial sncRNAs negatively correlate with enhanced lysosomal activity, suggesting a crucial interplay between mitochondrial RNA dynamics and lysosomal function in AD. Given the versatile tsRNA/tsRNA molecular actions in cellular regulation, our data provide insights for future mechanistic study of AD with potential therapeutic strategies.

Keywords: Alzheimer’s disease; lysosome; mitochondrial; small non-coding RNAs (sncRNAs); tRNA-derived small RNAs.

Figure 1

Age-dependent cognitive impairment in transgenic AD mice. (A) Schematic of the Y-maze Spontaneous alternation behavior assay showing an example of a correct (left) and incorrect (right) spontaneous alternation. (B) The alternation index in the Y-maze test in two- and six-month-old WT and AD mice (*** p < 0.001, n = 3–4 in each group). The alternation index (%) was calculated as (number of spontaneous alternations/maximum possible alternations) × 100. (C) Maximum alternation by WT and AD mice in Y-maze test at two-month and six-month age, respectively. (D) Schematic of Novel Object Recognition Test (NORT) protocol consisting of a habituation phase (Day 1), training phase (Day 2), and testing phase (Day 3). During the training phase, mice were exposed to two identical triangle-shaped objects. In the test phase, a novel object (cross-shaped object) was introduced to replace one triangle-shaped object. The time each mouse spent exploring the novel object (cross-shaped object) was recorded to assess recognition memory. (E) The recognition index during the testing phase in two- and six-month-old WT and AD mice (** p < 0.01, n = 3–4 in each group). The recognition index (%) was calculated as (total exploration time for the novel object during testing/total exploration time for the familiar object during testing) × 100. Data were analyzed by two-way ANOVA with Sidak post-hoc test for multiple comparisons.

Figure 2

PANDORA-seq revealed a tsRNA/rsRNA-dominant sncRNA landscape in the prefrontal cortex. Representative result of sncRNAs species detected by traditional-seq and PANDORA-seq, respectively. (A) SncRNA landscape revealed by traditional-seq; (B) SncRNA landscape revealed by PANDORA-seq. Pie charts depict the relative proportions of miRNAs, piRNAs, tsRNAs, rsRNAs, and other sncRNAs identified by each method. Bar graphs further illustrate the mean expression levels (RPM) of these sncRNA types across different size ranges. The error bars stand for the standard error of the mean. (traditional-seq n = 14 mice, PANDORA-seq n = 16 mice).

Figure 3

PANDORA-seq identified mitochondrial tsRNA/rsRNA as most sensitively dysregulated in AD mice. (A,B) Volcano plots of differentially expressed sncRNAs (DE-sncRNA) in the cortex of AD mice as compared with WT mice, revealed by traditional-seq (A) and PANDORA-seq (B). Colored dots represent the increased or decreased DE-sncRNAs with an FDR < 0.05 and FC > 2 as a cutoff threshold. (C,D) Probability density plots show the distribution of the log₂ fold change (log₂FC) of tsRNAs (C) and rsRNAs (D) in WT versus AD mice. (traditional-seq n = 7 mice in each group, PANDORA-seq n = 8 mice in each group).

Figure 4

Mapping analyses of representative mitochondrial tsRNAs in WT and AD mice brain prefrontal cortex. (A) The expression profile of mt-tsRNA-Gln across the length in WT and AD samples. (B) The expression profile of mt-tsRNA-His across the length in WT and AD samples. (C) The expression profile of mt-tsRNA-Leu across the length in WT and AD samples. (D) The expression profile of mt-rsRNA-16S (101–140 nt) across the length in WT and AD samples. The solid curves indicate the mean of RPM, while the shading represents the standard error of the mean. (n = 8 mice in each group).

Figure 5

AD cortex shows increased lysosome activity, negatively correlated with mitochondrial tsRNAs. (A) KEGG pathway enrichment analysis of differentially expressed genes in WT versus AD cortex. The vertical dash line indicates the significance level of α = 0.05. (B) The gene-set score of the lysosome pathway in WT and AD cortex (left); Heatmap representation of differentially expressed genes in the lysosome pathways (right). Red represents relatively upregulated expression, whereas blue represents reduced. (C) Relationship between the gene-set score of the lysosome pathway and the abundance of mt-tsRNA-Gln, mt-tsRNA-His, mt-tsRNA-Leu and mt-rsRNA-16S (101–140 nt). Each dot represents one sample (n = 16). The gray band stands for the 95% confidence interval. p means statistical significance while ρ means Spearman’s rank correlation coefficient. (n = 8 mice in each group).