microRNA expression patterns reveal differential expression of target genes with age

Abstract

Recent evidence supports a role for microRNAs (miRNAs) in regulating the life span of model organisms. However, little is known about how these small RNAs contribute to human aging. Here, we profiled the expression of over 800 miRNAs in peripheral blood mononuclear cells from young and old individuals by real-time RT-PCR analysis. This genome-wide assessment of miRNA expression revealed that the majority of miRNAs studied decreased in abundance with age. We identified nine miRNAs (miR-103, miR-107, miR-128, miR-130a, miR-155, miR-24, miR-221, miR-496, miR-1538) that were significantly lower in older individuals. Among them, five have been implicated in cancer pathogenesis. Predicted targets of several of these miRNAs, including PI3 kinase (PI3K), c-Kit and H2AX, were found to be elevated with advancing age, supporting a possible role for them in the aging process. Furthermore, we found that decreasing the levels of miR-221 was sufficient to cause a corresponding increase in the expression of the predicted target, PI3K. Taken together, these findings demonstrate that changes in miRNA expression occur with human aging and suggest that miRNAs and their predicted targets have the potential to be diagnostic indicators of age or age-related diseases.

Figure 1. Downregulation of miRNA expression in old individuals.

For each panel, one young 30- year-old participant and one old 64-year-old participant were screened for miRNA expression using a real-time RT-PCR based miRNome miRNA profiler kit as described in Materials and Methods. Different pairs of individuals were used for both miRNome 1 and miRNome 2. The heat map was generated using dChip software and indicates the fold log₁₀ relative change in expression in young versus old participants. The top upregulated and downregulated miRNAs are shown. Blue indicates downregulation of miRNA expression in the old individuals and red indicates upregulation of miRNA expression in the old individuals. Each patient was sex and race matched as indicated in Table 1. A detailed spreadsheet of the data is available as supplementary information (Table S1). * next to the microRNA refers to the minor form of the miRNA.

Figure 2. Upregulated and downregulated miRNAs in young and old individuals.

A Venn diagram showing miRNAs that expression was either upregulated >2-fold (A) or downregulated >2-fold (B) in the old participants compared to young participants from the two different miRNome analyses in Figure 1. Listed are the miRNAs that overlapped between these 2 different data sets. The complete miRNome analysis for the listed miRNAs is available (Table S2).

Figure 3. Real-time RT-PCR validation of changes in miRNA expression in young and old participants.

(A) Expression of miRNAs from the miRNome analysis were further validated in 14 young and 14 old subject PBMCs (see Table 1B for detailed demographic data) using real-time RT-PCR as described in Materials and Methods. The histograms show normalized averages ± SEM from duplicate experiments. *P<0.05 by both Student's t-test and three-way ANOVA. (B) miR-24 and miR-221 expression was examined in 14 young and 14 old individuals using a TaqMan microarray assay as described in Materials and Methods. (C)Age-associated miRNAs that have been implicated in various cancers. See recent reviews for details , , .

Figure 4. Putative target validation of miRNAs downregulated by age.

(A) A Venn diagram showing the overlap between predicted targets of cancer-related miRNAs that were found to be downregulated in older participants. All cancer-related miRNAs have in common the predicted targets PIK3R1 and NOVA1. (B) Relative mRNA expression of PI3K, c-Kit and H2AX from 14 young and 14 old participants using real-time RT-PCR. The histograms show the normalized means ± SEM. (C) Representative immunoblots showing protein expression of PI3K, c-Kit and H2AX in a young and old patient. Participant PBMCs were lysed and immunoblotted with anti-PI3K, anti-c-Kit and anti-H2AX antibodies and reprobed with anti-actin and anti-histone H3 antibodies as protein loading controls. The same individuals were used for each representative immunoblot in C. (D) The protein levels of PI3K, c-Kit and H2AX from PBMCs were quantified from immunoblots and normalized to the amount of actin or histone H3. The histogram represents the normalized relative mean ± SEM from 14 young and 14 old individuals for PI3K and 10 young and 10 old individuals for c-Kit and H2AX. **P<0.01, *P<0.05 by Student's t-test.

Figure 5. miR-221 regulates p85 expression.

(A) Hela cells were transfected with control (Ctrl) siRNA or (AS)miR-221 and 48 h later mRNA levels of miR-221 and p85 were analyzed by RT-qPCR using U6 and UBC for normalization, respectively. The histogram represents the mean ± SEM from three independent experiments. *P<0.05 by Student's t-test. (B) Forty eight hrs after transfection with the indicated small RNAs, protein lysates were used for Western blot analysis of p85α expression. β-actin signals served as protein loading controls. (C) Schematics of the p85α 3′UTR and the dual-luciferase plasmids used in (E). pLuc control plasmid expresses both the Renilla luciferase (RL) and firefly luciferase (FL). Constructs of the p85α 3′ UTR span either two (pLuc-p85 (2644)) or one (pLuc-p85 (4491)) of the predicted miR-221 sites (black bars). (D) The p85α mRNA was predicted to be targeted by miR-221 using prediction algorithims from Target Scan and microRNA.org. The position of each predicted site within p85α is indicated. (E) Hela cells were transfected with the dual-luciferase plasmids shown and either control siRNA, (Pre)miR-221 or (AS)miR-221. The ratio of the RL/FL activity is shown. *P<0.05 using one-way ANOVA and Tukey's post-hoc test for the indicated comparisons.