Circulating microRNA profile of long‐lived Okinawans identifies novel potential targets for optimizing lifespan and health span

Abstract Nonagenarians and centenarians serve as successful examples of aging and extended longevity, showcasing robust regulation of biological mechanisms and homeostasis. Given that human longevity is a complex field of study that navigates molecular and biological mechanisms influencing aging, we hypothesized that microRNAs, a class of small noncoding RNAs implicated in regulating gene expression at the post‐transcriptional level, are differentially regulated in the circulatory system of young, middle‐aged, and nonagenarian individuals. We sequenced circulating microRNAs in Okinawan males and females <40, 50–80, and >90 years of age accounting for FOXO3 genetic variations of single nucleotide polymorphism (SNP) rs2802292 (TT ‐ common vs. GT ‐ longevity) and validated the findings through RT‐qPCR. We report five microRNAs exclusively upregulated in both male and female nonagenarians with the longevity genotype, play predictive functional roles in TGF‐β, FoxO, AMPK, Pi3K‐Akt, and MAPK signaling pathways. Our findings suggest that these microRNAs upregulated in nonagenarians may provide novel insight into enhanced lifespan and health span. This discovery warrants further exploration into their roles in human aging and longevity.


| INTRODUC TI ON
Centenarians and nonagenarians demonstrate successful aging, a phenomenon described as aging devoid of advanced disease, impaired cognitive function and reduced physical ability (Engberg et al., 2009).In addition to exhibiting improved health span, centenarians surpass the average life expectancy by over two decades (Engberg et al., 2009).Across the globe, Okinawa ranks among the highest for both life expectancy and prevalence of centenarians, with a frequency of 6 centenarians per 10,000 individuals (Buchholz, 2021;Willcox, Willcox, et al., 2008).Long-lived Okinawans also demonstrate reduced signs of age-associated diseases, eliciting interest in identifying genetic and environmental contributors to the population's overall improved health and longevity (Bernstein et al., 2004).Furthermore, Okinawan centenarians and long-lived nonagenarians exhibit advanced regulation of biological mechanisms and homeostasis, and more specifically, display greater insulin sensitivity.
Their improved insulin sensitivity has been somewhat attributed to variations of the forkhead box 03 (FOXO3) gene (Willcox, Donlon, et al., 2008).The FOXO3 gene encodes a key regulator of the insulin-IGF-1 signaling pathway, central to metabolic homeostasis and notably deregulated with age (López-Otín et al., 2013).A single-nucleotide polymorphism (SNP) in the FOXO3 gene was previously identified to be linked with longevity in Okinawan centenarians and nonagenarians.Carriers of the G allele are associated with improved health at baseline (Willcox, Donlon, et al., 2008).These findings were supported by a separate study investigating FOXO3 SNPs in German centenarians and nonagenarians that concluded FOXO3 is also associated with human longevity (Flachsbart et al., 2009;Frankum et al., 2022).
In addition to evaluating the role of FOXO3 in longevity, there is growing support for the role of metabolic pathways-including Insulin-IGF-1 signaling-in aging.Thus, investigating the molecular and genetic factors that modulate metabolic processes could provide insight into the mechanisms that support extended lifespan in nonagenarians.A class of small noncoding RNAs known as microRNAs (miRNAs) repress the translation of sequence-compatible messenger RNAs (mRNAs) through destabilization or degradation of the target mRNA and/or through 3′UTR (untranslated region) binding (Nunes et al., 2022).Therefore, several pathways can be regulated through this mechanism, including longevity (Marchegiani et al., 2023).
Given the potential for miRNAs to be used as biomarkers of aging or anti-aging therapies (Jueraitetibaike et al., 2023), we were interested in identifying differentially expressed circulating miRNAs and their pathway interactions in Okinawan nonagenarians.To address this, we sequenced, and validated through qPCR, circulating miRNAs in Okinawans <40, 50-80, and ≥90 years of age accounting for differences in age, sex, and FOXO3 rs2802292 genotype (TT and TG).S1.This study was conducted following approval by the Ethics Committees from Tomishiro Central Hospital (H25R008), Fukushima Medical University (#30167).Hence, based on the inclusion-exclusion criteria, n = 76 participants were included in this study.A breakdown of number of participants and age distribution is presented in Figure 1.

| Participant inclusion and serum collection
To isolate serum from included participants, 2 mL of Dulbecco's phosphate buffered saline (DPBS) with calcium and magnesium (ThermoFisher Scientific, Waltham, Massachusetts, USA) were added to 2 mL of whole blood and mixed at room temperature.Three milliliter of Ficoll-Paque PREMIUM (GE Healthcare/Cytiva, Chicago, Illinois, USA) were added to a new 10 mL centrifuge tube.Four milliliter of the diluted blood sample was carefully layered on top of the Ficoll-Paque Premium, ensuring that there was no mixing between the layers, before being centrifuged at 400× g for 40 min with the brake turned off.After centrifugation, the upper layer containing the plasma was removed.

| RNA isolation
200 μL of serum were homogenized with 5 volumes of QIAzol Lysis Reagent, combined with chloroform and mixed vigorously to ensure subsequent phase separation.Samples were incubated for 3 min and centrifuged at 12,000× g at 4°C for 15 min.Following phase separation, RNA was extracted using the QIAGEN miRNeasy Serum/ Plasma kit (Hilden, Germany).All steps were performed in accordance with the manufacturer's protocol.

| Library Prep and small RNA sequencing
To prepare libraries for sequencing, 10.5 μL of total undiluted RNA was combined with the NEXTFLEX Small RNA Seq.kit V3 reagents as per the manufacturer's protocol (Perkin-Elmer, Waltham, Massachusetts, USA).Samples were purified using the gel-selection method in a 6% TBE-PAGE gel.Gel electrophoresis was performed at 145 V for approximately 30 min.Bands were visualized on a UV transilluminator following staining with SYBR Gold (Invitrogen, Waltham, Massachusetts) in accordance with the manufacturer's instructions.
Bands visualized at approximately 150 bp were excised, homogenized with disposable pestles, and incubated overnight at room temperature with agitation in the NEXTFLEX Elution Buffer (Perkin-Elmer, Waltham, Massachusetts).Homogenized gel fragments and eluent were then transferred to Corning® Costar® Spin-X® centrifuge tube filters (Glendale, Arizona, USA) following overnight incubation and centrifuged at 16,000× g for 2 min.The collected eluent was then purified using the QIAquick Gel Extraction Kit (QIAGEN, Hilden, Germany) as per the user-developed protocol by Sambrook et al. (1989).The final libraries were then outsourced for quality check, pooling, and small RNA sequencing via Illumina HiSeq 2x150 system (Admera Health).Based on QC results, at least n = 4 per sex and age group was selected for pooling/sequencing.Samples were multiplexed and run on 2 separate lanes for a total of 32 libraries.

| sRNA seq fold change and relative expression
Alignment and quantification of miRNA libraries was performed using cutadapt and bowtie2 (Langmead et al., 2019) as described previously (Rueda et al., 2015).Statistical analyses of differentially expressed miRNAs was performed using EdgeR (Robinson et al., 2010) on the R software (3.2.2). miRNAs with a p-value <0.05 and FC >2.0 were considered as upregulated while miRNAs with a p < 0.05 and FC < 0.50 were considered as downregulated.A summary of the differentially expressed miRNAs is shown in Figure 2a.A principal component analysis was performed to determine the variations in the datasets.This analysis established a PC1 of 20.8% and a PC2 of 9.7% (Figure 2b).Furthermore, volcano plots were derived from the sequencing datasets as shown in Figure S1.Based on these analyses, a narrowed list of miRNAs differentially expressed with age and/or changes in genotype was derived for further analysis through RT-qPCR using the full sample set of n = 76.

| cDNA Prep and quantitative real time PCR
To validate differentially expressed miRNAs derived post-small RNAseq using all samples from included participants (n = 76), cDNA was synthesized from total RNA using the Applied Biosystems™ TaqMan™ Advanced miRNA cDNA Synthesis Kit (ThermoFisher Scientific, Waltham, Massachusetts, USA) for subsequent quantitative real-time PCR (RT-qPCR).Each reaction step was performed using the Bio-Rad   S2).
RT-qPCR was normalized using miR-16 as the housekeeping gene with a maximum range of 3 miR-16 C T values between all samples.
Relative expression of each miRNA was then calculated using the −∆∆C T method as described previously (Noureddine et al., 2023;Piotrowski et al., 2021).

| Prediction of microRNA pathway interactions and gene targets
To evaluate predictive pathway associations for differentially expressed microRNAs validated through RT-qPCR, the DIANA Tool miRPath v3 was used to generate a comprehensive list of gene and pathway targets for each microRNA using the micro-T-CDS v5 database (Paraskevopoulou et al., 2013;Vlachos et al., 2015).To further demonstrate miRNA-gene interactions, MIENTURNET, a miRNAtarget enrichment and network-based analysis tool was utilized (Licursi et al., 2019).
These miRNAs are predicted to have functional roles in regulating pluripotency of stem cells, FoxO, Ras, Pi3K-Akt, MAPK, mTOR, TGFβ, p53, and AMPK signaling pathways as well as long-term depression, suggesting the middle-aged and elderly group demonstrate increased expression of nutrient-sensing pathways and onset of long-term depression when compared to the nonagenarian cohort (Table S7).
On the other hand, miR-22-3p, miR-450b-5p, miR-93-3p, and let-7i-3p were upregulated in the 50-80 age group (Table S6) and appear to target genes involved in similar pathways.However, the number of genes targeted are less than the genes targeted by the upregulated miRNAs, suggesting the overall outcome is likely upregulation of the pathways targeted by the downregulated miRNAs (Tables S7 and S8).
Furthermore, sRNA-seq analysis comparing the <40 group to individuals 50-80 years of age revealed downregulation of 8 miRNAs and upregulation of 20 miRNAs (Table S9) in the younger cohort.
Pathway analysis revealed the group of downregulated miRNAs possibly function by repressing genes involved in regulating pluripotency of stem cells and TGFβ signaling (Table S10).On the other hand, the upregulated miRNAs demonstrate predictive function in regulating FoxO, Ras, TGFβ, Pi3K-Akt, MAPK, mTOR, insulin signaling, and type II diabetes mellitus (Table S11).
Together, these results indicate that miRNA expression patterns greatly differ with age and could offer insight into the regulation of pathways involved in lifespan-extension.

| sRNA-sequencing reveals genotype and miRNA associations across Okinawan cohorts
Okinawan nonagenarians exhibit enhanced insulin sensitivity that has been attributed to the FOXO3 genetic variations.Previous studies have demonstrated a correlation between the G allele and improved health (Willcox, Donlon, et al., 2008).Hence, to account for genotypic influence on circulating miRNA expression patterns, we analyzed differentially expressed miRNAs in young, middle aged-elderly, and nonagenarian Okinawan individuals with either TT or TG alleles.

| D ISCUSS I ON AND CON CLUS I ON
miRNAs are important regulators of biological and metabolic processes (Kinser & Pincus, 2019).Prior studies have demonstrated varying roles of let-7 in aging, with a perceived increase in insulin resistance, as well as onset of cardiovascular and age-related diseases associated with overexpression of let-7 (Wang et al., 2022).However, loss of let-7 has conversely been shown to reduce lifespan in C. elegans (Xu et al., 2014).Hence, the role of this miRNA family in aging appears to be multifaceted.
Our findings demonstrate an increase in the let-7a and let-7b miRNAs encoded by the let-7 gene.Based on our predictive pathway analysis, these miRNAs target various components of the insulin/IGF-1 and mammalian target of rapamycin (mTOR) pathways.These pathways are notoriously de-regulated with age and have been widely implicated in promoting aging and onset of age-related diseases (López-Otín et al., 2013).Inhibition of insulin/IGF-1 and mTOR signaling pathways can extend lifespan with prior research demonstrating improved lifespan directly attributed to inhibition of mTOR (Johnson et al., 2013).Some studies have also established that let-7a-targeted regulation of important genes such as P66shc leads to delayed senescence in human dermal fibroblasts (Xu et al., 2014).Coupled with the apparent role in targeting nutrient sensing genes Igf-1, Igf-1-receptor (Igf1r), insulin receptor substrate (Irs) 1 and 2, Homer scaffold protein 1 (Homer1) as well as senescence-related genes Cdkn1a (p21 cip1 ) and Tgfβ receptor (Tgfβr), microRNAs let-7a and let-7b exhibit important regulatory contributions to aging and are uniquely upregulated in Okinawan nonagenarians, warranting further investigations into their roles in metabolism and longevity.

KEGG pathway p-value #genes #miRNAs
Signaling pathways regulating pluripotency of stem cells to inhibition of the Pi3K-AKT pathway, a nutrient sensing pathway typically associated with promoting aging through the activation of mTOR (Noureddine et al., 2023;Stallone et al., 2019;Wang et al., 2020).Hence, miRs-98-5p, −let-7a, and −let-7b's apparent roles in metabolism may contribute to the observed enhanced insulin sensitivity and overall improved health of the nonagenarian cohort.
In addition to metabolism, lifespan and health span are both impacted by the accrual of senescent cells with age.As a hallmark of aging, cellular senescence contributes to accelerated aging and chronic inflammation primarily through the senescence-associated secretory phenotype (SASP) (López-Otín et al., 2013).This stressresponse can be triggered by TGFβ signaling which involves the activation of the Smad transcription complex (comprised of Smad2 and Smad4) necessary for mediating the signaling cascade that promotes the transcription of genes related to cellular proliferation, senescence, and the production of the SASP (Tominaga & Suzuki, 2019).TGFβ also promotes radical oxygen species (ROS) production, DNA damage, as well as the suppression of telomerase activity both dependently and independently of Smad activation (Liu & Desai, 2015;Tominaga & Suzuki, 2019).Taken together, TGFβ signaling can contribute significantly to aging and age-associated diseases.Interestingly, miR-92a-1-5p, significantly upregulated in circulation in Okinawan nonagenarians, is predicted to target Smad2 post-transcriptionally, thereby eliciting its important potential function in regulating TGFβ signaling and cellular senescence.However, miR-92a-1-5p has also been predicted to target Prkaβ2, the gene encoding the human AMPK-β2 subunit-a nutrient-sensing gene that plays an important role in AMPK signaling (Ziegler et al., 2020).
AMPK activation/responsiveness typically declines with age, leading to reduced autophagy and increased oxidative stress (Stancu, 2015).from DIANA Tools and predicted gene targets were determined using the microT-CDS target prediction algorithm (Vlachos et al., 2015).For full list of pathways, refer to Table 1.For miRNA-gene interactions, refer to Figure S6.
Hence, the role of miR-92a-1-5p in aging requires further investigation specifically into its involvement in senescence as well as its potential regulatory function in AMPK signaling.
Similarly, miR-181a-2-3p is expected to be involved in regulating senescent cell survival through its ability to bind and inhibit the activity of serum response factor (srf).SRF binds the ETS (Erythroblast Transformation Specific) domain of transcription factor serum response element (SRE) to mediate gene expression of essential MAPK signaling genes (Whitmarsh et al., 1995).MAPKs promote key factors necessary for the persistence of senescent cells, such as growth arrest, cell survival, and the SASP, acting as important mediators of senescent cell survival (Anerillas et al., 2020).Taken together, upregulation of miR-92a-1-5p and miR-181a-2-3p could be contributing to the enhanced longevity of the nonagenarian cohort through regulation of cellular senescence and pathways involved in maintaining both the SASP and SnC survival.
mTOR has also been widely implicated in accelerating aging and aggravating age-related pathologies primarily through its ability to promote increased cell survival and proliferation coupled with inhibition of autophagy.Increased mTOR expression (through the activation of the Pi3K-AKT signaling cascade) can lead to protein aggregates and mitochondrial dysfunction with subsequent ROS production and cellular stress (Stallone et al., 2019).Hence, mTOR is a significant pathway to regulate and an important target for age-related interventions.Further, it has been well established that targeting mTOR successfully decelerates aging and reduces senescent cell accumulation (Weichhart, 2019).miR-100 targets and suppresses the translation of mTOR post-transcriptionally, therefore its therapeutical potential has been explored (Saccon et al., 2021).
Prior research has also demonstrated that miR-100 overexpression in mice attenuates high-fat diet-induced weight gain, liver steatosis, and metabolic syndrome (Smolka et al., 2021).These mice also have improved insulin sensitivity and energy expenditure, possibly mediated by miR-100.Of interest to this study, miR-100 was found to be uniquely upregulated in circulation of the Okinawan nonagenarian cohort, suggesting that these individuals exhibit improved regulation of mTOR and nutrient sensing pathways, a notion that may justify their overall improved metabolism and insulin sensitivity.
In short, our data demonstrates potential links between miRNAs let-7a, let-7b, miR-92a-1-5p, miR-98- Male and female participants (n = 320) ranging in age from 19 to 104 years were recruited during annual health examinations, following written informed consent, through Tomishiro Central Hospital (Tomishiro City, Okinawa, Japan) and affiliated clinics and facilities throughout the Okinawa prefecture.Subjects were recruited during nationally required annual health screening examinations, thereby mitigating potential bias toward healthier participants.Recruitment occurred between May 2018 and July 2019.Principal inclusion criteria focused on healthy individuals or individuals with non-severe medical conditions over the age of 18 years.Subjects were excluded from participation if they were (a) aged <18 years, (b) had a severe medical condition, such as diabetes mellitus with glycated hemoglobin A1c ≥ 8.5%; recent (within a year) onset of ischemic heart disease, cerebrovascular disease, and cancer, (c) were taking anti-diabetic, lipid-lowering or anti-hypertensive medications within 18 h prior to the blood sampling that potentially affect microRNA measurements, (d) exhibited severe dementia or an inability to comprehend the informed consent, (e) had a known genetic disease or disability, or (f) were restricted from participation by the subject's attending physician.Demographic characteristics of participants are shown in Table Number of male and female participants per age group.(b) Age distribution of males and females per age group.

F
Schematic representation of predicted pathway interactions of microRNAs upregulated in individuals >90 years of age, present only in those subjects with the FOXO3 longevity genotype (GT).Inhibitory effects of microRNAs depicted represent posttranscriptional regulation of gene expression.Figure was made with BioRender and KEGG pathways.Pathway interactions were derived span and lifespan observed in Okinawan nonagenarians based on the predicted regulatory functions of upregulated miRNAs further emphasizing the importance of metabolism and nutrient sensing pathways in aging.Our findings also support the correlation between FOXO3 genetic carriers of the G allele and improved metabolism, as observed by the genotype and miRNA associations predictively demonstrating improved regulation of nutrient sensing pathways in individuals with TG alleles versus TT alleles.Our study limitations include variable age distribution across males and females as well as sample size per group.Future considerations include narrowing age ranges per cohort to further specify changes in miRNA expression with age as well as verifying the miRNA pathway interactions presented in this study to solidify their potential pro-longevity functions.It would also be interesting to explore their capacity to act as biomarkers to predict regulation patterns of key aging pathways.Overall, our results provide a strong rationale for added investigation into the mechanistic functions of the miRNAs uniquely upregulated in circulation in Okinawan nonagenarians and their potential to be used in future therapies.AUTH O R CO NTR I B UTI O N S Concept and design, acquisition of data, analysis, interpretation of data, manuscript original draft, manuscript revision: Sarah Noureddine.Interpretation of data, analysis, manuscript original draft, manuscript revision: Augusto Schneider.Acquisition of data, analysis, manuscript revision: Sydney Strader.Data analysis, interpretation of data, manuscript original draft, manuscript revision: Xiang Zhu.Funding acquisition, analysis, manuscript revision: Joseph Dhahbi.Concept and design, acquisition of samples, cohort recruitment, manuscript revision: Richard Allsopp.Concept and design, acquisition of samples, cohort recruitment, manuscript revision: D. Craig Willcox.Interpretation of data, manuscript revision: Timothy A Donlon.Concept and design, acquisition of samples, cohort recruitment, manuscript revision: Michio Shimabukuro.Acquisition of samples, cohort recruitment, manuscript revision: Moritake Higa.Concept and design, acquisition of samples, cohort recruitment, manuscript revision: Makoto Suzuki.Concept and design, acquisition of samples, manuscript revision: Trevor Torigoe.Interpretation of data, manuscript revision: Sarah Ashiqueali.Concept and design, acquisition of samples, cohort recruitment, manuscript revision: Hariom Yadav.Concept and design, acquisition of samples, cohort recruitment, manuscript revision: Bradley J. Willcox.Concept and design, funding acquisition, acquisition of data, analysis, interpretation of data, manuscript original draft, manuscript revision: Michal M. Masternak.ACK N OWLED G EM ENTS This work was supported in part by the National Institute on Aging (R56AG074499 (MMM)), the National Institutes of Health (R56AG069676, R56AG064075, RF1AG071762, R21AG072379, U01AG076928, R21DE032197 HY) (R5R01AG027060 BW, RA, DCW and 1P20GM125526 BW, RA, DCW), the Department of Defense (W81XWH-18-PRARP AZ180098 (HY)), National Science Foundation (Award Number (FAIN): 2317758 (MM)) and the Ed and Ethel Moore Alzheimer's Disease Research Program of the Florida