RNY-derived small RNAs as biomarkers for atherosclerosis-related disorders

Abstract

The present invention concerns an in vitro method of diagnosis or prognosis of an atherosclerosis-related disorder by detecting a small Y RNA (s-RNY), as well as the use of an inhibitor of s-RNY as a medicament against atherosclerosis-related disorders. The invention also concerns a method for screening for a compound suitable for the treatment of an atherosclerosis-related disorder.

Claims

1. A method of treating an atherosclerosis-related disorder selected from the group consisting of atherosclerosis, chronic kidney disease, cerebrovascular disease, peripheral vascular disease, ischemic heart disease, carotid artery disease, and coronary artery disease in an individual in need thereof, which comprises: a) diagnosing or prognosing the individual with an atherosclerosis-related disorder by measuring the level of expression of at least one biomarker consisting of a small Y RNA (s-RNY) in a biological sample of said individual, wherein a level of expression of said biomarker higher than a reference value is indicative that the individual has developed or is at risk of developing an atherosclerosis-related disorder; and b) administering a therapeutically effective amount of an anti-atherosclerosis drug to said individual diagnosed or prognosed with an atherosclerosis-related disorder, wherein said anti-atherosclerosis drug is selected from the group consisting of bile acid sequestrant, niacin, statins, fibrates, probucol, an s-RNY inhibitor, and combinations thereof.

2. The method of claim 1, wherein step a) comprises: a) determining the level of expression of at least one biomarker consisting of a small Y RNA (s-RNY) in a biological sample of said individual; b) comparing the level of expression of said at least one biomarker with a reference value, wherein a level of expression of said biomarker higher than the reference value is indicative that the individual has developed or is at risk of developing an atherosclerosis-related disorder, and c) deducing for said comparison if the individual has developed or is at risk of developing an atherosclerosis-related disorder.

3. The method of claim 1, wherein in step a) said at least one biomarker is selected from the group consisting of s-RNY1-5p (SEQ ID NO: 7), s-RNY1-3p (SEQ ID NO: 8), s-RNY3-5p (SEQ ID NO: 9), s-RNY3-3p (SEQ ID NO: 10), s-RNY4-5p (SEQ ID NO: 11), s-RNY4-3p (SEQ ID NO: 12), and s-RNY5-3p (SEQ ID NO: 13) or variants thereof.

4. The method of claim 1, wherein said atherosclerosis-related disorder is coronary artery disease.

5. The method of claim 1, wherein said anti-atherosclerosis drug is an inhibitor of a s-RNY selected from the group consisting of s-RNY1-5p (SEQ ID NO: 7), s-RNY1-3p (SEQ ID NO: 8), s-RNY3-5p (SEQ ID NO: 9) and S-RNY4-5p (SEQ ID NO: 11).

6. The method according to claim 5 wherein said inhibitor of a s-RNY is a nucleic acid which specifically hybridizes to at least one sequence selected from the group consisting of SEQ ID NO: 7 (s-RNY1-5p), SEQ ID NO: 9 (s-RNY3-5p), SEQ ID NO: 8 (s-RNY1-3p) and SEQ ID NO: 11 (s-RNY4-5p).

Description

BRIEF DESCRIPTION OF THE FIGURES

(1) FIG. 1: (A) Bioinformatic analysis of the high-throughput small RNAs sequencing showing the distribution of different classes of small RNAs identified and the upregulation of small RNAs derived from RNYs. (B) (s-RNYs) in human primary macrophages stimulated with 1 M of staurosporine (STS) at the indicated time points.

(2) FIG. 2: (A) Bioinformatic analysis of small RNA-Seq data sets identifying a consensus sequence of small RNAs (the s-RNYs) derived from the RNYs in bone marrow derived-macrophages (BMDMs) after M-CSF withdrawal treatment for 36 hr. The bar graph on the top shows the localization of s-RNYs in RNYs. (B) Secondary structure models of mouse RNYs using RNAfold web server (http://rna.tbi.univie.ac.at/cgi-bin/RNAfold.cgi). Arrowheads indicate the mature s-RNYs, arrows indicate the potential cleavage sites of recombinant Ago2 in an in vitro processing assay. Putative hnRNP A1 binding sites are highlighted.

(3) FIG. 3: (A) Quantitative RT-PCR analysis of the indicated s-RNYs in BMDMs incubated for 18 hr with 10 mg/ml LTA (Lipoteichoic acid) alone or in combination with 0.25 M thapsigargin (Tg). The data were normalized by U2 snRNA. (B) Quantitative RT-PCR analysis of the indicated s-RNYs in BMDMs incubated for 28 hr with 0.25 M of thapsigargin (Tg) alone or thapsigargin in combination with the indicated concentrations of oxidized-LDL (oxLDL). The data were normalized by U2 snRNA and presented as mean and s.d. (n=8). (C) Quantitative RT-PCR analysis of the indicated s-RNYs in BMDMs incubated for 28 hr with 0.25 M Tg in combination with the indicated concentration of acetylated-LDL (acLDL). The data were normalized by U2 snRNA.

(4) FIG. 4: (A) Quantitative RT-PCR analysis of the indicated s-RNYs in BMDMs incubated for 18 hr with 0.25 M of thapsigargin (Tg) with BSA, 0.25 mM of the indicated fatty acids in complex with BSA, or thapsigargin plus the fatty acids. The data were normalized by U2 snRNA and presented as mean and s.d. (n=8). The unsaturated fatty acids were linoleic acid (LA) and oleic acid (OA), and the saturated fatty acids were stearic acid (SA) and palmitic acid (PA). Quantitative RT-PCR analysis of the indicated s-RNYs in control or ApoE.sup./ (B) or Ldlr.sup./ (C) aortic arches. The mice were fed with either chow diet or high cholesterol diet (HCD). The data were normalized by U2 snRNA and presented as mean and s.d. (n=8 for B and n=5 for C). Student's t-test: *P<0.05; **P<0.01.

(5) FIG. 5: Effectiveness of the knockdown mediated by SMARTpool siRNAs purchased from Thermo Scientific was assessed in BMDMs. After 48 hr from the transfection total RNA was isolated and analyzed by quantitative RT-PCR analysis. Data were normalized by U2 snRNA and presented as mean and s.d. (n=4). Student's t-test: **P<0.01.

(6) FIG. 6: Depletion of known regulators of small RNA biogenesis pathways and RNA metabolism shows that loss of La/SSB, Ro60, p53, Ago2, and hnRNP A1 (all in light grey) leads to a statistically significant reduction (Student's t-test: P<0.05) of the indicated s-RNYs, while depletion of Dicer leads to a significant induction (Student's t-test: P<0.05). BMDMs were left unstimulated or stimulated with 0.25 mM of PA for 18 hr, and total RNA was isolated and analyzed by quantitative RT-PCR. The data were normalized by U2 snRNA. Data are presented as mean and s.d. (n=4).

(7) FIG. 7: RNA immunoprecipitation analysis using the indicated antibodies and the indicated RNAs in BMDMs. BMDMs were left unstimulated or stimulated with 0.25 mM of PA for 18 hr and total RNA was isolated and analyzed by quantitative RT-PCR. Data are presented as mean and s.d. (n=6) and normalized with the IgG and the input.

(8) FIG. 8: (A) RNA immunoprecipitation analysis of Ro60 and the indicated s-RNYs in BMDMs. BMDMs were left unstimulated or stimulated with 0.25 mM of PA for 18 hr. RNA was isolated from immunoprecipitation and analyzed by quantitative RT-PCR. Data are presented as mean and s.d. (n=6) and normalized with the IgG and the input. Student's t-test: **P<0.01. (B) RNA immunoprecipitation of either Ago2 or La/SSB and the indicated s-RNYs in BMDMs. BMDMs were left unstimulated or stimulated with 0.25 mM of PA for 18 hr. RNA was isolated from immunoprecipitation and analyzed by quantitative RT-PCR. Data are presented as mean and s.d. (n=6). Student's t-test: **P<0.01.

(9) FIG. 9: (A) RNA immunoprecipitation of either Ago2 or IgG and the indicated RNYs in BMDMs. BMDMs were transfected with the indicated siRNAs and either left unstimulated or stimulated with 0.25 mM of PA for 18 hr. RNA was isolated from immunoprecipitation and analyzed by quantitative RT-PCR. Data are presented as mean and s.d. (n=6). Student's t-test: **P<0.01. (B) RNA immunoprecipitation analysis of GW182 and the indicated s-RNYs. BMDMs were left unstimulated or stimulated with 0.25 mM of PA for 18 hr. RNA was isolated from immunoprecipitation and analyzed by quantitative RT-PCR. Data are presented as mean and s.d. (n=6) and normalized with the IgG and the input.

(10) FIG. 10: (A) Venn diagram showing the potential s-RNYs target-mRNAs among the downregulated mRNAs detected from a microarray analysis of BMDMs overexpressing s-RNYs and control from 4 independent experiments. (B) Venn diagram showing the number of potential target-binding sites for each s-RNYs in both coding region and 3UTR of downregulated mRNAs of s-RNYs-overexpressing BMDMs.

(11) FIG. 11: Quantitative RT-PCR analysis of the potential s-RNYs-target mRNAs in BMDMs transfected with siAgo2, 2-OMe-RNA antisense oligonucleotides (AS) to s-RNYs, or control. BMDMs were left unstimulated (black column) or stimulated with 0.25 mM of PA for 18 hr, and total RNA was isolated and analyzed. Data were normalized by U2 snRNA and are presented as mean and s.d. (n=6).

(12) FIG. 12: (A) RNA duplex expected to result from base pairing of KFL2 mRNA with s-RNY1-3p. A-U and G-C base pairs are represented by solid lines; G-U wobble base pairs are represented by dotted lines, shown herein as SEQ ID NOS 58 and 59, respectively. (B) The sequence of the predicted target site of KFL2 mRNA from the indicated species, shown herein as SEQ ID NOS 60-63, respectively. The s-RNY1-3p seed complementary sequence is enclosed by a black box. Nucleotide conservation across the species is indicated by asterisks. (C) Relative luciferase activity of reporter constructs containing two either wt or mutant s-RNY1-3p binding sites (BS) from KFL2 mRNA sequence in NIH-3T3 cells transfected with either 2-OMe-RNA antisense oligonucleotide to s-RNY1-3p or control. Cells were left unstimulated or stimulated with 0.25 mM of PA for 18 hr. The data were normalized using Renilla activity. Data are presented as mean and s.d. (n=4). Student's t-test: **P<0.01.

(13) FIG. 13: Apoptotic data by fluorocytometry for BMDMs overexpressing s-RNYs, transfected with either 2-OMe-RNA antisense oligonucleotides to s-RNYs or control. BMDMs were stimulated with 0.25 mM of PA for 18 hr (A), 0.25 M thapsigargin (Tg) in combination with 100 M of oxLDL for 28 hr (B), or left unstimulated (C). Percentage of apoptotic and necrotic cells was determined by staining with Dapi and annexin V-Alexa568. Data are presented as mean and s.d. (n=3).

(14) FIG. 14: (A) Quantitative RT-PCR analysis of the indicated s-RNYs in BMDMs incubated with 0.25 mM of PA the indicated time. The data were normalized by U2 snRNA. (B) Quantitative RT-PCR analysis of the indicated s-RNYs from the medium of BMDMs. Cells were left unstimulated or incubated with 0.25 mM PA for 18 hr. Data were normalized by using the synthetic non-mammalian cel-miR-39, which was added to the samples before the RNA extraction. Data are presented as mean and s.d. (n=8).

(15) FIG. 15: Quantitative RT-PCR analysis of NOS2A, IL-6, IL-12b, and IL-1b transcripts in BMDMs overexpressing s-RNYs and transfected with either 2-OMe-RNA antisense oligonucleotides to s-RNYs or control. BMDMs were left unstimulated or stimulated with 100 ng/ml LPS for 12 hr, and total RNA was isolated and analyzed. Data were normalized by U2 snRNA and are presented as mean and s.d. (n=4). Student's t-test: *P<0.05; **P<0.01.

(16) FIG. 16: Quantitative RT-PCR of the indicated s-RNYs in the blood plasma of ApoE.sup./ (A) or Ldlr.sup./ (B) mice or control. Mice were fed with either chow diet or HCD for either 12 (ApoE.sup./ mice) or 20 (Ldlr.sup./ mice) weeks. Data were normalized by using cel-miR-39 and are presented as mean and s.d. (n=8 for ApoE.sup./ and n=5 for Ldlr.sup./ mice). Student's t-test: *P<0.05, **P<0.01.

(17) FIG. 17: (A) Significant different expression of circulating s-RNY1-5p in the serum of patients with Coronary Artery Disease (CAD, n=45) versus healthy control (CTR, n=45). Data derived from quantitative RT-PCR, normalized using cel-miR-39, and presented as mean and s.d. Student's t-test: ***P<0.001. (B) Receiver Operating Characteristic (ROC) curve for predicting CAD. The ROC curve was constructed using s-RNY1-5p based on the quantitative RT-PCR data (n=45). The area under the ROC curve is given and 95% confidence interval was calculated (0.90-0.99). The diagonal line indicates a test with an area under the ROC curve of 0.5.

(18) FIG. 18: Significant different expression of circulating s-RNY4-5p in the serum of patients with Coronary Artery Disease (CAD) versus control (CTR). (A) Data derived from quantitative RT-qPCR, Normalized by using cel-miR-39 and presented as mean and s.d. (CTR n=19; CAD n=30). (B) Area under the Receiver Operating Characteristic (ROC) curve for s-RNY4-5p based on the Quantitative RT-PCR data.

(19) FIG. 19: s-RNYs are expressed in the peripheral blood of patients with coronary artery disease. A model for atherogenic lipids-dependent generation of s-RNYs that act on a subset of mRNAs containing seed complementary sequences and exert signaling pathway effects.

(20) FIG. 20: Box plots of s-RNY1_5P according to coronary artery disease (CAD) status.

(21) FIG. 21: Normal density plots of crude and transformed s-rny1_5p.

(22) FIG. 22: Receiver-operator characteristic (ROC) curve for prediction of CAD with s-RNY1-5p.

(23) FIG. 23: ROC curve for prediction of CAD with hypertension, dyslipidemia, diabetes, smoking, school duration, ankle-arm index, ApoA1 and CRP.

(24) FIG. 24: ROC curves for prediction of CAD Comparison of basic cardiovascular risk factors model with full model (+s-RNY1_5P).

(25) FIG. 25: Odds ratio for s-RNY1_5P.

EXAMPLES

Example 1: Material and Methods

(26) Animals and Diets

(27) Animals were kept in a pathogen-free barrier facility and maintained in accordance with Institutional Animal Care and Use Protocol of University of Nice Sophia Antipolis in accordance with appropriate national regulations concerning animal welfare. The following mice were purchased from Charles River Laboratories (L'Arbresle, France): C57BL/6J, ApoE/ (B6.129P2-APOE/J), and Ldlr/ (B6.129S7-Ldlrtm1Her/J). High Cholesterol Diet (HCD) formula # TD02028 and TD96335 (Harlan Teklad) for ApoE/ or Ldlr/, respectively, were purchased from ssniff Spezialdiaten GmbH (Soest, Germany). ApoE/ and Ldlr/ male mice at 8 weeks of age were fed with either HCD or regular diet (chow diet) for 12 or 20 weeks, respectively. Aortic arches, heart, and blood plasma were dissected.

(28) Study Population and Data Collection

(29) Subjects were randomly selected from a large case-control study on coronary artery disease (CAD) referenced as GENES study (Genetique et Environnement en Europe de Sud) as previously described (Bouisset et al., 2012). Briefly, CAD male patients living in the Toulouse area (Southwestern France) and aged between 45 and 74 years, were prospectively recruited from 2001 to 2004, admitted in the Department of Cardiology of the Toulouse University Hospital, and referred for evaluation and management of stable CAD. Stable CAD was defined by a history of acute coronary syndrome and of coronary revascularization, documented myocardial ischemia, stable angina, or the presence on the coronary angiogram of a coronary stenosis of >50%. During the same period, healthy male controls, aged 45 to 74, were selected from the general population using the electoral rolls. Stratification by 10-age group was used to approximately match the age distribution of the controls with that of cases. All individuals underwent medical examination in the Department of Cardiology in the Toulouse University Hospital and completed standardized questionnaires covering age, medical history, socioeconomic and lifestyle variables such as educational level, smoking status and physical activity. Physical activity was categorized into five levels from 1 to 4: no physical activity (1), light physical activity for 20 min once a week (2), moderate physical activity for 20 min twice a week (3), intense physical activity for 20 min 3 times a week or more (4). They also underwent a standardized medical examination with anthropometric and clinical measurements and provided fasting blood samples. Blood samples were analyzed for serum total cholesterol, high-density lipoprotein-cholesterol (HDL-C), triglycerides (TG), glucose, -glutamyltransferase (-GT) and sensitive C-reactive protein (CRP) with enzymatic reagents on an automated analyser (Hitachi 912, Roche Diagnostics, Meylan, France). LDL-cholesterol (LDL-C) was calculated using the Friedewald formula, with VLDL-cholesterol (VLDL-C) (g/L)=TG (g/L)/5 as long as TG concentration was below 4 g/L (Genoux et al., 2011). Lipoproteins containing ApoB and ApoE (LpB:E) were assayed by a specific immuno-electrodiffusion assay (Sebia, France). In the present work, 45 CAD patients and 45 age-matched control individuals were randomly drawn for serum s-RNYs measurements.

(30) In the tables 3 and 4, data are presented as percentage for qualitative variables or as mean with standard deviation for quantitative ones. Qualitative variables were compared with .sup.2 test (or Fisher exact test when necessary). The mean values of quantitative variables were compared to Student's t-test. Shapiro-Wilks was used to test the normality of distribution of residuals and Levene's test to check the homogeneity of variables. When the basic assumptions of Students t-test were not satisfied, the data were logarithmically transformed or subjected to a Wilcoxon Mann Whitney's test. Associations of s-RNYs with clinical characteristics and biological markers were tested using Spearman's rank correlations. Analyses were two-tailed and p<0.05 was considered to be significant. These statistical analyses were carried out using the SAS statistical software package 9.2 (SAS Institute, Cary, Ill., USA) and with STATA statistical software (release 11.2, Stata Corporation, College Station, Tex., USA). ROC test was calculated using GraphPad Prism (version 5.01, La Jolla, Calif. 92037, USA).

(31) Ethics Statement

(32) The study protocol was approved by the local Ethics Committee of the hospital of Toulouse (CHU Toulouse/INSERM, file 1-99-48, February 2000) and the national commission for data processing and freedoms (N 900165). Written informed consent was obtained from all participants involved in the study. Biological collection was constituted according to the principles expressed in the Declaration of Helsinki and registered under number DC-2008-403 at the Ministry of Research and at the Regional Health Agency.

(33) Reagents

(34) LPS, LTA, Tg, STS, BSA, PA, stearic acid, oleic acid, and linoleic acid were purchased from Sigma (Saint-Quentin Fallavier, France). oxLDL was purchased from Clinisciences (Nanterre, France).

(35) Primary Macrophages

(36) Bone marrow cells were collected from femurs and tibias of ten-week-old male C57BL/6J mice by flushing with sterile medium as previously described (T. Ruggiero, M. et al (2009) FASEB J 23:2898-2908). To differentiate into macrophages, bone marrow cells (10.sup.6) were plated in 10-cm plates in 7 ml of BMDM medium (DMEM supplemented with 20% low-endotoxin fetal bovine serum, 30% L929-cell conditioned medium, 1% I-glutamine, 1% Pen/Strep, 0.5% Na pyruvate, and 0.1% -mercaptoethanol), and fed with 2.5 ml of fresh medium every 2 days for 7 days. Human primary macrophages were prepared from peripheral blood monocytes obtained from healthy donors (Jacquel, A., et al. (2012). Blood 119, 4527-4531) in agreement with the French legislation on human biomedical research. All participants provided written informed consent attesting they had received all the information they needed about the study and they agreed in accordance with appropriate national regulations.

(37) Cell Transfection and Immunoblotting

(38) BMDMs and NIH-3T3 cells (ATCC) were transiently transfected for 48 hr with Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. siRNAs were transfected at the final concentration of 24 nM, while 2-OMe-RNA antisense oligonucleotides to s-RNYs (Eurogentec) were transfected at the final concentration of 100 nM. When required, cell lysates were incubated at room temperature with RNase A (10 mg/ml, Ambion) for 30 min. Three-hundred micrograms of protein was immunoprecipitated with Protein A-Dynabeads (Invitrogen)-coupled antibodies for 16 hr at 4 C. with rotation. Immunoprecipitates were washed four times with lysis buffer and resuspended in SDS protein loading buffer. Proteins were subjected to SDS-PAGE, electroblotted onto PVDF membranes, and probed with different antibodies as indicated.

(39) Small RNA Sequencing

(40) Ten micrograms of total RNAs were subjected to the small RNA preparation procedure, and the libraries were sequenced on either Illumina or Solidi platforms. For the analysis, the 3 adaptor sequence was removed using FASTX, and the clipped reads were aligned to the genome database and DeepBase53 using BLAT and Novoalign.

(41) Fluorescence In Situ Hybridization and Immunofluorescence

(42) Fluorescence in situ hybridization coupled with immunofluorescence was carried out as previously described (Hu, Q., et al. (2012). Nature structural & molecular biology 19, 1168-1175). BMDMs were fixed in 10% neutral formalin and hybridized with FAM-labeled antisense LNA probe. The slides were then incubated with an human anti-GW182 antiserum to stain P-bodies (Eystathioy, T., et al. (2003). Journal of molecular medicine (Berlin, Germany) 81, 811-818). The probe used in this study was purchased from Exiqon (Denmark). For immunofluorescence, aortic sinus from ApoE.sup./ mice were fixed in 10% neutral formalin, penetrated with 20% sucrose in PBS, and embedded in OCT compound. Serial 10-m sections of the aortic sinus were stained with a rat Mac3 and a mouse monoclonal hnRNP A1 antibodies. Slides were coverslipped in Vectashield Mounting Medium with Dapi (Invitrogen). The results were analyzed on a Leica DM5500B microscope with a HAMAMATSU camera ORGA-ER. LNA probe sequence is detailed in Table 1.

(43) TABLE-US-00001 TABLE1 OligonucleotidesusedforRT-PCR,cloningandNorthernblotanalyses. Forwardprimer Reverseprimer Mouseandhuman 5-GTCGTATCCAGTGCAGGG s-RNY1-5p(RT) TCCGAGGTATTCGCACTGGAT ACGACATTGAG-3 (SEQIDNO:36) mouses-RNY3-5p 5-GTCGTATCCAGTGCAGGG (RT) TCCGAGGTATTCGCACTGGAT ACGACAACACC-3 (SEQIDNO:37) humans-RNY3-5p 5-GTCGTATCCAGTGCAGGG (RT) TCCGAGGTATTCGCACTGGAT ACGACAGTTGT-3 (SEQIDNO:38) humans-RNY4-5p 5-GTCGTATCCAGTGCAGGG (RT) TCCGAGGTATTCGCACTGGAT ACGACAGTTCT-3 (SEQIDNO:39) s-RNYsuniversal 5-GTGCAGGGTCCGAGGT-3 primer(qPCR) (SEQIDNO:40) Mouse/humans- 5-TGGTCCGAAGGTAGTGAGT-3 RNY1-5p(qPCR) (SEQIDNO:41) Mouses-RNY3-5p 5-TTGGTCCGAGAGTAGTGGT-3 (qPCR) (SEQIDNO:42) Humans-RNY3-5p 5-TCCGAGTGCAGTGGTGTTTA-3 (qPCR) (SEQIDNO:43) Humans-RNY4-5p 5-GGTCCGATGGTAGTGGGTTAT-3 (qPCR) (SEQIDNO:44) wts-RNY1-3pBS 5-AAGGCGCATCTGCGTACACAC fromKLF2mRNA ACAGGTGAGAAGCCTAAGGCG (cloning) CATCTGCGTACACACACAGGT GAGAAGCCT-3 (SEQIDNO:45) mutants-RNY1-3p 5-AAGGCGCATCTGCGTACACAC BSfromKLF2 ACAGAGTGAGGACCTAAGGCG mRNA CATCTGCGTACACACACAGAG TGAGGACCT-3 (SEQIDNO:46) Mouseand 5-ACTCACTACCTTCGGACCA-3 humanss-RNY1- (SEQIDNO:47) 5p(Northernblot) Mouseand 5-AGTCAAGTGCAGTAGTGAG-3 humanss-RNY1- (SEQIDNO:48) 3p(Northernblot) MouseRNY1loop 5-TTCAATCTGTAACTGACTG-3 (Northernblot) (SEQIDNO:49) Mouses-RNY3-5p 5-ACCACTACTCTCGGACCAA-3 (Northernblot) (SEQIDNO:50) Mouses-RNY3-3p 5-CTGGTCAAGTGAAGCAGTG-3 (Northernblot) (SEQIDNO:51) Humans-RNY4-5p 5-CCCACTACCATCGGACCAG-3 (Northernblot) (SEQIDNO:52) U6snRNA 5-CGTTCCAATTTTAGTATATGT (Northernblot) GCTGCCGAAGCGAGCAC-3 (SEQIDNO:53)

(44) Microarray Analysis

(45) For mRNA profiling analysis BMDMs were collected from 4 C57BL/6J males and transfected with siRNAs against the terminal loop of RNYs and control. 48 hr after transfection, total RNA was isolated by using RNAeasy kit (Qiagen). RNA samples were labeled with Cy3 dye using the low RNA input QuickAmp kit (Agilent) as recommended by the supplier. 400 ng of labeled cRNA probe were hybridized on 860K high density SurePrint G3 gene mouse GE 860K Agilent microarrays. Normalization of microarray data was performed with the Limma package available from Bioconductor (http://www.bioconductor.org) using the quantile methods. Means of ratios from si RNYs-treated versus control were calculated. Differentially expressed genes were selected based on at least a modulation of 1.2 fold change between BMDMs overexpressing s-RNYs and control, and a statistically significant Student t-test (p<0.05).

(46) Apoptosis Assay

(47) Apoptosis was assayed in BMDMs by staining with Alexa 488-conjugated annexin V and Dapi as previously described (Jacquel, A., et al., 2012. Blood 119(19): 4527-4531). The number of annexin V- and Dapi-positive cells was counted and expressed as a percent of the total number of cells from duplicate wells.

(48) Luciferase Reporter Assay

(49) NIH-3T3 cells (80% confluence in 96-well plates) were transfected with Lipofectamine 2000 according to the manufacturer's instructions. Luciferase reporter assay was performed as previously described (Repetto, E., et al. (2012). PLoS genetics 8, e1002823).

(50) RNA In Vitro Processing and UV-Crosslinking

(51) .sup.32P-labeled RNAs were synthesized and used as substrates for either in vitro processing assays or UV-crosslinking experiments as previously described (Trabucchi, M., et al (2009). Nature 459, 1010-1014).

(52) RNA Immunoprecipitation

(53) RNA immunoprecipitation was performed as previously described (Trabucchi, M., et al (2009). Nature 459, 1010-1014) with minor modifications. Briefly, cells lysates were immunoprecipitated with Protein A-Dynabeads (Invitrogen)-coupled antibodies at 4 C. overnight. Total RNA was prepared using Trizol (Invitrogen), and analyzed by quantitative RT-PCR. The primer sequences are detailed in Table 1.

(54) Northern Blot

(55) Total RNA was isolated from cells using Trizol (Invitrogen), resolved on 10% polyacrylamide-urea gels, and electroblotted onto HyBond N+ membranes. Membranes were hybridized overnight with radiolabeled DNA antisense oligonucleotides to s-RNYs in ExpressHyb solution (Clontech). After hybridization, membranes were washed three times with 2SSC and 0.05% SDS, twice with 0.1SSC and 0.1% SDS, exposed overnight to imaging screens, and analyzed using a Storm 860 PhosphorImager. The same blot was hybridized (upon stripping in boiling 0.1% SDS) with three distinct probes, including control U6 snRNA (Table 1).

(56) mRNA and s-RNY Expression Analysis

(57) RNA expression by quantitative RT-PCR was carried out by using standard procedures. Briefly, for mRNA, total RNA was isolated from cells using Trizol and cDNA was synthesized with a random hexonucleotides using Superscript III (Invitrogen). For s-RNYs detection by quantitative RT-PCR a stem-loop quantitative RT-PCR method was used according to Chen et al. (Chen, C., et al. (2005). Nucleic acids research 33, e179). Briefly, total RNA was isolated from cells, tissue, or immunoprecipitation using Trizol (Invitrogen) according to the manufacturer's instructions. RNA extraction from extracellular medium, blood plasma, and serum was performed as previously described (Turchinovich, A., et al (2011). Nucleic acids research 39, 7223-7233). Briefly, 1.2 ml of Trizol LS were added to 0.4 ml of blood plasma, serum or media. Then (i) 5 pg of synthetic miRNA-39 from Caenorhabditis elegans (cel-miR-39) were added as a spike-in control for purification efficiency; (ii) 1.2 l of glycogen (10 mg/ml) were added to enhance the efficiency of RNA column binding. Purification of extracted total RNA was performed with miRNeasy columns (Qiagen) according to the manufacturer's instructions. RNA was eluted in 30 l of RNase-free water. cDNA was synthesized using a hairpin long oligonucleotide. This method allows the detection of the s-RNYs derived from only the 5 end of the precursor by quantitative RT-PCR analysis, namely the s-RNY1-5p and s-RNY3-5p from mouse and s-RNY1-5p, s-RNY3-5p, and s-RNY4-5p from human. Quantitative RT-PCRs using Sybr Green and TaqMan (Invitrogen) for cel-miR-39 were performed on a StepONE system (Applied Biosystem). The primer sequences are detailed in Table 1.

(58) Recombinant Proteins and Antibodies for Western Blot

(59) Recombinant Ago2 and hnRNP A1 were purchased from Sino Biological Inc. and CliniSciences, respectively, while recombinant KSRP was a gift from Dr. R. Gherzi. Mouse monoclonal anti-hnRNP A1 (4B10), mouse monoclonal anti-c-Myc (9E10), goat anti-actin (1-19), and IB (H-4) antibodies were purchased from SantaCruz. Rabbit anti-La/SSB and anti-Ro60 (TROVE2) antibodies were purchased from GmbH. Anti-Ago2 antibody was purchased from Wako Chemicals while rabbit anti-KLF2 and anti-NOS2A antibodies were from Millipore. Anti-KSRP antibody was a gift from Dr. R. Gherzi. Mouse monoclonal anti-p53 (1C12), rabbit anti-cleaved Caspase-3 (Asp175), rabbit anti-caspase 3 (8G10), rabbit anti-P-p38 (9211), rabbit anti-p38 (9212), mouse anti-P-JNK (9255), and rabbit anti-P-JNK (9252) antibodies were purchased from Cell Signaling. Human anti-GW182 was a gift from Dr. M. J. Fritzler. Rat anti-mouse Mac-3 antibody (M3/84) was from BD Pharmingen. The immunofluorescence was revealed by using fluorescent secondary antibodies: Alexa Fluor 488-conjugated anti-rat IgG (H+L) antibody (Cell Signaling) for anti-mouse Mac3, Alexa Fluor 594-conjugated anti-mouse IgG (H+L) antibody (Molecular Probes) for anti-hnRNP A1, and Alexa Fluor 594-conjugated anti-human IgG (H+L) antibody (Invitrogen) for GW182.

(60) Candidate siRNA Screen

(61) SMARTpool siRNAs for the screen were purchased from Thermo Scientific (Illkirch, France). BMDMs were plated in 24-well plates and each SMARTpool siRNA was singularly transfected in three wells. 48 hr from the transfection cells were either left untreated or stimulated with 0.25 mM of PA for 18 hr, and total RNA was isolated and analyzed by quantitative RT-PCR. The experiment was repeated 4 times and statistically analyzed using Student's t-Test.

(62) Our screen included enzymes involved in small RNA biogenesis, such as Dicer, Drosha, Ddx-5, Ago1, 2, 3 and 4, Piwil1, 2 and 4, and Elac1 (also known as tRNase Z 1) as well as RNA-binding protein components of the small RNA processing machineries such as Dgcr8, Trbp-2, hnRNP A1, Ilf3 (also known as NF90), Xpo5, Ksrp, Lin28, Zc3h12a (also known as MCPIP1), Ews, p53, and Fus-Tls. In addition, genes implicated in small RNA stability were knocked down such as Zcchc11 (also known as TUTase 4), Zcchc6 (also known as TUTase 7), Xrn2, Exosc2, and RNAseL.

(63) siRNAs and Modified-RNA Oligonucleotides

(64) siRNA against mouse p53 and Ro60 were purchased from Ambion. Custom siRNAs or modified-RNA oligonucleotides were synthesized by Eurogentec: mouse Ago2 5-GUUGUAUUGUUUAGCGAUU-3 (SEQ ID NO: 22) mouse hnRNP A1 5-GCACUAGCCAUCUCUUGCUUC-3 (SEQ ID NO: 23) mouse La/SSB 5-UUCCUUUAAAUCUUCCACC-3 (SEQ ID NO: 25) mouse Dicer 5 UUCAGCUCGAUGGAUAUGGUG 3 (SEQ ID NO: 55) mouse si-RNA against RNY1 terminal loop 5-CAGUCAGUUACAGAUUGAA-3 (SEQ ID NO:56) mouse si-RNA against RNY3 terminal loop 5-CAACCAGUUACAGAUUUCU-3 (SEQ ID NO:57) mouse 2-OMe-RNA antisense to s-RNY1-5p 5-UUGAGAUAACUCACUACCUUCGGACCAGCC-3 (SEQ ID NO: 26) mouse 2-OMe-RNA antisense to s-RNY1-3p 5-AAGACUAGUCAAGUGCAGUAGUGAGAAG-3 (SEQ ID NO: 27) mouse 2-OMe-RNA antisense to s-RNY3-5p 5-UAAACACCACUACUCUCGGACCAACC-3 (SEQ ID NO: 28).

Example 2: Pro-Apoptotic and Atherogenic Stimuli Induce the Expression of RNY-Derived Small RNAs in Macrophages

(65) A high-throughput small RNAs sequencing has been performed in primary macrophages stimulated with either staurosporine (STS) or withdrawal of Macrophage-Colony Stimulating Factor (M-CSF) treatment. By mapping the sequencing data to small RNAs databases, including miRNAs, piRNAs, snoRNAs, riRNAs, and tRNAs using the Novoalign program (FIG. 1A), it was shown that apoptotic macrophages highly expressed RNY-derived small RNAs (FIG. 1B). 3 RNY-derived small RNAs were found in mouse and 6 in human varying in length from 24 to 34 nt which map to the end and stem regions of the RNYs (FIGS. 1B, 2A and 2B). These small RNAs will be referred hereinafter as s-RNYs, with the mouse s-RNY1-5p being previously classified as piRNA (piRNABank, http://pirnabank.ibab.ac.in/) and the human s-RNY5-3p as miRNA (Meiri, E., et al. (2010) Nucleic acids research 38, 6234-6246; Schotte, D. (2009) Leukemia 23, 313-322). RNY genes count four copies in human (RNY1, 3, 4, and 5) and two in mouse genome (RNY 1 and 3), and their sequence is well conserved in vertebrates. Northern blot analysis confirmed that the expression of s-RNYs (s-RNY1-5p, s-RNY1-3p, s-RNY3-5p and s-RNY4-5p) was induced in human primary macrophages upon treatment with either STS or withdrawal of M-CSF (data not shown).

(66) Using the stem-loop RT-qPCR method previously described (Mestdagh, P. et al (2008) Nucleic acids research 36, e143), the expression of the s-RNYs derived from the 5 end of RNYs has been investigated in macrophages treated with the pro-apoptotic stimulus lipoteichoic acid (LTA) from the Gram-positive bacteria Staphylococcus aureus in combination with an endoplasmic reticulum (ER) stressor, such as thapsigargin (Tg). ER stress renders macrophages susceptible to apoptosis in the face of other pro-apoptotic stimuli. LTA/thapsigargin treatment induces the expression of s-RNYs in bone marrow-derived macrophages (BMDMs) (FIG. 3A), suggesting a general mechanism of s-RNY induction in dying macrophages.

(67) A number of lipids and lipoproteins associated with atherosclerotic diseases conspire with ER stress to cause macrophage apoptosis and progression of atherosclerotic lesions. To determine whether s-RNYs are induced in macrophages stimulated with athero-relevant stimuli, BMDMs have been treated with oxLDL, alone or in presence of thapsigargin. Whereas low dose of oxLDL or thapsigargin alone had no effect, higher level of oxLDL or both reagents together synergistically induced s-RNY expression (RT-PCR analysis for s-RNY1-5p and s-RNY3-5p (see FIG. 3B) and Nothern blot analysis for s-RNY1-3p (data not shown)). Similar results were observed when ER-stressed BMDMs were incubated with acetylated-LDL (acLDL) (FIG. 3C). Different types of fatty acids have been tested then to determine whether they could also trigger s-RNYs induction in macrophages, and if so, whether such an effect was amplified by ER stressor. As shown in FIG. 4A, the unsaturated fatty acids, such as oleic acid and linoleic acid did not induce s-RNY expression when given alone or in combination with thapsigargin. However, when the saturated fatty acids, such as palmitic (PA) and stearic acids, were given alone s-RNY expression was induced, and this effect was increased markedly in the setting of thapsigargin-induced ER stress (RT-PCR analysis for s-RNY1-5p and s-RNY3-5p (see FIG. 4A) and Nothern blot analysis for s-RNY1-3p and s-RNY1-5p (data not shown)). Moreover, to test whether the s-RNYs are induced in animal models for atherosclerosis, ApoE.sup./, Ldlr.sup./, and control mice have been compared. ApoE.sup./ and Ldlr.sup./ mice were fed with either chow diet or an atherogenic diet containing 1.3% or 1.25% of cholesterol, respectively. As shown in FIGS. 4B and 4C, quantitative RT-PCR from microdissected aortic arches revealed that s-RNYs were significantly upregulated in ApoE.sup./ and Ldlr.sup./ mice compared to control, and that high cholesterol diet (HCD) significantly increased their expression levels. Therefore, these results strongly indicate that s-RNY expression is induced in macrophages exposed to various types of apoptotic stimuli, including those that are relevant to atherosclerosis development. Moreover, s-RNYs are expressed in aortas of mouse models for atherosclerosis, raising the possibility that they might play a role in regulating lesion development.

Example 3: Ago2- and hnRNP A1-Dependent Processing of RNYs in Lipid-Laden Macrophages

(68) Consistent with the observation that the induction of s-RNY expression (by M-CSF withdrawal or addition of oxLDL in combination with Tg, or PA in BMDM; or addition of STS on cultured human primary macrophages) is always accompanied by a reduction of RNY (Nothern blot analyses performed for s-RNY1-3p/RNY1, s-RNY1-5p/RNY1, s-RNY3-5p/RNY3, and s-RNY4-5p/RNY4; data not shown), a processing upregulation of RNYs has been found in vitro using total cell extracts from PA treated-BMDMs compared to unstimulated control (analyses performed for RNY1/s-RNY1-3p and s-RNY1-5p, and RNY3/s-RNY3-5p; data not shown), overall suggesting that s-RNY expression is modulated at the post-transcriptional level. Because RNYs form a hairpin secondary structure rather similar to that of miRNA precursors (FIG. 2B), an involvement of small RNA processing machinery components in s-RNY maturation has been hypothesized. To validate this hypothesis, a candidate siRNA screen has been performed in PA-treated BMDMs using smart-pool siRNAs to knockdown RNAse and RNA-binding proteins implicated in small RNA biogenesis, as well as Ro60 and La/SSB as RNA-binding proteins associated to the RNYs and forming the so called Ro ribonucleoprotein (RNP) complex (FIG. 5). Among the saturated fatty acids, PA is by far the most potent promoter of macrophage foam cell and apoptosis in atherosclerosis. As shown in FIG. 6, depletion of La/SSB, Ro60, p53, Ago2, and hnRNP A1 significantly decreased the induction of s-RNY expression, compared to Gapdh and non-targeting siRNA controls, as evaluated by quantitative RT-PCR from BMDMs stimulated with PA. In all these cases, s-RNY expression was statistically reduced by more than half. Surprisingly, it has been observed a significant increase of s-RNY expression levels by knocking down Dicer. Northern blot analysis from PA-treated BMDMs transfected with different siRNAs from those used in the siRNA screen, confirmed that the levels of s-RNYs were reduced by 70-90% upon knockdown of Ago2, hnRNP A1, Ro60, La/SSB, and p53 while increased upon Dicer knockdown (data not shown). Importantly, it has been noticed that upon Ago2 or hnRNP A1 knockdown in PA-treated BMDMs the decrease of s-RNY level induction was accompanied by an increase of the RNYs compared to control (Nothern blot analyses performed for s-RNY1-3p, s-RNY1-5p, and s-RNY3-5p; data not shown), suggesting that both Ago2 and hnRNP A1 may directly control the maturation of s-RNYs in lipid-laden macrophages. Conversely, knockdown of Ro60 and La/SSB decreased both s-RNYs and RNYs (data not shown), suggesting that these two proteins may stabilize RNYs and/or s-RNYs, as it was previously proposed. In support to a direct role of Ago2 and hnRNP A1 in regulating RNY processing, both RNYs but not Gapdh (negative control) were immunoprecipitated with antibodies against Ago2 or hnRNP A1 in PA-treated BMDMs while anti-La/SSB and anti-Ro60 antibodies did immunoprecipitate RNYs from both stimulated and unstimulated cells (FIG. 7). However, they did not immunoprecipitate with an anti-p53 antibody. Together, these results strongly indicate that both Ago2 and hnRNP A1 may directly control RNYs processing in lipid-laden macrophages.

(69) Among the proteins found in the siRNA screen, only Ago2 is an enzyme. Ago2 has an endoribonuclease activity for double-stranded RNAs, is the catalytic component of the RISC and is involved in miR-451 processing. Ago2 is associated with different classes of small non-coding RNAs, including miRNAs and siRNAs. To establish whether Ago2 directly interacts with RNYs, recombinant protein and synthetic radiolabeled-RNYs (RNY1, RNY3) have been used in UV-cross-linking assays, which indicated a direct Ago2-RNYs interaction, while the TNF-ARE RNA, used as negative control, did not cross-link (data not shown). These data demonstrated the direct loading of the RNYs into Ago2 and raised the possibility that Ago2 might directly catalyze the maturation of s-RNYs. To check this hypothesis an in vitro processing assay has been performed using naked RNYs and the recombinant Ago2. Processing assay with RNY1 and RNY3 shows that recombinant Ago2 can cleave both RNYs but was unable to generate the final mature form, as speculated taking into account the predicted secondary structure of RNYs (data not shown and FIG. 2B). However, this experiment alone cannot exclude the occurrence of different cleavage sites in vivo due to a protein complex associated with Ago2, as one would expect from the well-established biochemical properties of Ago2. Indeed, Ago2 depletion in BMDM extracts stopped the RNY processing activity while the addition of the recombinant Ago2 was sufficient to rescue the blockade and to generate the full maturation of s-RNYs, as it has been demonstrated for RNY3 (data not shown).

(70) Interestingly, it has been recently reported that both RNY-associated proteins, La/SSB and Ro60, are integral components of Ago2 complex. In particular, previous reports demonstrated that La/SSB is both involved in miRNA processing and turnover of RISC by promoting the release of cleaved mRNA. Coimmunoprecipitation experiments in NIH-3T3 cells revealed that Ago2 interacted with La/SSB in both untreated- and PA-treated cells, while with Ro60 only upon PA treatment (data not shown), indicating that Ago2 is associated to the Ro RNP complex in lipid laden-cells. Moreover, antibodies against Ago2, La/SSB, or Ro60 immunoprecipitated s-RNYs from PA-stimulated BMDMs (FIGS. 8A and 8B), indicating the formation of a protein complex containing s-RNYs. In conclusion, these data strongly demonstrated a s-RNY maturation that proceeds with an Ago2-mediated cleavage.

(71) Next, the possibility that the single-stranded RNA binding protein hnRNP A1 favors the Ago2-dependent processing of RNYs in PA-treated macrophages has been explored. hnRNP A1 is a nucleo-cytoplasmic shuttling protein with roles in many aspects of RNA metabolism. It has been recently reported that hnRNP A1 is involved in miRNA processing by specifically binding to the single-stranded terminal loop structure of selected miRNA precursors to finely modulate the biogenesis. In particular, it binds to miRNA precursors to induce a change in the secondary structure creating a more favorable cleavage site for Drosha. The data reveal that hnRNP A1 coimmunoprecipitated with Ro60 (data not shown) and that anti-hnRNP A1 antibody immunoprecipitated both RNYs in PA-stimulated BMDMs (FIG. 7). To establish whether hnRNP A1 interacts directly with RNYs, recombinant protein and in vitro transcribed RNY have been used in UV-cross-linking assays, which indicated a direct hnRNP A1-RNYs interaction (RNY1 and RNY3; data not shown) while the RNA binding protein KSRP, used as negative control did not cross-link to RNY1 or RNY3 (data not shown). The pri-let7-a, was used as a positive control, crossed-link to KSRP (data not shown). Interestingly, both RNY1 and RNY3 share some similarities with the hnRNP A1 binding sequence to the terminal loop of miR-18a precursors (FIG. 2), indicating that, as for miRNA precursors, hnRNP A1 binding to RNYs may create a more favorable cleavage site. Indeed, the presence of hnRNP A1 promotes the Ago2 cleavage (data not shown). Furthermore, because hnRNP A1 knockdown in PA-treated BMDMs abrogated the interaction of Ago2 with both RNY1 and RNY3 (FIG. 9A), it was concluded that hnRNP A1 favors the association of Ago2 to RNYs in lipid-laden macrophages. Therefore, these data indicate that hnRNP A1 is a key regulator of RNYs processing in lipid-laden macrophages based on its affinity binding to RNYs. Upon binding, hnRNP A1 could optimize the positioning/recruitment of Ago2 through a protein-protein interaction. Indeed, in PA-stimulated BMDMs, a RNAseA independency of the interaction between Ago2 and hnRNP A1 has been demonstrated (data not shown). In support of this conclusion, hnRNP A1 cytoplasmic levels dramatically increased during PA treatment of BMDMs (data not shown) and in Mac3-positive macrophages in aortic sinus of ApoE.sup./ mice (data not shown). The cytoplasmic localization of hnRNP A1 would promote the interaction to Ago2 and then the processing of RNYs. Overall, these results indicate that in lipid-laden macrophages the processing of RNYs is mediated by the Ago2/hnRNP A1 complex.

Example 4: Post-Transcriptional Regulation of Target mRNAs by S-RNYs

(72) Given that s-RNYs are associated with Ago2 (FIG. 8B), the possibility that they might exert a post-transcriptional control of gene expression comparable to that of miRNAs has been explored, to target mRNAs by recognition of complementary sequences. It has been first determined whether s-RNYs localize to the Processing bodies (P-bodies), which serve as sites for RNA degradation mediated by small RNAs, such as miRNAs, piRNAs, and riRNAs. A fluorescence in situ hybridization has been performed followed by immunostaining (immuno-FISH) using a FAM conjugated-LNA probe to s-RNY1-5p together with immunofluorescent staining of GW182, as marker for the P-bodies. Although, the LNA probe used here can recognize both RNY1 and s-RNY1-5p, a change in the LNA-staining pattern has been noticed in BMDMs upon PA stimulus. In untreated BMDMs the RNY1 signal was diffused in all cytoplasm, while in PA-treated BMDMs the RNY1/s-RNY1-5p signal was a cytoplasmic dot staining pattern similar to that of GW182 (data not shown). Notably, the RNY1/s-RNY1-5p signal partially colocalized with GW182 in PA-treated BMDM (data not shown). Most importantly, RNA immunoprecipitation analysis demonstrated that GW182 interacts with s-RNYs in PA-treated BMDMs (FIG. 9B). Therefore, these data indicate that s-RNYs localize to the P-bodies in lipid-laden macrophages.

(73) To determine the existence of direct targets of s-RNYs, microarray experiments were performed from BMDMs overexpressing s-RNYs and control. Because siRNA against the terminal loop of RNY1 or RNY3 generates s-RNYs associated to Ro60 mimicking the induction of the processing that occur in lipid-laden macrophages and apoptotic macrophages (demonstrated for s-RNY1-3p, s-RNY1-5p, and s-RNY3-5p; data not shown), this strategy has been used instead of overexpressing ectopic mimic s-RNYs. This approach would avoid any off-target effects due to the overexpression of ectopic s-RNYs. Direct s-RNY target mRNAs were expected to be enriched among the mRNAs that are dowregulated in BMDMs overexpressing s-RNYs compared to control.

(74) The 57 mRNAs statistically downregulated of at least 1.2 fold change in s-RNYs-overexpressed BMDMs were analyzed for the presence of potential binding sites for s-RNYs using Sylamer bioinformatic program (van Dongen, S., et al (2008) Nature methods 5, 1023-1025). Because miRNAs use a seed sequence of 6-8 nt in their 5 end to target both the 3 untranslated region (3UTR) and the coding region (CDS) of mRNAs, it has also been searched for at least 7 nt of sequence complementary between s-RNYs 5 end sequence and both 3UTR and CDS of the downregulated genes. As expected a significant enrichment of s-RNYs targets was observed in the downregulated mRNAs, about 56% (32 mRNAs) (FIGS. 10A and 10B). Consequently, genes regulated by s-RNYs would be affected in PA-treated macrophages transfected by siAgo2 or s-RNYs inhibitors. Indeed, the expression of at least 20/32 mRNAs (about 62%) were Ago2 and s-RNYs dependent, as evaluated by quantitative RT-PCR in PA-treated BMDMs transfected with siAgo2, 2-OMe-RNA antisense oligonucleotides to s-RNYs, or control (FIG. 11; and data not shown). These data indicate that the altered mRNA expression profile of BMDMs expressing s-RNYs exhibits a signature that could be mediated by a direct target of mRNA by s-RNYs.

(75) Interestingly, in the list of the potential s-RNYs direct target-mRNAs, some genes have been clearly demonstrated to be important in regulating the pathogenesis of atherosclerosis, the cell survival and the anti-inflammatory response Fos, Krppel-like factor 2 (KLF2), Zfyve28, Rhomboid family-1 (Rhbdf1), Fgfr1, Lysyl oxidase (Lox) and Nr4a1 (also known as Nur77). In particular, KLF2 is a transcription factor having an anti-apoptotic and an anti-inflammatory role in macrophages. KLF2 is downregulated in atherosclerotic lesions of ApoE/ and Ldlr/ mice compared to control and mice bearing the double knockout for ApoE and KLF2 develop more severe and faster atherosclerotic lesions compared to control, indicating a protective role of this gene. Western blot experiments confirmed that the downregulation of KLF2 expression in PA-treated BMDMs was indeed rescued by the knockdown of Ago2 or of s-RNYs (data not shown), indicating a major role played by s-RNYs in regulating the expression of this gene. According to bioinformatic predictions, there is a single s-RNY1-3p binding site in KLF2 mRNA (FIG. 12A). Sequence conservation across vertebrates was strongest in the segment showing complementary to the seed region of s-RNY1-3p (FIG. 12B). To assess whether KLF2 mRNA is directly targeted by s-RNY1-3p, a reporter construct has been made containing 2 copies of the wild-type (wt) sequence of KLF2 having the binding site for s-RNY1-3p, and the mutant one cloned downstream to a luciferase CDS. Importantly, PA treatment of NIH-3T3 cells repressed luciferase activity of the reporter construct containing the wt sequence which was rescued by the knockdown of s-RNY1-3p (FIG. 12C). In contrast, the mutant reporter construct was not affected. Altogether, these data indicate that s-RNY expression affects the modulation of gene expression in lipid-laden macrophages by directly regulating at the post-transcriptional level a critical subset of mRNAs for the pathogenesis of atherosclerosis development, as it has been demonstrated for KLF2.

Example 5: s-RNYs Regulate Apoptosis and Inflammatory Response in Lipid-Laden Macrophages

(76) In accordance with previous reports (Scull, C. M., and Tabas, I. (2011). Arteriosclerosis, thrombosis, and vascular biology 31, 2792-2797; Seimon, T. A et al. (2010). Cell metabolism 12, 467-482), FIGS. 13A and 13B show an apoptosis activation in macrophages treated with pro-atherogenic stimuli, such as PA or Tg in combination with oxLDL. Hence, the effects of s-RNYs on BMDMs apoptosis using flow cytometry have been evaluated. As shown in FIG. 13A, induction of s-RNYs maturation by using siRNA against the terminal loop of RNYs causes an upregulation of the apoptosis mediated by PA treatment, which is rescued by co-transfecting the 2-OMe-RNA antisense oligonucleotides (SEQ ID NO: 26, SEQ ID NO: 27 and SEQ ID NO: 28) to s-RNYs. Moreover, a downregulation of the apoptosis is observed when the endogenous s-RNY expression is knocked down in BMDMs treated with PA or oxLDL (FIGS. 13A and 13B). Notably, by promoting s-RNY maturation the apoptosis is induced in untreated BMDMs, which is also rescued by co-transfecting the 2-OMe-RNA antisense oligonucleotides to s-RNYs (FIG. 13C). Furthermore, 2-OMe-RNA antisense oligonucleotide to s-RNYs inhibits the activation of caspase-3 in PA-treated BMDMs (data not shown) while s-RNYs overexpression is able to activate it in unstimulated BMDMs (data not shown). Therefore, these results indicate that atherogenic stimuli-induced s-RNYs are an intrinsic component of the machinery regulating apoptosis in lipid-laden macrophages.

(77) Given that PA treatment promotes both apoptosis and inflammatory response in macrophages by regulating signaling pathways downstream Toll-like receptors (TLR) activation, including c-Jun NH2-terminal kinase (JNK), NF-B, and p38, it has been hypothesized that s-RNYs may regulate PA mode of action in macrophages by ultimately modulating these signaling pathways. Time course experiments revealed that PA enhanced p38 activation, which depends on p38 phosphorylation, starting by 9 hr while NF-B activation, which depends on IB degradation, after 9 hr of treatment (data not shown). As previously reported, free fatty acids treatment of BMDMs did not significantly enhance the JNK activation upon phosphorylation (data not shown). Interestingly, s-RNY expression was induced by 9 hr of PA treatment in BMDMs (FIG. 14A) suggesting a timing correlation between s-RNYs induction and cell signaling activation, and a possible role of s-RNYs in regulating them. In fact, the inventors have demonstrated that NF-B and p-38 activation are both reduced upon s-RNYs knockdown in PA-treated BMDMs (data not shown), indicating that s-RNYs promote the activation of both signaling pathways. Consistent with this conclusion, s-RNYs overexpression in BMDMs strongly increased the PA-induced expression of NF-B-target genes, as it was demonstrated for nitric oxide synthase 2 (NOS2A) (data not shown). Moreover, the effect of inducing s-RNY maturation on the expression regulation of death cytokines and nitric oxide activated by NF-B pathway has been tested in BMDMs treated with the general macrophage activator lipopolysaccharide (LPS). As shown in FIG. 15, s-RNYs enhanced a significant upregulation of the mRNAs encoding for NOS2A, interleukin (IL)-6, IL-12b, and IL-1b. Therefore these data indicate that s-RNYs activate p38 and NF-B pathways to ultimately promote cell death and pro-inflammatory response in macrophages.

Example 6: s-RNYs are Upregulated in the Blood of Mouse Models for Atherosclerosis and in Patients with Coronary Artery Disease

(78) Significant amounts of miRNAs have been recently found in extracellular body fluids, including blood, urine, saliva, and semen. Some circulating miRNAs in the blood have been successfully revealed as biomarkers for several human disorders, such as many cancers, cardiovascular diseases, and brain and liver injuries. To evaluate whether s-RNYs may be considered as novel biomarkers for atherosclerosis-derived diseases, the presence of s-RNYs in the extracellular environment has first been determined. For that purpose, RNA was isolated from the medium of PA-treated BMDMs and s-RNY expression levels were measured by quantitative RT-PCR (FIG. 14B). PA treatment resulted in a 100-fold enrichment of s-RNYs in extracellular environment compared to control. Like miRNAs, s-RNYs are also strikingly stable in cell culture medium. Indeed, Northern blot from a combination of medium and debris of BMDMs treated with STS at different time points revealed that s-RNYs were significantly stable in the extracellular space (Northern blot analysis showing the stability of s-RNY1-5p in BMDMs stimulated with 1 M of staurosporine (STS) over 72 hours; data not shown). This result may indicate that s-RNY stability is likely caused by their association with protein complex, microvesicles, exosomes, apoptotic bodies, or lipoproteins as it has been recently proved for extracellular miRNAs. In accordance with previous reports, Northern blot results show that non-miRNA species, such as U6 snRNA, are extremely unstable in the extracellular environment (data not shown).

(79) To test whether s-RNYs are significantly present in the blood of animal models for atherosclerosis s-RNY expression levels has been compared in the blood plasma of ApoE.sup./, Ldlr.sup./, and control mice. The animals were fed with either chow diet or HCD. As shown in FIG. 16, quantitative RT-PCR from blood plasma revealed that s-RNYs were significantly upregulated in ApoE.sup./ and Ldlr.sup./ mice compared to control, and even more when mice were fed with HCD.

(80) Therefore, because of these data it can be logically anticipated that s-RNYs can be found in significant amounts in human body fluids of patients with atherosclerosis-related diseases. To validate this hypothesis, the expression of s-RNYs were measured in 45 male patients with stable Coronary Artery Disease (CAD) and 45 age-matched healthy normolipemic male subjects, in the context of a case-control study (Bouisset F. et al. (2012) The American journal of cardiology 110, 197-202). Metabolic and clinical variables of CAD and control individuals are summarized in Table 2.

(81) TABLE-US-00002 TABLE 2 Clinical and biological characteristics of the study population Cases (n = 45) Controls (n = 45) P Age (year) 59.6 (8.2) 59.2 (9.3) 0.84 Waist (cm) 97.0 (10.4) 97.0 (13.1) 0.86 BMI (kg/m2) 26.8 (4.0) 27.8 (5.4) 0.29 Triglycerides (g/L).sup.a 1.59 (0.74) 1.21 (0.73) 0.02 Total cholesterol (g/L) 2.16 (0.41) 2.20 (0.40) 0.70 LDL-C (g/L) 1.39 (0.40) 1.41 (0.34) 0.82 HDL-C (g/L) 0.48 (0.15) 0.55 (0.13) 0.003 apoB (g/L) 1.12 (0.24) 1.06 (0.21) 0.20 apoA-I (g/L) 1.24 (0.25) 1.51 (0.24) 0.001 LpB:E (mg/L) 85.6 (44.8) 45.3 (45.6) 0.001 -GT (IU/L) 53.3 (34.1) 40.0 (32.0) 0.06 CRP.sup.b (mg/L) 5.8 (3.8) 2.2 (2.8) 0.003 Physical activity.sup.c 1.9 (0.8) 2.3 (0.9) 0.03 s-RNY1-5p.sup.d 2.09 (0.79) 0.39 (0.57) 0.001 Data are expressed in mean (SD) or %; -GT: -glutamyltransferase; CRP: C-Reactive Protein; BMI: Body Mass Index. .sup.alog transformed data; .sup.bgeometric mean; .sup.clevel of physical activity 1 to 4; .sup.dsquare root value.

(82) Among metabolic markers, total or LDL-cholesterol (LDL-C) were not significantly different between the two cohorts, probably reflecting effects of lipid-lowering drugs in CAD patients. However, CAD patients displayed higher levels of pro-atherogenic lipids such as triglycerides and lipoproteins containing ApoB and ApoE (LpB:E), increased CRP and lower atheroprotective high-density lipoprotein (HDL) markers such as HDL-cholesterol (HDL-C) and apoA-I levels. Body mass index (BMI) and waist circumference were not significantly different between cases and controls but physical activity was lower in cases and 84.4% of current or former smokers were found in cases against 55.6% in control individuals. Altogether, these clinical and biological characteristics reflect an increased prevalence of classical cardiovascular risk factors in CAD patients. A dysregulation of the expression of s-RNYs has been investigated in serum from CAD patients compared to healthy controls. Data show a significant upregulation of s-RNY1-5p expression in CAD patients (p<0.001, Table 2 and FIG. 17A). s-RNY4-5p was also detected in human serum and found unregulated in serum CAD patients compared to control individuals (p<0.001, Table 3 and FIG. 18A) while s-RNY3-5p was not detected in both CAD patients and control. Importantly, correlation between s-RNY1-5p levels and individual metabolic parameters and cardiovascular risk markers were investigated (Table 4).

(83) TABLE-US-00003 TABLE 3 Clinical and biological characteristics of the study population (s-RNY4-5p) Cases (n = 30) Controls (n = 19) p Age (year) 59.2 (7.7) 59.1 (8.6) 0.98 Waist (cm) 98.0 (11.8) 101.0 (15.1) 0.52 BMI (kg/m2) 27.4 (4.7) 29.5 (6.8) 0.25 Triglycerides (g/L).sup.a 1.56 (0.76) 1.56 (0.93) 0.97 Total cholesterol (g/L) 2.15 (0.38) 2.35 (0.47) 0.10 LDL-C (g/L) 1.37 (0.35) 1.52 (0.42) 0.20 HDL-C (g/L) 0.46 (0.17) 0.52 (0.12) 0.14 apoB (g/L) 1.14 (0.25) 1.08 (0.26) 0.49 apoA-I (g/L) 1.23 (0.28) 1.55 (0.26) 0.001 LpB:E (mg/L) 73.2 (43.7) 45.9 (58.6) 0.07 -GT (IU/L) 55.9 (39.5) 45.7 (40.2) 0.49 CRP.sup.b (mg/L) 6.9 (3.7) 2.6 (4.6) 0.006 Physical activity.sup.c 1.9 (0.8) 2.1 (0.9) 0.34 s-RNY4-5p.sup.d 4.43 (1.00) 2.75 (0.99) 0.001 Data are expressed in mean (SD) or %; -GT: -glutamyltransferase; CRP: C-Reactive Protein; BMI: Body Mass Index. .sup.alog transformed data; .sup.bgeometric mean; .sup.clevel of physical activity 1 to 4; .sup.dsquare root value.

(84) TABLE-US-00004 TABLE 4 Sperman rank correlation coefficients between s-RNYs and some metabolic parameters and cardiovascular risk factors in the study population s-RNY1-5p.sup.# s-RNY4-5p.sup.# (n = 90) p (n = 49) p Age (year) 0.01 0.91 0.22 0.13 Waist (cm) 0.02 0.83 0.08 0.60 BMI (kg/m2) 0.10 0.35 0.14 0.34 Triglycerides (g/L) 0.24 0.02 0.33 0.03 Total cholesterol (g/L) 0.07 0.48 0.06 0.69 LDL-C (g/L) 0.07 0.55 0.03 0.84 HDL-C (g/L) 0.37 0.001 0.27 0.06 apoB (g/L) 0.17 0.11 0.19 0.19 apoA-I (g/L) 0.56 0.001 0.39 0.006 LpB:E (mg/L) 0.36 0.001 0.52 0.001 -GT (IU/L) 0.28 0.009 0.34 0.02 CRP (mg/L) 0.43 0.001 0.47 0.001 Physical activity.sup.a 0.21 0.05 0.34 0.02 .sup.#Spearman rank correlation; -GT: -glutamyltransferase; CRP: C-Reactive Protein; BMI: Body Mass Index. .sup.alevel of physical activity 1 to 4

(85) s-RNY1-5p was positively correlated with pro-atherogenic lipids (triglycerides and LpB:E) and negatively with HDL markers (HDL-C and apoA-I). Environmental factors, such as physical activity and an inflammatory condition, as documented by elevated CRP, displayed positive correlation with RNY1-5p (Table 4). Same correlations were found for s-RNY4-5p (Table 4). Overall, these data indicate a strong correlation between s-RNYs and CAD risk, as it was attested by the high score of the area under the ROC curves (FIGS. 17B and 18B), supporting the hypothesis that s-RNY1-5p and s-RNY4-5p measurement might improve CAD risk prediction.

(86) The upregulation of RNY-derived small RNAs (in particular s-RNY1-5p) in patients with coronary artery disease was confirmed on a larger cohort of patients (n=263) (see Table 5 and FIGS. 20-25).

(87) TABLE-US-00005 TABLE 5 Demographic, clinical and biological characteristics of patients and controls Case Control t-test or mean SD mean SD CHI2 n = 263 n = 514 p s-RNY1-5p 10.42 12.33 1.32 1.67 0.001** (RNY1-5p).sup.1/5 1.49 0.28 0.94 0.25 0.001 (RNY1-5p).sup.1/6 1.39 0.22 0.95 0.22 0.001 (RNY1-5p).sup.1/7 1.33 0.18 0.95 0.19 0.001 Quartiles RNY1-5p (%) 0.001 Q1: <0.474 0.4 37.9 Q2: 0.474-1.618 5.3 34.6 Q3: 1.619-4.701 28.5 23.4 Q4: >4.701 65.8 4.1 Total cholesterol (g/L) 2.02 0.38 2.24 0.38 0.001 HDL cholesterol (g/L) 0.43 0.12 0.55 0.13 0.001 LDL cholesterol (g/L) 1.25 0.34 1.45 0.32 0.001 Serum triglyceride (g/L) 1.68 0.89 1.21 0.77 0.001* Apolipoprotein A1 (g/L) 1.23 0.22 1.50 0.24 0.001 Apolipoprotein B (g/L) 1.04 0.22 1.08 0.22 0.03 ApoC3 (mg/L) 33.8 12.9 30.8 12.7 0.002 ApoE (mg/L) 99.4 56.0 71.5 46.5 0.001* LpBC3 (mg/L) 16.4 10.6 14.1 10.5 0.006 LpBE (mg/L) 76.4 56.7 48.6 46.1 0.001* Lp(a) (g/L) 0.47 0.44 0.30 0.37 0.001* IF1 (mg/L) 0.43 0.14 0.52 0.15 0.001 Gamma glutamyl 62.8 68.6 45.2 56.8 0.001** transferase (IU/L) High sensitivity C reactive 17.2 29.7 3.1 5.1 0.001** protein (mg/L) Fasting glucose (mmol/l) 5.93 2.01 5.43 1.06 0.19** Serum insulin (UI/L) 15.9 19.5 10.0 8.1 0.001** HOMA-IR 4.2 4.9 2.6 2.8 0.001** Adiponectin (g/mL) 5.6 4.3 7.0 4.4 0.001 Waist (cm) 99.3 10.7 95.3 9.9 0.001 Waist to hip ratio 0.99 0.04 0.96 0.05 0.001 Body mass index (kg/m.sup.2) 27.5 4.0 26.9 3.6 0.04 Systolic blood pressure 137.0 20.2 137.4 16.5 0.82** (mmHg) Diastolic blood pressure 83.5 11.2 82.7 7.9 0.55** (mmHg) Resting heart rate 63.7 11.5 62.8 9.2 0.79** (beat/mn)*** Body Fat (bioelectrical 28.2 5.4 26.2 5.1 0.001 impedance %) Age (year) 60.3 8.0 59.0 8.3 0.04 Schooling (number of years 9.6 3.0 13.1 4.3 0.001** spent at school) Current tobacco 3.4 10.1 2.3 7.0 0.22** consumption (number cig) Tobacco consumption 41.5 37.9 17.2 21.3 0.001** number (pack year) Physical activity score 1.86 0.67 2.23 0.82 0.001** Alcohol consumption 28.8 28.5 23.8 24.1 0.09** (g/day) Wine 25.8 26.4 20.4 21.8 0.02** Schooling (number of years 0.001 spent at school) % <9 42.2 17.5 9-10 32.7 9.0 >10 25.1 73.5 Treatment for hypertension 44.1 19.6 0.001 (%) Treatment for diabetes (%) 23.2 5.2 0.001 Treatment for dyslipidemia 63.5 23.3 0.001 (%) Smoking (%) 0.001 current 18.3 14.6 former 65.4 50.4 never 16.3 35.0 Ankle-arm index 0.9 (%) 33.6 1.6 0.001 Physical activity (%) 0.001 no 27.4 18.5 medium 61.6 45.9 high 11.0 35.6 Alcohol consumption (%) 0.03 no 17.1 12.4 <40 g/day 54.4 64.3 40 g/day 28.5 23.3 Metabolic syndrome 48.5 17.7 0.001 (NCEP ATP-III %) *analyses performed on log transformed data **Kruskal-Wallis test

(88) The level of expression of these small RNAs was also checked in 514 healthy individuals and it was found that the higher expression of s-RNY1-5p is associated with high levels of CRP, LDLc, ApoB and lower levels of ApoA. 24 patients infected by bacteria, therefore having high CRP but no atherosclerostis, show intermediate levels of s-RNY1-5p expression.

(89) Taken together, the highest levels of s-RNY1-5p are specifically associated to patients with atherosclerosis.

(90) FIGS. 22-24 respectively display ROC curve for prediction of CAD based on (i) s-RNY1-5p, (ii) basic cardiovascular risk factors (hypertension, dyslipidemia, diabetes, smoking, school duration, ankle-arm index, ApoA1 and CRP), or (iii) s-RNY1-5p combined with basic cardiovascular risk factors,

(91) A ROC curve is a graphical representation of the sensitivity (or true positive rate) against the false positive rate (i.e. [1specificity], specificity being the true negative rate) of a marker-based test. A ROC space is defined by sensitivity and (1-specificity) as x and y axes respectively. The best possible prediction method would yield a point in the upper left corner or coordinate (0,1) of the ROC space, representing 100% sensitivity (no false negatives) and 100% specificity (no false positives). A completely random guess would give a point along a diagonal line (the so-called line of no-discrimination) from the left bottom to the top right corners. The diagonal divides the ROC space. Points above the diagonal represent good classification results (better than random), points below the line poor results (worse than random). The Area Under the Curve (AUC) of a ROC curve may be calculated. The higher the AUC, the higher the diagnostic accuracy of the diagnostic marker.

(92) For combined analysis of markers, a new virtual marker Z is calculated based on a linear combination of the levels of the individual markers. Z is calculated as follows: Z= a.sub.i[Marker.sub.i] where a.sub.i are calculated coefficients and [Marker.sub.i] are individual levels of marker (optionally in normalised units). The values of the a.sub.i coefficients are determined in order to maximize the Area Under the Curve (AUC) of the ROC curve for the selected marker combination. Determination of the coefficient values may be readily achieved using for instance mROC program or any other program implementing an algorithm for maximising the AUC of ROC which may be used for multivariate ROC.).

(93) The results shown on FIG. 24 indicate that s-RNY1-5p improves the diagnostic accuracy of CAD when combined with basic cardiovascular risk factors.