Page 79 - Haematologica July

Exosomal miRNAs in AA and MDS
filtering to extract genes that were exclusively present in SAA, MDS, or SAA-responders, or present in both MDS and SAA or SAA-responders (Figure 5B and C, Online Supplementary Table S5 and Online Supplementary Figure S5). Common genes targeted by the 7 shared miRNAs but not present in MDS were selected for analysis, and the top 20 pathways were reported (Figure 5D). miR-4651 and miR-126-5p were used for pathway analysis in SAA- responders, after removing the common targeted genes from MDS and SAA at diagnosis (Online Supplementary Figure S5C). The analysis revealed involvement of several intracellular functions, related to cell cycle, DNA damage response, intracellular signaling, and metabolic pathways.
Discussion
We investigated exosomal miRNA profiles in SAA and MDS in order to find potential biomarkers for diagnosis and disease progression in BM failure syndromes. Based on screening of 372 miRNAs in the discovery set of plas- ma exosome samples, a custom miRNA PCR plate was designed, including 42 miRNAs for validation in a larger cohort. The analysis revealed 25 exosomal miRNAs that were uniquely or commonly present in SAA and/or MDS patients; they were involved in several biological func- tions, such as HSC differentiation.
Recently published work from our laboratory describes circulating miRNAs using whole plasma samples of AA and MDS patients;9 however, no distinctive signatures have been reported for exosomal miRNAs to date. In our current study, we identified exosomal miRNAs exclusive- ly present in SAA (miR-532-5p), MDS (14 miRNAs), and SAA responders to IST (miR-4651) patients, or common to SAA and MDS (7 miRNAs), and to SAA responders to IST and MDS (miR-126-5p). The miRNAs we identified have not been reported to be different in AA and/or MDS plas- ma samples. However, circulating miRNAs are composed of passively released nucleic acids from different cell types,15,36 and actively secreted in exosomes. Exosomal miRNAs are not randomly loaded into vesicles; rather, mature miRNAs are specifically sorted by diverse mecha- nisms and based on sequence, reflecting the tissue of ori- gin.17,20,24 Because they are tissue-specific and stable under different conditions, exosomal miRNAs have been pro- posed as better biomarkers. Their significance in AA and MDS has not yet been examined. Plasma circulating miRNAs have been related to extracellular pro-inflamma- tory signaling pathways, such as Toll-like receptor or tumor necrosis factor α.9 Dysregulated exosomal miRNAs in our cohort were associated with numerous intracellular functions.
Circulating miR-532-5p (unique in SAA) is related to response to chemotherapy in AML, and associated with decreased expression of Runt-related (RUNX) 3 protein in melanoma.37,38 miR-196a and miR-196b (present in our SAA and MDS cases) are differentially expressed in the long-term HSC compartment compared to more mature populations, and miR-196 induces the promotion of HSC differentiation through the inhibition of Homeobox (HOX) family members.39 miR-19b (decreased in our SAA and MDS cases) together with other miRNAs promotes the differentiation of progenitor cells into lymphoid pre- cursors through inhibition of PTEN.39 Additionally, cytokines modulate miRNA expression, such as miR-196
by interferons during hepatitis C infection.40 Based on this, increased exosomal miR-532-5p and miR-196, and decreased miR-19b in our cohort of SAA and MDS patients may indicate attempted expansion of the stem and progenitor cell compartments. miR-1180-3p has an anti-apoptotic function through inhibition of the BCL pro- tein family;41 this miRNA was decreased in our SAA and MDS, patients more apoptotic in marrow with active dis- ease. Other exosomal miRNAs present in our cohort do not have known functions in hematopoiesis or the immune response.
There is increasing evidence of essential roles of circu- lating and exosomal miRNAs in modulating cancer cell metabolism, the immune response, and altering the local and distant tumor environments.12,17,42 miR-126-5p and its 3’ UTR counterpart are regulators of angiogenesis, cell metabolism, and glucose homeostasis, and they are down-regulated in many tumors. By targeting the PI3K/AKT/mTOR pathway, overexpression of miR-126 results in reduction of proliferation, driving malignant and normal HSCs into quiescence.43,44 Decreased exosomal miR-126 is not only related to increased proliferation of leukemic cells and differentiation44 but also decreases autophagy in cancer cells, leading to accumulation of dam- aged mitochondria and reactive oxygen species.44,45 In SAA, increased amino acid biosynthesis has been report- ed.46 It is plausible that miR-126-5p modulates metabolic programming during BM failure and recovery, especially as miR-126-5p was decreased in MDS and SAA respon- ders to IST. Pathway analysis of predicted target genes using miRWalk 2.0 and IPA software revealed that down- stream signaling could be regulated by exosomal miRNAs, as for MAPK, p38, JAK/STAT, and ERK. These pathways can be activated by many different extracellular stimuli, such as cytokines and growth factors, and exosomal miRNAs by interfering with this signaling could impair intracellular responses.12,17 However, miRNAs from circu- lating exosomes may be derived from different tissues, such as BM or T cells. Therefore, additional functional studies and exosomal miRNA profiling in BM samples will resolve the origin and biological functions of these miRNAs during active disease.
Utilization of circulating and exosomal miRNAs in clin- ical practice is uncertain, mainly due to technical issues such as data normalization, exosome preparation, and RNA extraction.9,47-49 For circulating miRNAs, normaliza- tion is performed using small nucleolar RNAs (snoRNAs) and other methods such as GeNormPlus, NormFinder, and global mean of miRNA expression. For exosomal miRNAs, there are no clear guidelines for data normaliza- tion. RNU6 used for fresh samples for miRNA normaliza- tion is not a reliable endogenous control for frozen sam- ples.47 The mean expression value is often applied for nor- malization47 but employing only one method could result in missing significant differences.48 In our study, we nor- malized our data using the geometric mean of 2 control miRNAs within each group: one snoRNA (SNORD61) and one miRNA (miRNA-211-5p) homogeneously expressed in the discovery set. This approach achieved high concor- dance across plates and for replicates. Ultracentrifugation remains the standard assay for extracellular vesicles isola- tion and enrichment, as no commercial kit achieves a high-yield exosome with more than 99% purity and therefore allows exclusion of lipoproteins. Treating extracted exosomes with RNase or proteinase K can
haematologica | 2018; 103(7)
1157