Upon cleavage of the pre-miRNA by Dicer, the resultant double stranded miRNA associates with the RNA-induced silencing complex (RISC) [29], an incompletely characterized protein complex which includes the RNase argonaute 2 (Ago2) [30]. Once associated, the miRNA duplex is separated into a 5′ strand (guide strand) which is integrated into the RISC [31, 32] and a 3′ strand (passenger strand) [29]. The RISC then uses the guide strand to target mRNA though base pairing with complementary sequences in their 3′ untranslated region (UTR) [33]. Once bound, the mRNA will either: (1) be degraded by Ago2 in the setting of a near perfect match with the miRNA or (2) be prevented from association with ribosomes resulting in inefficient translation in the setting of a partial match with the miRNA [33] (Fig. 10.2). In each case, the net effect is a reduction of protein expression of the targeted gene. It was initially believed that the 3′ strand was degraded by RISC upon separation of the miRNA duplex [29], however, subsequent work has demonstrated that these passenger strands can be functionally active in mRNA targeting as well [34]. Accordingly, miRNA are now designated with a standard nomenclature such that the 5′ strand is given the ‘5p’ suffix while the 3′ strand is given ‘3p’.

The efficiency by which an individual miRNA can recognize and inhibit a target mRNA strand is determined by the complementary match between the miRNA and sequences in the 3′ UTR of the mRNA. Nucleotides 2–7 of the miRNA, known as the miRNA seed, appear to be particularly important as perfect pairing within the seed has been shown to be important for the recognition of most mRNA [34]. Additional pairing outside of the seed including at nucleotide 8 or nucleotides 13–17 of the miRNA have also been shown to increase the efficacy by which it inhibits mRNA translation [35]. Binding site location on the mRNA can also influence a miRNA’s ability to repress translation. For instance, binding sites in the 3′ UTR of the mRNA result in more efficient inhibition than do binding sites in translated regions. Further, binding sites that are distant from the center of long UTRs or which are located in regions with concentrated A-U sequences appear to be more susceptible to miRNA targeting [35]. Finally, emerging research has also suggested the existence of miRNA binding sequences within the 5′ UTR of mRNA potentially expanding the number of genes subject to miRNA regulation [36].

MiRNA are ubiquitous across cell types and across species of both the plant and animal kingdoms. They have been identified among all cell lineages and can be traced back to pluripotent stem cells where they have been shown to play important regulatory roles in the processes of self-renewal and pluripotency [37]. However, miRNA are not exclusively confined to the intracellular space. Indeed, miRNA have been identified in a variety of biofluid compartments including plasma, pleural fluid, and bronchoalveolar lavage (BAL) fluid. Analyses of each compartment suggest that while there are a considerable number of miRNA that are common to multiple different biofluids, some biofluids contain miRNA profiles that are unique [38].

Plasma is perhaps the best-studied biofluid and has provided many insights into how miRNA exist extracellularly. The initial observation by Valadi et al. that cells could transfer both mRNA and miRNA between each other suggested that miRNA could exist outside of cells at least transiently. These investigators discovered that cells could package RNA, including miRNA, into exosomes which were then released by the cells [39]. Ultimately, analyses of plasma exosomes have revealed a large number of contained miRNA [40]. When the miRNA expression profile of these exosomes was compared to a similar profile from peripheral blood mononuclear cells (PBMC), there were detectable differences [41] suggesting that either: (a) PBMCs were not the sole source of circulating miRNA in the plasma or (b) exosomal miRNA contents differ from that of their donor cells. Subsequent work has demonstrated that both explanations are correct. A variety of cell types have been shown to release exosomal miRNA into the plasma including platelets, endothelial cells, endothelial progenitor cells, and lymphocytes [42–45]. Further, evidence suggests that these cells may release a variety of different vesicles including exosomes, microparticles, microvesicles, and apoptotic bodies [46], each of which may contain miRNA. Additionally, miRNA expression in donor cells and their daughter exosomes often differ [39, 45, 47–49]. The mechanism behind this is incompletely understood and may be related to “selective exportation” of miRNA into the exosomes or differences in the decay kinetics between different miRNA [50, 51].

In addition to vesicle-associated miRNA, investigators have also identified extracellular miRNA which exist outside of membrane-derived particles. This so-called non-vesicle-associated miRNA population appears to be largely protein bound with the RISC component, Ago2, as the predominant extracellular chaperone [52]. When quantified by PCR, this protein bound population appears to account for >90 % of circulating miRNA [52, 53]. Analyses of both the vesicle-associated and non-vesicle-associated miRNA populations have demonstrated that each population is comprised of a distinct pattern of miRNA suggesting that the export systems for each population may differ from each other [54]. Some have suggested that the extracellular protein bound miRNA may derive from the inadvertent release of miRNA during cell death [52], however, this or alternative export mechanisms have not been confirmed.

A third population of extracellular miRNA has been discovered which is bound to circulating high density lipoprotein (HDL) [55]. Similar to theAgo2-bound fraction, the miRNA expression profile in this population was distinct from the vesicle-associated population and has even appeared to differ between disease state and controls. However, the proportion of circulating miRNA which is bound to HDL appears modest raising the question of its functional importance [56]. The different forms of extracellular miRNA are depicted in Fig. 10.3.

Regardless of whether extracellular miRNA are being shuttled inside of exosomes or by proteins, they demonstrate stability both in vivo and in ex vivo storage. Ribonucleases (RNases) in both the bloodstream and the environment commonly degrade large molecular weight RNAs; however, miRNA are typically resistant to this enzymatic cleavage [57]. A portion of this resistance is likely conferred by their association with membrane-derived vesicles or carrier proteins as these may shield the miRNA from enzyme exposure. Indeed, nucleophosmin 1 (NPM1) is a RNA binding protein that has also been implicated as a miRNA shuttle protein and has been shown to protect miRNA from degradation by RNase A [54]. Interestingly, when serum miRNA were exposed to a variety of harsh conditions including boiling, multiple freeze/thaw cycles, low/high pH, and extended storage, they demonstrated remarkable stability suggesting that they may be inherently resistant to degradation as well [57].

Several studies have demonstrated that miRNA can be measured in the alveolar compartment through analysis of BAL fluid [38, 58, 59] and over 200 individual miRNA have been identified in this compartment to date [38]. Unlike some biofluids, BAL does not appear to contain miRNA that are unique only to that compartment but rather it appears to express a subpopulation of miRNA that are found in other compartments including plasma. It is not clear, however, if BAL contains fewer individual miRNA than plasma because certain miRNA are selectively filtered out before entering the alveolar space or because bronchoalveolar cells express and release miRNA in different patterns than circulating or endovascular cells. It is also worth acknowledging that new miRNA are continuously being discovered; thus, miRNA that are unique to the alveolar space with physiologic relevance to pulmonary disease may exist but have yet to be identified. At least two studies have compared expression patterns of miRNA in the BAL between control humans and humans with disease. The first study examined exosomal miRNA expression in both asthmatics and healthy controls and identified a signature of 16 miRNA whose expression levels allowed for classification of the subjects with asthma [58]. The second study reported differential expression patterns of total BAL miRNA between patients with dyspnea and healthy controls and identified correlations between BAL miRNA levels and pulmonary function testing [59].

Since the discovery of extracellular miRNA, investigators have sought to determine whether it is a mechanism by which cells can communicate with each other through manipulation of target cell gene expression. Indeed, numerous examples have now been described in which membrane-derived vesicles containing miRNA have been taken up by recipient cells leading to altered target gene expression and/or cellular function [39, 43, 45, 47, 60–62]. Hergenreider et al. [62] found endothelial cells placed under shear stress released miRNA-filled vesicles which, in turn, were internalized by cocultured vascular smooth muscle cells resulting in alterations in their gene expression. Similarly, Zernecke and colleagues demonstrated that apoptotic endothelial cells can release vesicular apoptotic bodies which contain, among others, miR-126 and are internalized by adjacent endothelial cells. Through a series of elegant experiments using apoptotic bodies from control and miR-126 deficient mice, they showed that delivery of miR-126 induced expression of the chemokine CXCL12 and attenuated the development of atherosclerosis [43].

Non-vesicle-associated miRNA may also serve as a mechanism for cell-to-cell communication. For example, HDL-bound miRNA can be effectively delivered to hepatocytes in vitro with resultant alterations in hepatocyte gene expression [55]. This mode of delivery is dependent upon scavenger receptor class B type I, however, and may not occur in all cell types. Indeed, similar uptake has not been demonstrated in endothelial cells, smooth muscle cells, or PBMCs suggesting that HDL-miRNA complexes play only a limited role in cell-to-cell transport of miRNA [56]. Although they represent the largest fraction of extracellular miRNA, it is unclear at this time if Ago2-bound miRNA can be internalized in a manner similar to HDL-bound miRNA or vesicle-associated miRNA. The C. Elegans transmembrane channel, SID-1, facilitates the uptake of double stranded RNA including long hairpin miRNA precursors [63, 64]. However, while mammalian homologs of SID-1 exist [65], it is unclear if they can import precursor or mature miRNA, particularly when bound to a protein. Recent investigations which demonstrate that miRNA can bind to and activate extracellular toll-like receptors (TLR) suggest a potential alternative mechanism by which Ago2-bound miRNA can facilitate cell-to-cell communication [66, 67]. Further investigation into the potential roles of non-canonical actions of extracellular miRNA in cell-to-cell communication is warranted.