Intercellular Transport of MicroRNAs

Abstract

Extracellular microRNAs (miRNA) are present in most biological fluids, relatively stable, and hold great potential for disease biomarkers and novel therapeutics. Circulating miRNAs are transported by membrane-derived vesicles (exosomes and microparticles), lipoproteins, and other ribonucleoprotein complexes. Evidence suggests that miRNAs are selectively exported from cells with distinct signatures that have been found to be altered in many pathophysiologies, including cardiovascular disease. Protected from plasma ribonucleases by their carriers, functional miRNAs are delivered to recipient cells by various routes. Transferred miRNAs utilize cellular machinery to reduce target gene expression and alter cellular phenotype. Similar to soluble factors, miRNAs mediate cell-to-cell communication linking disparate cell types, diverse biological mechanisms, and homeostatic pathways. Although, significant advances have been made, miRNA intercellular communication is full of complexities and many questions remain. This review brings into focus what is currently known and outstanding in a novel field of study with applicability to cardiovascular disease.

I. Introduction

MicroRNAs (miRNAs) are small non-coding RNAs that provide post-transcriptional regulation of gene expression and control of many metabolic and physiological processes associated with cardiovascular disease^1–6. Most annotated genes are predicted targets of miRNAs, including many key regulators of cholesterol metabolism and cardiovascular function^4–7. Intracellular miRNAs have proven to be critical mediators in the response to cellular stress, disease, and environmental stimuli⁸. Extracellular miRNAs are a new class of cellular messengers as they stably exist in most biological fluids, including blood, urine, cerebral spinal fluid, saliva, semen, and breast milk^9–16. Selectively exported and functional in recipient cells, extracellular miRNAs are now recognized as regulatory signals in cell-to-cell communication^17–19. Very much like soluble factors, extracellular miRNAs provide intercellular gene regulation and phenotypic control. However, the delivery of miRNA cassettes or clusters of miRNAs may have a greater capacity to impact a more diverse set of genes and biological pathways than cytokines or hormones. Similar to soluble factors, extracellular miRNAs likely function to regulate gene expression in both macro and microenvironments. Here we review the current biology of miRNAs as intercellular messengers and the phenotypic outcome of small RNA communication.

II. Extracellular miRNAs

a. Carriers of Extracellular miRNAs

Then classified as low molecular weight RNA, extracellular small RNAs were observed in blood as early as 2004²⁰; however, miRNA profiling of human plasma/serum was not completed until 2008, when Lawrie et al. found miRNAs in serum of lymphoma subjects and Mitchell et al. found extracellular miRNAs to be stable in human plasma/serum^{9, 21}. Strikingly, miRNAs were reported to be functional extracellular signaling molecules in plants more than a decade ago^{22, 23}. Nonetheless, the global plasma miRNA signature is composed of individual sub-classes of miRNA carriers. Membrane-derived vesicles, lipoproteins, and ribonucleoprotein complexes all have been found to transport extracellular miRNAs^24–26. Exosomes and microparticles (MPs) are two distinct classes of membrane-derived vesicles, differentiated by their biogenesis and secretory mechanisms¹⁸. Exosomes, observed as vesicles ~40–100 nm in diameter, are initially formed by the inward budding of the plasma membrane into multivesicular bodies within endosomes²⁷. Exosomes and their miRNA content are released into the extracellular compartment upon the fusion of endosomes with the plasma membrane^{18, 27}. Conversely, MPs are generally larger vesicles (100–4000 nm) that result from the outward budding and blebbing of the plasma membrane²⁸. During apoptosis, cells can release even larger MPs or apoptotic bodies, which also transport specific sets of miRNAs²⁹. Originally described as platelet dust, MPs were first observed in plasma more than 45 yrs. ago³⁰.

Extracellular miRNAs are also transported by lipoproteins, namely high-density lipoproteins (HDL) and low-density lipoproteins (LDL), which are both highly abundant in plasma²⁶. Whereas exosomes and MPs are composed of a bilayer-phospholipid shell and hydrophilic core, lipoproteins consist of a single layer of lipids, a hydrophobic core, and are defined by specific structure-function apolipoproteins. Evidence suggests that the exosomal, HDL, and LDL-miRNA signatures are distinct (low correlation), although some miRNAs are found in all sub-classes, including miR-223²⁶. Argonaute 2 (AGO2), the main functional component of the cytoplasmic miRNA ribonucleoprotein complex (miRNP), is observed bound to extracellular miRNAs both in and out of membrane-derived vesicles^31–33. Biophysical studies suggest that miRNAs are associated with protein complexes 50–300 kDa in size, which includes AGO2 as well as other ribonucleoproteins^{31, 32}. Recently, viral surface antigen particles were also found to transport specific miRNAs, as hepatitis B surface antigen particles were found to contain AGO2-bound miRNAs³⁴.

In vitro studies suggest that most cell types secrete exosomes or MPs, including neurons, inflammatory, muscle, and tumor cells^{25, 35–39}. However, the majority of circulating MPs and exosomes in vivo are likely secreted by platelets^{40, 41}. In 2002, exosomes were first reported to transfer information between cells^{27, 42}; however, it was not until a seminal study in 2007 that exosomes were found to contain miRNAs²⁵. Differential miRNA profiles associated with membrane-derived vesicles have been described for many pathophysiologies, including cardiovascular disease^43–45. Currently, it is unknown if the abundances of specific miRNAs carried on extracellular protein complexes are also altered with disease. Likewise, differential HDL-miRNA signatures were observed in humans and mice with hypercholesterolemia²⁶. Most interesting, many of the differential miRNAs associated with cardiovascular disease, miR-150, miR-223, miR-92, are candidate signaling molecules as they have each been reported to alter gene expression upon transfer to recipient cells^{24, 26, 46}.

b. Cellular miRNA Export

miRNA-based intercellular communication is composed of three critical processes. First, miRNAs must be selectively and actively secreted from cells and packaged into appropriate carriers. Second, miRNAs must be protected from circulating RNAses and transferred to targeted or receptor-specific recipient cells. Third and most importantly, miRNAs must retain the ability to recognize and repress mRNA targets within recipient cells. Although, little is currently understood about how miRNAs are selectively exported multiple observations suggest that the process is selective and regulated as some miRNAs are found to be only exported and not retained in the parent cell and specific signaling pathways have been found to regulate cellular miRNA release^{26, 45, 47,48–50}. Most importantly, miRNA profiles of extracellular vesicles and lipoproteins are not representative of their parent cell-type, but are distinct sets of miRNAs^{24, 25, 47}. Secondly, miRNA profiles of total plasma/serum and the individual profiles of distinct miRNA carriers in plasma/serum are surprisingly consistent amongst individuals²⁶. Furthermore, each biological fluid compartment contains distinct miRNA signatures. Collectively, these observations support the selective export hypothesis, in that cells actively secrete specific miRNAs due to cellular signals or environmental cues. Extracellular vesicles released from tumor cells were found to contain miRNAs not present in the parent cell, which suggest that some miRNAs may be transcribed only to be exported^{48, 49}. Conversely, some cellular miRNAs may not be transcribed or processed in a given cell, but may be present through exogenous delivery by extracellular vesicles. Multiple studies have found that the ceramide pathway regulates the export of cellular miRNAs^{26, 47, 50–52}. Neutral sphingomyelinase 2 (nSMase2), the rate limiting enzyme in the conversion of sphingomyelin to ceramide, is a critical regulator of exosome biogenesis and secretion⁵³. Inhibition of nSMase2 resulted in decreased export of specific miRNAs to exosomes, but increased the export of miR-223 to HDL^{26, 50–52}. Previously, miR-223 was found to be enriched in non-vesicle sub-classes²⁴, which is supported by the observation that miR-223 is one of the most abundant and consistent miRNAs found on HDL²⁶. Nevertheless, nSMase2, which resides on the inner leaflet of the plasma membrane⁵⁴, likely serves in some capacity to govern extracellular miRNA release to differential carrier sub-classes.

c. Transfer of Functional miRNAs

Endogenous transfer of functional miRNAs was first reported in 2009, when apoptotic bodies were found to transfer functional miR-126 to recipient cells²⁹. Smaller vesicles (exosomes) were found to transfer of functional miRNAs in 2010 when Epstein-Barr virus infected parent cells were reported to secrete exosomes containing viral miRNAs that altered target genes in recipient dendritic cells^55–57. The exogenous miRNA targeting was readily differentiated from endogenous miRNA gene regulation⁵⁷. Most importantly, these studies set-forth the concept of miRNA intercellular communication. In parallel, MPs were found to transfer miR-150 to recipient endothelial cells and enhance migration²⁴. This study, and others that followed, identified a novel miRNA-based communication network between inflammatory cells and the vascular endothelium. It has been proposed that extracellular miRNAs are especially adept at mediating gene regulation amongst cells in a microenvironment⁵⁸. In this regard, sub-intimal inflammation and atherosclerotic lesions, which harbor complex multicellular microenvironments, may be a nexus of miRNA-based signaling. Multiple studies have found that endothelial cells are particularly efficient recipient cells for miRNA transfer^{24, 46, 59, 60}. However, a recent study has demonstrated that endothelial cells also secrete extracellular miRNAs, as vesicles originating from stimulated endothelial cells were found to transfer miR-143/miR-145 and alter vascular smooth muscle cell phenotype⁶¹.

d. Intercellular Communication

Although cell-to-cell communication though miRNA transfer may be more perfectly suited for microenvironments, lipoproteins and extracellular vesicles are found in plasma and likely reach many distant tissues. In addition to cardiovascular disease, many other pathophysiologies likely use circulating carriers to modulate systemic homeostasis or propagate the disease. Evidence suggests that cancer likely utilizes circulating miRNAs and their carriers to promote tumorgenesis and malignancy^{46, 60, 62}. Multiple reports have found that tumor cells communicate with endothelial cells through the transfer of pro-angiogenic miRNAs that leads to cell migration, angiogenesis, tumor growth, and malignancy⁶⁰. Other studies have also identified miRNAs secreted from tumor cells that drive tumorgenesis and cell invasiveness upon delivery⁴⁶. Recently, miRNAs within exosomes were found to activate Toll-like receptors within endosomes leading to oncogenic phenotypes in recipient cells and tumor growth⁶².

Adipocytes may use miRNAs for both local and systemic communication. Large adipocytes were found to transfer miR-16, miR-222, miR-27a, and miR-146b to small adipocytes which stimulated lipid storage⁶³. The miRNA-induced lipid storage may not be contained locally and may serve as a strong metabolic signal due to various stimuli. In addition, adipocytes and fat may communicate with inflammatory cells though miRNA transfer⁶⁴. Most interesting, mothers likely utilize exosomes to shuttle miRNAs to offspring in utero through release of exosomes into maternal circulation by chronic villi⁶⁵. In addition, embryonic stem cells were found to transfer miRNAs to fibroblasts⁶⁶.

III. Role in Cardiovascular Disease

Extracellular miRNAs are not inherently insensitive to RNase degradation, as detergents, proteinases, and sonication of extracellular carriers has been reported to render extracellular miRNAs sensitive to digestion^{9, 24, 32, 51, 63, 67}. As such, extracellular carriers likely protect miRNAs from circulating RNases. Due to extracellular miRNA’s noninvasive accessibility and remarkable stability they hold great potential for disease biomarkers. Altered extracellular profiles have been reported for cardiovascular disease, diabetes, cancer, nonalcoholic steathepatitis, and multiple other pathophysiologies^{43–45, 68, 69}. At this time, rapid analysis of extracellular miRNA is not feasible/ available and purification of the lipid-based carriers is time consuming. Furthermore, future studies will have to identify altered miRNAs within specific sub-classes and demonstrate representative ranges of health and disease in the global profile. Nevertheless, miRNAs likely hold enormous value in diagnostics, early prediction strategies, and pharmacologic assessments.

To date two reports have been published where communication via miRNAs is directly studied in a model of cardiovascular disease^29,60. Both studies assess the effects of endothelial-secreted microRNAs on the development of atherosclerosis in mice. In the first study by the Weber laboratory²⁹, the authors show that miR-126 is secreted by endothelial cells via apoptotic bodies, that function as a “danger signal” for other endothelial cells, that obtain anti-atherosclerotic properties through the miR-126-mediated repression of RGS16. The second study described the transfer of miR-143 and miR-145 from endothelial cells that were exposed to atheroprotective blood flow to smooth muscle cells⁶⁰. In smooth muscle cells, miR-143 targets ELK1, PHACTR4 and SSH2 and miR-145 targets KLF4, CAMK2d, CFL1, PHACTR4, SSH2. The suppression of these targets induces a contractile atheroprotective phenotype in the receiving smooth muscle cells.

Extracellular miRNAs and their lipid-based carriers may also be used for novel treatment strategies. By gaining a better understanding and decoding of the unique signals and comprehending the physiological response, we may aim to enhance or attenuate specific messages for therapeutic approaches. For example, specific miRNAs or inhibitors may be added to specific lipoproteins or extracellular vesicles as a strategy to alter the extracellular miRNA signature to gain a therapeutic advantage. Reconstituted HDL, recombinant apolipoprotein A-I (ApoA-I), and ApoA-I mimetic peptides each are currently being tested in clinical trials to increase HDL-cholesterol levels and enhance the benefits of HDL⁷⁰. Future studies may look to customize reconstituted particles with unique cassettes of miRNAs for therapeutic gains. Multiple groups have already used lipoproteins and/or extracellular vesicles to delivery anti-cancer drugs, modified siRNAs, and enriched miRNAs^71–73. One could also utilize naturally atheroprotective extracellular vesicles such as apoptotic bodies from endothelial cells enriched with miR-126, which were previously found to stabilize plaques in atherogenic mice²⁹, or extracellular vesicles generated by endothelial cells exposed to atheroprotective stimuli that contain a complex cocktail of beneficial miRNAs, including miR-143 and miR-145⁶¹.

In 2010, it was reported that breast milk contains exosomes with functional miRNAs¹⁰. Authors proposed then that lipid-based carriers may transfer functional miRNAs to offspring by crossing the gut lumen^{10, 71, 74}. A subsequent study found that miRNAs found in foods are indeed passed through the gut and enter circulation, as plant miRNAs were observed to be transported to the liver⁷¹. Nevertheless, both studies support the hypothesis that ingested miRNAs, likely protected by lipid-based carriers, retain functionality and regulate genes in recipient tissues.

Future advances into the role of extracellular miRNAs may include the packaging of miRNAs into lipid-based carriers for dietary supplementation during gestation or nutrition. miRNA intercellular communication likely controls many facets of cardiovascular disease and novel therapeutic strategies manipulating these networks have tremendous applicability in cardiovascular sciences. Many individuals suffer from ischemic events and morbidity with normal lipid values and statin therapy. Decoding and manipulating the miRNA intercellular message provide a unique opportunity to advance our current approaches.

In summary, analysis of distinct sub-classes of miRNA carriers has revealed specific miRNA profiles in health and disease. Basic biology of these carriers has provided many fundamental insights into the mechanisms that govern miRNA export, transport, and delivery. However, much is still to learn about the selective process of miRNA export and the physiological outcome of each miRNA signaling network. Nevertheless, a novel field of study has emerged that has great potential for the detection, prevention, and treatment of cardiovascular disease.

IV. Open Questions?

a. Functional Relevance in vivo

Although many in vitro studies have provided fundamental insights into the machinery and mechanisms that govern the export, transport, and delivery of miRNAs from cell-to-cell, little is understood about the exact nature of extracellular miRNAs in health and disease in vivo. One of the most outstanding questions in the field concerns the necessary payload or concentration of miRNAs that is needed to alter cellular physiology or phenotype. Many of the lipid-based carriers only transport a small number of specific miRNAs in relatively low abundance. In vitro studies have clearly shown that carriers can transfer enough miRNAs to repress gene expression and alter phenotype; however, it remains to be determined what the actual physiological flow of information requires in vivo. Extracellular vesicles collected from stimulated endothelial cells were found to reduce atherosclerosis in Apoe^−/− mice, which confirmed an earlier study that also demonstrated the potential of miRNA intercellular communication in vivo^{24, 61}. A more biochemical question relates to the cellular usage of transferred miRNAs. Currently, it is not known if cells have differential sensitivities or utilization of extracellular miRNAs compared to endogenous miRNAs transcribed and processed within the cell. To our knowledge, studies have yet to demonstrate that extracellular miRNAs are indeed loaded into cellular AGO2-miRNP complexes.. Furthermore, it is unknown if extracellular miRNAs regulate miRNA machinery in the recipient cell through targeting miRNA biogenesis pathways or if delivered miRNAs alter recipient cell miRNA transcription or stability.

b. Selective miRNA Export

Maybe the most important facet of intercellular communication is the selective export of specific miRNAs from the cell. Current evidence suggests that the export profile is not representative of the parent cell and is distinct in abundance and content, although very little is understood about the selection and export process. Multiple studies have found that miRNP complexes associate with multivesicular bodies within endosomes likely in part of a secretory or turnover process^{75, 76}. Less is understood how selective miRNAs are partitioned and released in outward budding MPs. Precursor miRNA hairpins have been found within microparticles⁷⁷, but in most cases it appears that extracellular miRNAs are single stranded²⁶. The current model of miRNA biogenesis suggests that single stranded extracellular miRNAs must have been completely processed, unwound, and likely loaded into AGO2-miRNP complexes, only to escape during selection and/or exportation. Because not all extracellular miRNAs are bound to AGO2, it is currently unknown if AGO2 is necessary for export or if miRNAs break free from AGO2 after cellular export. Since AGO2 is not present on lipoproteins it is very likely that miRNAs can be secreted from cells independently of miRNP complexes²⁶. The biogenesis, delivery mechanisms, and sources for each circulating carrier is likely different. Most importantly, it is unknown if AGO2-miRNA or other ribonucleoprotein complexes contain lipids. Likewise, it is unknown if extracellular non-vesicle AGO2-miRNAs are transferred to recipient cells and /or alter gene expression. It is certainly possible that some extracellular ribonucleoprotein complexes may not act as cellular messengers and not participate in intercellular communication.

c. Role of Extracellular miRNAs in Disease

Cellular miRNAs have proven to be powerful mediators in response to metabolic cues and cellular stress. Nevertheless, future studies are required to fully understand the role extracellular miRNAs play in adaptive and maladaptive responses to disease. Evidence suggests that there are consistent extracellular miRNAs signatures in health, which may be a novel way to quantify health and differentiate disease. Each miRNA carrier sub-class likely has its own distinct miRNA signature in health and disease; however, some miRNAs are certainly shared amongst the carriers, including miR-223 which was found in both exosomes and HDL²⁶ It is likely that vesicles and particles exchange proteins, lipids, and other factors; therefore, it is possible miRNAs could be exchanged or transferred between similar and disparate carrier types.

Another remaining question to the cell-to-cell transfer hypothesis concerns the physical delivery of miRNAs to the cytoplasm or miRNP complexes of recipient cells. The HDL receptor, SR-BI, uses a selective core uptake process for HDL cholesterol ester delivery⁷⁸. The delivery of HDL-miRNAs is dependent on SR-BI expression and function; however, it is currently unknown if selective core uptake is distinct from the miRNA delivery process. Similarly, it is not known if SR-BI mediates the uptake of LDL-miRNAs or if the LDL receptor-mediated endocytosis delivers functional LDL miRNAs to recipient cells. Membrane–derived vesicles have been observed to transfer miRNAs through membrane fusion and endocytosis^79–82. It is likely that membrane fusion could directly dump miRNAs into the cytosol⁷⁹; however, miRNAs taken up by endocytosis still must cross a phospholipid bilayer to enter the cytosol for target recognition and functionality. As such, there are likely many currently unresolved proteins that act as sensors, carriers, and transporters within the cell membrane of endosomes/lysosomes.

Figure 1. Schematic of miRNA intercellular communication.

Secreted miRNAs are transported by extracellular carriers and transferred to recipient cells where theyalter gene expression. Membrane-derived exosomes are released from cells as endosomes containing multivesicular bodies (MVB) fuse with the plasma membrane. Microparticles (MP) are released from the plasma membrane through outward budding. Neutral sphingomyelinase 2 (nSMase2) governs the release of extracellular miRNAs. Exosomes and microparticles deliver extracellular microRNAs (miRNA) through endocytosis and membrane fusion. Delivery of high-density lipoproteins (HDL) miRNAs is dependent upon scavenger receptor B1 (SR-BI). Argonaute 2 (AGO2) is the main structure-function protein within the miRNA ribonucleoprotein complex where miRNAs recognize and bind to mRNA targets. LDL, low-density lipoproteins.

Table 1.

Functional miRNAs in intercellular communication.

miRNA	Secreting Cell	Recipient Cell	Carrier	Targeted Genes	Effect	Reference
miR-150	Monocytes	Endothelial cells	Microparticles	c-Myb	Increased migration	24
miR-223	n/a	Hepatocytes	HDL	RhoB, EFNA1	n/a	26
miR-126	Endothelial cells	Endothelial cells	Apoptotic Bodies	RGS16	Anti-atherosclerotic	29
miR-143	Endothelial cells	Smooth muscle cells	Exosomes/microparticles	ELK1, PHACTR4, SSH2	Anti-atherosclerotic	61
miR-145	Endothelial cells	Smooth muscle cells	Exosomes/microparticles	KLF4, CAMK2d, CFL1, PHACTR4, SSH2	Anti-atherosclerotic	61
miR-16	Large adipocytes	Small adipocytes	Microparticles	n/a	Lipid synthesis	63
miR-222	Large adipocytes	Small adipocytes	Microparticles	n/a	Lipid synthesis	63
miR-27a	Large adipocytes	Small adipocytes	Microparticles	n/a	Lipid synthesis	63
miR-146b	Large adipocytes	Small adipocytes	Microparticles	n/a	Lipid synthesis	63

HDL, high-density lipoproteins; RhoB, ras homolog gene family, member B; EFNA1, ephrin-A1; RGS16, regulator of G-protein signaling 16; ELK1, E twenty-six (ETS)-like transcription factor; PHACTR4, phosphatase and actin regulator 4; SSH2, protein phosphatase slingshot homolog 2; KLF4, krueppel-like factor 4; CAMK2d, calcium/calmodulin-dependent protein kinase type II delta chain; CFL1, cofilin 1; n/a, not available.