Diabetes, microRNAs and exosomes: Les liaisons dangereuses

Type 2 diabetes mellitus (T2DM) is a complex, multisystem disease that represents one of the most common metabolic disorders worldwide and is widely being recognized as one of the main health-threats of this century [1] and [2]. It is estimated that diabetes affects around 8% of the US population [3] and that by 2030 will affect over 400 million people worldwide, with over 80% of cases being specifically T2DM [4]. Cardiovascular death is prevalent in diabetic subjects. Diabetes can lead to cardiomyopathy, macrovascular and microvascular disease, thus contributing to ischemic heart disease and heart failure. In the diabetic heart, structural alterations in the myocardium (fibrosis, hypertrophy) go together with microangiopathy and an impaired potential for vascular regeneration [5]. In this clinical picture, some level of communication between the different cellular components is obviously expected. It is immediate, and already well described, that vascular disease leading to acute ischemia (i.e. a heart attack) or a chronically insufficient myocardial perfusion can damage the cardiac myocytes by starving them of oxygen and nutrients. However, the concept that cardiomyocytes may govern endothelial cell (EC) function has not been similarly investigated.

In this issue of the Journal of Molecular and Cellular Cardiology, Wang et al. describe a mechanism by which cardiac myocytes signal to ECs by using microRNA (miR) delivery through exosomes. Interestingly, the exosomes produced by cardiac myocytes derived from the heart of a T2DM rat model have a negative impact on co-cultured ECs; while the ones from healthy rats appear protective. The authors identify a culprit of the diabetes effect in miR-320, which was already found to be elevated in microvascular ECs prepared from the heart of diabetic rats [6].

miRs are short non-coding RNAs, each of which is capable of post-transcriptionally regulating a variety of target genes. Since their recent discovery, miRs have captured the interest of basic and clinical scientists, who are trying to exploit them as new therapeutic targets and biomarkers. miRs have already been found to be involved in angiogenesis [7] and to be deregulated in the diabetic endothelium and heart (reviewed in [8]). For example, we showed that miR-503 increases in ECs exposed to high glucose in culture and to diabetes in an in vivo mouse model of limb ischemia, and that it inhibits EC proliferation and angiogenesis [9]. Conversely, the Mayr group showed that the pro-angiogenic and vascular protective miR-126 is decreased in ECs exposed to diabetes [10]. Moreover, a potential role for miRs in vascular-cell-to-vascular-cell as well as in intra-cardiac communication is emerging [11] and [12]. Extracellular vesicles (EVs) are the new toy of the miR scientist family. What makes EVs attractive is that miRs exploit them to maintain a biologically-active form during safe trips out of the parent cell. In EVs, miRs cannot be attacked by ribonucleases and can even be transferred to new cells where they can take command of the posttranscriptional expression regulatory machinery. The most studied EVs are microparticles and exosomes.

Exosomes are lipid membrane-bound vesicles, 30 to 100 nm in diameter, which arise from inward budding of the late endosome with subsequent formation of multivesicular bodies. They are released into the extracellular space when the endosomes fuse with the cell membrane. First described in mammalian reticulocytes [13], exosomes are released by different cell types, including cardiovascular cells, and can be isolated in conditioned cell media and in bodily fluids. Initially thought to be merely cellular debris, emerging research has suggested that exosomes serve as a messenger of intercellular communication, with the potential to target distant cells and deliver their enclosed cargo of proteins, lipids, mRNA and miRs [14]. The contents of exosomes are thought to be reflective of, at least in part, their underlying cell of origin [15] and, as an extension, the health status of the parent cell. Secreted exosomes can interact with cells in close proximity, or be released into the systemic circulation, and exosomes from a specific cell of origin may be able to be selectively internalized by only specific target cells [16], although the molecular mechanisms surrounding this remain largely undetermined.

In the cardiovascular field, recent studies have shown that exosomes extracted from peripheral blood-derived plasma are a rich source of miRs. However, the specific source of these miR-rich exosomes, their mechanisms of release and their functions are yet to be clearly defined. It is interesting to note, however, that cardiomyocytes have been suggested to secrete exosomes, with in vitro reports using primary rodent cardiomyocytes providing this evidence [17]. Work on adult cardiomyocytes demonstrated that they released exosomes containing the cytosolic heat shock protein HSP60, especially under hypoxia, which could have a negative effect on surrounding cells [18]. Furthermore, pathological aberrations in the local milieu, such as fever and pH imbalance, may alter the rate of exosome secretion as well as their protein content [19]. The profiles of cardiac exosome contents differ significantly from other types of exosome described in the literature [20]. The cardioprotective miR-214 has previously been shown to be upregulated in the heart after ischemia [21]. miR-214 is secreted via exosomes from human ECs [22] and is known to induce proangiogenic responses in endothelium-derived “exosome-recipient” ECs [22]. Moreover, Hergenreider et al. found that EVs (the majority of which were of the exosomal size range) from human umbilical vein ECs (HUVECs) can target smooth muscle cells (SMCs) in vitro [23]. Specifically, HUVECs were transfected with a Caenorhabditis elegans miR not found in humans (cel-miR-39) and then either (1) co-cultured with human aortic SMCs, or (2) the EVs secreted from HUVECs were isolated and added to an SMC culture. In both cases, the SMCs were able to express cel-miR-39, with expression increasing with an increasing duration of exposure, indicating continuing vesicular transfer [23]. Furthermore, vesicles from HUVECs placed under shear-stress demonstrated an enrichment of miR-143/145 – which are miRs that are known to control SMC phenotype – and when co-cultured with SMCs, they caused repression in known target genes, confirming functionality in the vesicle-transferred miRs [23]. A recent report by Bang et al. described another example of exosome-mediated intercellular communication, this time directly relevant to the myocardium, investigating the role of miR-21* in cardiac hypertrophy [24]. miR-21* – which is the non-dominant passenger strand of miR-21, and is usually degraded in the cytoplasm – was found to be selectively packaged into exosomes in cardiac fibroblasts and shuttled to cardiomyocytes, where it promotes hypertrophy [24].

Here, Wang et al. point to a mechanism of exosomal transfer of miR-320 from cardiomyocytes to ECs in the setting of type 2 diabetes [25]. miR-320 has previously been shown to be elevated in cardiac microvascular ECs isolated from T2DM rats, although seemingly conflicting reports found that miR-320 was downregulated in ECs cultured under conditions of high glucose [26]. The present study used cardiomyocyte and EC co-cultures to elucidate the interplay between these two cell types, specifically using cardiomyocytes isolated from either type 2 diabetic Goto–Kakizaki (GK) or non-diabetic rats and a line of cardiac ECs (MCECs) [25]. Compared to MCECs cultured alone, GK cardiomyocyte and MCEC co-culture resulted in a reduced EC functional capacity (proliferation and migration) [25]. Interestingly, co-culture of cardiomyocytes from a Wistar rat (WT) non-diabetic control with MCECs improved proliferation and migration, suggesting that under normal conditions ECs may benefit from growing in proximity to cardiomyocytes. They then went on to study the exosomes from primary cardiomyocyte cultures, demonstrating first that the size and molecular markers of exosomes from GK and WT cardiomyocytes are similar. Incubating ECs with the exosomes derived from diabetic cardiomyocytes resulted in impaired proliferation, migration and tube formation, whereas myocyte-derived exosomes from non-diabetic rats improved these responses, suggesting a direct effect of the exosomes on angiogenic mechanisms. As diabetic microangiopathy is associated with EC apoptosis [27], it would have been important to study the impact of the “diabetic exosomes” on EC death. Diabetic cardiomyocyte-derived exosome transfer to MCECs was accompanied by an increase in miR-320 expression and a reduction in the expression of the previously-validated targets HSP20 [28], IGF-1 [6] and the transcription factor ETS2 [29]. This response was significantly reduced when the co-culture was treated with GW4689, an inhibitor of exosome formation and release, suggesting that exosomes are responsible for the changes seen. Further experiments demonstrated reduced miR-320 expression in MCECs cultured under high glucose conditions to mimic diabetes in vitro, highlighting that glucose itself does not have a direct effect on miR-320 expression in this cell type, and confirming previous reports [26]. Finally, the inhibitory effects of diabetic cardiomyocyte-derived exosomes on angiogenic mechanisms were suppressed when miR-320 expression was inhibited in the parent cell line. In fact, the authors found that miR-320-reduced exosomes actually promoted MCEC migration and tube formation compared to control exosomes. Wang et al. consequently propose that healthy (non-diabetic) cardiomyocytes can use exosomal miR and protein transfer to maintain normal myocardial angiogenesis, and that expressional changes that occur in diabetes result in an altered exosome profile with anti-angiogenic functions [25].

miR-320 aside, exosomes encapsulate many specific proteins, including enzymes, and miRs that are reflective of the content of their parent cells [30], [31] and [32]. It is therefore possible, if not likely, that other exosomal contents that are cargoed to ECs may additionally contribute to the impaired angiogenesis in the myocardium in diabetes. Wang et al. suggest that a possible therapeutic strategy to improve angiogenesis may be to use artificially-engineered exosomes that include large amounts of angiogenesis-promoting miRs or proteins. This is a fascinating concept which might translate to regenerative medicine applications.

While the mechanism of exosome-mediated angiogenesis inhibition by miR-320 in diabetic conditions appears convincing in vitro, it remains to be seen whether miR-320-rich exosomes impair myocardial angiogenesis in vivo. Indeed, these data, and others from work on cultured primary cardiomyocytes, should be interpreted with care, given that the secretory capacity of any cell in vitro may differ from the reality in vivo, and that some exosomes harvested from conditioned media may in fact originate from contaminating cells of the primary cell line, especially fibroblasts. Finally, a note of caution should be added as currently there is no method allowing the preparation of pure exosome from any sort of biological material. The exosome preparations used in this and any other study would have been “contaminated” to a greater or lesser extent by EVs of different sizes, as well as by non-exosomal protein contaminants [33]. This should be taken into account when interpreting results.

In summary, Wang et al. present novel findings relating to the exosomal transfer of miR-320 from diabetic cardiomyocytes to ECs, and the consequent anti-angiogenic effects in vitro. Further study in the in vivo setting is required to delineate the efficacy and functional relevance of this mechanism. Moreover, methodological advancements of the systems used to purify and detect exosomes will be important to allow progress in the field.

Sources of funding
British Heart Foundation, National Institute of Health Research (NIHR)Bristol Cardiovascular Biomedical Research Unit and a Leducq Foundation Transatlantic Networks of Excellence grant in vascular microRNAs (MIRVAD).

Acknowledgment
SS is a British Heart Foundation (BHF) PhD student; GDA is BHF Chair in Cardiac Surgery and a National Institute of Health Research (NIHR) senior investigator; CE is a BHF senior research fellow. The Authors are part of the NIHR Bristol Biomedical Research Unit in Cardiovascular Medicine.