Keywords
microRNA - platelet function - biomarker
Introduction
Platelets play a key role in maintaining hemostasis[1] and vascular integrity; they are also involved in regulating inflammation,[2] immunity,[3] and tumor metastasis.[4] Platelets' main role is stopping bleeding during vascular lesions by accumulating
at the site of vessel injury, thanks to their adhesive properties. Contact with the
subendothelial components exposed in damaged vessels promotes platelet activation,
aggregation, and thrombus growth. Additionally, platelet activation induces phosphatidylserine
exposure at the membrane surface, activating the coagulation cascade and formation
of a stable clot.[5] Platelets' ability to promote hemostasis—platelet reactivity (PR)—is a variable
phenotype in individuals taking or not taking aspirin,[6]
[7] and may be associated with bleeding or thrombotic events. The determinants of PR
have yet to be fully elucidated, but family-based studies point to a genetic origin.[8]
In recent decades, miRNAs' roles in the regulation of various biological processes
and the progression of several diseases have gradually emerged. miRNAs are noncoding
sections of RNA approximately 22 nucleotides long; they regulate gene expression at
the post-transcriptional level by degrading mRNA or repressing its translation. Increasing
numbers of clinical trials have shown correlations between PR and levels of circulating
miRNA, revealing several platelet-derived miRNAs as putative biomarkers of platelet
function or the recurrence of ischemic events.[9]
[10]
[11]
[12]
[13]
[14] Nevertheless, a causal link between circulating miRNA and platelet function or thrombus
formation remains elusive, and mechanisms involved in miRNA-mediated regulation of
platelet function are poorly understood. This may be due to the many miRNA–mRNA duplexes
potentially involved and the variety of biological pathways that could be either upregulated
or downregulated by miRNA levels. Validating candidate miRNAs identified as being
regulators of platelet function in association studies is a major challenge, and of
utmost importance to select them as potential biomarkers and to identify the underlying
mechanisms.[15]
miRNA Biogenesis
In humans, miRNA loci are mainly located in intronic regions, in both transcription
coding or non-coding units. Approximately 10% of all miRNAs are encoded by exons.[16] The genes that encode for miRNAs are transcribed by either RNA polymerase II or
III into primary miRNA transcripts (pri-miRNA), depending on the miRNA loci and their
upstream promoters.[17]
[18] Pri-miRNA consists of a stem and a terminal loop flanked by two single-stranded
segments. The single-strand–double-strand junction of the pri-miRNA hairpin is cleaved
by Drosha, an RNase III endonuclease coupled to the DiGeorge syndrome critical region
8 (DGCR8) at approximately 11 bp from the pri-miRNA stem, which allows the release
of the precursor miRNA (pre-miRNA). Exportin-5 forms a complex with the guanine triphosphate
(GTP)-bound form of the small nuclear guanine triphosphatase (GTPase) Ran (RanGTP).
This complex recognizes the pre-miRNA structure,[19] protects pre-miRNA from degradation, and promotes the export of properly processed
pre-miRNA into the cytoplasm.[19]
[20] In the cytoplasm, the double-strand pre-miRNA is cleaved by the RNase III endonuclease
Dicer, resulting in the generation of two partially complementary strands of an miRNA,[20] each with its own targets and regulating specific biological functions.[21]
Mature miRNA recognizes six to eight nucleotides located on the 3′-untranslated region
(3′-UTR) of mRNA.[22] Moreover, one miRNA can target several mRNAs and regulate several biological functions.
Conversely, one mRNA can have a binding sequence for multiple miRNAs; therefore, the
same gene may be regulated by several miRNAs. The miRNA–mRNA duplex guides the binding
of the Ago2 protein in an RNA-induced silencing complex,[23] allowing mRNA cleavage, mRNA decay, and/or translational repression ([Fig. 1]).
Fig. 1
miRNA biogenesis. miRNAs are mainly transcribed from the intronic region by polymerase II/III. The
pri-miRNAs is processed into pre-miRNAs by the Drosha–DGCR8 complex and is then exported
to the cytoplasm by exportin-5 for further maturation by the Dicer protein. Mature
miRNA binds to a seed sequence by imperfect complementarity on the 3′-untranslated
region (3′-UTR) of an mRNA. The miRNA–mRNA duplex is stabilized by Ago2 into an miRNA-induced
silencing complex (RISC), thus triggering mRNA translational repression.
Platelets contain the necessary machinery for mRNA translation and protein synthesis.[24] In response to their activation, platelets synthesize proteins by signal-dependent
translation, using newly spliced mRNA and mature mRNA.[25] Platelets contain more than 500 miRNAs inherited from their parent cell, the megakaryocyte.[26] Therefore, it seems reasonable to speculate that these miRNAs have a fine-tuning function in the modulation of PR.
Circulating miRNAs
After stimulating various cells, including cell fragments such as platelets, circulating
miRNAs release their intracellular content into the extracellular medium. Current
understanding points to microvesicles being the main transporters of circulating miRNAs[27] since miRNAs' profiles in the blood are close to those of microvesicles.[28] However, the composition of this pool of extracellular miRNAs also includes miRNAs
bound to proteins. A recent study reported that miRNAs are more abundant in non-vesicular
than in vesicular fractions,[29] and that these vesicle-free miRNAs were associated with the Ago2 protein.[30] The authors concluded that Ago2 as a miRNA carrier could account for all vesicle-free
miRNAs; however, other miRNA complexes can also exist in plasma, such as with nucleoplasmin
or high-density lipoproteins.[31]
[32] The mechanism governing extravesicular biogenesis, transport, and cargo uptake in
recipient cells is not entirely understood and is reviewed elsewhere.[33]
Since approximately 45% of microvesicles circulating in the blood are released from
platelets,[28] the circulating miRNA profile may reflect the platelet miRNA content and provide
information on platelet activation status in real time. This concept is illustrated
by the impact of antiplatelet drugs on circulating miRNA,[34] meaning that the profile of circulating miRNA may provide important information
on PR in diseases where platelets play major roles, such as in the progression of
atherosclerosis,[35] cancer metastasis,[36] and inflammation.[3]
Associations between miRNAs and Platelet Reactivity in Cardiovascular Patients
Associations between miRNAs and Platelet Reactivity in Cardiovascular Patients
Several studies have addressed correlations between circulating miRNAs and PR or cardiovascular
events,[14]
[34]
[37] promoting the concept of using circulating, platelet-derived microRNAs (e.g., miR-223,
miR-126, miR-197, miR-150, miR-21, miR-96, or miR-204) as potential biomarkers for
the recurrence of cardiovascular events ([Fig. 2]).[15]
[38] It is of note that miR-223–3p and miR-126–3p, among the most abundant platelet-derived
miRNAs, have been particularly studied.
Fig. 2
Platelet-derived miRNA. Selected platelet-derived miRNAs and their experimentally validated mRNA targets
(in brackets) associated with atherosclerosis, platelet reactivity, and cancer. For
miR-197, no validated target related to these diseases was identified.
MiR-223 may mediate its effect on platelet function via the regulation of several
genes. For example, miR-223 represses P2Y12 mRNA translation, the ADP receptor targeted by antiplatelet drugs (thienopyridines
and ticagrelor).[24]
[39] Increased exosomal miR-223 was shown to be associated with the occurrence and severity
of acute ischemic stroke and short-term outcomes.[40] In patients with acute coronary syndrome, levels of circulating miR-223 in plasma
were statistically correlated with the PR determined using vasodilator-stimulated
phosphoprotein phosphorylation flow cytometry after clopidogrel or dual-antiplatelet
therapy. Moreover, a lower level of circulating miR-223–3p was the only independent
predictor of poorer response to treatment, as determined using a PR index.[39]
[41] These results indicated that miR-223 might be useful in predicting disease severity
and assessing responsiveness to antiplatelet agents.
Zampetaki et al showed that diabetes was associated with a reduced level of miR-126
in plasma,[42] which may reflect a platelet dysfunction in diabetic populations. A study of patients
with type 2 diabetes showed a direct correlation between platelet activation, as assessed
by measuring soluble P-selectin, and the level of circulating miR-126. In addition,
the administration of aspirin induced a decrease in the level of circulating miR-126
in this population.[43] Another study, of patients with diabetes mellitus, showed that miR-126 exhibited
antithrombotic properties by regulating post-transcriptional tissue factor expression.
Optimizing antidiabetic treatment using metformin or sulfonylurea was associated with
the upregulation of miR-126 in plasma, which correlated with reduced thrombogenicity.[44] However, this association may not be true in nondiabetic patients.[45] In recent years, several hypotheses have emerged regarding the underlying mechanisms
mediated by miR-126's impact on platelet function. These include the modulation of
ADAM9, a mediator of platelet binding to collagen,[37] and of PLXNB2, a transmembrane protein known to support thrombus formation in mice.[46]
Other miRNA–mRNA duplexes have been described as putative regulators of PR, including
miR-96–VAMP8, described as having a role in the heterogeneity of PR.[47]
Platelet-Derived miRNAs and Atherosclerosis
Platelet-Derived miRNAs and Atherosclerosis
Accumulated evidence has indicated that many miRNAs, including platelet-derived miRNAs,
can affect the initiation and progression of atherosclerosis and the development of
its complications via endothelial cell (EC) function regulation, including the maintenance
of vascular integrity (miR-126), EC proliferation (miR-126, miR-17, miR-18, miR-20a),
migration (miR-218, miR-106b-25 cluster), and senescence (miR-34, miR-217, miR-146a;
see [Fig. 2]).[48]
miRNAs are important regulators of lipoprotein homeostasis and lipid accumulation.
Among them, miR-223 represses genes involved in cholesterol biosynthesis (HMGS1, SC4MOL) and HDL uptake (SRB1).[49] The miR-33 family targets the HDL–reverse cholesterol transport pathway.[50]
[51]
Overexpression of miR-146a can delay both inflammatory response and oxidized low-density
lipoprotein accumulation by inhibiting activation of the toll-like receptor 4–dependent
signaling pathway.[52] MiR-30c has been shown to decrease the production of apoB-containing lipoproteins
(very low density lipoprotein, low-density lipoprotein) by targeting both the microsomal
triglyceride transfer protein (MTTP), a protein essential for the lipidation of nascent apoB, and lysophosphatidylglycerol
acyltransferase 1 (LPGAT1), which reduces de novo lipogenesis.[53]
miRNAs can also play roles in the regulation of proinflammatory response. The inhibition
of endogenous miR-125a-5p expression has been correlated with an increased inflammatory
cytokines tumor growth factor-β level, tumor necrosis factor-α, interleukin 2 (IL-2),
and interleukin 6 (IL-6).[54]
Additionally, other miRNAs, including miR-126–3p, miR-17–3p, and miR-31, regulate
the development of vascular inflammation by controlling the expression of adhesion
molecules, such as vascular cell adhesion molecule 1, intercellular adhesion molecule,
and E-selectin, respectively.[54]
[55] Other miRNAs facilitate calcification by targeting Ets1 (miR-125b) or, on the contrary,
inhibit calcification (miR-29a and miR-29b) by suppressing the expression of ADAMTS-7.[56]
[57] Furthermore, microRNAs play roles in plaque stability, with either stabilizing effects
(miR-222, miR-24, miR-26a, miR-27b, let-7e) or destabilizing effects (miR-21, miR-181,
miR-712, miR-29).[58] Finally, several microRNAs, including miR-155–5p, miR-483–5p, and miR-451a, have
been reported as potential biomarkers for the early identification of plaque rupture.[59]
Among the microRNAs identified as playing roles in the development of atherosclerosis,
several are found in platelets and in platelet-derived microvesicles, including miR-223,
miR-126, miR-24, miR-155, and miR-21. Therefore, platelet-derived microRNAs may contribute
to the development of atherosclerosis at every stage leading to the formation of atherosclerotic
plaques.
Platelet-Derived miRNA and Cancer
Platelet-Derived miRNA and Cancer
Besides their potential role in PR and cardiovascular diseases, activated platelets
and platelet-derived microvesicles are suggested to be implicated in the progression
of cancers.[60] Cancer patients are also at an increased risk of platelet-driven venous thromboembolism.[61]
Several miRNAs enriched in platelet-derived microvesicles can target both tumor suppressor
genes and oncogenes (in multiple cancer types), among them miR-27a, miR-24, miR-155,
miR-195, let-7a/b, and miR-223.[60]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69] The platelets and platelet-derived microvesicles of patients with non-small-cell
lung cancer contain higher levels of miR-223 than those of healthy individuals. The
incubation of A549 human lung cancer cells with platelet-derived microvesicles has
been shown to result in the rapid delivery of miR-223 into those cells, thus inhibiting
the EPB41L3 tumor suppressor and promoting A549 cell invasion.[67]
However, although platelet-derived microvesicles are generally considered to participate
in cancer progression, recent data suggest that they may have a cancer-suppressive
effect. Platelet-derived microvesicles infiltrate solid tumors in humans and mice
and transfer their RNA content, including miRNAs promoting tumor cell apoptosis. MiR-24
was a significantly present species in this transfer. In another model, where human
platelet-derived microvesicles were transfused to mice, mitochondrial mt-Nd2 and Snora75—two small, noncoding, nucleolar RNAs—were identified as direct targets of platelet-derived
miR-24. These target RNAs were suppressed in platelet-derived, microvesicle-treated
tumor cells, resulting in mitochondrial dysfunction and tumor growth inhibition.[60]
Analytical Pitfalls in Measuring Circulating miRNAs
Analytical Pitfalls in Measuring Circulating miRNAs
miRNAs are highly stable in biological fluids[70]
[71] and could therefore be interesting for developing a new type of biomarker. Over
recent decades, numerous studies have investigated using circulating miRNAs as potential
biomarkers,[37]
[39]
[72]
[73]
[74] raising several issues regarding sample types, extraction and quantification procedures,
and clinically significant cut-off values.[75]
Sample selection is important. Indeed, serum samples may reflect the total miRNA content
in platelets and miRNAs derived from the microvesicles of other blood cells that can
be activated during the coagulation process. miRNA quantification using plasma samples
is probably more likely to reflect basal platelet function status and to be better for evaluating the risk of ischemic event
recurrence; thus, plasma samples are preferred.[75]
[76] Freezing and thawing cycles can lyse residual cells and release all of their miRNA
content, biasing miRNA quantification. Therefore, plasma samples should be prepared
using a double centrifugation step, and measuring platelet-specific or leukocyte-specific
proteins (e.g., ITGA2B or PTPRC, respectively) can validate the absence of any residual
cell contamination.[77]
The miRNA extraction procedure is usually performed using a TRIzol-based method followed
by (or not) a standardized column extraction process.[39] This only enables a small amount of miRNA to be extracted, however; so, a preamplification
step is usually performed to optimize subsequent quantification. The ratio between
sensitivity and the number of targets differs according to the quantification technique.
Indeed, a complete miRNA profile can be evaluated using high throughput techniques
such as RNA sequencing,[37] whereas custom-designed locked nucleic acid (Exiqon)[37]
[78] and NanoString Technologies[73]
[74] enable the quantification of approximately 100 miRNAs targeted in one sample. The
most sensitive technique remains the TaqMan-based quantitative polymerase chain reaction
on complementary DNA,[75]
[79] which is considered the gold standard. Because only a few miRNAs can be evaluated
at once, however, candidate miRNAs must have been identified in a previous study.
Finally, although the normalization procedure remains challenging, it is of utmost
importance for comparing samples and studies. A wide range of strategies has been
developed,[80] including using exogenous spike-in (e.g., UniSP6[14] and cel-miR-39[81]), endogenous, small, non-coding RNA (e.g., RNU6[82]) or endogenous miRNA(s). The spike-in is useful for evaluating the quality and reproducibility
of the extraction and quantification process, whereas endogenous RNA or miRNAs are
more adapted to normalize the level of quantified miRNA. However, endogenous normalizers
from the same class of nucleotide sequence should be preferred. Endogenous miRNA normalizers
differ according to the sample used—for example, miR-638, miR-93, and miR-484 are
used for plasma,[83]
[84]
[85] whereas miR-23a, let7a, and miR-1260 are preferred for serum.[78]
[86]
[87] Using a combination of stable normalizers, determined using dedicated tools (e.g.,
GeNorm[88] and NormFinder[89]), enhances the efficiency of the normalization.[87]
Functional Validation of miRNAs
Functional Validation of miRNAs
Although several miRNAs have been pinpointed as potential biomarkers of PR or the
recurrence of ischemic events, the functional validation of candidate miRNAs as true
regulators of platelet function remains a crucial step. It will be the cornerstone
criterion formally linking a specific miRNA and platelet function regulation.[15] Platelets have a very short lifespan and only remain functional for a few hours
after blood sampling. This precludes the modulation of their miRNA content via transfection;
thus, alternative strategies should be used. Several in vitro models are available
using megakaryocyte lines (MEG01, DAMI, or K562) or megakaryocytes from hematopoietic
stem cells (CD34) differentiated into platelet-like structures (PLSs). The latter
model has the advantage of producing PLSs that are functionally close to human platelets.[46]
[90] However, this model's main limitation is the small number of cells recovered, which
means that it is not compatible with all of the conventional methods dedicated to
the study of platelet function. Modulation of the miRNA content in these in vitro
models can be done using lipofection or nucleofection to transfect the megakaryocytes
with synthetic miRNA mimics or miRNA inhibitors. An alternative method uses lentiviral
transduction, where the lentiviral vector genome is integrated into the host genome,
leading to stable miRNA expression.
Several in vivo models are available, including zebrafish and mouse models that have
hemostatic systems close to that of humans. The zebrafish model is widely used. The
transparency of zebrafish embryos and larvae between 24 hours and 5 days after fertilization
facilitates the visualization of their vessels and real-time thrombocyte accumulation
after laser injury.[15]
[45] Modifications in thrombocyte miRNA content can be obtained by genetic modifications.
Inserting a sequence coding for the miRNA of interest, under the control of CD41 promoter,
results in the generation of a transgenic line that only overexpresses miRNA in thrombocytes
without affecting the levels of other miRNAs.[45]
The mouse is a model of choice for studies related to hemostasis. It allows the performance
of aggregation tests, monitoring the exposure of CD62P and the expression of activated
GPIIb–IIIa on platelets using flow cytometry, and the assessment of tail bleeding
time. Several gene-editing strategies can be used, including transgenesis, homologous
recombination using the Cre/loxP,[91]
[92]
[93] and the TALEN[94] and CRISPR[95] systems that can successfully induce miRNA deletions. The transduction of a lentiviral
vector specifically designed to stably overexpress a specific miRNA can be injected
into mice.[96] Bidirectional miRNA reporter lentiviral vectors have shown promising results for
post-transcriptional (miRNA-restricted) repression in hematopoietic stem cells in
mice via the insertion of target sequences for hematopoietic stem cell–specific miRNA
into the lentiviral cassette.[97]
In contrast, a transient modulation of a miRNA level can be obtained by systemic injection
of locked nucleic acid, a chemical inhibitor.[37]
Generating transgenic lines has obvious advantages for the functional investigation
of a miRNA's impact on PR and hemostasis. However, animal models must be used cautiously
due to the possible non-conservation of miRNA–mRNA duplexes, which can differ from
miRNA's function in humans.
Identification and Validation of miRNA Targets
Identification and Validation of miRNA Targets
The in silico identification of target miRNAs using databases such as Targetscan,
MiRanda, or DIANA-microT represents the first step in deciphering the mechanisms linking
miRNA and platelet function. However, different databases may identify variable numbers
of putative targets. For example, the Targetscan database (Release 7.2) finds 423
putative targets for miR-223, whereas the MiRanda database finds 3,485. These databases
are also known to provide high numbers of false-positive results[98]; therefore, any miRNA–mRNA duplexes predicted in silico must be validated. The reporter
gene assay is one of the most widely used methods to identify an miRNA's binding site
on the target gene's 3′-UTR and to validate the direct interaction between an miRNA
and its target. In addition to P2Y12, calpain-2 (known to regulate the platelet signaling
pathway)[99] has also been validated as a direct target of miR-223–3p.[100] Several other predicted targets of miR-223–3p have been validated using this method,
including PDGFRβ,[101]
[102] EPB41L3,[67] Septin-2, and Septin-6[103]; however, their direct role in platelet function remains poorly understood.
Conclusion
There is increasing evidence of associations between levels of specific miRNAs and
PR and the recurrence of ischemic events.
The significant methodological differences between different clinical association
studies, in terms of their biological samples, miRNA extraction methods, quantification,
and clinical settings, make functional validation of the utmost importance when attempting
to provide a causal link and a mechanistic insight into the role of candidate miRNAs
in regulating PR.[15] These functional validation data are more limited but are critical to selecting
the most promising miRNAs as biomarkers for diseases where platelets play a major
role.
Taken together, in the near future, the profile of a patient's circulating miRNA could
be evaluated for the biomarkers of various platelet-related diseases and contribute
to the emergence of personalized medicine.
