Using miR sequencing of KD acute and convalescent whole blood samples as a discovery approach, we identified 6 differentially expressed miRs (miRs-143, -199b-5p, -618, -223, -145 and -145*) that were validated in an independent cohort by qRT-PCR. Among these, miR-145 was uniquely expressed in acute KD vs. acute adenovirus infection. Importantly, the level of miR-145 did not significantly correlate with peripheral white blood cell counts, suggesting that this miR mediates biologic processes distinct from processes associated with elevation of these cell counts. Predicted targets of miR-145 are highly enriched for genes that play critical roles in the TGF-β signaling pathway.
Role of miR-143/145 in Vascular Biology
miR-143 and -145 are encoded in close proximity to each other on human chromosome 5 (≈1.7Kb apart) and share critical roles in the differentiation of neutrophils 
, stem cells 
, and VSMC 
. In acute KD samples, levels of miR-143 and miR-145*, but not levels of miR-145, correlated with mature neutrophil counts. In the vasculitis of KD, neutrophils participate in the early destruction of the media and are detected in the damaged coronary arterial wall from autopsies of KD patients who died within 14 days after fever onset 
. Although miR-143/145 is usually considered to be a bicistronic locus under the control of a common promoter 
, the existence of an internal regulator for miR-145 transcription has been suggested in stem cells and could be operating in other cell types 
. Levels of miR-143 and miR-145* were highly correlated with each other but showed only weak correlation with miR-145 ().
Recently researchers have focused on the importance of miR-143/145 in VSMC phenotype switching 
. LacZ reporter gene experiments have shown that miR-143/145 is initially expressed in the developing murine heart at an early embryonic stage (E8.5 to E9.5) and that the expression becomes restricted to SMCs of various organs including the aorta, heart, and coronary arteries during subsequent fetal stages (E16.5) 
. VSMC, unlike cardiomyocytes and skeletal muscle cells, retain plasticity and can switch from an α-SMA expressing contractile phenotype (regulated by miR-145) to a proliferative, synthetic phenotype 
. α-SMA positive/smoothelin negative myofibroblasts, whose generation is influenced by TGF-β, participate in the destruction of the arterial wall in acute KD and it is intriguing to speculate that miR-145 modulates the generation of myofibroblasts from VSMC 
. Another feature of the vascular injury in acute and subacute KD is the development of myointimal proliferation, which is known to be regulated by miR-145 
Several lines of evidence suggest that animal models of non-inflammatory vascular injury or human diseases characterized by chronic vascular inflammation are associated with low levels of miR-145. miR-145 expression was down-regulated in tissues at the site of experimental injury-induced lesions in animal models 
and in human aortic aneurysms 
. Similarly, levels of miR-145 in the plasma or serum of patients with stable coronary artery disease (CAD) were low compared to healthy controls 
. Thus, conditions associated with chronic vascsular inflammation are associated with low levels of miR-145. In contrast, levels were high in our acute KD subjects and lower in the convalescent phase when inflammation had resolved. We postulate that miR-145 down-regulates TGF-β signaling, thus leading to resolution of the myofibroblast-mediated acute inflammation.
Although miR-145 is predicted to regulate transcript levels of several genes in the TGF-β pathway, including SMAD2, 3 and 4 (Table S6
), confirmatory luciferase reporter gene experiments have not yet been performed. Studies of the target genes of miR-145 may deepen our understanding of the role of miR-145 in KD pathogenesis. miR-145 also affects VSMC phenotype by targeting calcium/calmodulin-dependent protein kinase II, delta (Camk2d) 
and angiotensin converting enzyme (Ace) 
. Genetic variation in CAMK2D 
and ACE 
has been linked to KD susceptibility and outcome.
miRs can be transported from cell to cell by small (~100 nm) extracellular vesicles and can block mRNA translation in recipient cells 
. Thus, miRs carried by extracellular vesicles may play a significant role in disease pathogenesis 
. Communication between endothelial cells and VSMCs by extracellular vesicles carrying miR-143/145 has been reported 
. We found that approximately 40% of miR-145 in plasma was contained within small extracellular vesicles (). Identifying the cellular origins and targets of these vesicles may contribute to our understanding of the role of miR-145 in KD. The use of circulating miRs and miRs in extracellular vesicles as biomarkers has been pursued for cancer and cardiovascular disease 
. Future studies with expanded cohorts of KD and febrile control patients will explore whether miR-145 contained within extracellular vesicles may serve as a biomarker for acute KD.
We identified three additional miRs that were differentially expressed in whole blood between the acute and convalescent stages of KD and were validated in independent patient cohorts. miR-199b is regulated by the transcription factor NFAT, which has been implicated in KD pathogenesis 
.This miR is increased in the setting of heart failure in humans 
. miR-23a ,which is also regulated by NFAT and up-regulated during cardiac hypertrophy 
, was differentially expressed based on the sequencing data (), but was not confirmed by qRT-PCR in the independent cohort (). While clinically significant ventricular dysfunction is rare in acute KD, it is well-documented that all KD patients have some degree of myocardial inflammation during the acute phase 
. In our analysis, it is interesting to note that the subject with highest miR-199b-5p levels by qRT-PCR had a low ejection fraction (56%) during the acute phase, suggesting more severe myocardial inflammation (data not shown).
miR-223 was also significantly differentially expressed between acute and convalescent samples. This miR is reported to be expressed in neutrophils, eosinophils, monocytes and platelets 
. miR-223 acts to fine-tune granulocyte production by the bone marrow and the inflammatory response by modulating the proliferation of granulocyte progenitors and maturation of neutrophils 
. In our study there was weak correlation between miR-223 expression and absolute neutrophil count (r
0.53) but not with counts of eosinophils, monocytes, or platelets. miR-223 expression levels showed strong correlation with miR-143 and miR-145*, suggesting that these miRs may be regulated by a shared mechanism ().
miR-618, another differentially expressed miR, has a polymorphism (rs2682818 at Chr. 12q21.31) on the complementary strand in the seed region 
. However, the mature sequence is not affected by this polymorphism and the functional significance is unknown. None of the KD genetic studies published to date has identified an association between KD and this locus on Chr. 12.
Pathway analysis of the predicted target genes for the six differentially expressed miRs identified the TGF-β signaling pathway. Although most miRs down-regulate translation of their target genes, enhanced gene expression regulated by miRs has been reported 
. In zebrafish, miRs play an important role in stabilizing intracellular processes by buffering expression fluctuations of opposing genes in the TGF-β pathway 
. It is intriguing to speculate that the abundance of miR targets in the SMAD 2, 3, 4 canonical pathway might lead to reduced signaling and to preferential shunting down the non-canonical MAPK pathway that has been recently implicated in other aneurysm syndromes 
. Experiments by Long and Miano using human coronary artery smooth muscle cells demonstrated the time-dependent increase in miR-145 levels and increase in non-canonical TGF-β pathway molecules, pERK and pp38MAPK, by Western blotting, following TGF-β stimulation 
. In addition, work by Zhu et al. showed increased levels of miR-145 and decreased levels of SMAD3 following TGF-β stimulation of human fibrobalsts 
. In KD we have demonstrated increased signaling through the canonical TGF-β pathway in the coronary artery wall 
. In a discovery analysis of the miR-ome in acute and convalescent KD, we discovered increased levels of miR-145 and other miRs that also negatively regulate TGF-β signaling (, Figure S3
). We postulate that miR-145 acts as a negative regulator of TGF-β signaling through the canonical pathway and thus may contribute to the down regulation of inflammation in the arterial wall.
Previously, we demonstrated that transcripts for TGFBR3 and SMAD3 were reduced during acute KD by microarray analysis and reduced expression of TGFBR3 was confirmed by qRT-PCR 
. Although these transcripts are predicted to be targets of miR-223, -618, -143 and -145, we did not see a correlation between these miRs and transcript levels of TGFBR3
in whole blood from acute KD patients (data not shown). It may be that analysis of changes in the peripheral blood does not necessarily reflect the status in the arterial wall, where these miRs may be acting to negatively regulate these molecules.
Strength and Limitation
We recognize several strengths and limitations to our study. This is the first study to analyze miR sequences in KD. The novel findings of differentially expressed miRs enriched for the TGF-β pathway further emphasizes the important of this pathway in KD pathogenesis. This is also the first study to examine extra-cellular vesicles in acute KD. The study demonstrated that miRs mediating KD pathogenesis may be associated with extracellular vesicles. The role of these vesicles in inter-cellular communication is a new intriguing avenue of research. The small sample size is a limitation in this type of study and may impact its general applicability. However, sequencing results were validated using a different method in two independent cohorts of KD and adenovirus-infected subjects. Another limitation is the use of RNA extracted from whole blood, which contains diverse cell types, including smooth muscle progenitor cells, extracellular vesicles and Argonaut protein-complexed miRs 
. Therefore the cellular source of the differentially expressed miRs is unclear. We did not see any correlation between aneurysm formation and miR-145 levels. This may be related to sampling in the peripheral blood vs. in the arterial wall.
In summary, our study demonstrates that elevated miR-145 levels are found in whole blood from patients with acute KD. To the extent that many of the predicted targets of miR-145 are involved in TGF-β signaling, we propose that miR-145 downregulation of the canonical TGF-β pathway may have an important role in recovery from the acute vasculitis. These results lend further weight to the importance of the TGF-β signaling pathway in acute KD.