In the present study we have analyzed the potential effect of DOAC treatment on miR-27a-3p expression in naïve AF patients, a subgroup of miR-CRAFT study participants, starting oral anticoagulation with rivaroxaban, apixaban or dabigatran. To the best of our knowledge, this is the seminal study to study miR-27a-3p expression over time in DOAC treated AF patients. Our results show that 7 days of DOAC therapy did not alter miR-27a-3p expression either in pooled population or after stratifying patients by DOAC, sex, bleedings and comorbidities.

miR-27a, generating the mature miR-27a-3p and miR-27a-5p microRNAs, has been characterized as an oncogenic miR [13] and, more recently, as a contributing factor to toxicity incidence to fluoropyrimidines via inhibition of dihydropyrimidine dehydrogenase encoding gene (DPYD) [14]. miR-27a-3p is the major isoform targeting nearly 1,500 genes versus 251 targets of miR-27a-5p isoform [10]. Beyond its role in cancer, miR-27a-3p targets span different molecular pathways, including the coagulation cascade. More specifically, among the proteins implicated in blood coagulation pathway, miR-27a-3p regulates the expression of TFPI (miRDBD target score 72), TF (miRDBD target score 57) and TFPI2 (miRDBD target score 56) [10]. TFPIs directly inhibit factor Xa, the target of rivaroxaban and apixaban, and TF-FVIIa, preventing thus the formation of an active prothrombinase and regulating the generation of thrombin [15, 16]. Additionally, TFPIs modulate the severity of a wide variety of bleeding and clotting disorders [17]. TF, the initiator of blood coagulation, triggers the initiation of thrombin formation [18]. Given the crucial role of both TFPIs and TF on coagulation homeostasis maintenance, the rational of our hypothesis in the present study is that the DOAC inhibiting effect on factor Xa and thrombin results to an amplification reaction towards anticoagulation that mechanistically involves the altered expression of miR-27a-3p and the modulation, thus, of TFPIs and TF expression.

The results of our study show that 7 days of DOAC treatment do not alter the expression of miR-27a-3p either in the pooled patient population or per drug of DOAC drug class. Since this is the seminal study on the potential effect of DOACs on mir-27a-3p expression, no further comparisons can be made with the published literature. In the field of AF, however, the potential use of miRs as circulating biomarkers of the disease increasingly gains attention [19]. miR-27a-3p was highlighted as a co-DEG AF-related stroke in a bioinformatic gene analysis seeking for potential biomarkers and therapeutic targets of this condition [20]. More recently, in a study aiming to investigate the regulatory effect of lncRNA GAS5 on the electrical remodeling of neonatal rat cardiomyocytes induced by rapid pacing, Xi et al. have shown that the overexpression of GAS5 attenuated electrical remodeling and that miR-27a-3p had a direct negative regulatory effect on lncRNA-GAS5 [21]. Though this is still a naïve field, the potential role of miR-27a-3p on AF incidence merits further investigation. Additionally, the heterogeneity between various DOACs on their pleiotropy [1], prompted us to assess whether there is a different effect on miR-27a-3p expression between thrombin inhibitor and factor Xa inhibitors, as well as among different types of factor Xa inhibitors. Our results do not provide evidence for such an effect, however, results should be replicated in larger studies.

Recent evidence suggests that sex-differences exist in AF pathophysiology [22]. Additionally, a sex-specific expression of miRs has been recognized in both physiological and pathological processes [23]. For miR-27a-3p, sex-specific correlations that were dependent on cancer stage were observed in female patients [24]. To reduce sex-confounding in our study, we have compared the baseline miR-27a-3p expression between male and female AF patients. We have found that female patients presented at baseline with higher miR-27a-3p expression compared to male patients and a reduction of its expression after 7 days of DOAC treatment was found only in female patients, however none of these differences reached statistical significance. To further study the potential effect of sex on miR-27a-3p expression, larger studies are necessary.

Similarly to sex, and according to published literature, it cannot be excluded that other common comorbidities may interfere with miR-27a-3p expression. Some evidence exists linking the expression of miR-27a with hypertension [25, 26], type 2 diabetes [27, 28] and dyslipidemia [29, 30]. In our study, we did not find a different in miR-27a-3p expression in either situation, suggesting that hypertension, type 2 diabetes and dyslipidemia are not confounding factors in the study population.

Our study has several strengths. miR-CRAFT is a well-controlled study, and the study population is balanced for two major confounders of miR expression, cancer and insulin treatment. For cancer, it is well known that the expression of several miRs can be up- or down-regulated and oncology is currently the field of medicine that benefits the most from the characterization of miRs that can be used as prognostic markers of disease. Increasing evidence shows that miR-27a plays a vital role in tumor biology acting either as an oncogene or a tumor suppressor gene depending on the type of cancer [31] and that miR-27a-3p is abnormally expressed in various types of cancers [32]. Recently, it has been shown that insulin treatment induced the downregulation of miR-27a-3p expression in human granulosa-like tumor cell line [33]. These confounders are both patient exclusion criteria to reduce confounding [11]. Moreover, both the intraday and intraindividual day-to-day variances of miR expression are currently scarcely studied. For selected miRs, such as miR-125a, miR-146a, miR-155, let-7e and miR-106a evidence proposes that no common circadian variances exist in each miR expression in plasma, however, intraindividual day-to-day variances in miR expression in plasma were noticed [34]. In our study, in order to reduce the possibility of such variability, we have used strict criteria for sampling timing, and all blood samples on 7 days of DOAC treatment were drawn in the morning and prior to the scheduled drug dosage [11]. Additionally, in our study other comorbidities did not affect miR-27a-3p expression. miR-CRAFT study is adequately powered to detect a 1.5 fold-change in miR expression with an 80% average power within groups.

However, it should be acknowledged, that several limitations also exist. Mainly, the results should be interpreted with caution, especially in sub-group analyses due to low sample size, to avoid false negative results. The design of the study is focused on the effect of DOACs on miR expression and, thus, other endpoints such as the role of miR-27a-3p in AF-related coagulopathy and thrombosis risk cannot be assessed. We herein report results of the effect of DOAC treatment on miR-27a-3p expression at 7 days of therapy. Since miRs respond quickly to stimuli such as drug treatment, we did not further seek for differences at 28 days. However, the lack of association in early response found in our study cannot exclude a potential late response effect of DOAC treatment on miR-27a-3p. Moreover, despite the finding that DOACs do not alter miR-27a-3p expression, TF and TFPI could be both altered by DOAC treatment irrespectively of miR-27a-3p regulation pathway; however, this analysis is beyond the scopes of our study.