Microbiota Thrombus Colonization May Inuence Athero-ThrombosIs in Hyperglycemic Patients With ST Segment Elevation Myocardial Infarction (STEMI). MARIANELLA Study

Objectives. We examined the association of the coronary thrombus microbiota and relative metabolites with major adverse cardiovascular events (MACE) in hyperglycemic patients with ST segment elevation infarction (STEMI). Methods. We undertook an observational cohort study of 146 rst STEMI patients treated with primary percutaneous coronary intervention (PPCI) and thrombus-aspiration (TA). Patients were clustered, based on admission blood glucose levels, in hyperglycemic (>140 mg/dl) and normoglycemic (<140 mg/dl). We analyzed gut and thrombus microbiota in all patients. Moreover, we assessed TMAO, CD40L and von Willebrand Factor (vWF) in coronary thrombi. Cox regressions were used for the association between Prevotellaspp and TMAO terziles and MACE. MACE endpoint at 1 year included death, re-infarction, unstable angina. we observed a signicantly different prevalence of both Prevotellaspp between with (n=56) and those with normal glucose levels (n=90). The abundance of Prevotella increased in hyperglycemic vs normoglycemic patients whereas the contrary was observed for Alistipes. Interestingly, coronary thrombus, the content of Prevotella was associated with admission blood glucose levels (p<0.01), thrombus dimensions (p<0.01), TMAO, CDL40 (p<0.01) and vWF (p<0.01) coronary thrombus contents. Multivariate Cox-analysis disclosed a reduced survival in patients with high levels of Prevotella and TMAO in coronary thrombus as compared to patients with low levels of Prevotella and TMAO, after 1-year follow up. statistically signicant. The study population was equally distributed for age (p = 0.807). To evaluate the independent contribution of hyperglycaemia in thrombus Prevotella and TMAO levels, we performed a correlation between Prevotella and TMAO levels in coronary thrombus and blood glucose level during STEMI. Risk adjusted Cox-regression analysis curves showing survival from severe disease through days of hospitalization. Cox models were adjusted for; age, diabetes, hypertension, diabetes, new diabetes, dyslipidemia, cigarette smoking, BMI, systolic blood pressure, total cholesterol, HDL and LDL-cholesterol levels, triglycerides, TIMI scores, lesions of RCA and LM, dual antiplatelet therapy and ejection fraction. To investigate the effects of Prevotella and TMAO thrombi expressions on cardiovascular endpoints, we evaluated STEMI outcomes at 1-year follow-up thrombus contents did not change despite a similar severity of atherosclerotic disease at baseline and asimilar glycemic control both at discharge and at follow-up. These observations suggest that on admission hyperglycemia may play a pivotal role in the increase of coronary thrombus burden, during STEMI, favoring thrombus colonization and proliferation by Prevotella.

monooxygenase 3 (FMO3) (10). Among the genera of gut microbiota, Prevotella has showed the strongest association with TMAO production (11). Baseline measures of 53 participants in a controlled feeding study showed that those with a Prevotella enterotype had higher plasma TMAO concentrations than those with a Bacteroides enterotype (12). Moreover, recent data showed the involvement of both Prevotella (14) and TMAO (13) in the arterial thrombosis. To date, evidence about the possibleinteraction between hyperglycemia, microbiota and its metabolitesin the development and progression of coronary thrombus in ST-elevation myocardial infarction (STEMI) patients is lacking. To de ne a more robust evidence of gut microbiota signature and TMAO in coronary thrombus development and progression during AMI, we assessed both gut and thrombus microbiota in hyperglycemic (blood glucose levels >140 mg/dl at admission) and normoglycemic patients with STEMI. Furthermore, we explored associations between thrombus microbiota and TMAO levels, and composite endpoints (MACE= death, re-infarction, unstable angina) in cohorts at 1-year.

Patients and study design
This was a multicenter observational prospective cohort study aimed to investigate the relationship between gut microbiota signature, TMAO thrombus levels, and 1-year outcomes in STEMI patients. We examined an observational cohort of consecutive patients with rst STEMI treated with PPCI and thrombus aspiration (TA), between February 2016 and September 2019 at the Department of Cardiology of Cardarelli Hospital in Naples Italy, at Department of Cardiology of the University of Campania "Luigi Vanvitelli" Italy, and at Department of Cardiac Surgery of the SS. Annunziata Hospital, Sassari, Italy. All patients with onset of symptoms <12 h and at least 1-mm ST-segment elevation in 2 or more contiguous limb leads or at least 2 mm in 2 or more contiguous precordial leads or left bundle branch block were considered eligible for PPCI. Coronary angiography was performed either via the radial or femoral artery.
The culprit lesion was identi ed and crossed with an angioplasty guidewire. Manual TA was performed at the discretion of the operator, due to technical aspects (e.g., type and number of stents, use of any other devices), and with consideration of angiographic selection criteria (e.g., the presence of a visible thrombus on angiography), followed by conventional PCI to the culprit vessel. We considered eligible for the study all patients with: correspondence between ECG ndings and suspected culprit artery; a minimum visual estimate of 50% stenosis in the culprit artery, and feasibility of performing TA, as judged by the treating physician; age of 18 years or greater; presentation to the cardiac catheterization laboratory for PPCI in the setting of rst STEMI. Patients with left ventricular ejection fraction less than 25%, with previous myocardial infarction or previous PPCI and/or coronary by-pass grafting, or who had received brinolytic therapy were instead excluded from the study.
Based on the admission blood glucose levels, patients were clustered in two groups: (a)hyperglycemic STEMI patients (glucose levels >140 mg/dl), as suggested by the position statement of the American Heart Association (ADA) (15) and normoglycemic STEMI patients (glucose levels <140 mg/dl). The study is in accordance with the principles outlined in the 1976 the Declaration of Helsinki and its later amendments for use of human tissue or subjects. The Institutional Review Board of University of Campania "Luigi Vanvitelli" Italy approved the protocol.

Procedures
Routine analyses were obtained on admission before coronary angiography and before the initiation of full medical therapy. Stool samples were instead obtained during the rst day after admission. Treatment was classi ed as ''thrombus aspiration'' when TA was attempted as suggested by the operator, irrespective of procedure success. Thrombus grade was classi ed in accordance with Sianos et al. (16) into: Grade 0 (G0), no angiographic characteristics of thrombus present; Grade 1 (G1), possible thrombus present, with the following angiography characteristics: reduced contrast density, haziness, irregular lesion contour, or a smooth convex meniscus at the site of total occlusion suggestive, but not diagnostic, of thrombus; Grade 2 (G2), de nite thrombus with largest dimension ≤½ the vessel diameter; Grade 3 (G3), de nite thrombus, with largest linear dimension >½ but <twice vessel diameter; Grade 4 (G4), de nite thrombus, with the largest dimension ≥2 vessel diameters; Grade 5 (G5), total occlusion, unable to assess thrombus burden due to total vessel occlusion. Detailed recommendations about TA procedure were further provided. The procedure was done by interventional cardiologists' investigators. All patients were treated with adenosine (given 120 mcg as a fast bolus followed by 2 mg in 2 minutes) and with bolus infusion of Abciximab (0.25 mg/kg IV bolus). Patients undergoing TA received the same treatment also after thrombus aspiration. Aspiration had to be started before crossing the lesion, with a minimum of two syringes (40 mL) of aspirate recommended. Investigators were appropriately trained to ensure that the guide catheter was engaged with the coronary ostia when removing the thrombectomy catheter. Finally, the guide catheter was aspirated after thrombectomy to avoid embolization of either air or thrombus from the guide catheter.
Stool sample collection and fecal microbiome analysis electrophoresis and then puri ed using AMPure Beads XP (cat. n. A63880, Beckman Coulter, Inc.). The amplicons quality was assessed by 1,6 % agarose gel electrophoresis, product visible as a band with a molecular weight of approximately 840 bp. A 6-bp barcode sequence was added to the ends of both forward and reverse primers in the PCR "index". PCR "index" products were examined and then puri ed as described above. Amplicons targets were normalized to 10nM and used to prepare the library for sequencing. The library was processed using kit MiSeq Reagent Nano Kit v2 (500-cycles) (cat. n: MS-103-1003, Illumina Inc.), following the user guidelines for the system MiSeq Illumina ("Denature ana Dilute Libraries Guide"-MiSeq System). The library was corrected by adding 5% of Phix Control v3 (cat. n. FC-110-3001, Illumina Inc.) and sequenced using an Illumina MiSeq platform. The amplicons were pooled, and sequencing was conducted at Ames Center (Naples, Italy), with an Illumina MiSeq sequencing system (Illumina, San Diego, CA, USA). Obtained data were processed with the MicrobAT (Microbiota Analysis Tool) system of the SmartSeq S.r.l.
16S rRNA gene sequence analysis. Sequence analysis of low quality with a lower quality score of 20, a length less than 50 bp, or any misalignments to the primer or barcode containing chimeras were removed.

Thrombus analysis
Thrombi obtained from matched STEMI patients were analyzed for microbiota colonization, TMAO levels and pro-thrombotic molecules. All analyses were performed by 3 independent pathologists, blinded to the patients' characteristics. After TA, the specimens were frozen in liquid nitrogen, stored at -20 °C and processed within 48 hours. After defrosted, the specimens were cut perpendicular to the long axis into two halves. The rst half was used for the following ELISA analysis: CD40L (Soluble-Human ELISA Kit, Invitrogen); von Willebrand Factor(VWF) (Human von Willebrand Factor ELISA Kit -ab223864); TMAO (TMAO elisa kit: Human Trimethylamine-N-oxide Reductase ELISA Kit MBS7254766, MyBiosource). A portion of the other half of the specimen was instead analyzed for microbiota colonization as previously described for stool samples.
In vitro studyof bacterial growth.
In order to analyze the growth response of Prevotella and Bacteroides toward different concentrations of glucose, bacterial strains were grown as previously described (17). Brie y, Prevotella and Bacteroides from their glycerol stocks were rstly cultured on blood agar plates, then single colonies were picked individually and cultured in an anaerobic chamber at 37°C overnight in Gifu Anaerobic Medium (OD 600 1.0-1.5). Bacterial cultures were centrifuged at 1300 × g for 5 min and culture pellets were washed. Twenty microliters of the culture were then added to 2 ml medium supplemented with different concentrations of D-glucose, 7 mM, 14mM, 22 mM and 32mM. Measures of optical density at 600 nm wavelength at 0, 12h, 24h, 36h, 48h, 60h and 72 h were performed in order to assess bacterial growth rates. All experiments were performed in triplicate. One-way analysis of variance with Tukey's test was performed using GraphPad Prism (version 6 for Windows, GraphPad Software, La Jolla, CA, USA).
Blood glucose control in emergency wards.
After coronary angiography procedures, all hyperglycemic patients with blood glucose > 180 mg/dl were treated with intensive glucose control to keep blood glucose levels between 140 and 180 mg/dl, as previously described (18). Continuous insulin infusion of 50 IU Actrapid HM (Novo-Nordisk) in 50 ml NaCl (0.9% using a Perfusor-FM-pump) was started only when blood glucose levels exceeded 180 mg/dl, and adjusted to keep blood glucose between 140 and 180 mg/dl. When blood glucose fell <140 mg/dl, insulin infusion was tapered and eventually stopped. After the start of insulin infusion protocol, a glycemic control was provided every hour in order to obtain three consecutive values that were within the goal range. The infusion lasted until achievement of stable glycemic goal (140-180 mg/dl) for at least 24 h. Once the glycemic goal was preserved for 24 h, the infusion was stopped, and subcutaneous insulin was initiated. Insulin was given as short-acting insulin before meals and long-acting insulin in the evening throughout the whole hospitalization period. As for the full medical therapy, the protocol stated that the use of concomitant treatment should be as uniform as possible and in accordance with evidence-based international guidelines for STEMI (19).

Follow-up
After discharge, all patients were managed and followed quarterly for 12 months after event, as outpatients, to perform clinical evaluation, routine analyses and cardiovascular evaluation (ECG, exercise ECG, echocardiography), alongside with the objective to keep HbA1c levels <7% in diabetic patients.

Outcomes
The outcome: clinical outcomes at 1-year follow-up were major adverse cardiovascular events (MACEs). MACEs are de ned as a composite of cardiovascular death, non-fatal acute coronary syndrome and heart failure.

Statistical analyses
After appropriate assessment of their normal distribution by the Shapiro-Wilk test, differences between continuous variables were analyzed either with the parametric Student t-test or the non-parametric Mann- differences in the troponin levels, in the median time between symptom onset and the start of angiography procedure at hemodynamic unit, in the number of coronary vessel diseases and in the number of patients with post-procedural TIMI 3 were observed (Table 1).
In-hospital treatments and glucose control. During hospitalization, hyperglycemic patients were treated according to the most recent international guidelines (23), with high proportions of patients receiving platelet inhibitors and antithrombotic agents before, during and after the procedure. In particular, aspirin was administered to 98% of patients and statins to 86% in both groups. Beta-blockers were given to 81% of patients, of whom the 43% received an ACE-inhibitor in both groups. Among the whole study population, 88% of patients were treated with a combination therapy of thienopyridine and aspirin. After PPCI, the glycemic goal was preserved for 24 h (mean glycemia 160 ±23 mg/dl). In the 10% of patients, blood glucose <70 mg/dl both with and without symptoms occurred during insulin infusion. At hospital discharge, both fasting and post-prandial plasma glucose levels were: fasting, 133±31 mg/dl, postprandial 177±19mg/dl. In addition, the rate of ACE-inhibitors, aspirin, beta-blockers, statins, and thienopyridines use increased as compared to that on admission both in normoglycemic and hyperglycemic patients (Table 1).

Fecal microbiota
Differences between gut microbiota of hyperglycemic and normoglycemic STEMI patients were observed (Figure 1). Some bacterial taxa in the intestinal microbiota were found signi cantly more abundant in  (Figure 2). At the genus level, the highest rates belong to the genus Prevotella, with a 5-fold increase, while Bacteroides disclosed a 7-fold reduction in hyperglycemic as compared to normoglycemic patients ( Figure 2). Examination of the prevalence of speci c genera and bacterial species present in thrombus revealed that few genera can be signi cantly different between hyperglycemic and normoglycemic patients ( Figure 2). Alistipes was absent in hyperglycemic thrombi, whilst it occurred in normoglycemic patients (7.3%). Finally, thrombus Prevotella content was signi cantly associated with admission blood glucose levels (P=0.001) ( Figure 2).
In-vitro study, effects of high glucose concentration on growth curve of Prevotella DJF RP53 and Alistipes. Due to the major modi cation of Prevotella DJF RP53 and Alistipes in hyperglycemic as compared to normoglycemic thrombi, we assessed the effects of high-glucose concentration on Prevotella DJF RP53 and Alistipes growth curve on soil. Results clearly indicated that Prevotella DJF RP53 growth curve was accelerated and increased in soil containing high-glucose concentration, as compared with normal glucose concentration (P=0.012). On the other hand, high-glucose concentration slowed and decreased the Alistipes growth curve on soil, as compared normal glucose concentration on soil (P=0.016) (Figure 3).

Coronary thrombus analysis in hyperglycemic and normoglycemic patients
The volume and size of the thrombi, and the number of retrieved fragments were signi cantly higher in hyperglycemic as compared to normoglycemic patients (P<0.001) (Supplementary gure 3). Thrombi composition displayed higher CD40L, VWF and TMAO levels as compared to normoglycemic patients (Supplementary gure 2). Moreover, microbiota genus analysis revealed a marked presence of Prevotella in hyperglycemic thrombi which,instead, was almost absent in normoglycemic thrombi (Supplementary gure 2). Moreover, Prevotella thrombus levels were signi cantly associated with the size of thrombi, and TMAO, CDL40 and VWF levels during STEMI (Figure 4).  Figure 5). STEMI outcomes at 1-year follow-up strati ed by Prevotella and TMAO level terziles in coronary thrombi of hyperglycemic and normoglycemic patients were further assessed to translate molecular results in clinical endpoints. Cox multivariable regression analysis reported a lower survival from events among patients with higher Prevotella and TMAO levels ( Figure 5). Interestingly, we observed an increased risk association with Prevotella and TMAO terziles, consistent regardless of the dual antiplatelet therapy during the follow-up, thus suggesting a reduction of effectiveness of anti-platelet therapy in patients with high levels of Prevotella and TMAO thrombus levels.

Discussion
In the present study we investigated for the rst time the relationship between thrombus Prevotella colonization and recurrent acute coronary syndrome and death among hyperglycemic patients with rst STEMI. Moreover, we analyzed the relationship between thrombus microbe colonization and thrombus size, TMAO content, as well as thrombus levels of CD40L and VWF. We observed signi cant differences between gut and thrombus microbiota of hyperglycemic and normoglycemic STEMI patients. Fecal and thrombus Prevotella contents were more abundant in hyperglycemic than normoglycemic patients and, interestingly, were also signi cantly associated with admission blood glucose levels. On the contrary, Alistipes contents were lower in fecal samples and absent in thrombus samples from hyperglycemic patients, while well represented in normoglycemic patients.
These ndings raise two important considerations. First, hyperglycemia on admission, in the setting of acute coronary syndrome, may suggest a substantial change in both gut and coronary thrombus microbiota compositions. In fact, fecal and thrombus samples from hyperglycemic patients showed higher contents of Prevotella and low Alistipes contents as compared to samples from normoglycemic patients. This hypothesis is supported by data from our in vitro study. Indeed, the high-glucose concentration, mimicking the hyperglycemic milieu, accelerated and increased Prevotella DJF RP53 growth curve and slowed and decreased Alistipes growth curve on soil. In addition, during acute coronary syndrome, hyperglycemia may favor thrombus colonization, thus driving intestinal barrier permeability (20). In this setting, hyperglycemia and GLUT2-dependent transcriptional reprogramming alter both tight and adherence junction integrity of intestinal epithelial cells (20). Consequently, hyperglycemia-mediated barrier disruption leads to the systemic in ux of microbioma and an enhanced systemic dissemination. Moreover, treatment of hyperglycemia, intestinal epithelial-speci c GLUT2 deletion, or inhibition of glucose metabolism restore barrier function and bacterial containment (20). Thus, our data suggest that the increase in the intestinal barrier permeability among patients with hyperglycemia may favor thrombus contamination in a coronary syndrome setting.
The role of both Prevotella and Alistipes in the metabolic and cardiovascular diseases has been already reported by previous studies (21).Fecal and gingival Prevotella increased both in diabetic and hyperglycemic patients as compared to healthy subjects (22,23). These studies not only demonstrated a positive association of Prevotella gingival content with higher blood glucose levels, though also a signi cant contribution of Prevotella gingival content to the development of disease condition, via increased TMAO production (23). Accordingly, data from recent studies (11) suggest that Prevotella, among the microbialgenera, showed the strongest association with TMAO levels. On the contrary, abundance of stool Alistipes has been associated with low levels of TMAO (24). There is considerable evidence linking the metaorganismal TMAO pathway to the development of atherosclerosis and thrombosis in animal models and incident cardiovascular disease risks. In particular, the proatherogenic effects of microbe-dependent TMAO formation include enhanced macrophage cholesterol accumulation and subsequent foam cell formation, proin ammatory changes in the artery wall, and promotion of both platelet hyperactivity and enhanced arterial thrombosis potential. (25). Thus, our data suggest a potential antagonistic effect between Alistipes and Prevotella, as shown by the absence of Alistipes and higher contents of Prevotella in hyperglycemic coronary thrombus, in the setting of acute coronary syndrome.
This trend has been commonly seen in other heart conditions, such as atrial brillation and heart dysfunction (26,27). Moreover, to date, there has been no evidence about the possible role of Prevotella thrombus colonization in outcomes of high-risk patients (e.g., those found in hyperglycemic patients during STEMI), and about the speci c pathways transducing thrombus Prevotella stimuli in the modulation of coronary thrombus burden.
The novelty of our study is represented by the evidence that Prevotella coronary thrombus content is remarkably increased and it is associated with high thrombus burden, TMAO, CDL40, and vWF in hyperglycemic patients as compared with normoglycemic patients with STEMI. Interestingly, Prevotella coronary thrombus contents were associated with admission blood glucose levels. These data suggest that high thrombus burden observed in hyperglycemic patients could represent a result of the Prevotella effects on increased TMAO levels and on the consequent effects on CDL40 and vWF thrombus contents, as reported in experimental studies that clearly underlined the role of TMAO in increasing coagulation (13). In addition, several studies (28) suggest that fecal over-abundance of Prevotella is associated with an increase in insulin-resistance. The role of Prevotella in the human microbiome is controversial, and the relative abundance of Prevotella in general and P. copri is highly dependent on diet and lifestyle (29)(30)(31). We can speculate that hyperglycemia could activate the following harmful vicious circle: I) hyperglycemia promotes systemic diffusionof Prevotella; II) Prevotella overabundance associates with increased TMAO levels; III) high TMAO levels increase systemic IR; IV) increased insulin-resistance in turn increases hyperglycemia ( Figure 6). This hypothesis may have a clinical implication since patients with higher Prevotella and TMAO thrombus contents were more likely to experience subsequent cardiovascular events and death. The associations observed between Prevotella and TMAO thrombus contents and cardiovascular events and death remained signi cant also after adjustment for potentially confounding clinical factors. Consistently with such evidence, in a contemporary sample of patients with STEMI treated with PPCI and TA in routine practice, we observed a higher cumulative incidence of 1-year mortality and adverse cardiovascular outcomes, in patients with higher thrombus Prevotella and TMAO contents. In fact, the 1-year prognosis of patients with STEMI and III terziles of Prevotella (<2.72%) and TMAO (>60 ng/ml) was signi cantly worse, with a 12.1% rate of death or re-admission for cardiovascular diseases, as compared with prognosis of patients with STEMI and I terziles of Prevotella (>1.25%) and TMAO (>36 ng/ml), which disclosed a 1.9% rate of death or re-admission due to cardiovascular diseases at follow-up. TMAO levels are also affected by diet in a microbiome-driven manner, and lower levels were associated with consumption of plant-based diets (32) although an anticipated reduction of TMAO levels was not observed upon intervention with high-vegetables Mediterranean diet in obese subjects (33).Interestingly, the 1-year follow-up results based on Prevotella and TMAO thrombus contents did not change despite a similar severity of atherosclerotic disease at baseline and asimilar glycemic control both at discharge and at follow-up. These observations suggest that on admission hyperglycemia may play a pivotal role in the increase of coronary thrombus burden, during STEMI, favoring thrombus colonization and proliferation by Prevotella.
However, our study has some limitations. First, the observational prospective study design and the relatively small sample size. In addition, no power calculation could be performed, due to the lack of previous studies in this setting. As a consequence, the enrollment of patients in each group was arbitrary.

Conclusion
In conclusion, our data suggest that the coronary thrombus environment is more likely susceptible to a microbial colonization which might stimulate a pro-coagulatory response in acute coronary syndrome. These ndings add new elements to the pathogenesis of ACS and suggest novel therapeutic targets. Availability of data and materials: data and materials are fully available. intervention in overweight and obese subjects lowers plasma cholesterol and causes changes in the gut microbiome and metabolome independently of energy intake. Gut. 2020 Jul;69 (7):1258-1268. Tables   Table 1. Baseline clinical characteristics, angiographic and procedural data in overall study-population.