Productive cytomegalovirus infection is associated with impaired endothelial function in ST-elevation myocardial infarction

Abstract

Background:

An association between productive cytomegalovirus infection and atherosclerosis was shown recently in several trials, including a previous study of ours. However, the mechanism involved in this association is still under investigation. Here, we addressed the interaction between productive cytomegalovirus infection and endothelial function in patients with ST-elevation myocardial infarction (STEMI).

Methods:

We analyzed the presence of cytomegaloviral DNA in plasma and endothelial function in 33 patients with STEMI and 33 volunteers without cardiovascular diseases, using real-time polymerase chain reaction and a noninvasive test of flow-mediated dilation.

Results:

Both the frequency of presence and the load of cytomegaloviral DNA were higher in plasma of STEMI patients than those in controls. This difference was independent of the other cardiovascular risk factors (7.38 [1.36–40.07]; p=0.02). The results of the flow-mediated dilation test were lower in STEMI patients than in controls (5.0 [2.65–3.09]% vs 12.5 [7.5–15.15]%; p=0.004) and correlated negatively with the cytomegaloviral DNA load (Spearman R =−0.407; p=0.019) independently of other cardiovascular risk factors.

Conclusions:

Productive cytomegalovirus infection in STEMI patients correlated negatively with endothelial function independently of other cardiovascular risk factors. The impact of cytomegalovirus on endothelial function may explain the role of cytomegalovirus in cardiovascular prognosis.

Introduction

The role of immune system activation in atherosclerotic plaque destabilization and rupture, the major cause of acute coronary syndrome, is well established^1,2. However the triggers of immunoactivation in atherosclerosis remain controversial. Among the possible proposed agents are various bacterial and viral infections^3–6, whose chronic persistence corresponds to chronic inflammation within plaques^7,8. The impact of herpesviruses and in particular of cytomegalovirus (human herpesvirus type 5) on atherosclerosis has been shown in animal models^8–11. However, clinical studies based on sero-epidemiological data or viral DNA presence within the artery wall were far less conclusive^12–14. At the same time, investigation of a productive cytomegalovirus infection by viral DNA in blood plasma and viral RNA in peripheral blood mononuclear cells found a strong positive correlation to the development of acute coronary syndrome^15,16. In the current study, we investigated the mechanisms of impact of cytomegalovirus on a major atherosclerotic complication, STEMI. Earlier in in vitro studies it had been shown that cytomegalovirus infects endothelial cells and activates T lymphocyte migration into plaques, causing endothelial dysfunction^17–20, which worsens the cardiovascular prognosis of STEMI patients^21,22. We found that productive cytomegalovirus infection negatively correlates with endothelial function in STEMI patients. This correlation may explain the role of cytomegalovirus in cardiovascular prognosis that was found in the previous studies^23–25.

Materials and Methods

A total of 33 volunteers without cardiovascular diseases (14 men and 19 women, mean age 49.9±7.6 years) and 33 patients with STEMI (28 men and 5 women, mean age 58.6±10.2 years), diagnosed according to the current guidelines²⁶, were enrolled in our study. Detailed inclusion and exclusion criteria are presented in Table 1.

Table 1.

Inclusion and exclusion criteria for STEMI patients and volunteers without cardiovascular diseases

	Inclusion criteria	Exclusion criteria
STEMI patients	STEMI diagnosis at the moment of admission according to the guidelines	For STEMI patients: • Time from symptoms onset > 24 hours • Bleeding or necessity of blood/plasma transfusion • Thrombolytic therapy • Cardio-pulmonary resuscitation • Cardiogenic shock • Infusion of nitrates For all participants: • Age below 40 • Acute infections • Neoplasms • Pregnancy
Volunteers without cardiovascular diseases	No signs of possible atherosclerosis according to treadmill-test, echocardiography, and ultrasound of peripheral arteries

STEMI – ST-elevation myocardial infarction.

In both groups we analyzed traditional cardiovascular risk factors according to the established criteria: sex, age, smoking, dyslipidemia (fasting LDL-cholesterol level>160 mg/dL or total cholesterol level>200 mg/dL and HDL-cholesterol level<40 mg/dL), obesity (BMI≥30 kg/m²), arterial hypertension (systolic blood pressure≥140 mmHg or diastolic blood pressure≥90 mmHg), diabetes mellitus (glycated hemoglobin>6.5% or fasting glucose level>7 mmol/l)²⁷, and elevated high sensitive CRP level (hs-CRP>2 mg/l)²⁸

The study protocol and participants’ consent were approved by the Interuniversity Committee of Ethics; all participants signed their written informed consent to participate in this study.

Cytomegaloviral DNA measurement

Peripheral blood samples were obtained from patients at admission before coronary angiography and from controls at the time of examination. Blood was centrifuged at 3,000 g for 15 minutes. We extracted viral DNA from plasma using the viral DNA spin protocol QIAamp UltraSense Virus kit. Detection of cytomegalovirus was performed by means of quantitative real-time PCR on a Bio-Rad thermal cycler CFX 96 C1000 using primer/probe sets for cytomegalovirus and human endogenous retrovirus 3 (ERV-3)²⁹ (Table 2). Load of cytomegalovirus was expressed as copy number of cytomegaloviral DNA per 1 ml of plasma. The specificity of the PCR reaction was confirmed with purified cytomegalovirus as positive control and with human cord blood DNA as negative control. The threshold level of 1,000 copies of cell-free cytomegaloviral DNA per 1 ml of plasma was used as a categorical parameter of the presence of highly productive cytomegalovirus infection³⁰.

Table 2.

Sequences of primers/probes

Sequence Name	Sequence	Five Modification	Three Modification
ERV-3 Probe	Aacatgggagaccaatggccatggg	Quasar 705*	BHQ
ERV-3 Forward	Gttgcttcatgttatgtctgtgg
ERV-3 Reverse	Ggcggttagtgtgaaattatcttg
CMV Probe	Tacctggagtccttctgcgagga	CAL Fluor Red 610*	BHQ
CMV Forward	Aaccaagatgcaggtgatagg
CMV Reverse	Agcgtgacgtgcataaaga

^*Quasar 705 and Cal Fluor Red 610 are fluorescent dyes.

ERV-3 - human endogenous retrovirus 3; CMV – cytomegaloviral; BHQ - black hole quencher.

Measurement of flow-mediated dilation

We examined flow-mediated dilation of the brachial artery using a high resolution ultrasound procedure according to the current guidelines³¹. The flow-mediated dilation test was performed in STEMI patients at admission before coronary angiography and in controls without cardiovascular diseases at the time of examination. The threshold level of 10% for the result of the flow-mediated dilation test was used as a categorical parameter of the presence of endothelial dysfunction³².

Statistical analyses

The data obtained in the present study were not normally distributed and are represented as medians and interquartile ranges (IQR). For the comparison of two groups we used the Mann-Whitney rank test. For the analysis of categorical parameters, frequency of endothelial dysfunction and presence of highly productive cytomegalovirus infection we used the Chi-square test with 2×2 contingency tables. For the analysis of correlations, Spearman’s coefficient of correlation was estimated. For the age distribution we made the assumption of its normality and analyzed it using the t-test for two groups. We also performed multiple binomial logistic regression analysis. For this, we used dichotomized age ≥55 years for males and ≥65 years for females and hs-CRP level ≥2 mg/L²⁸ as well as other binominal factors as predictors. The results for regression analysis are presented as odds ratio (OR) ±95% confidence interval (CI). We performed statistical analysis using Statistica 12.0. Values of p<0.05 were considered statistically significant.

Results

We found that the control group was different from the STEMI group in age/sex characteristics as well as in traditional cardiovascular risk factors such as smoking, hypertension and hs-CRP level (Table 3). Therefore we assessed further data using multiple logistic regression analysis.

Table 3.

Clinical characteristics of patients with STEMI and volunteers without cardiovascular diseases

	STEMI patients	Volunteers without cardiovascular diseases	p-value
N	33	33
Age (years), mean ± SD	58.6 ± 10.2	49.9 ± 7.6	0.0028*
Men, N (%)	28 (84.9)	14 (42.4)	0.0003*
Smoking, N (%)	24 (72.7)	5 (15.2)	0,0000*
Hypertension, N (%)	21 (63.6)	10 (30.3)	0.0067*
Diabetes mellitus, N (%)	5 (15.2)	1 (3.0)	0.0868
Dyslipidemia, N (%)	9 (27.3)	4 (12.1)	0.1218
Obesity, N (%)	10 (30.3)	15 (45.5)	0.2045
hs-CRP level (mg/l), median [IQR]	4.14 [1.75–8.15]	1.65 [0.62–2.93]	0.0004*

Data are presented as mean ± SD, medians [interquartile ranges] or as number (percentage) of patients.

^*Differences are statistically significant at p < 0.05.

STEMI – ST-elevation myocardial infarction; hs-CRP –high sensitive C-reactive protein level.

Analysis of cytomegalovirus infection

To evaluate productive viral infection, we measured cytomegaloviral DNA in plasma³³. In our measurements cord blood samples were negative for the presence of cytomegalovirus, therefore we analyzed all the positive results for cytomegaloviral DNA load. However, we set additionally an arbitrary level of 1,000 copies of viral DNA per 1 ml of plasma as a threshold for the presence of highly productive cytomegalovirus infection³⁰.

The frequency of presence of highly productive cytomegalovirus infection was higher in plasma of STEMI patients than in plasma of volunteers without cardiovascular diseases (60.6% vs. 24.2%, respectively; p=0.028). The amount of cytomegaloviral DNA was also significantly higher in STEMI patients than in volunteers without cardiovascular diseases (1,638.64[0–2,543.42] vs. 0[0–910.79]; p=0.011) (Figure 1).

Figure 1. Cytomegaloviral DNA loads in controls and STEMI patients.

Presented are the medians and interquartile range of the amounts of cytomegaloviral DNA copies per 1ml of plasma for volunteers without cardiovascular diseases and patients with STEMI.

* Differences are statistically significant at p < 0.05.

Since our groups differ in several traditional risk factors, we analyzed whether productive cytomegalovirus infection correlate with these factors (Table 4). We found that plasma cytomegaloviral DNA load was higher in hypertensive subjects (1,688.1[0–2,631.8] vs. 0[0–967.0] copies/ml; p=0.013). However, the presence of highly productive cytomegalovirus infection did not correlate with the other cardiovascular risk factors. Only when we analyzed separately cytomegalovirus-positive cases we found a positive correlation of cytomegaloviral DNA load with patients’ age (N=35, Spearman R=0.456; p=0.006). To establish whether productive cytomegalovirus infection is an independent predictor for STEMI or depends on the cardiovascular risk factors, different between two groups, we performed a multiple binomial logistic regression analysis of STEMI development with such predictors as age, sex, smoking, hypertension, and elevated hs-CRP level and with presence of highly productive cytomegalovirus infection as an additional predictor. We found that presence of cytomegalovirus and smoking independently predict the development of STEMI in the multiple regression model (OR[95%CI], respectively, 7.38[1.36–40.07], p=0.021; 17.17[2.93–100.62]; p=0.002), while age, sex, hypertension, and elevated hs-CRP level showed a non-significant, but also substantial, increase in the odds ratios of STEMI development (Figure 2).

Table 4.

Interactions between cytomegaloviral DNA load or presence and cardiovascular risk factors in all participants

Risk factor	Presence of the risk factor		Absence of the risk factor		p-value(load) / p-value(presence)
	CMV DNA copies/ml, Me [IQR]	CMV presence, N (%)	CMV DNA copies/ml, Me [IQR]	CMV presence, N (%)
Male sex	409.6 [0–2,311.2]	18 (42.9)	35.6 [0–1,942.6]	10 (41.6)	0.827/0.925
Smoking	1,185.7 [0–2,327.6]	15 (51.7)	0 [0–1,688.1]	13 (35.1)	0.212/0.176
Hypertension	1,688.1 [0–2,631.8]	19 (61.3)	0 [0–967.0]	9 (25.7)	0.013*/0.004*
Diabetes mellitus	844.0 [0–1,728.8]	3 (50.0)	228.9 [0–2,335.7]	25 (41.7)	0.853/0.694
Dyslipidemia	445.0 [0–2,311.4]	6 (46.2)	71.1 [0–2,310.5]	22 (41.5)	0.702/0.761
Obesity	0 [0–1,554.0]	8 (32.0)	910.8 [0–2,327.6]	20 (48.8)	0.158/0.181

Presented are the amounts of cytomegaloviral DNA copies per 1 ml of plasma and the percentages of cytomegaloviral DNA presence in plasma in all participants, with the presence or absence of different cardiovascular risk factors. Data are presented as medians [interquartile ranges] of copies of cytomegaloviral DNA and as number (percentage) of cytomegalovirus-positive patients.

^*Differences are statistically significant at p < 0.05.

CMV – cytomegaloviral; DNA – deoxyribonucleic acid.

Figure 2. Multiple logistic regression analysis of cardiovascular risk factors and presence of cytomegalovirus in STEMI patients and controls.

Presented are the odds ratios for binomial cardiovascular risk factors (with age ≥ 55 years for males and ≥ 65 years for females, and hs-CRP level ≥2 mg/l) and the presence of productive cytomegalovirus infection (>1,000 copies of DNA/ml of plasma) in the multiple regression model for comparison of STEMI patients and volunteers without cardiovascular diseases.

* Differences are statistically significant at p < 0.05.

Analysis of endothelial function

We evaluated endothelial function using the flow-mediated dilation test. In STEMI patients, the flow-mediated dilation test results were significantly lower than those in the control group (5.0[2.65–3.09]% vs 12.5[7.5–15.15]%; p=0.004) (Figure 3). As well, the frequency of endothelial dysfunction was also significantly higher in STEMI patients than in volunteers without cardiovascular diseases (78.8% vs. 36.4%; p=0.0005).

Figure 3. Flow-mediated dilation test results in controls and STEMI patients.

Presented are the medians and interquartile ranges of the flow-mediated dilation test results in volunteers without cardiovascular diseases and patients with STEMI.

* Differences are statistically significant at p < 0.05.

Since our groups of participants differed in several risk factors, we analyzed whether flow-mediated dilation test results correlated with them. We found that flow-mediated dilation test results were lower in hypertensive subjects (5.8[0–11.0]% vs 11.4[6.1–15.6]%; p=0.016), and correlated negatively with patients’ age and hs-CRP level (respectively, N=66, Spearman R=−0.355; p=0.004; N=66, Spearman R=−0.409, p=0.0007). There was a tendency for lower flow-mediated dilation test results in men, which was in accordance with previous studies³⁴; however, in our study it did not reach statistical significance (Table 5). In our participants endothelial function also did not correlate with smoking, diabetes mellitus, dyslipidemia, or obesity.

Table 5.

Interactions between endothelial function and cardiovascular risk factors in all participants

Risk factor	Presence of the risk factor	Absence of the risk factor	p-value
	Flow-mediated dilation test result (%), Me [IQR]
Male sex	7.7 [0–13.6]	11.5 [6.2–16.3]	0.052
Smoking	8.2 [0–13.6]	10.0 [4.6–14.0]	0.286
Hypertension	5.8 [0–11.0]	11.4 [6.1–15.6]	0.016*
Diabetes mellitus	2.2 [0.5–8.1]	9.2 [3.1–14.0]	0.258
Dyslipidemia	7.8 [3.1–13.5]	10.0 [2.2–14.0]	0.924
Obesity	11.4 [4.6–14.6]	7.5 [0–13.5]	0.252

Data are presented as medians [interquartile ranges] of flow-mediated dilation test results.

^*Differences are statistically significant at p < 0.05.

To establish whether endothelial dysfunction is an independent predictor for STEMI or depends on the other risk factors, we performed a multiple binomial logistic regression analysis of STEMI development with endothelial dysfunction as an additional predictor. We found that endothelial dysfunction and smoking independently predict the development of STEMI in the multiple regression model (OR[95%CI], respectively, 8.76[1.49–51.52], p=0.016; 27.46[4.16–181.12]; p=0.0006), while age, sex, hypertension, and elevated hs-CRP level showed non-significant changes in the odds ratio of STEMI development (Table 6).

Table 6.

Multiple logistic regression analysis of cardiovascular risk factors and endothelial dysfunction impact for STEMI patients and controls comparison

Factor	Odds ratio	−95% CI	+95% CI	p-value
Age	3.41	0.68	17.16	0.137
Male sex	1.00	0.19	5.22	0.997
Smoking	27.46	4.16	181.12	0.001*
Hypertension	2.26	0.34	14.93	0.399
Elevated hs-CRP level	1.40	0.23	8.51	0.714
Endothelial dysfunction	8.76	1.49	51.52	0.016*

Presented are the odds ratios for binomial cardiovascular risk factors (with age ≥ 55 years for males and ≥ 65 years for females, and hs-CRP level ≥2 mg/l) and the presence of endothelial dysfunction (flow-mediated dilation test result <10%) in the multiple regression model for comparison of STEMI patients and volunteers without cardiovascular diseases.

^*Differences are statistically significant at p < 0.05.

STEMI – ST-elevation myocardial infarction; hs-CRP –high sensitive C-reactive protein level.

Correlation of productive cytomegalovirus infection with endothelial function

We assessed the correlation between cytomegaloviral DNA loads and flow-mediated dilation test results in both groups together and separately. We found that cytomegaloviral DNA level negatively correlates with flow-mediated dilation test results in all participants (Figure 4), as well as in patients with STEMI (respectively, N=66, Spearman R=−0.360, p=0.003; N=33, Spearman R=−0.407, p=0.019). However, cytomegaloviral DNA loads in controls didn’t correlate significantly with flow-mediated dilation test results.

Figure 4. Correlation between cytomegaloviral DNA loads and flow-mediated dilation test results.

Presented is the correlation with 95% confidential interval according to Spearman Rank test between cytomegaloviral DNA loads and flow-mediated dilation test results in all participants.

* Correlation is statistically significant at p < 0.05.

In view of the correlations of flow-mediated dilation test results and loads of cytomegaloviral DNA with cardiovascular risk factors found in our study, we investigated whether productive cytomegalovirus infection is an independent predictor for endothelial dysfunction or depends on the other risk factors. We found that presence of cytomegalovirus independently predicts the development of endothelial dysfunction in the multiple regression model (OR[95%CI], 3.48 [1.06–11.37]; p = 0.039), while age, sex, hypertension, and elevated CRP level didn’t show significant changes in the odds ratios (Table 7).

Table 7.

Multiple logistic regression analysis of impact of cardiovascular risk factors and presence of cytomegalovirus for comparison of patients with and without endothelial dysfunction

Factor	Odds ratio	−95% CI	+95% CI	p-value
Age	2.42	0.66	8.92	0.185
Male sex	2.65	0.76	9.17	0.125
Hypertension	0.51	0.12	2.11	0.351
Elevated hs-CRP level	2.36	0.70	8.00	0.167
Presence of cytomegalovirus	4.10	1.19	14.05	0.025*

Presented are the odds ratios for binomial cardiovascular risk factors (with age ≥ 55 years for males and ≥ 65 years for females, and hs-CRP level ≥2 mg/l) and the presence of productive cytomegalovirus infection (>1000 copies of DNA/ml of plasma) in the multiple regression model for comparison of patients with and without endothelial dysfunction (defined as flow-mediated dilation test result <10%).

^*Differences are statistically significant at p < 0.05.

hs-CRP –high sensitive C-reactive protein level.

In summary, we found that (i) the presence and the load of cytomegaloviral DNA in plasma was higher in STEMI patients than in controls, independently of cardiovascular risk factors; (ii) the presence of endothelial dysfunction was higher in STEMI patients than in controls, independently of cardiovascular risk factors; (iii) cytomegaloviral DNA load independently negatively correlates with flow-mediated dilation test results.

Discussion

The destabilization of atherosclerotic plaques and development of recurrent cardiovascular events are strongly associated with activation of the immune system^1,2,35,36. The perturbation of T cell repertoire with higher expression of effector memory and activation markers was found not only in blood of patients with cardiovascular diseases but also within their atherosclerotic plaques³⁷.

These data may indicate the presence of a specific antigen(s) towards which these lymphocytes are reactive^3–6. Herpes viruses are one of the main candidates causing persistent immunoactivation within plaques due to their ubiquity and ability to cycle between dormancy and replication^7,8. Thus, it has been shown that accumulation of intermediate-and late-differentiated T effector memory cells in blood is associated with cytomegalovirus infection^38–40.

The first evidence of a relation of herpes viruses to atherosclerosis was found early in 1980^s in chickens with Marek’s herpesviral disease⁹ and was later confirmed through use of various other animal models^8,10,11. Moreover, in a number of experimental studies infection of vascular endothelial cells by cytomegalovirus was associated with the progression of atherosclerosis. Thus, replication of cytomegalovirus in endothelial cells and macrophages⁴¹ trigger the release of proinflammatory cytokines, adhesion molecules, and matrix metalloproteinases, as well as cellular death^11,42–46. While cytokines and adhesion molecules cause CD4 and CD8 T lymphocyte and platelet migration to the subendothelium^44,46–48, matrix metalloproteinases cause plaque fibrotic cap destabilization⁴⁹, altogether leading to the development of acute coronary syndrome. Moreover, in experimental studies it was found that cytomegalovirus infection can lead to severe endothelial dysfunction^50,51 and altered response to oxidized lipids in the subendothelium⁵², leading to atherosclerosis progression.

However, the transition from experimental settings to clinical studies resulted in a contradiction. Analysis of the cytomegaloviral DNA distribution in the human vascular wall revealed a relatively low prevalence of viral DNA in the walls of the aorta and coronary arteries, both in patients who died from atherosclerosis¹² and in those who died from other causes⁵³. However, in further studies, cytomegaloviral DNA was found more often in arteries of patients with different complications of atherosclerosis^4,54,55 independently of the presence of the other infectious agents⁵⁶.

Evaluation of the immune response to cytomegalovirus infection initially didn’t reveal an effect of seropositivity to cytomegalovirus on the development of atherosclerosis^13,14,63. However, the results of sero-epidemiological studies were also controversial, with further studies, that found a positive correlation between the increased titer of antibodies to cytomegalovirus and the development of atherosclerosis^{23–25,57,58}, cardiovascular events^59,60 and mortality^61,62, even when adjusted to other traditional cardiovascular risk factors. These controversies in sero-epidemiological studies as well as the low level of cytomegaloviral DNA found in atherosclerotic plaques and later in peripheral blood mononuclear cells of patients with cardiovascular diseases⁶⁴ could be explained by the occasional detection of the latent infection in these studies, rather than productive viral infection, while only the latter correlates with plaque destabilization^65,66.

Accordingly, it was shown that only the detection of the products of replication of cytomegalovirus (viral RNA) or isolation of viral DNA in blood plasma could be reliable signs of productive infection³³. S. Gredmark et al. found that the amounts of early RNA genes of cytomegalovirus in peripheral blood mononuclear cells of patients with acute coronary syndrome were higher than those of patients with chronic coronary artery disease and of healthy volunteers¹⁵.

In our previous work, we evaluated cytomegaloviral DNA in plasma and showed that productive cytomegalovirus infection is more common in patients with acute coronary syndrome than in patients with chronic coronary artery disease and healthy volunteers. Moreover, in that work we also found a positive correlation between the cytomegaloviral DNA load and T-lymphocyte differentiation within the atherosclerotic plaques of patients with cardiovascular diseases¹⁶. This result is in agreement with the studies on differentiation of effector T lymphocytes in blood of patients infected with cytomegalovirus^{19,20,38–40}.

Here, we showed that the presence of cytomegaloviral DNA in plasma of STEMI patients was significantly higher than in volunteers without cardiovascular diseases. These data confirm the results of our previous study and are also in agreement with a meta-analysis, which shows a significant correlation between the concentration of antibodies to cytomegalovirus and the risk of developing of cardiovascular diseases, both in prospective and retrospective analysis²⁴.

Also, we found a significant prevalence in the number of copies of cytomegaloviral DNA in older hypertensive patients, which correspond to the previous studies that showed the positive correlation of antibodies to cytomegalovirus and cytomegaloviral microRNA with arterial blood pressure, especially in elderly patients^38,67,68. Moreover, one of the possible mechanisms of the development of hypertension in patients with cytomegalovirus infection could be the development of endothelial dysfunction with the synthesis of proinflammatory cytokines and renin by endothelial cells infected with cytomegalovirus⁶⁹.

Endothelial function plays an important role in patients with atherosclerosis⁷⁰. Severe endothelial dysfunction in patients with STEMI compared with healthy volunteers has been reported⁷¹. Here, we showed that flow-mediated dilation test results were an independent risk factor of STEMI, corresponding to the results of other trials^21,22,72. Also, in this study, we confirmed that endothelial function is impaired more significantly in older patients with elevated levels of hs-CRP⁷³.

An interaction between endothelial dysfunction and cytomegalovirus infection was initially found in clinical studies of healthy volunteers^74,75. However, these data were not consistent⁷⁶. Later, J. Simmonds and P. Petrakopoulou found a negative correlation between cytomegaloviral DNA in peripheral blood mononuclear cells and endothelial function in patients after heart transplantation^77,78.

We demonstrated here for the first time the negative correlation between productive cytomegalovirus infection and endothelial function in patients with STEMI (alone and together with the volunteers without cardiovascular diseases). This result could be explained by the impact of cytomegalovirus infection on the synthesis of an inhibitor of endothelial NO-synthase and the blockade of NO-synthase activation^17,18. Also, the effect of cytomegalovirus on endothelial function can be caused by the release of effector Т cells with receptors to fractalkine^19,20 – a chemokine, causing endothelial cell activation and apoptosis, abolished by antibodies to these receptors⁷⁹. In a multivariate analysis, we confirmed that the identified correlation persisted, regardless of sex, hypertension, CRP level, and patient age - factors that were significantly associated with production of cytomegalovirus and endothelial function.

In summary, we found that cytomegalovirus infection is strongly associated with endothelial dysfunction in STEMI patients.

Our work has significant limitations: the frequencies of presence of some risk factors such as diabetes mellitus, obesity, and dyslipidemia in both groups were so low that there were perhaps too few events to support these variables in the simple or multiple regression model; also, while the release of cytomegalovirus into the blood is accepted as evidence of productive infection we do not know which specific cells produce cytomegalovirus and whether these cells are localized in plaques or at other sites.

Although we found an association between the cytomegalovirus in plasma and endothelial dysfunction, it is possible that other pathogens, not studied in the present work, may also associate with this pathological state. In general, our results support a hypothesis that acute coronary events are synchronized with replication of cytomegalovirus, and this viral reactivation contributes to endothelial dysfunction. Therefore, it is conceivable that preventing the reactivation of cytomegalovirus may enhance endothelial function and affect the outcome of coronary artery disease.

Future studies should shed more light on the relation of cytomegalovirus infection to endothelial dysfunction in STEMI in particular and to cardiovascular diseases in general and may lead to new treatment strategies.

1. Productive cytomegalovirus infection is present more often in patients with ST-elevation myocardial infarction than in volunteers without cardiovascular disease independently of cardiovascular risk factors.

2. Productive cytomegalovirus infection correlated negatively with endothelial function in STEMI patients.

3. The correlation of cytomegaloviral DNA in plasma with endothelial dysfunction is independent of other cardiovascular risk factors.