Novel Strategies to Combat CMV-Related Cardiovascular Disease

Cytomegalovirus (CMV), a ubiquitous human pathogen that is never cleared from the host, has long been thought to be relatively innocuous in immunocompetent adults, but causes severe complications including blindness, end-organ disease, and death in newborns and in immuno-compromised individuals, such as organ transplant recipients and those suffering from AIDS. Yet even in persons with intact immunity, CMV infection is associated with profound stimulation of immune and inflammatory pathways. Carriers of CMV infection also have an elevated risk of developing cardiovascular complications. In this review, we define the proposed mechanisms of how CMV contributes to cardiovascular disease (CVD), describe current approaches to target CMV, and discuss how these strategies may or may not alleviate cardiovascular complications in those with CMV infection. In addition, we discuss the special situation of CMV coinfection in people with HIV infection receiving antiretroviral therapy, and describe how these 2 viral infections may interact to potentiate CVD in this especially vulnerable population.


INTRODUCTION
Human cytomegalovirus (CMV) is a beta-herpesvirus that is common worldwide, infecting about 50% of the general population in the United States and in Europe, and is particularly prevalent among those who have low socio-economic status or receptive sexual intercourse (women and men who have sex with men) [1,2]. Many adults are initially infected with CMV during childhood or early adulthood, and the incidence continues to rise throughout adulthood, by approximately 1% annually [3]. In some resource-poor areas, such as sub-Saharan Africa and parts of Asia, CMV infection rates approach 100% [4][5][6]. Primary CMV infection elicits robust innate and adaptive immune responses and can cause a febrile mononucleosis and hepatitis but is subclinical for most healthy individuals. Like the other herpesviruses, CMV is never fully cleared from the infected host and persists either in a true latent form or in a state of low-level replication made possible by multiple immune evasion mechanisms [7]. Reactivations from latency that can lead to virus transmission occur periodically and are triggered by various inflammatory stimuli and other physiologic stressors; however, they are largely asymptomatic and are mostly self-limited. Importantly, and unusually, immunity to CMV does not readily confer protection from super-infection, and secondary infection with other strains of CMV is not uncommon [8][9][10][11].
Active CMV infections can be identified by the presence of CMV nucleic acids and proteins in mucosal secretions. The frequency of CMV DNA shedding in the genital tract and saliva can vary substantially across different studies and is strongly dependent on the detection methods and cohort characteristics including geographical location [12,13]. Such asymptomatic CMV shedding is important both for horizontal transmission and for the interactions of CMV with co-pathogens, adaptive immune responses, and innate immune activation. CMV is a large virus (approximately 236 kb) and is highly immunogenic-upwards of 10% of all memory T cells may be reactive to CMV antigens in persistently-infected individuals [14]-and many viral proteins have evolved to combat and evade host immunity [13]. When an infected individual has a compromised immune system, shedding of CMV DNA increases dramatically. In individuals who are profoundly immunocompromised (eg, AIDS or after transplant), CMV replication can be uncontrolled and lead to end-organ diseases such as pneumonitis, retinitis, hepatitis, or hemorrhagic colitis [7]. Congenital CMV infection is a substantial problem, particularly in resource-limited settings [15,16], and can lead to hearing and vision loss, among other complications [15,[17][18][19].
While long thought to be relatively innocuous in immunocompetent adults, CMV infection is becoming more appreciated as an active contributor to a variety of complications. For example, in adults, CMV infection is associated with the development of cardiovascular diseases (CVD) including atherosclerosis, ischemic heart disease, and myocardial infarction, as well as cardiovascular death [20][21][22]. A recent meta-analysis of 10 prospective studies including over 30,000 participants determined exposure to CMV was associated with a 22% increased relative risk of developing CVD [20], and a comprehensive study of people with HIV infection (PWH) found that having anti-CMV antibodies (CMV seropositivity) was an independent risk factor for CVD and cerebrovascular diseases [23]. In this review, we discuss our current understanding of the mechanistic pathways by which CMV infection might promote CVD development in the general population, and, as a special case, in PWH. We also describe several innovative strategies to target CMV and how they might be applied to reducing CVD onset and severity.

Systemic Inflammation
Persistent systemic inflammation is a risk factor for many complications, including CVD. Indeed, several systemic inflammatory mediators, such as high serum or plasma levels of interleukin (IL)-1b, IL-6, C-reactive protein (CRP), as well as tumor necrosis factor (TNF) receptor II (TNFR2), D-dimer, soluble CD14, and soluble CD163 have been shown to be associated with multiple cardiovascular outcomes and/or mortality [24][25][26]. Recently, in a large study of people at risk for CVD, treatment with the IL-1b antagonist canakinumab was shown to reduce systemic inflammation (ie, IL-6 and CRP levels) and also reduce the frequency of cardiovascular events even without lowering circulating lipid concentrations [27,28]. IL-1b is produced by multiple cell types early after CMV infection or injury [29,30]. While IL-1b can induce CMV reactivation from latency in cell culture and in mouse models [31,32], evidence also suggests that CMV employs several mechanisms to counter IL-1b activity-potentially limiting its antiviral effects to allow for more robust infection [33][34][35][36].
IL-6 levels are highly correlated with CVD risk [37-39], and targeting IL-6 activity with tocilizumab has been shown to improve indices of endothelial dysfunction in participants at high risk for CVD [40,41]. Acute CMV infection is associated with IL-6 expression both in vitro and in vivo [42][43][44][45]. During latency, plasma IL-6 levels remain elevated, even in those without active CMV replication, suggesting active CMV replication is not necessary for ongoing IL-6 expression [46,47]. Accordingly, inhibiting CMV replication with ganciclovir in CMV seropositive critically ill individuals had no effect on IL-6 levels [47].
Pharmacologically targeting either IL-1b or IL-6 results in decreased CRP levels, consistent with CRP expression being downstream of their activities [27,28,48]. CRP can bind to low-density lipoproteins (LDL), allowing macrophages to take up the LDL without further modifications such as oxidation, and studies have shown increased inflammation following CRP treatment [49][50][51]. In a study of nearly 1000 participants with coronary artery disease, elevated CRP levels coupled with CMV seropositivity were associated with mortality, whereas neither elevated CRP levels without CMV nor CMV seropositivity without elevated CRP showed such an association [ [25,63]. As discussed in more detail below, we found that CMV infection was associated with elevated levels of soluble TNFR2 in PWH compared to CMV-uninfected PWH [64], and a small interventional study demonstrated that inhibiting CMV replication with valganciclovir resulted in a significant reduction in plasma levels of soluble TNFR2 in PWH [65]. Taken together, these data from multiple studies suggest that CMV infection (and likely continued low-level viral replication) is a critical determinant of persistent systemic inflammation, especially in those who are at risk of developing inflammation and CVD from other factors.

Monocyte Activation
The stimulatory effect of CMV infection on the activity of immune cells has been well documented [66,67]. Myeloid cells, including monocytes and macrophages, have long been known to be major contributors to CVD, and myeloid cells are some of the most abundant cells within plaque tissues [68,69]. In atherosclerosis, monocytes enter the nascent plaque from the circulation where they encounter proinflammatory cytokines, take up oxidized lipoproteins via scavenger receptors, and differentiate into cholesterol-laden foam cells [70]. The combination of infiltration and cell retention results in macrophage accumulation in the plaques.
The latent CMV reservoir is retained mainly within CD34+ hematopoietic stem cells-including monocyte precursor cells-in the bone marrow [71][72][73]. Circulating monocytes harbor latent CMV that turns on viral gene expression upon differentiation into tissue macrophages [66]. The proinflammatory microenvironment in the developing plaque may thus induce CMV replication, because in vitro experiments have demonstrated that endothelial cell (EC) co-culture and oxidized LDL (oxLDL) can synergistically induce CMV gene expression [66]. CMV infection reciprocally promotes expression of the scavenger receptor CD36 and enhances oxLDL uptake in a monocyte/macrophage cell line [74]. CMV infection also increases production of IL-1β, IL-18, TNF, and interferon (IFN)-g by infected monocytes and macrophages in vitro [61,75]. These phenomena are of particular interest because of the proatherogenic effects of these cytokines in atherosclerosis. Mechanistically, CMV could promote early atherogenesis through direct effects on CMV-laden monocytes as they enter the nascent plaque, release proinflammatory cytokines, acquire oxidized lipoproteins, and differentiate into foam cells.
Intriguingly, CMV encodes several chemokine receptor homologs that induce signaling pathways in the infected cell. One of these is US28, a G-protein-coupled receptor that can bind to the chemokines CCL2, CCL5, CCL7, and CX3CL1, and which may contribute to the development of CMV-associated CVD [76,77]. The binding of US28 to CX3CL1, also known as fractalkine, may be particularly important. CX3CL1 is abundantly expressed by vascular ECs, is upregulated by inflammation, is detected in atherosclerotic plaque tissues, and mediates both cell adhesion and chemotaxis of cells that express the fractalkine receptor (CX3CR1) [78,79]. In addition, CX3CL1 binding to US28 induces macrophage migration [80]. As US28 expression is also necessary for the maintenance of CMV latency [81], circulating monocytes harboring latent CMV should have constitutive expression of US28. Thus, it is conceivable that latently infected cells might be more susceptible to the influence of CX3CL1 signals and traffic more readily to sites of endothelial dysfunction, such as atherosclerotic plaques. As the infected monocyte is activated and differentiates, CMV could reactivate from its latent state, spread to surrounding ECs and smooth muscle cells, and further potentiate atherogenesis.

T-cell Stimulation
Primary CMV infection drives expansion of CMV-specific CD4 and CD8 T cells. The functionality of these cells, and ultimately their capacity to prevent CMV recurrence, can be predicted by their expression of transcription factors, including T-bet and eomesodermin [82,83]. Unlike many other viral infections in which the T-cell response is substantially reduced during the contraction phase and is maintained as a low-level memory response, latent CMV infection often drives the continued expansion of activated T cells, a process known as memory inflation. In humans, this phenomenon can result in a surprisingly large proportion of all memory T cells being specific for CMV antigens, upwards of 10% in the general population and even higher in PWH and the elderly [14, [84][85][86]. Although these CMV-specific T cells retain their functionality, and are not exhausted per se, they are incapable of either clearing CMV infection from the body or preventing CMV superinfections.
CD4 and CD8 T cells that express the fractalkine receptor (CX3CR1) have been associated with adverse cardiovascular outcomes [87][88][89][90]. As described above, fractalkine (CX3CL1) may drive the infiltration and retention of T cells, other CX3CR1+ cells (ie, macrophages and NK cells), and CMV-infected cells expressing US28 into the plaques where they may colocalize at sites of EC activation, interact, and contribute to atherogenesis [78][79][80]. We have recently shown that CMV infection is associated with an increase in the proportion of peripheral blood T cells that expresses CX3CR1 [91], and we and others have shown that CMV-specific CD8 T cells are enriched for CX3CR1 expression [91][92][93][94]. CX3CR1+ CD8 T cells are potent effector cells; they are enriched for the cytolytic enzymes granzyme B (GzB) and perforin [94,95] and for production of IFN-g and TNF following T-cell receptor (TCR)-mediated stimulation [93]. TNF release may be an important link between CMV-specific CX3CR1+ T cells and CVD: we have shown that TCR-activated CX3CR1+ CD8 T cells can induce the procoagulant tissue factor (TF) on the surface of monocytes in a TNF-dependent manner [96], and TF-expressing monocytes have been shown to promote coagulopathy in a non-human primate model of HIV infection [97]. Although less abundant than their CD8 T-cell counterparts, CX3CR1+ CD4 T cells are also enriched for TNF production [98][99][100], and our group recently demonstrated that CX3CR1+ CD4 T cells can migrate toward TNF-activated EC in vitro in a CX3CL1-dependent manner [101]. Accordingly, CMV-specific CD4 T cells in general [102], and CMV-specific CX3CR1+ CD4 T cells in particular [99], have been linked to increased carotid intima media thickness in PWH. Furthermore, T cells from CMV seropositive donors provoke endothelial damage, and EC upregulation of adhesive proteins and CX3CL1 [103,104]. Thus, it is plausible that CMV infection results in an immune environment that is primed to exacerbate CVD via trafficking of CX3CR1+ T cells to sites of endothelial dysfunction, which are enriched in CX3CL1. The T cells can mediate further damage to the endothelium via cytokine and lytic granule release and interact with colocalized cells such as monocytes to initiate inflammatory and coagulation pathways. Whether these interactions and subsequent atherogenesis require specific recognition of CMV peptides is unclear, but evidence suggests that CMV nucleic acids and proteins are abundant in plaque tissues (see below). Following this, it is unclear if the cell-mediated mechanisms that may promote cardiovascular risk in CMV seropositive individuals are fundamentally different than those found in CMV seronegative people, or if they are just more abundant.

CMV and Endothelium
The importance of the endothelium in the development and progression of atherosclerosis has long been recognized [105]. The condition of the endothelium determines the adhesive status of the vessel, vessel tone and permeability, the possibility for thrombus formation, and other features. The endothelial state can in turn correlate with the clinical complications of atherosclerosis [106][107][108][109][110]. Smoking, severe mental stress with high catecholamine emission, hemodynamic forces, oxLDL, IFNg, TNF, and other factors can influence the health and function of the endothelium [111][112][113][114][115][116].
Numerous studies have revealed that CMV infection is a key determinant of endothelial status. Direct CMV infection causes vascular EC injury [117,118], enhancing adhesiveness, and EC damage caused by CMV leads to the synthesis and expression of CX3CL1, intercellular adhesion molecule (ICAM)-1, vascular cell adhesion molecule (VCAM), von Willebrand factor (vWF), and other molecules [119][120][121][122][123]. CMV infection also results in EC apoptosis and downregulation of endothelial nitric oxide synthase (eNOS) expression [104,119,121,122,124], which may explain the experimentally impaired endothelium-dependent and endothelium-independent arterial vasodilation driven by CMV [125,126]. Thus, CMV infection increases the permeability of the endothelium, the recruitment and trans-endothelial migration of monocytes [117], and possibly of CX3CR1+ T cells. CMV has recently been shown to affect the expression of matrix metalloproteinases in endothelial and smooth muscle cells, influencing atherosclerotic plaque stability [61, 75,127]. Moreover, CMV infection-associated inflammation leads to an imbalance in the thrombogenic/antithrombogenic function of the endothelium, increasing platelet adhesion to the endothelium as well as the risk of thrombus formation [122,128].
That CMV directly infects the vascular walls in vivo has been demonstrated in clinical studies of individuals suffering from inflammatory eye diseases [129]. CMV seropositivity has been linked to impaired vascular function, assessed from response to bradykinin [130] and brachial artery flow-mediated dilation (FMD) [131], but these results are somewhat controversial [132,133]. In a study of children after heart transplantation, detectable CMV DNA in peripheral blood cells was associated with a significant impairment in FMD [134]. CMV infection in adult heart transplant recipients was associated with impaired endothelial function, assessed from coronary flow reserve [135]. These effects might be due in part to modulation of eNOS activity by CMV infection. Specifically, CMV infection results in an elevated level of asymmetric dimethylarginine (an endogenous eNOS inhibitor), which abolishes eNOS activation, in the serum of heart transplant patients [136]. Endothelial function was also assessed from angiography in coronary artery atherosclerosis patients with different titers of IgG antibodies against the coronary artery disease-related infections CMV, Chlamydia pneumoniae, Helicobacter pylori, hepatitis A virus, and herpes simplex virus (HSV)-1. Simultaneous detection of antibodies against several pathogens significantly correlated with deteriorated endothelium-dependent and endothelium-independent vasodilation in tests using adenosine and acetylcholine in both patients with atherosclerosis and volunteers [137]. In our recent study, we analyzed productive CMV infection using detection of CMV DNA in blood plasma of individuals with acute myocardial infarction and volunteers without CVD. We found a significantly higher frequency of presence and load of CMV DNA in patients with acute myocardial infarction than in controls [138]. Furthermore, when we assessed endothelial function in these patients, we for the first time showed a significant negative correlation between CMV DNA in plasma and the results of the FMD test in patients with acute myocardial infarction [139].
CMV infection is also closely related to increased production of the proinflammatory cytokine IL-15, which also activates T cells, enhances mitochondrial activity, and promotes the intracellular accumulation of cytolytic molecules such as GzB and perforin, as shown in our studies and in those of others [91,101,[140][141][142]. In our recent report investigating the role of CMV infection on the development of CD57+ CD4 memory T cells, we found that IL-15 treatment in vitro upregulated intracellular GzB and perforin and surface expression of CX3CR1 by both CD57+ and CD57-CD4 memory T cells, but more so for CD57+ cells [101]. IL-15 has also been shown to promote the maintenance of CMV-specific inflationary memory T cells in a mouse model [142], and it has been linked to cardiovascular disease-serum IL-15 levels are elevated in patients with coronary artery disease, peripheral artery disease, and hypertension compared to controls [143,144]. In addition, IL-15 is induced by a Western diet and exacerbates atherosclerotic lesion development in the LDL receptor knockout mouse model [145]. IL-15 protein is detected within atherosclerotic plaques in mice and humans, seemingly colocalized with macrophages [146,147], and we recently showed that oxLDL exposure can trigger IL-15 production by monocytes in vitro [148]. Thus, IL-15 is a likely contributor to CVD, and the pathways that drive its expression in CMV seropositive individuals require further exploration.
As described above, CMV infection of the endothelium triggers active recruitment of peripheral mononuclear cells [117,124], as well as the transfer of CMV antigens to T cells via exosome-like extracellular particles [149]. Therefore, it has been suggested that CMV infection of the endothelium results in secretion of viral proteins and mature viral particles, which are then presented by the neighboring ECs to smooth muscle cells [150]. As a result, chemokines actively recruit myeloid cells and T cells to the endothelium, leading to perpetuation of inflammation in the plaques and atherosclerosis progression. Extracellular vesicles (EVs) significantly increase in myocardial infarction [151]; this may be one of the mechanisms allowing CMV to spread among endothelial and immune cells. Over the last few years, the transfer of virus-derived genetic material between cells by EVs has been reported, which in some cases promoted further infection of the target cells with mature viral particles [152,153] or even resulted in productive infection mediated by EVs [154,155]. A similar mechanism has been examined in detail for the transmission of HIV, influenza virus, hepatitis B and C viruses, HSV-1, and HSV-2 [156], and recently the same pathway was demonstrated for CMV transfer in blood [150]. Furthermore, statins, which significantly decrease EV shedding in blood [157], can also suppress production of CMV DNA in cultured cells [158] and abolish the effect of CMV on the expansion of atherosclerosis in mice [159], supporting a role for EVs in CMV production and shedding.

ANIMAL MODELS
Human CMV is highly species-specific, so current animal models of CMV infections and pathophysiology utilize related viruses. Indeed, much of the foundational animal work linking CMV and CVD has been performed in rat models. In a model of photochemically-induced femoral artery injury, rats infected with CMV at the time of injury or 2 weeks later exhibited increase media or neointima thickness, respectively, consistent with the interpretation that CMV infection stimulates cell proliferation in the vasculature [160]. Lemstrom and colleagues reported that CMV infection early after aortic allograft transplantation accelerated atherosclerosis development [161], which could be delayed by ganciclovir administration [162]. In a subsequent study, acute CMV infection was associated with significant EC proliferation and increased intimal thicken-ing, and mononuclear cell infiltrates around CMV-infected cardiac allografts exhibited both early and late CMV antigen expression [163]. Murine CMV has also been shown to be associated with CVD-in fact, mice latently infected with CMV demonstrate vascular dysfunction, particularly as the mice age [126]-and atherosclerotic lesions in CMV-infected apolipoprotein E (ApoE) knockout mice are larger and more advanced than those in uninfected ApoE knockout controls [164]. Monkey studies of CMV pathogenesis, which might be most informative given the similarity to human physiology, are unfortunately hindered by the high prevalence of CMV infections in most non-human primate colonies.

CMV AND HIV -A SPECIAL CASE
Although CMV and HIV are very different at the molecular level, both viruses persist lifelong, and almost all individuals infected with HIV are also infected with CMV (>90%). However, unlike CMV, which has been infecting humans since our species arose, HIV is a far younger virus, infecting humans for only the past hundred years or so, and currently an estimated 40 million people are living with HIV infection worldwide. In the absence of therapy, infection with HIV-1 results in a progressive loss of immune function marked by depletion of CD4 T cells, leading to opportunistic infections and malignancies characteristic of AIDS. Before the advent of antiretroviral therapy (ART), CMV disease was the most common viral complication of AIDS, leading to devastating blindness in many PWH. The introduction of ART in 1995-1996 profoundly reduced the incidence of AIDS-related CMV end-organ diseases [165]. Notably, the substantial overlap between the HIV epidemic and CMV prevalence does not appear to be coincidental. In addition to having similar risk factors for acquisition, accumulating data show that HIV and CMV infections enhance each other's pathogenesis not only by facilitating replication and transmission but also by exacerbating comorbidities, including CVD [12, 23].
In a large longitudinal study of PWH with (n=5119) or without (n=992) CMV coinfection, CMV seropositivity was linked to an increased risk of adverse clinical events, and the cardiovascular outcomes were the most frequent events associated with CMV [23]. HIV infection is also associated with elevated levels of systemic inflammatory mediators (such as IL-6 and interferon-inducible protein [IP]-10) and with chronic immune activation, as evidenced by persistently activated platelets, monocytes, and T cells [166]. The extent to which these effects are due to HIV infection, or are due to coinfections such as CMV, is not known. That CMV infection does contribute is clear however, as we and others have shown that CMV seronegative PWH lack the HIV-associated CD8 T-cell expansion [64,167] and have significantly reduced plasma levels of IP-10, soluble TNFR2, and D-dimer compared to CMV seropositive PWH [64].

Antiviral Drugs
Since CMV infection is a major complication in immunosuppressed individuals, targeting CMV with antivirals has been tried in several clinical settings, including in allogeneic hematopoietic stem cell transplant (HSCT) recipients and in PWH, but any cardiovascular benefits of these treatments are unclear. A randomized controlled study showed reductions in CMV shedding and in CD8 T-cell activation when PWH were treated for just 8 weeks with the anti-CMV drug valganciclovir [168]. Because this trial was small (n=30) and of relatively short duration, analysis of CVD outcomes was not possible, although soluble TNFR2 levels, which are associated with cardiovascular outcomes in PWH [25], were reduced following valganciclovir treatment [65]. Letermovir is a CMV DNA terminase complex inhibitor that is FDA-approved for prophylaxis of CMV infection and disease in adult CMV-seropositive allogeneic HSCT recipients [169][170][171]. Letermovir has superb antiviral activity against CMV with an excellent safety profile and little cross-resistance with other antivirals. Letermovir demonstrated significant benefit compared to placebo measured by time to clinically significant CMV disease post-HSCT (18.9% vs 44.3% cumulative rate; stratified log-rank test, 2-sided P-value <0.0001). Overall, letermovir has shown promising clinical efficacy and is generally well tolerated, thus providing a favorable new option in the prophylaxis of CMV infection and disease. Some limitations of letermovir are that the drug has no significant activity against other herpesvirus or non-human CMV, and it has a relatively low barrier for resistance. Our group is currently developing a clinical trial to test the safety and antiviral efficacy of letermovir in CMV-seropositive PWH. Although this trial will not be powered to detect if letermovir treatment reduces CVD, we plan to measure indices of arterial inflammation and levels of CVD risk biomarkers, including soluble TNFR2, which is strongly linked to cardiovascular outcomes and to CMV coinfection in PWH [25].

Adoptive Cell Therapies
One of the most exciting areas of current research is the development of adoptive cell therapies. This term encompasses a wide range of clinical approaches that span from HSCT to the more recent developments of chimeric antigen receptor (CAR) T cells. In addition to its benefits for treating malignancies, HSCT is the only strategy that has resulted in functional cure of HIV infection [172,173]. However, given the inherent dangers in the procedure and the lack of scalability, HSCT is unlikely to be a widely applied strategy for treatment of any viral infection, especially one as prevalent as CMV. In that sense, more directed approaches, such as CAR T and CAR NK cells, are more promising. CAR cells express engineered transmembrane constructs that contain the antigen-binding variable portion of an antibody extracellularly (to confer specificity) and the CD3z signaling portion of a TCR receptor intracellularly (to confer functionality), often with additional costimulatory elements included, thereby allowing the engineered cell to bypass traditional TCR-MHC restrictions [174,175]. CAR T cells that have specificity for CMV glycoprotein B (gB) have been developed and seem to limit CMV infection in vitro and in a humanized mouse model by cytokine release but not cytotoxicity [176][177][178][179]. Unfortunately, CAR T cells carry considerable risk, including that of cardiovascular complications [180,181], and so it is unlikely anyone but the most vulnerable would be eligible for adoptive cell therapy with CMV-specific CAR T cells.

Immunotherapy
Immunotherapy is another exciting area of advancement in cancer treatment that is being applied for infectious diseases. While the term "immunotherapy" refers to any agent that targets the host response as opposed to the tumor or pathogen (such as with chemotherapy or antivirals), most often the term is used to described therapies (usually antibodies) that inhibit components of immune checkpoint pathways. The FDA has approved immunotherapy regimens that target the immune checkpoint molecules programmed death (PD)-1, PD-ligand 1 (PD-L1), and cytotoxic T lymphocyte-associated protein 4 (CTLA-4), alone or in combination, and therapies that target other pathways are currently in development [182,183]. Many CMV-reactive T cells express PD-1, although their functionality does not seem to be impaired [100,184]. In general, these therapies inhibit the molecular pathways that actively limit T-cell function. Unfortunately, immunotherapy can also lead to considerable immune-related adverse events [185], and as with adoptive cell therapy, it is unlikely that immunotherapy against CMV will gain widespread use in healthy individuals. However, when someone who is CMV seropositive requires immunotherapy, it is possible that anti-CMV activity could be a welcome 'side-effect' of treatment, ie, PD-1 blockade could potentially license T cells to kill CMV-infected monocytes that they might otherwise ignore.
We should note that CMV encodes some of its own inhibitory agents that are potentially targets for immunotherapy. For instance, the viral UL18 protein is a homolog of class I major histocompatibility complex (MHC-I) molecules and can inhibit the function of cells that express the leukocyte Ig-like receptor, subfamily B (LILRB)-1 protein [186]. LILRB-1 expression is found on about half of the CMV-specific CD8 T cells, and blocking LILRB-1 activity increases IFN-g production [187]. Since LILRB-1 binds UL18 with more than 1000-fold higher affinity than host MHC-I molecules [188], UL18 expressed by CMV-infected cells could be a potent inhibitor of immune responses in vivo [189]. Therefore, blocking LILRB-1/UL18 interactions might confer anti-CMV specificity with limited immunopathology and be a useful strategy to target CMV and its associated comorbidities.

CMV Vaccines
The development of a CMV vaccine is a top priority due to its potential cost-effectiveness and associated public health benefits. Prophylactic vaccines might prevent the acquisition of CMV infection, thereby reducing overall infectious burden, and therapeutic vaccines might reduce reactivation from latency, virus shedding and transmission, and CMV-associated morbidities such as CVD. However, there are many challenges facing vaccine development including (1) complex immune evasion mechanisms, (2) unclear immune correlates of protection, and (3) a narrow range of CMV hosts which limits the value of animal models. In spite of these limitations, several types of CMV vaccine candidate, including live-attenuated, DNA, vectored, and peptide vaccines, have been developed or are currently under development, and have been reviewed in detail elsewhere [190,191]. Recently, a modified vaccinia Ankara (MVA)-based vaccine (Triplex), which encodes 3 full-length CMV antigens-pp65 (UL83), IE1-exon4 (UL123), IE2-exon5 (UL122)-obtained excellent safety and immunostimulatory results in a phase 1 trial with healthy adults [192] and in a phase 2 trial of over 100 adult HSCT recipients at high risk for CMV reactivation [193]. A randomized placebo-controlled clinical trial to test the safety and efficacy of Triplex in PWH is being developed by our group. While the trial will not be powered to assess clinical outcomes such as a reduction in cardiovascular outcomes, we plan to measure whether vaccination elicits changes in plasma levels of soluble inflammatory markers associated with CVD.

Cardiovascular Benefits of Anti-CMV Therapies
Most of the focus of anti-CMV therapies has understandably been to prevent acute congenital and transplant-associated infections, as these situations represent the most immediate need for interventions. However, given the association of CMV infection and cardiovascular risk, it is tempting to consider CMV as a target for novel strategies to treat and/or reduce CVD. The high prevalence of CMV within atherosclerotic plaques of patients undergoing endarterectomy [138,194,195], suggests that active CMV replication might be a contributor to plaque development and/or stability. Thus, these approaches could have a direct positive effect by protecting the vascular endothelium and smooth muscle cells that become infected with CMV. Alternatively, targeting CMV could have an indirect effect by reducing the proinflammatory milieu that is central to CVD risk. Of course, these outcomes are not mutually exclusive, and it may be a combination of these effects that ultimately proves protective. As CVD is already the leading cause of mortality worldwide [196], and expected to increase in prevalence [197], finding novel targets to alleviate disease burden is critical. Targeting CMV infection just might serve as such an opportunity.

CMV AS VACCINE VECTOR
In addition to the advancements in anti-CMV drugs, vaccines, and other therapeutic approaches, another promising area of research has been to utilize CMV as a vaccine vector. This is an appealing strategy for a number of reasons, including its robust immunostimulatory potential, its lifelong persistence, and its ability to resist immune clearance in CMV-infected hosts [198,199]. Indeed, preclinical studies have demonstrated that CMV vectors are highly efficacious against challenge with a wide array of pathogens including viruses (SIV), bacteria (Mycobacterium tuberculosis), parasites (Plasmodium knowlesi), and even cancers [200][201][202][203][204][205][206][207]. Current efforts are underway to translate this technology for safe usage in human trials [199,206,208,209]. This potent vaccine efficacy seems to be linked to the development of vaccine antigen-specific effector memory CD8 T cells, as opposed to antibody or central memory T-cell responses [200,210], a finding that is consistent with the important role of CD8 T-cell surveillance in limiting superinfections with CMV [211]. Interestingly, much of the vaccine efficacy seems to be due to the elicitation of CD8 T cells that recognize nonclassical MHC-E alleles [212], and how these observations will manifest in human studies remains to be seen. Nevertheless, caution must be taken to ensure that recipients of CMV-based vaccine constructs do not acquire increased risk of developing cardiovascular complications.

CONCLUSION
Although traditional risk factors (smoking, diabetes, etc.) are certainly important drivers of CVD, it is becoming increasingly evident that persistent CMV infection is a contributor to a variety of cardiovascular complications, even in individuals without overt virus expression. Asymptomatic reactivation from latency is likely far more common than appreciated and promotes inflammation and activation of T cells and monocytes. Early in life, this response is possibly protectiveevidence suggests latent CMV infection may promote influenza vaccine efficacy and clearance of bacterial coinfections [213,214]-yet over time, this benefit is lost and the CMV-associated inflammation proves detrimental. When coupled with direct infection of vascular endothelial and smooth muscle cells, CMV becomes a key factor in cardiovascular pathogenesis. Given the high prevalence of latent CMV infection worldwide, innovative strategies to target the virus, like those described in this review, may be instrumental in helping to alleviate the tremendous burden of CVD.

POTENTIAL CONFLICTS OF INTEREST
The authors declare that no financial conflicts of interest exist for this work.