OUP user menu

Cytomegalovirus: recent progress in understanding pathogenesis and control

V.C. Emery
DOI: http://dx.doi.org/10.1093/qjmed/hcr262 401-405 First published online: 22 December 2011


Cytomegalovirus continues to be an important pathogen in a variety of patient groups especially the neonate and the transplant recipient, and has implicated in a range of pathologies including inflammatory disease and in contributing to early death in ageing populations. This review will focus on advances in understanding the virus–host interaction and options for the new therapeutic control measures.


Cytomegalovirus (CMV) has the largest genome of any human herpes virus with ∼230 Kb of genetic information, which encodes a proteome of at least 167 unique viral proteins.1 Clinical strains of CMV posses additional genes not found in the laboratory-adapted strains and many of these genes are important in pathogenesis of infection.1 CMV is a relatively successful pathogen in so far as it infects and establishes latency in ∼70% of the global human population with seroprevalence highly correlated with socioeconomic group, ethnicity and geographic region.

The virus devotes a substantial proportion of its genetic content to the control of host processes presumably because its ability to remain latent in bone marrow progenitor cells and to replicate in a broad range of cells and tissues necessitates an intimate ability to ensure the virus and host remain in perfect harmony.2 The virus encodes a number of proteins that manipulate the Class I and Class I human leuckocyte antigen (HLA) response, interfere with natural killer cell (NK) cell activities, control and manipulate the cell cycle, inhibit apoptotic pathways and modulate inflammatory pathways including the matrix metalloproteinase pathway and cellular adhesion molecules. Based upon data from the Rhesus CMV model, the Class I HLA manipulation genes may serve as facilitators of reinfection.3

Control of replication in the immunocompetent host is mediated through a robust CD4 and CD8 T-cell response,4 through NK cells and via antibodies that recognize key surface glycoproteins such as gB and gH either singly or, as recently described, as part of multiprotein viral surface complexes.5 Historically, the T-cell immunocompromised host has been in the group where CMV infection has exerted its full pathogenic effects leading to a range of pathologies. In congenital infection, pathological consequences include microcephaly leading to poor mental development and sensorineural hearing loss6; whereas in the transplant recipient both direct effects, such as CMV hepatitis, pneumonitis, gastrointestinal disease and prolonged fever are evident as well as indirect effects, such as acute and long-term graft rejection, especially accelerated coronary artery disease after heart transplantation.7 A model for the pathogenesis of CMV after solid organ transplantation is shown in Figure 1. Although the mechanistic basis for the indirect effects have not been fully elucidated, there is an increasing body of evidence from human studies, in vitro studies, and small animal models that CMV intimately contributes to these pathologies and using more sensitive methodologies viral DNA can be directly detected in affected tissues arguing that the term ‘indirect effects’ may be misleading.8 In contrast to the range of CMV diseases observed in the neonate and transplant recipient, human immunodeficiency virus (HIV)-1-infected patients, prior to widespread use of highly active anti-retroviral therapy (HAART), predominantly suffer from CMV retinitis as their first episode of CMV disease occurring when CD4 T-cell counts fell <50 cells/Ul although in the absence of CMV therapy development of further pathologies was relatively common including peripheral and central nervous system disease and adrenalitis. The reasons for these differences in the pathogenetic pattern between different immunocompromised hosts are unknown but it is possible that without a fully functional immune system some pathologies associated with CMV are not manifest (see below).

Figure 1.

A model for CMV pathogenesis after solid organ transplantation. The donor organ harbours a small number of cells with latent infection (red dots), which become activated through the effects of the proinflammatory environment on the major immediate early promoter shortly after transplant. Subsequent local spread of virus in the infected organ ensues over the next 7 days, which may then spread through the blood to infect other target organs, which contributes to the overall level of CMV DNAemia. If left untreated, these high levels of replication will be associated with the direct effects of CMV infection. In addition, early graft infection may contribute to acute organ malfunction, occurrence of other opportunistic infections and also long-term graft and patient survival. GI: Gastrointestinal.

In addition to the classical T-cell immunocompromised host, there is increasing evidence that CMV can also be pathogenic in other non-T-cell compromised patients. For example, it has been shown recently that in patients with the late-onset primary antibody deficiency (common variable immune deficiency disease), the combination of CMV replication in target organs such as gut and kidney together with a hyper-reactive CD8 T-cell immune response can combine to yield substantial tissue inflammation.9 Interestingly, this inflammatory disease can be reduced through deployment of anti-CMV therapy using ganciclovir and by inhibiting tumour necrosis factor (TNF)-α through antibody therapy with infliximab. These data indicate that the pathologies we observe with CMV may be a consequence of both viral-mediated destruction of cellular systems and the host immune response against infection. An exaggerated immune response to CMV has also been observed in chronic lymphocytic leukemia patients and is associated with shortened time to death.10 In addition, in the elderly, there is controversial data linking CMV with an immune risk phenotype that disposes individuals to early death compared to those without the CMV immune risk phenotype,11 and CMV infection may also portend a poor prognosis in patients who are in intensive care units (ICUs).12 Studies to define whether intervention in the ICU setting to reduce CMV load (see below) can improve patient outcome are awaited.

In all patients groups studied to date, viral load (as a surrogate of CMV replication) has been shown to be intimately associated with pathogenesis.13 Thus, at critical viral loads the probability of a patient suffering disease is substantially increased, although longer periods of lower level replication with a similar cumulative viral load can also be a risk factor for CMV disease.14 These observations coupled with the demonstration suggest that CMV replication in vivo is highly dynamic and have led the way to improvements in the management of CMV through the deployment of prophylaxis or pre-emptive therapy.

Immune control of CMV in the immunocompromised host

Significant progress has been made in understanding how the immunocompromised host attempts to control CMV replication. Utilization of Class I HLA multimers together with intracellular cytokine assays has revealed that patients who fail to control CMV replication have poorly functional CD4 and CD8 T-cell responses especially with respect to mono- and polyfunctional cytokine production.1517 However, very few studies have identified prognostic immune markers for future disease whereas many studies have provided evidence of phenotypic markers that are modulated and associated with periods of active replication including PD-1 and perforin levels.17 In the HIV infected individual, it has been shown that subsequent to the initial loss of CMV-specific CD4 T-cells there is a subsequent loss of functionality of the CD8 T-cell population specific to CMV leading to a poorly functional set of cells that are either unable to kill targets effectively or, more likely, unable to proliferate effectively in the absence of CD4 help.18 These data concur with the observations on CD8 T-cell function in transplant recipients summarized above. Since T-cells are critical in the control of CMV replication, many groups have been developing adoptive immunotherapy approaches to allow control of CMV especially after stem cell transplantation. Recent technological progress enables such cells to be made more rapidly and cost-effectively and various clinical trials are underway to assess their efficacy at modulating CMV infection after stem cell transplantation.19

Antiviral and vaccine-mediated control of replication

To date, the mainstay for control of CMV replication and treatment of established CMV disease has been through the deployment of the nucleoside analogue ganciclovir and its prodrug valganciclovir. The use of these agents in prophylactic and pre-emptive treatment settings has impacted substantially on the burden of CMV disease after transplantation, especially in high-risk transplant populations.20 However, the drug has in vivo toxic side effects and post-prophylaxis disease remains an important problem in the transplant setting. Extending prophylaxis from 100 to 200 days in high-risk renal transplant recipients has been shown to decrease the incidence of post-prophylaxis CMV infection and disease by 35% and 50%, respectively, and is cost-effective.21 Trials of a new CMV drug based upon a benzimidazole nucleus (Maribavir) that inhibits the CMV UL97 protein kinase (which also activates ganciclovir to its monophosphate) have been disappointing with a phase 3 study in stem cell transplant patients showing no benefit over placebo22 and a phase 3 study in liver transplant patients being stopped prematurely by an independent data safety monitoring board on the basis of no obvious benefit. The underlying reasons for the failure of this drug in the clinical setting have not been fully determined, although the effective drug levels attained in vivo may have contributed to be disappointing results since the drug is highly protein bound. Other compounds that have shown promise include cyclopropavir (a nucleoside analogue that requires UL97 for phosphorylation and ultimately targets the CMV DNA polymerase), a lipid conjugated form of cidofovir (CMX001) that is ∼100-fold more potent than its parent cidofovir and does not seem to have the renal toxicity associated with cidofovir. The most recent addition to the stable is AIC246, a molecule that targets the CMV maturation complex and has shown excellent results in phase 1 studies and is likely to enter phase 2 studies in early 2012 (reviewed in Emery and Milne).23

CMV vaccines have had a renaissance in recent years partly propelled by the success of a recombinant glycoprotein B vaccine produced by Sanofi-Pasteur in seronegative women of child bearing age.24 Vaccination of these women prevented infection in 50% of vaccines and the same vaccine has been shown to be highly immunogenic in patients awaiting a kidney or liver transplant and to reduce markers of viral replication in the post-transplant period.25 A number of other vaccines are being developed but the alphavirus-based replicon containing both B- and T-cell targets (gB and pp65-IE1 proteins, respectively) elicits impressive antibody and T-cell responses in healthy subjects26 while a CMV DNA vaccine incorporating gB and pp65 has shown sufficient promise in phase 2 development that it will enter phase 3 evaluation in the near future.27 These vaccines are now being pursued through a licensing arrangement with Novartis (the alphavirus vaccine) and Astellas (the DNA vaccine).

Concluding comments

The last decade has seen an impressive enhancement of our knowledge relating to CMV pathogenesis and virus–host interactions. The next decade will continue to enhance this understanding but will see the introduction of new antiviral agents and potentially vaccines to control this important and ubiquitous pathogen.


Work in the laboratory of VCE is supported by an MRC Centre grant and the Wellcome Trust.

Conflict of interest: None declared.


View Abstract