Cytomegalovirus in Renal Transplantation

Daniel C. Brennan^*

Mark A. Schnitzler^†

^*Department of Internal Medicine, Renal Division, and Pharmacoeconomic Transplant Research, Barnes-Jewish Hospital, Washington University School of Medicine, St. Louis, Missouri 63110, and ^†St. Louis University Center for Outcomes Research, St. Louis, Missouri 63104

INTRODUCTION

Cytomegalovirus (CMV) continues to be a common cause of morbidity and mortality in transplant recipients. It has shifted from being overtly to insidiously lethal. Even with effective prophylactic and preemptive treatment strategies, it is the most concerning viral agent in transplant recipients. CMV disease has been associated with the two most common causes of late graft loss: cardiovascular disease and chronic rejection. The interaction between human leukocyte antigen (HLA), CMV, and impact on transplant outcomes is increasingly understood. Two-HLA-DR mismatches are associated with an increased incidence of CMV disease and poorer allograft survival. The incidence and morbidity of CMV infection and disease is probably a reflection of immunosuppressive strategies, HLA-DR matching, and ability to diagnose and monitor for CMV. New tests have been developed which include the antigenemia assay, the hybrid capture assay, RNA detection by nucleic acid sequence-based amplification (NASBA), and DNA polymerase chain reaction (PCR). New insights have also been reported in the diagnostic, monitoring and treatment of the emergent problem of resistance. Nevertheless, the Holy Grail of CMV is in the development of an effective vaccination against this serious viral pathogen.

DISCOVERY OF CYTOMEGALOVIRUS

CMV was first isolated from the salivary gland and kidney of two dying infants with cytomegalic inclusion bodies and reported in 1956 (1). Two other labs isolated CMV at about the same time. Thus, CMV was initially called “salivary gland virus” or “salivary gland inclusion disease virus.” In 1960, Weller et al proposed the use of the term “cytomegalovirus.” Klemola and Kaarianinen first described CMV mononucleosis, the principal presentation of previously healthy individuals, in 1965. CMV was first isolated in a renal transplant recipient in 1965.

VIROLOGY OF CYTOMEGALOVIRUS

CMV is a member of the genus Herpesvirus and belongs to the family Herpesviridae (2). There are currently eight known human herpes viruses (HHVs) (Table 28.1). The HHVs are further divided into three subfamilies: the
Alphaherpesvirinae, the Betaherpesvirinae, and the Gammaherpesvirinae. The Alphaherpesvirinae include human simplex virus (HSV) 1 and 2, and varicella zoster virus (VZV). The Betaherpesvirinae include CMV, HHV-6 and HHV-7. The Gammaherpesvirinae include Epstein-Barr virus (EBV) and HHV-8. Morphologically the herpes viruses are indistinguishable from one another. The complete virion is 150 to 200 nm in diameter, icosahedral in shape, and consists of an inner core, a capsid, and an envelope (Fig. 28.1). Human CMV is the largest known virus. The inner core (genome) of the CMV virus is a 64-nm linear double-stranded DNA molecule. There is little genetic homology between human CMV and CMV of other species. The capsid is 110 nm in diameter and consists of 162 protein capsomers. The envelope contains lipoproteins and at least 33 structural proteins, some of which are glycosylated (glycoproteins). The glycoproteins determine the strain of CMV, are used for cellular entry of the virus, and are the targets of virus neutralizing antibody.

TABLE 28.1. Human herpes virus and disease

Virus	Disease and possible disease associations
Herpes simplex 1	Herpes labialis
Herpes simplex 2	Herpes genitalis
Varicella zoster virus	Chickenpox
Epstein-Barr virus	Infectious mononucleosis
	Burkitt lymphoma
	Nasopharyngeal carcinoma
	Posttransplant lymphoproliferative disorder
Cytomegalovirus	Cytomegalovirus disease
	Salivary gland virus disease
Human herpes virus 6	Roseola (exanthem subitum)
	Transplant encephalopathy
	Multiple sclerosis
Human herpes virus 7	Roseola
Human herpes virus 8	Kaposi sarcoma
	Body cavity-based lymphoma
	Sarcoidosis

CMV is a labile virus and readily inactivated by lipid solvents, pH below 5, heat (37°C for 1 hour or 56°C for 30 minutes), and ultraviolet light for 5 minutes. It can survive on environmental surfaces for several hours. CMV can be stored at 4°C for a few days without loss of infectivity. Storage at −70°C without loss of infectivity is possible for several months. CMV can be stored at −190°C (liquid nitrogen) indefinitely.

White blood cells and CD13-positive cells in particular are the principal reservoir where CMV is harbored (3). It has been detected in most tissues in the body and may remain latent. The virus enters the host cell by fusion of the virus envelope with the cell membrane or via phagocytosis. Infectious particles are first detectable by electron microscopy 1 to several days after inoculation; however, viral DNA and protein can be detected in the infected cell before viral assembly is complete. Virus particles are made and assembled in the nucleus and attain an envelope by budding through the inner nuclear membrane. From there the particles go through the trans-Golgi network where the virus particle becomes pathogenic through proteolytic cleavage of a consensus furin site to form glycoprotein B (gB). Furin was the first proprotein convertase to be identified. It is mainly localized in the trans-Golgi network. This endoprotease is capable of cleaving precursors of a wide variety of proteins, including growth factors, serum proteins such as proteases of the blood-clotting and complement systems, matrix metalloproteinases, receptors, bacterial exotoxins and viralenvelope glycoproteins (4). Glycoprotein B is the CMV UL55 gene product and the predominant human CMV envelope glycoprotein (5). Because defective virus particles are non-infectious, this process of virion formation may be amenable to development of potential vaccines (see below).

FIG. 28.1. Schematic of cytomegalovirus (CMV). CMV shows a similar composition to all human herpes viruses.

CMV replication produces immediate-early (IE), early, and late CMV antigens. IE antigens appear in the nucleus of CMV infected cells 1 to 3 hours after infection and remain present even in latent infection. IE antigen gene products direct production of both viral and cellular genes. Early antigens appear in the cytoplasm or membrane about 3 hours after infection. Early antigen gene products direct viral DNA synthesis. Late antigens appear in the nucleus and cytoplasm within 6 to 24 hours after infection. Late antigen gene products direct production of structural nucleocapsid proteins. IE and early antigens are virus-induced nonstructural proteins and appear before DNA synthesis. This is important because the mechanism of action of ganciclovir, foscarnet, and cidovir (the three most common agents used for treatment of
CMV) is through interruption of DNA synthesis. Late antigens are virally encoded structural proteins and appear after DNA synthesis, so their appearance is sensitive to the common antiviral agents. Because of this, monitoring late antigen levels may be more relevant than monitoring IE or early antigen levels when assessing response to therapy.

The CMV antigens have a number of other effects. The IE gene product upregulates transcription and expression of interleukin-2 (IL-2) and the IL-2 receptor (6). The CMV IE gene also prevents the inhibitory effect of cyclosporine on IL-2 gene transcription (7). It may do this in part by sustained NF-κB activity and transactivation of promoters containing NF-κB enhancer sequences. The IE and early antigen gene products also upregulate adhesion molecules such as intracellular adhesion molecule-1 (ICAM-1) and lymphocyte functioning antigen-3 (LFA-3). This upregulation is blocked by ultraviolet irradiation, which destroys viral infectivity. ICAM-1 and LFA-3 upregulation are not blocked but actually enhanced with treatment with ganciclovir and foscarnet (8). Thus, CMV infection can initiate endothelial activation and rejection despite effective treatment with ganciclovir, foscarnet, or cidofovir.

The CMV virus has developed a number of mechanisms to subvert host defenses. One of these is the ability to downregulate expression of Major Histocompatibility Complex (MHC) class I molecules, which may allow evasion of, and recognition by cytotoxic T lymphocytes (9). Human CMV proteins also block the transporter associated with antigen processing (TAP), retain MHC class I molecules in the endoplasmic reticulum (ER), and recycle nascent class I heavy chains back to the cytosol. Because MHC class I molecules, acting through various receptors, signal inhibitory messages to natural killer (NK) cells, CMV-infected cells (which lack MHC class I molecules) should be more susceptible to attack by NK cells. However, CMV-infected cells express CMV gpUL18, a class I homolog that may allow for evasion of NK lysis of infected cells (9). Another CMV antigen, gpUL40, has been shown to upregulate the nonclassical MHC class I molecule HLA-E. HLA-E inhibits NK cell-mediated lysis by binding to a C-type lectin receptor on NK cells (10). CMV can also evade eradication by production of a viral IL-10 homolog [11]. IL-10 blocks proinflammatory cytokine synthesis and suppresses the ability of macrophages to serve as antigen-presenting or costimulatory cells.

DEMOGRAPHICS

While symptomatic CMV infection occurs in 20% to 60% of all transplant recipients and is a significant cause of increased morbidity and mortality in this population (12,13), latent infection occurs in 60% to 90% (14). The incidence may be somewhat lower in kidney transplant recipients who do not receive antilymphocyte therapy, high-dose mycophenolate, or the potent combination of mycophenolate and tacrolimus without adequate antiviral prophylaxis. Retrospective studies in CMV seronegative transplant recipients of a kidney from a CMV seropositive donor have shown that the use of mycophenolate in combination with cyclosporine and prednisone increases the incidence of symptomatic CMV but does not increase the incidence of primary CMV infection or severity of infection (15,16). The U.S. Multicenter FK506 kidney transplant trial demonstrated no difference in the incidence of symptomatic or asymptomatic CMV infection between patients receiving cyclosporine versus tacrolimus (17). Two phase III, multicenter, randomized, double-blind, placebo-controlled trials with Daclizumab, a humanized anti-IL-2 receptor monoclonal antibody, showed no significant difference in the incidence of CMV infection, disease, or severity of disease (18). These results were also reported for basiliximab, another humanized anti-IL-2 receptor monoclonal antibody (19). It is important to emphasize, that in both the U.S. Multicenter FK506 kidney transplant trial and the phase III trials with Daclizumab approximately 10% to 60% of the patients received CMV prophylaxis with acyclovir, ganciclovir, or CMV IgG. Sirolimus, a new immunosuppressive agent, has not been shown to increase the incidence of CMV infection when compared to cyclosporine (20,21).

The wide range in the reported incidence of infection and disease results from the varying intensities of immunosuppression used and the frequency and methods used to detect CMV infection. Historically, concern has mainly focused on avoiding CMV infection in the CMV D+/R− group because this group has been at greatest risk for severe “primary” infection during the first 3 months posttransplant. However, the indirect effects of CMV infection on graft and patient survival have been increasingly recognized in recent years. Our own analyses of data from the United States Renal Data System (USRDS) and United Network of Organ Sharing (UNOS) showed that by 3 years, it is the D+/R+ group and not the D+/R− group that has the worst graft and patient survival (22,23). The reason for this is not entirely clear but may reflect the prevalence of multiple CMV virotypes and that the D+/R+ patients may have a double CMV exposure with reactivation of differing latent donor and recipient CMV. CMV can predispose to rejection as well. Subclinical CMV may also mimic and or predispose to late acute rejection. Late subclinical infections are common and associated with relatively rapid graft loss (24).

EFFECT OF CYTOMEGALOVIRUS DISEASE AND HLA-DR MATCHING ON GRAFT SURVIVAL

More recent data show that it is the D+/R− group that has the worst graft survival. This recent shift may have resulted from the widespread use of prophylaxis recently (25). Prophylaxis is very effective in D+/R+, but D+/R− patients still have a high incidence of CMV-disease after prophylaxis. In our own single-center study, despite 3 to 6 months of oral ganciclovir prophylaxis, development of even mild CMV disease was associated with worse long-term graft survival (Fig. 28.2) (26). The degree of HLA-DR matching
was shown to have a significant effect upon the development of CMV disease and allograft survival despite prophylaxis (Fig. 28.3) (26). At 5 years, allograft survival was significantly decreased among those with CMV disease and zero matches versus patients with disease and one or two HLA-DR matches (16% and 76%, respectively). These results suggest that to enhance allograft survival consideration should be given to HLA-DR matching plus CMV status in the allocation of kidneys and that longer periods of prophylaxis may be warranted (see below).

FIG. 28.2. In patients with functioning grafts at 6 months, cytomegalovirus (CMV) disease was associated with 56.8% graft survival compared to 79.1% graft survival in patients without CMV disease (P < 0.001). (From Schnitzler M, Lowell J, Hmiel S, et al. Cytomegalovirus disease after prophylaxis with oral ganciclovir in renal transplantation: The importance of HLA-DR matching. J Am Soc Neph 2003;14:780-785, with permission.)

CYTOMEGALOVIRUS AND OTHER DISEASE ASSOCIATIONS

CMV has been associated with both atherosclerosis and chronic rejection, and the two most common causes of late graft loss are cardiovascular death and chronic rejection. Chronic rejection is also known as chronic allograft nephropathy and is characterized by myointimal thickening, a form of atherosclerosis. Although the evidence of CMV causing chronic rejection has been inconsistent due to confounding factors like acute rejection and treatment associated with increased incidence of CMV, results of a cohort study involving 1339 patients revealed that graft survival was decreased for those patients who suffered from episodes of acute rejection and CMV disease in comparison to those suffering from acute rejection only (27). Other studies have associated CMV antigens and DNA with chronic histologic changes in renal allograft biopsies of patients with CMV infection (28).

Acute rejection is a major risk factor for chronic rejection and acute rejection has been associated with CMV. In a cohort study involving 192 renal transplant patients, CMV disease was found to be a risk factor for developing acute rejection (29).

FIG. 28.3. In patients who had a cytomegalovirus (CMV) disease diagnosis and who had functioning grafts through 6 months posttransplant, 5-year survival was 87.5% for two HLA-DR matches, 61.1% for one HLA-DR match, and 16.2% for zero HLA-DR matches (P = 0.002). There was no significant difference in graft survival after 6 months in patients diagnosed with CMV disease between one and two HLA-DR matches for a combined graft survival of 75.9% (P = 0.47), which was not significantly different from the graft survival in all patients without a CMV disease diagnosis (P = 0.55). Thus, in cases with grafts surviving at least 6 months, patients with zero HLA-DR matches and a diagnosis of CMV disease compared to all others had a 395% increase in the risk of graft loss by 5 years posttransplant (P<0.001) (From Schnitzler M, Lowell J, Hmiel S, et al. Cytomegalovirus disease after prophylaxis with oral ganciclovir in renal transplantation: The importance of HLA-DR matching. J Am Soc Neph 2003;14:780-785, with permission.)

Latent CMV infection has been associated with a markedly increased rate of restenosis (1,200% higher) after coronary angioplasty in nontransplant seropositive individuals compared to seronegative individuals (30). Furthermore, the histologic lesion of coronary restenosis is diffuse atherosclerosis characterized by myointimal hyperplasia resembling chronic rejection.

Additionally, Humar et al recently reported that kidney transplant recipients with CMV disease had overall more cardiac complications than patients with no history of CMV disease. Congestive heart failure, arrhythmias, and documented vessel occlusion by angiogram were significantly more common in patients with CMV disease, although myocardial infarction, angina and cardiac arrest were not significant (31). In a subanalysis of sera from 3,168 nontransplant Canadian patients in the Heart Outcomes Prevention Evaluation (HOPE), exposure to CMV but not to C. pneumoniae, H. pylori, or hepatitis A virus was associated with a slight excess risk of subsequent myocardial infarction, stroke, or CV death (32). CMV is felt to affect atherosclerosis in part by blocking the antiproliferative effects of p53 (33). However, more recent evidence suggests that the
myointimal increase is not from proliferation, but rather migration of smooth muscle cells (SMCs) mediated by virally encoded chemokine receptors that require tryosine kinases for expression (34). SMC cell migration could be blocked by tryosine kinase inhibitors and suggests another potential future therapeutic strategy. Human CMV infection may also cause atherosclerosis and chronic rejection by increasing oxidized-low-density lipoprotein (LDL) uptake by vascular SMCs (35). This process is mediated by the IE gene, IE72, and does not require viral replication, which may explain why angioplasty patients latently infected with CMV had higher rates of restenosis after angioplasty.

CMV has been associated with several other vascular injuries that may explain the poorer survival among patients with CMV infection. One such vascular injury associated with CMV is transplant glomerulopathy. However, the frequency and clinical significance of this lesion are uncertain (36). The lesion of the hemolytic uremic syndrome/thrombotic microangiopathy (HUS/TTP) is one of the more common vascular pathologies associated with CMV and may be confused with or present with cyclosporine or tacrolimus toxicity (37, 38, 39). CMV associated HUS/TTP may respond to immunoglobulin infusion (39).

CMV has also been associated with transplant renal artery stenosis. In a study of over 900 renal transplant recipients, 75 were diagnosed with renal artery stenosis via angiography (40). Patients with stenosis were paired with a control individual matched for age, sex, year of transplant, and number of grafts, but without renal artery disease. Definitive evidence of CMV infection was significantly associated with stenosis (36 vs. 12 control patients, P<0.001). Thus the more subtle effects of CMV on vascular biology may ultimately be more important than acute CMV disease itself as a cause of patient and graft loss.

TABLE 28.2. Diagnosis of cytomegalovirus

Method	Source	Comment
Histopathology	Tissue	Detects inclusion bodies, very insensitive, requires biopsy.
Serology	Serum	Several techniques available for detection of IgG and IgM. ELISA is most common, but others include complement fixation and latex agglutination. Useful to detect past exposure (IgG). Not useful for early diagnosis of acute infection. Failure to detect IgM in presence of viruria indicates a poor prognosis.
Conventional tube culture	Any	Long development time (1-3 weeks). Subject to technical complications from “cytotoxicity.” Positive test indicates active CMV. Sensitivity decreased with a delay of processing exceeding 6 hours. Necessary for plaque assay resistance testing.
Shell vial assay	Any	Rapid development time (1-2 days). Semi-quantifiable. Less labor intensive than antigenemia assay. Positive test indicates active CMV. Sensitivity decreased drastically with a delay of processing exceeding 6 hours.
Antigenemia assay	Blood	Rapid (same day). Semiquantifiable. Labor intensive, requires specially trained technician. Sensitivity decreased with a delay of processing exceeding 6 hours.
PCR and PCR TaqMan or LightCycler	Any	Fairly rapid (same day), can be batched. Most sensitive of all tests. Positive test indicates active CMV but not necessarily symptomatic CMV (latent CMV rarely detectable). Allows quantification. PCR of DNA is not adversely affected by delays in transportation.
Hybrid capture assay	Any	Rapid. Detects viral DNA, less sensitive than PCR. Quantifiable.
Branched DNA	Any	Less sensitive. Reproducible.
NASBA	Any	Detects viral RNA, highly sensitive
In situ DNA hybridization	Tissue	Requires special techniques, may localize CMV DNA
Ig, immunoglobulin; ELISA, enzyme-linked immunosorbent assay; CMV, cytomegalovirus; PCR, polymerase chain reaction; NASBA, nucleic acid sequence-based amplification.