Cytomegalovirus seropositivity is significantly associated with mycosis fungoides and Sézary syndrome: Presented in part at the 43rd annual meeting of the American Society of Hematology, Orlando, FL, December 8-11, 2001, and at the annual meeting of the American Academy of Dermatology, New Orleans, LA, February 24-28, 2002.

Cutaneous T-cell lymphomas (CTCLs) are clonal T-cell malignancies with primary cutaneous manifestations. The most common variant of CTCL is mycosis fungoides (MF), an indolent lymphoma characterized histologically by atypical epidermotropic CD4⁺helper/CD45Ro⁺ memory T-cell clones. Sézary syndrome (SS) is the erythrodermic, leukemic variant of MF.1 MF is hypothesized to arise through persistent antigenic stimulation, leading to an accumulation of skin-homing T cells that are defective in apoptotic programmed cell death.2 3 Although viruses are known to provide chronic immune stimulation, cytomegalovirus (CMV) has not previously been identified as a risk factor for MF/SS.

Human cytomegalovirus (CMV) is a member of the herpes family of viruses. Primary infection is followed by latency. Polymorphonuclear leukocytes, T cells, endothelial cells, and salivary glands have all been shown to harbor CMV in latent form.4 Active CMV infection causes characteristic changes in specific lymphocyte populations, ie, CD8⁺ T-cell lymphocytes and natural killer (NK) cells.5,6 CMV has an array of immune escape strategies for establishing life-long latency. CMV inhibits major histocompatibility complex (MHC) class I expression within infected cells and impairs MHC class II–dependent antigen presentation by macrophages.7 It also encodes proteins that interfere with the presentation of viral peptide antigens to T cells.

CMV is a common opportunistic pathogen in both the transplant and human immunodeficiency virus–infected (HIV⁺) populations, reflecting the T-cell deficiency in these patients.8Patients with CTCL also have profound immunodeficiency in later stages of the disease with diminished T-cell function and disappearance of the CD8 cytotoxic population.9 Infectious syndromes resulting from CMV in immunocompromised hosts include pneumonitis, gastroenteritis, hepatitis, pancreatitis, encephalitis, retinitis, and bone marrow suppression. CMV-induced skin lesions are nonspecific and generally localized and ulcerative or are generalized and maculopapular.10 In patients receiving solid-organ transplants, CMV infection causes direct injury to transplanted organs and contributes to both acute and chronic transplant rejection.11 CMV has been implicated in the development of accelerated atherosclerosis in the coronary arteries of transplanted hearts.12 In healthy individuals, infection manifests most commonly as a self-limited mononucleosislike syndrome or is entirely asymptomatic. The seropositivity rate in the general population is generally increased with parity and age.

This was a retrospective study of new patients with MF and SS undergoing study evaluation at the MD Anderson Cancer Center Skin Center from 1992 to the present. All patients with CTCL had a viral serology panel drawn as part of their baseline routine evaluation, and all gave written consent to have tissue and data used for research studies. Control subjects were bone marrow donors who gave consent to be prospectively evaluated for CMV seropositivity at baseline and who were retrospectively analyzed for seropositivity data. Patients and control subjects are grouped as seropositive or seronegative. Staging of patients with CTCL is done by the tumor, node, metastases (TNM) classification method.13 The demographics of patients and controls are shown in Table1.

Serology methods for measuring CMV immunoglobulin G (IgG) used by the MD Anderson Diagnostic Laboratory are the commercially available ELISA (enzyme-linked immunosorbent assay) and MEIA (membrane enzyme immunoassay). HIV, human T-cell leukemia virus type I and II (HTLV-I and II, and Epstein-Barr virus (EBV) serologies are also measured by using ELISA.

Chi-square analyses using 2 × 2 tables were performed to compare CMV seropositivity in all CTCL patients, in early-stage patients, and in late-stage patients with healthy bone marrow control subjects. In order for a chi-square value to be significant with one degree of freedom, and P = .05, that value must be more than 3.841.

In addition, chi-square analysis and independent samples ttest were used to compare CMV seropositivity in CTCL patients aged 55 years or younger with bone marrow donors aged 55 years or younger.

The demographics of the patients with CTCL between 1992 and the present who had prospective evaluations done at baseline for CMV seropositivity are shown in Table 1. Of 116 MF/SS CTCL patients, 113 (97.4%) had positive CMV-IgG serologies, including 59 men and 57 women, and are equally distributed in our population. Only 3 CTCL patients were found to be seronegative. Those were 2 women and 1 man, stages IA, IIB, and IVB, respectively (Table2). Of 53 patients with early-stage MF (IA, IB, IIA), 52 were CMV seropositive (98.1%). Patients with late-stage MF/SS (IIB-IVB) had a 96.8% (61 of 63) seropositivity rate. In contrast, the seropositivity rate among healthy bone marrow donors evaluated at MD Anderson since 1988 was 57.3% (757 of 1322).

MF/SS patients as a group had a significantly higher rate of CMV seropositivity compared with healthy control subjects (χ²_05(df=1) = 71.79). Compared with controls, the rate of CMV seropositivity in patients with early-stage MF is also significantly higher (χ²_05(df=1) = 34.98). Significantly higher CMV seropositivity was also noted among patients with late-stage MF/SS when compared with control subjects (χ²_05(df=1) = 38.9). By independentt test, the incidence of CMV seropositivity is significantly higher in all stages of MF compared with healthy bone marrow donor control subjects (P = .001).

We also examined serologies for other viruses such as HIV-1 and viruses implicated in CTCL and HTLV-I and II.14 One hundred thirteen patients with MF/SS were tested for HIV, and all were negative by ELISA. Of 114 patients tested for HTLVI/II, only one was positive. Only 13 patients had EBV serologies drawn, but all were seropositive (Table3).

Recognizing that the rate of CMV seropositivity increases with age, we also compared CMV seropositivity rates in CTCL patients with CTCL patients that were aged 55 years or younger with age-matched, healthy bone marrow donors aged 55 years or younger. Again, CMV seropositivity was significantly higher among “young” CTCL patients compared with healthy bone marrow donors (χ²_05(df=1) = 20.4). Thirty of 32 (93%) CTCL patients younger than age 55 were CMV seropositive in contrast to only 591 of 1103 (53.6%) bone marrow donors aged 55 or younger.

This study shows for the first time that more than 97% of patients at all stages of CTCL (MF and SS) have evidence for previous infection with human CMV. The seropositivity rates are significantly higher compared with healthy donor control subjects from bone marrow transplant donors and are high even in the subset of young patients. Surprisingly, the rate of seropositivity of CTCL patients exceeded that of patients who are historically considered to be in high-risk groups, including homosexual men and patients with HIV/AIDS (Table4). Patients with CTCL also develop profound immunodeficiency with disease progression, and their gradual loss of the CD8⁺cytotoxic T-cell population is associated with poor prognosis and has been attributed to the overgrowth of the malignant clone.9Apoptosis of CD8⁺ tumor-infiltrating lymphocytes may also be mediated by tumor cells with Fas ligand (FasL) expression, resulting in immunosuppression with time.27 

Cytomegalovirus is a latent virus and has been implicated in bone marrow suppression and with fatal pneumonia in our transplant population.28 CMV seropositivity is a negative independent prognostic factor for overall survival among patients receiving bone marrow transplants.26 Because CMV seropositivity increases with age, we also examined a subset of the patients and found that the seropositivity rate among younger CTCL patients was 93%, with only 2 of 32 patients seronegative. In fact, only 3 patients in the entire cohort were negative for CMV. We also examined the rate of Epstein-Barr virus seropositivity in the same group of patients. Of the 13 patients in the group who were tested, all 13 were positive.

Viruses have been implicated in a variety of autoimmune diseases and may share host antigens (molecular mimicry), resulting in inappropriate immune stimulation against host proteins. Autoimmune diseases are almost always associated with certain human histocompatibility antigens thought to confer disease susceptibility or be correlated with the development of specific antibodies.29 Epstein-Barr virus infection has been previously implicated in several kinds of lymphomas, including Hodgkin disease, Burkitt lymphoma, NK-T nasopharyngeal tumors, peripheral T-cell lymphomas, and posttransplantation lymphoproliferative disorders.30,31 EBV is also associated with autoimmune diseases such as childhood lupus erythematosus and rheumatoid arthritis.32 HTLV-1 is found in adult T-cell lymphoma and leukemia, and it has been implicated by several studies in CTCL.33-35 Others have not been able to confirm this finding.36 37 In this study, all except one of the patients with CTCL had negative HTLV-1 serology studies at baseline.

We have previously shown that patients with MF and SS have a significant association with major histocompatibility class II HLA-DR5 (w11) and HLA-DQB*03 alleles.38 Staphylococcal superantigens were also commonly associated with erythrodermic MF/SS patients in our previous study and can stimulate proliferation of CTCL cells.3,39 The HLA-DR5 allele is also associated with melanoma progression, scleroderma, Hashimoto thyroiditis, and alopecia areata,40 all characterized by a T-cell–mediated immune response. Of interest, the DR5 antigen is also exquisitely sensitive to stimulation by staphylococcal superantigen.41 HLA alleles may also influence the incidence of CMV infection and subsequent seropositivity within a population group.42 

A chlamydia-related factor found to stimulate proliferation of Sézary cells led to the finding of chlamydia organisms within MF lesions and has been suggested to contribute to the pathogenesis of MF/SS as a persistent antigen.43 Chlamydia organisms and also CMV have been implicated in inflammation initiating coronary artery disease.44 However, in one study of angiography patients in Spain, 97% had CMV seropositivity regardless of the presence of atherosclerotic plaques at catherization.45 A prospective, well-designed case-control study of healthy postmenopausal women with (n = 122) and without (n = 244) coronary artery disease (CAD) also found no significant differences in seropositivity rates for CMV, Chlamydia, andHelicobacter pylori.46 Thus, there is conflicting data on the role of infectious agents in mediating the chronic inflammation that is associated with both MF and atherosclerosis.

Risk factors for CMV infection reported in the literature include age, number of children, and immune status. Patients with HIV or allogeneic stem cell transplants (ie, T-cell deficiencies) are highly susceptible to CMV infection and may succumb to it.8 In patients with CTCL, there was no evidence for active CMV infection. In addition, most of the patients were not immunosuppressed and would be expected to have a rate of seropositivity equal to age-matched healthy controls. However, CMV IgG seropositivity is highly associated with MF and SS even in the earliest stages of the disease, when compared with healthy control subjects as well as various historical groups reported in the literature that are listed in Table 4. One variable is that the median age of our patient population was higher than all control groups, including bone marrow donors and patients with HIV and CAD. However, when we considered the CTCL patients younger than age 55 separately, 93% (30 of 32) were CMV seropositive and also were significantly higher than age-matched controls. Of the 28 patients with stage IA MF and only minor skin involvement, 27 were positive for CMV, making it unlikely that the cutaneous disease predisposed these patients to serologic conversion.

Of the 3 seronegative patients, 2 presented before age 25. One was a 21-year-old black woman with stage IVB MF whose course was characterized by persistent Staphylococcus aureus and mucocutaneous candidiasis infections. Her CTCL has been refractory to multiple treatment modalities. Another seronegative female patient presented with stage IIB rapidly progressive disease and succumbed at age 23 during bone marrow transplantation from staphylococcal sepsis. The third seronegative patient was a 59-year-old man with stage IA disease who had long-term remission with conservative topical therapy.

When compared with historical control groups reported in the literature (Table 4), patients with CTCL exceed seropositivity rates of virtually all groups reported. We were intrigued to find that CMV seropositivity was higher than in high-risk groups such as HIV⁺ men (86%),23 health care workers caring for children (44.5%), immunocompromised patients (43.6%),18 and patients with coronary artery disease (54.5%).24 Even patients with early-stage MF with normal immune systems and minimal skin involvement have significantly higher CMV seropositivity rates than do control subjects, and they are similar to late-stage patients who tend to become immunocompromised with disease state and therapy. The significance of this observation is not clear. Latent CMV and/or EBV infection could play a potential role in the susceptibility to or pathogenesis of CTCL. They could provide chronic antigen stimulation, induce T-cell proliferation, and adversely affect apoptosis of skin-homing memory/helper T-cells.