High Association of Human Herpesvirus 8 in Large-Plaque Parapsoriasis and Mycosis Fungoides FREE

Human herpesvirus 8 (HHV-8), the most recently identified member of the human herpesvirus family, is the etiologic agent of Kaposi sarcoma (KS), primary effusion lymphoma, and multicentric Castleman disease (MCD).^1- 3 In addition, HHV-8 DNA sequences have been found in association with other skin diseases such as angiosarcoma and pemphigus vulgaris, but the role of the virus in these diseases is largely unconfirmed and remains controversial.⁴ Some years ago, several studies provided strong evidence against a role for HHV-8 in the pathogenesis of mycosis fungoides (MF), the most common subtype of cutaneous T-cell lymphoma (CTCL).^5- 6

In the October 2005 issue of the Archives, Trento and colleagues⁷ investigated the prevalence of HHV-8 infection in an Italian population of 22 patients with lymphoproliferative diseases such as large-plaque parapsoriasis (LPP) and MF. Lesional skin samples, plasma, and peripheral blood mononuclear cells were analyzed for the presence of HHV-8. As the predominant result, the authors observed a highly significant association of HHV-8 infection in LPP compared with MF. Consequently, they speculated that HHV-8 could be causally involved in the pathogenesis of LPP and might therefore be a distinct clinical entity separate from MF. The 2 conditions are generally believed to represent different clinical stages of the same disease. This essential observation encouraged us to investigate the presence of HHV-8 in lesional skin of German patients with LPP or MF.

A total of 86 paraffin-embedded lesional skin specimens, including 53 from patients with lymphoproliferative skin diseases, were evaluated in this retrospective study. Patients had LPP (n = 14), MF (n = 31), and lymphomatoid papulosis (LyP) (n = 8). Diagnosis was based on clinical, histopathologic, immunopathologic, and molecular biological characteristics as detailed by the International Society for Cutaneous Lymphoma.⁸ When available, a detailed review of the clinical characteristics and pretreatment status of all patients with LPP and MF was performed. As positive controls, we used 23 tissue samples of classic and AIDS-associated KS, a tumor that is unambiguously associated with HHV-8. As negative controls, samples from 10 patients with atopic dermatitis (AD) were included. All patients had been diagnosed and treated at the Department of Dermatology, Ruhr University Bochum (observation period, January 1, 1999, through December 31, 2006). The protocol of the study was approved by the ethics review board of Ruhr University Bochum.

The presence of HHV-8 DNA was analyzed by amplification of a sequence of open reading frame (ORF) 26 by a nested polymerase chain reaction (PCR) protocol similar to the protocol reported by Trento et al.⁷ Identical PCR primers were used.⁷ To prevent possible cross-contamination between samples during the PCR procedure, each paraffin block was cut using a new disposable microtome blade. For DNA extraction, tissue samples were deparaffinized by adding 1 mL of xylene. After 10 minutes, tissue samples were spun, washed twice with 1 mL of 100% ethanol, and kept overnight in 180 μL of lysis buffer containing proteinase K at 56°C. The DNA was isolated using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany) following the manufacturer's protocol. The PCR products were checked by 2% agarose gel electrophoresis and ethidium bromide staining. The PCR reaction mixture consisted of 400 ng of DNA, PCR buffer containing 15mM magnesium dichloride, 30 pmol/μL of each primer, 200 μM of each deoxyribonucleoside triphosphate, and 2.5 U of HotStar Taq Plus DNA Polymerase (QIAGEN) in a total volume of 100 μL. Amplification was performed in a thermal cycler (TRIO-Thermoblock; Biometra, Göttingen, Germany). The external PCR product was a single-band fragment of 233 base pairs (bp). The nested PCR analysis was performed using 0.5 μL of the first-round product with the inner primer set. The nested PCR product was a fragment of 138 bp. The HHV-8 DNA PCR was independently performed in 2 different laboratories (Bochum and Cologne).

Epstein-Barr virus (EBV) DNA was determined by real-time PCR in a LightCycler (Roche Applied Science, Mannheim, Germany) using hybridization probes as described previously.⁹ For sample preparation, the QIAamp DNA Mini Kit was used, and the forward and backward primers described by Stöcher et al⁹ were slightly elongated 5′ by 2 nucleotides each (forward, 5′-AGGGTGGTTTGGAAAGC-3′; backward, 5′-AACAGACAATGGACTCCCTTAG-3′).

Sections were deparaffinized in xylol and washed with 100% and 96% ethanol as well as 70% and 50% propanol for 5 minutes each and rinsed with demineralized water. After washing with Target Retrieval Solution (DAKO Cytomation, Hamburg, Germany) at a pH of 6.0 for 20 minutes, sections were stored in a water bath at 96°C. Sections were stained with a rat monoclonal antibody against the latency-associated nuclear antigen (LANA) encoded by ORF 73 of HHV-8 (Advanced Biotechnologies Inc, Columbia, Maryland) at a dilution of 1:3000 for 30 minutes at 25°C in a DAKO Autostainer (DAKO Cytomation). For EBV protein analysis, a mouse monoclonal antibody against the latent membrane protein 1 at a dilution of 1:50 (DAKO Cytomation) was used. After washing with Wash Buffer (DAKO) for 2 minutes, incubation with streptavidin alkaline phosphatase was performed for 30 minutes. Chromogen red (DAKO) was used for visualization before counterstaining with hematoxylin and mounting in Mowiol (Roche Molecular Biochemicals, Mannheim, Germany).

Where appropriate, the 2-sided Fisher exact test for tabulated categorical data was used to analyze the differences between the negative control group (AD specimens) and the other diseases investigated with regard to the frequencies of HHV-8 positivity. A P value of .05 or lower was considered statistically significant.

A total of 53 patients with lymphoproliferative skin diseases were included in this retrospective study. An overview of representative patients with available data on age, stage of MF, duration of disease, and previous topical and systemic therapy is provided in Table 1. In all included patients, the duration of disease was at least 1 year (range, 1-30 years), and all of them had previously been treated with topical and/or systemic agents.

Lesional skin of patients with LPP and MF as well as control specimens were analyzed for the presence of HHV-8. A significant association of HHV-8 DNA prevalence was found in LPP as well as in MF. Seven of the 8 tissue samples of LPP (87%) (P = .02) and 7 of the 10 skin specimens of MF (70%) (P = .03) tested positive for the presence of HHV-8 DNA. In all specimens of LPP and MF, the viral sequence was already detectable after the first round of PCR and was finally confirmed by nested PCR (Figure 1A). As expected, all KS tissue specimens tested positive for HHV-8 DNA, and compared with LPP and MF, the single band fragment at 233 bp showed a much stronger positive result in KS indicating a larger amount of viral DNA (Figure 1B). All specimens from patients with AD and LyP tested negative in the first PCR round and in the nested PCR reaction. In contrast to the PCR results, all 14 samples of LPP and all 31 samples of MF tested negative for HHV-8 using the immunohistochemical marker for LANA encoded by ORF 73. A strong positive nuclear staining for HHV-8 ORF 73 was detected in lesional skin of all patients with KS, whereas tissue samples of LyP and AD tested negative. All immunohistochemical and PCR results are detailed in Table 2. Representative immunohistochemical images from HHV-8 analysis are provided in Figure 2.

In addition, EBV analysis was performed in patients with LPP and MF. By using real-time PCR, all tissue specimens analyzed (5 samples of LPP and 10 samples of MF) tested negative for EBV DNA. Similarly, 4 paraffin-embedded samples from patients with LPP and 25 samples of MF tested negative for EBV using the immunohistochemical marker latent membrane protein. Samples from a patient with Burkitt lymphoma, which were used as control specimens, showed strongly positive EVB findings. Representative immunohistochemical images for EBV analysis are depicted in Figure 3.

Human herpesvirus type 8 is a gamma herpesvirus isolated from KS lesions for the first time in 1994.¹ The 140-kilobase genome of HHV-8 can be divided into 4 major subtypes (A-D), and these can be further differentiated into 14 distinctive variants. Human herpesvirus type 8 carries at least 11 ORFs that encode homologues to cellular proteins involved in signal transduction, cell cycle regulation, inhibition of apoptosis, and immune modulation.¹⁰ Thus, HHV-8 has the genetic armamentarium of an oncogenic virus. A sexual mode of transmission has been confirmed for HHV-8, although an obvious nonsexual horizontal mode of transmission exists in endemic regions.¹¹ Seroprevalence of HHV-8 ranges from 100% in some African countries to 15% or 20% in the Mediterranean area and decreases to approximately 2% to 5% in northern Europe and in the United States.¹²

Besides KS, primary effusion lymphoma, and multicentric Castleman disease, several diseases have been reported to be associated with HHV-8 (eg, angiosarcoma and pemphigus vulgaris), but subsequent studies failed to demonstrate the presence of HHV-8 in these conditions.^13- 17 The detection of HHV-8 in a lymphomalike germinotropic disorder¹⁸ and in primary effusion lymphoma with T-cell phenotype¹⁹ led to the assumption that HHV-8 might also be present in CTCL. However, initial investigations failed to demonstrate an association.^5- 6,20 Moreover, HHV-8 was undetectable in MF lesions of 2 patients with concomitant KS, whereas viral DNA was found in the KS tissue of these patients.²¹

In contrast, Trento et al⁷ recently reported the detection of HHV-8 infection in skin lesions and peripheral blood of 100% of patients with LPP and in skin lesions and peripheral blood of 25% of patients with MF. Based on these findings, the authors speculated that LPP might be an entity separate from MF. In the present study, we confirmed the high association of HHV-8 with LPP. However, in contrast to the findings of Trento et al,⁷ most MF tissue samples in our study also harbored low amounts of HHV-8 DNA. The observation of a high association of HHV-8 with LPP might be explained by geographic regions of high seroprevalence of HHV-8, such as Italy or East Africa.¹⁰ However, all included subjects in the present study were of German descent and living in Germany, where the HHV-8 prevalence is considered to be low (<5%).¹² Moreover, although not documented in the patient's files, high-risk sexual behavior and preferences that could explain the high association of HHV-8 seem to be rather unlikely in the investigated cohort. Thus, further epidemiologic studies are required to define the mode of transmission of HHV-8 infection in patients with CTCL.

Importantly, all previous studies that failed to detect HHV-8 in MF specimens used skin samples from patients naïve to any kind of therapy.^5- 6 In the present study, however, all individuals had been previously treated with topical and/or systemic immunosuppressive agents. The high prevalence of positive findings for HHV-8 DNA in our study population might be explained by a new or recent infection or reactivation of HHV-8 in individuals who had undergone extensive treatment and/or who had an advanced stage of the disease, which is associated with altered immunocompetence. Similar observations have been made in patients with common variable immunodeficiency, in which an overproduction of inflammatory cytokines enhances the replication of HHV-8 and promotes tumor progression.²² None of the analyzed specimens in the present study harbored EBV, a B-lymphotropic gamma herpesvirus that is widespread in all human populations and that presents together with HHV-8 as a dual infection in some body cavity lymphomas or AIDS-related, non-Hodgkin lymphomas.²³ The lack of EBV and the high prevalence of HHV-8 DNA in LPP and MF might rather reflect a causal relation than a reactivation or secondary colonization of HHV-8 in CTCL. None of the tissue samples of LyP included in the present study carried HHV-8 DNA, thus confirming the results of previous investigations.²⁴

Interestingly, while our PCR analysis revealed HHV-8 at a high rate in LPP and MF specimens, immunohistochemical findings were negative. Similar discrepancies between DNA detection and protein expression have also been reported by Hammock et al²⁵ and Wheat et al,²² who investigated HHV-8 in benign vascular tumors and in lymphoproliferative disorders such as B-cell lymphomas, respectively. For example, liver specimens from patients with common variable immunodeficiency and granulomatous lymphocytic interstitial lung disease tested positive by nested PCR for HHV-8 DNA but negative for LANA by immunohistochemical staining, possibly reflecting the relative insensitivity of immunohistochemical analysis to detect very low amounts of HHV-8.²² Molecular genetic techniques such as PCR are more sensitive than LANA immunohistochemical analysis.²⁵ By contrast, highly sensitive PCR techniques have also led to a recent controversy. For example, initial studies using these methods suggested a biological association between HHV-8 and multiple myeloma. However, most experts now believe these to be false-positive reactions.²⁶ Nevertheless, the high occurrence of HHV-8 DNA in patients with LPP and MF and the absence of HHV-8 in patients with LyP and AD would not indicate false-positive findings in the present study.

In conclusion, this retrospective study demonstrated the presence of HHV-8 DNA in 2 clinical variants of CTCL (LPP and MF) in patients living in Germany, where the general HHV-8 seroprevalence is considered to be low. Possibly, in the context of systemic or local impairment of immunocompetence secondary to advanced disease or to aggressive and/or long-term disease treatment, HHV-8 acts as a cofactor in the progression of CTCL. Based on the methods of detection used in this study, no conclusion can be drawn on the etiologic and pathogenetic role of HHV-8 in these diseases. Because the virus was detected only by PCR and not by immunohistochemical analysis, we were unable to show whether HHV-8 is located in endothelial cells, keratinocytes, or neoplastic T cells. Given the limitations of this study and the discrepancies in other investigations, the etiologic and pathogenetic role of HHV-8 in CTCL and its precursors requires further research.

Correspondence: Alexander Kreuter, MD, Department of Dermatology, Ruhr University Bochum, Gudrunstrasse 56, D-44791 Bochum, Germany (a.kreuter@derma.de).

Accepted for Publication: August 21, 2007.

Author Contributions:Study concept and design: Kreuter and Gambichler. Acquisition of data: Kreuter, Bischoff, Skrygan, Wieland, and Stücker. Analysis and interpretation of data: Kreuter, Wieland, Brockmeyer, and Altmeyer. Drafting of the manuscript: Kreuter and Gambichler. Critical revision of the manuscript for important intellectual content: Kreuter, Bischoff, Skrygan, Wieland, Brockmeyer, Stücker, and Altmeyer. Statistical analysis: Gambichler. Administrative, technical, and material support: Bischoff, Skrygan, Wieland, and Stücker. Study supervision: Kreuter, Brockmeyer, and Altmeyer.

Financial Disclosure: None reported.