Association of Primary Pleural Effusion Lymphoma of T-Cell Origin and Human Herpesvirus 8 in a Human Immunodeficiency Virus–Seronegative Man

Primary effusion lymphoma (PEL) has been recently identified as a distinct subtype of non-Hodgkin lymphoma associated with infection of the neoplastic lymphoid cells by the Kaposi sarcoma-associated herpesvirus/human herpesvirus 8 (HHV-8).1,2 Primary effusion lymphoma has characteristic clinicopathologic features, including initial presentation as a lymphomatous effusion usually in the absence of a detectable tumor mass, occurs mostly in human immunodeficiency virus (HIV)–positive men, and has a morphologic structure that bridges large cell immunoblastic and anaplastic large cell lymphoma (ALCL).1,2 Neoplastic lymphoid cells are B cells with a peculiar phenotype. They usually lack surface immunoglobulin and B-cell–associated antigens such as CD19 and CD20 and express CD45, CD30, and antigens associated with late stages of B-cell differentiation such as CD138.1,2 Genotypic analysis of PEL has revealed clonal immunoglobulin gene rearrangements in all cases.1,2 We describe herein a unique association of a primary effusion large T-cell lymphoma with HHV-8 infection occurring in an HIV-negative patient.

An 87-year-old man, native of Algeria, was admitted for dyspnea and prominent right-sided pleural effusion. Clinical examination and whole body computed tomographic scanning failed to show lymphadenopathy or masses. The results of skin examination were normal. The white blood cell count was 9 × 10⁹/L, with a hemoglobin level of 15 g/dL and a platelet count of 170 × 10⁹/L. The serum lactate dehydrogenase level was within normal range. Results of serologic tests for HIV, Epstein-Barr virus (EBV), human T-lymphotropic virus 1, and human T-lymphotropic virus 2 were negative. The right pleural fluid cell count was 3 × 10⁹/L, with a protein level of 43 g/L and a red blood cell count of 60 × 10⁹/L.

Smear preparations from the pleural effusion were stained with May-Grünwald-Giemsa. Immunophenotyping was performed on acetone-fixed cell smears from the pleural effusion. The alkaline phosphatase anti-alkaline phosphatase method (Dako, Glostrup, Denmark) was used with the following antibodies directed against human cytokeratin KL1 and MNF116: CD45, CD2, CD3, CD5, CD19, CD20, CD79a, CD138 (Syndecan-1), and CD30 antigens; anaplastic lymphoma kinase (ALK1) protein; EBV latent membrane protein 1; and TIA-1. All antibodies were purchased from Dako, except KL1 (Immunotech, Marseille, France), CD138 (clone B-B4; Serotek, Oxford, England), and TIA-1 (Coulter, Marseille, France).

DNA was extracted from pleural effusion samples and peripheral blood lymphocytes. T-cell receptor γ chain gene analysis was performed by polymerase chain reaction (PCR) and analyzed on denaturing gradient gel electrophoresis as previously described.3 For the immunoglobulin gene rearrangement study, 2 PCR analyses were performed. In one PCR analysis, the 5′ sense oligonucleotide matched homologous sequences with the framework 2 region (FR2: GC(C/T) (C/T)CC GG(A/G) AA (A/G)GT CTG GAG TGG); in the second PCR analysis, it matched homologous sequences within the framework 3 region (FR3: ACACGGC(C/T)(G/C)TGTATTAC TGT). In the 2 PCR analyses, the 3′ antisense oligonucleotide hybridized the J_H region: ACCTGAGGAGACGGTGACC. For FR3/J_H analysis, 20% of the amplified products were run on a 10% polyacrylamide gel. For FR2/J_H analysis, 1% of the amplified products were run on an ABI prism 310 genetic Analyzer (Applied Biosystems, Courtaboeuf, France). Results were analyzed using Genescan analysis software (Applied Biosystems).

Detection of HHV-8 DNA was performed in duplicate using a single-tube, nested PCR method for the amplification of a 233-base pair (bp) product of the orf 26 region.4 The HHV-8 PCR products were analyzed by gel electrophoresis and hybridized onto a specific oligonucleotide probe using a DNA enzyme immunoassay (Gen-Eti-K DEIA, Dia Sorin, Italy).5 The EBV DNA was sought by PCR using primers specific for a fragment of 153 bp located in the BamHI WYH region of the EBV genome.

Anti–HHV-8 antibodies were sought using a standardized indirect immunofluorescent assay, according to the manufacturer's protocol (HHV-8 immunoglobulin G [IgG] indirect immunofluorescent assay kit, Biotrin International Ltd, Dublin, Ireland).

The smear preparations showed noncohesive large to very large lymphoid cells with abundant basophilic cytoplasm. They contained large, pleomorphic, round to multilobated or kidney-shaped nuclei that often disclosed prominent nucleoli (Figure 1, A). A few cells showed eccentric nuclei, surrounded by a prominent clear perinuclear Golgi zone (Figure 1, A). Mitotic figures were abundant. Apoptotic neoplastic cells were noted. Tumor cells were negative for cytokeratin (KL1, MNF116). Most neoplastic lymphoid cells were strongly positive for CD30 and CD3 (Figure 1, B) and showed weak staining for CD45, whereas they were negative for CD19, CD20, and CD79a B-cell antigens. The syndecan (CD138) was also absent except on a few mature plasma cells. Tumor cells did not express CD2, CD5, ALK1, or TIA-1. The PCR–denaturing gradient gel electrophoresis analysis of T-cell receptor γ chain gene rearrangement showed the presence of a predominant T-cell clone within an oligoclonal T-cell expansion in the pleural lymphocyte population, whereas no clonal population was detected in the peripheral blood lymphocytes (Figure 2). No clonal rearrangement of IgH chain gene was found in the pleural effusion lymphocytes. The search for EBV infection was negative both by immunocytochemistry and PCR analysis. The HHV-8 DNA sequences were detected by PCR analysis in the pleural effusion cells, whereas no HHV-8 sequences were detected in the peripheral blood lymphocytes (Figure 3). Patient's serum contained anti–HHV-8 antibodies.

Based on clinical, morphologic, and viral features, the present case resembles HHV-8–related BCBL/PEL with the important exception that the neoplastic lymphoid cells display the phenotype and genotype of T cells.1,2 The patient presented with the typical clinical features of PEL (eg, diffuse spreading growth pattern along serous membrane without lymphadenopathies or tumor masses) and the peculiar characteristics of HIV-unrelated PEL (eg, old age and negativity for EBV). The morphologic findings of single large atypical cells with immunoblastic to anaplastic features, as well as positive immunostaining for CD45 and negativity for epithelial cell markers, helped to establish the diagnosis of large cell lymphoma. Neoplastic lymphoid cells were also strongly positive for CD30 activation marker, which is commonly detected in PEL. Serologic test results showed that our patient was infected with HHV-8, which could be related to his geographic origin. Indeed, he originated from North Africa, where the prevalence of antibodies to HHV-8 is relatively high. The HHV-8 DNA sequences were detected in the lymphomatous pleural effusion. Although PCR analysis do not allow the identification of HHV-8–infected cells, the detection of HHV-8 sequences within pleural effusion but not within peripheral blood lymphocytes strongly supports the theory that HHV-8 infection could be related to the lymphoma cell population.

In the present case, phenotypic and genotypic findings disclosed the T-cell origin of lymphoma cells. Indeed, neoplastic cells were negative for CD19, CD20, CD79a, CD138, and B-cell antigens and did not exhibit clonal IgH rearrangement by PCR analysis. By contrast, they strongly expressed CD3, and a clonal rearrangement of T-cell receptor γ chain gene was found in the neoplastic pleural effusion but not in the peripheral blood lymphocytes. Primary effusion lymphoma exhibiting a T-cell phenotype was reported only once, in an HIV-seropositive male patient.6 In the latter case, neoplastic lymphoid cells expressed various T-cell–specific antigens, including CD2, CD3, CD5, and CD7, and no B-cell markers, and both T-cell receptor and immunoglobulin gene rearrangements.6 In our case, several cytologic and phenotypic features could be suggestive of ALCL. However, ALCLs commonly present with systemic disease, and isolated pleural effusion is uncommon. In addition, ALCLs are usually associated with a t(2;5)(p23;q35) translocation that leads to an abnormal expression of ALK, a feature that was not observed in our case.7 Moreover, TIA-1, a cytotoxic marker frequently detected in ALCL, was not present.7 

Whether this case may be included or not within the PEL entity, the association of a pleural T-cell lymphoma with HHV-8 infection raises the question of the possible pathogenic role of HHV-8, not only on B cells but also on T cells. Human herpesvirus 8 has a strong tropism for B cells. However, it has been shown to also infect, although at a lower rate, monocytes/macrophages and T cells from patients with Kaposi sarcoma.8,9 In addition, HHV-8 is related phylogenetically to EBV, a lymphotropic gammaherpes virinae. The oncogenic role of EBV was initially restricted to B-cell lymphomas, but has since been extended to several T-cell and NK cell lymphomas.10 It is noteworthy that HHV-8 sequences have been detected in a few cases of angioimmunoblastic lymphadenopathy, a disease now considered a T-cell lymphoma.9 

In conclusion, we report herein a primary pleural T-cell lymphoma, occurring in an HIV-negative, elderly man. This lymphoma fulfilled the clinical, morphologic, and virologic criteria for PEL except that it was of T-cell origin. This case report raises the question of the possible occurrence of T cells as the target of malignant transformation associated with HHV-8 infection.