Presented in part at the 1996 meeting of the United States and Canadian Academy of Pathology, Washington, DC.
Kaposi's sarcoma-associated herpesvirus (KSHV) DNA sequences have been identified in approximately 95% of Kaposi's sarcoma (KS) lesions and primary effusion lymphomas (PELs), suggesting a pathogenetic role for this virus in these lesions. However, KSHV has also been identified in a variety of specimens, including lymph nodes, peripheral blood B cells, semen, and prostate tissue, with varying frequencies. This suggests that KSHV, like Epstein-Barr virus, may be ubiquitously distributed. To evaluate further the clinical spectrum of KSHV infection and define better the prevalence of this virus in lymphoid tissues in the general population, the authors examined a wide spectrum of benign lymphoid proliferations occurring in human immunodeficiency virus (HIV)-negative individuals.
One hundred eight lymphoid lesions were examined for the presence of KSHV by polymerase chain reaction (PCR) amplification using primers to open reading frame (ORF) 26. Positive cases were confirmed by Southern blot hybridization using an internal oligonucleotide probe and by PCR amplification using primers to ORF 74 and ORF 75 of the virus.
Only 4 (4%) of 108 specimens were KSHV positive. Three positive lymph node specimens were taken from patients with multicentric Castleman's disease (3 of 11 total cases of Castleman's disease; 3 of 5 total cases of multicentric Castleman's disease). The remaining case was a lymph node showing paracortical hyperplasia, taken from a patient with systemic lupus erythematosus.
Sequences of a new herpesvirus were recently isolated from a Kaposi's sarcoma lesion in an HIV-positive individual by Chang et al.1 This new virus, which is known as Kaposi's sarcoma-associated herpesvirus (KSHV)1 and has also been referred to as human herpesvirus-8 (HHV8),2 shows sequence homology to both Epstein-Barr virus (EBV) and herpesvirus saimiri.1, 3 Since its discovery, KSHV DNA sequences have been identified in more than 95% of HIV-associated, classical, African, and transplant-related Kaposi's sarcomas.1, 4-9 They have also been identified in virtually all cases of a distinct subtype of lymphoma known as primary effusion (or body cavity-based) lymphoma,3, 10-12 which has an indeterminant phenotype, B-cell genotype, and grows mainly as an effusion. The identification of KSHVsequences in these lesions implies that KSHV is etiologically associated with their development. However, KSHV DNA sequences have been also identified by PCR amplification in a high percentage of cases of HIV-related and HIV-unrelated multicentric Castleman's disease13-16 and with varying frequencies in lymph nodes exhibiting nonspecific hyperplasia from both HIV-positive and HIV-negative individuals,1, 15-19 semen and prostate tissue of HIV-negative individuals,19 and peripheral blood B cells from HIV-positive individuals with and without Kaposi's sarcoma.9, 13, 18-22 These findings raise the possibility that, like EBV, KSHV may be distributed ubiquitously. In addition, while KSHV DNA sequences have been sought in a variety of clinical settings (particularly in patients with HIV infection, patients with Kaposi's sarcoma of all types, and some patients with multicentric Castleman's disease), the prevalence of KSHV in the HIV-negative population has only been evaluated in a limited number of reactive lymph nodes.15-18 To evaluate further the clinical spectrum of KSHV infection and confirm its association with some disease entities, we examined a wide spectrum of benign lymphoid proliferations occurring in HIV-negative individuals, using polymerase chain reaction (PCR) technology.
MATERIALS AND METHODS
One hundred eight cases were selected from among those routinely processed at or submitted in consultation to The New York Hospital-Cornell Medical Center, Columbia-Presbyterian Medical Center, and New York University Medical Center. In all cases, the histologic diagnosis was a nonmalignant lymphoid proliferation. Clinical information was obtained from the patients' medical records and/or the attending physicians. The patients included 45 males, 62 females, and 1 of unknown gender. They ranged in age from 4 to 84 years; the median and mean age was 40 years. The patients were HIV-seronegative and/or did not exhibit any risk factors for HIV infection. Of the 108 patients, 7 (5 men and 2 women, age 20-50 years) had been tested for HIV and were found to be negative. Nine cases were diagnosed prior to March 1985 when HIV testing first became available; thus, these patients were not tested for HIV. Of the remaining patients, 18 were younger than 20 years and 28 were older than 50 years; therefore, these patients were unlikely to be infected with HIV.
The 108 benign lymphoid proliferations investigated for the presence of KSHV were obtained from a variety of anatomic sites, including lymph nodes (70), skin (10), parotid gland (6), ocular adnexa (5), tonsils or adenoids (4), spleen (4), lung (4), breast (1), thyroid (1), cervix uteri (1), epigastric tissue (1), and soft tissue (1). Because of the propensity of Kaposi's sarcoma to involve the skin, a relatively large number of skin lesions, accounting for 9% of the specimens, were included in this study.
A wide spectrum of pathologic entities was examined for the presence of KSHV (Table 1). This included lesions classified as lymphoid hyperplasia (36), extranodal pseudolymphoma (17), localized Castleman's disease (6), multicentric Castleman's disease (5), granulomatous inflammation (10, including lymphogranuloma venereum and cat-scratch disease), necrotizing lymphadenitis (9, 7 of which exhibited the histologic features of Kikuchi's disease), angioimmunoblastic lymphadenopathy (7), inflammatory pseudotumor (5), viral lymphadenitis (4, including those with infectious mononucleosis), toxoplasmic lymphadenitis (4), dermatopathic lymphadenitis (3), and sinus histiocytosis with massive lymphadenopathy (2). The cases categorized as lymphoid hyperplasia were subclassified as follicular (14), atypical (11), paracortical (2), progressive transformation of germinal centers (1), and nonspecific (8).
Table 1. Benign Lymphoid Proliferations Examined for the Presence of Kaposi's Sarcoma-Associated Herpesvirus
PTGC: progressive transformation of germinal centers.
Sinus histiocytosis with massive lymphadenopathy
Genomic DNA was extracted from frozen tissue blocks or from cryopreserved cells by digestion with proteinase K, which was followed by extraction with phenol-chloroform and precipitation with ethanol.23 Alternatively, genomic DNA was obtained in some cases by a salting-out procedure, which does not require organic extraction.24 In cases where only formalin fixed, paraffin embedded tissue was available (11 cases), DNA was obtained by extraction of 3 5-mm paraffin tissue sections, as previously described.25 Briefly, the tissue sections were cut and placed in microfuge tubes and deparaffinized with xylene and 100% ethanol; this was followed by digestion with proteinase K (200 mg/mL) in a digestion buffer containing 50mM Tris pH 5.0, 1mM ethylenediamine tetraacetic acid, and 0.5% Tween 20 at 37 °C overnight. The proteinase K was then inactivated by heating the samples at 95 °C for 8 minutes. The quality of the paraffin extracted DNA from the 11 cases for PCR amplification was evaluated by amplifying p53 gene exon 6, 7, or 8. The 11 cases included lesions classified as Castleman's disease (4), Kikuchi's disease (2), atypical lymphoid hyperplasia (1), cat-scratch disease (1), inflammatory pseudotumor (1), paracortical hyperplasia (1), and progressive transformation of germinal centers (1).
Oligonucleotide Primers and Probes
All the oligonucleotides used for PCR amplification and hybridization were synthesized by the solid-phase triester method. The sequences of the oligonucleotides used for PCR amplification and for hybridization of open reading frame (ORF) 26, ORF 75 (membrane antigen), and ORF 74 (G-protein coupled receptor gene) have been previously published.1, 26 In cases that were KSHV positive, the lymphoid lesions were also examined for the presence and subtype of EBV by PCR. The primers used in these cases included sets for EBNA 2, EBNA 3C, and EBER regions were derived from published sequences.27 In addition, where sufficient DNA was available, KSHV-positive cases were also examined for the presence of HIV by PCR. The primers used were derived from previously published sequences.28
PCR, Hybridization, and Sequencing
The conditions for the PCR reaction of ORF 26 were as previously described.1 Identical conditions were used for PCR studies using primers to ORF 75 and ORF 74. Hybridization studies using an internal oligonucleotide probe to either the ORF 26 or the ORF 74 amplification products were performed as previously described.1 Where sufficient DNA was available, DNA sequencing of positive cases was performed. In these cases, the ORF 26 PCR products were subjected to direct sequence analysis with an AB1373A automated DNA sequencer (Applied Biosystems, Inc., Foster City, CA). PCR amplifications for the presence and subtype of EBV and for the presence of HIV were performed as previously described.28, 29
PCR Analysis for the Presence of KSHV in the Benign Lymphoid Proliferations
All 108 lesions were examined for the presence of KSHV by PCR using primers to ORF 26 of the virus. Four of the 108 benign lymphoid proliferations (4%) were found to contain KSHV sequences. Three of the positive lymphoid lesions were cases of multicentric Castleman's disease (Cases CD1, CD2, and CD11; Fig. 1 (144K)); the remaining case was a lymph node exhibiting paracortical hyperplasia (Case PH1) with changes suggestive of viral lymphadenitis (Figs. 2 (112K) and 3 (9K)). These findings were confirmed by hybridization of the ORF 26 PCR products using an internal oligonucleotide probe. In addition, PCR analysis using primers to ORF 74 and ORF 75 of KSHV was also performed in the positive cases, to confirm the results obtained using the ORF 26 primers and exclude the possibility of PCR contamination. PCR amplification using primers to ORF 75 of KSHV resulted in an amplification product in all four ORF 26-positive cases, i.e., Cases CD1, CD2, CD11 (Fig. 4 (6K)), and PH1. In addition, where sufficient DNA was available, PCR amplification using primers to ORF 74 in the three cases of multicentric Castleman's disease resulted in an amplification product; PCR amplification of Case PH1 using these primers could not be performed due to insufficient DNA. Furthermore, because of the high incidence of KSHV in previously published reports of multicentric Castleman's disease in HIV-negative individuals, all 11 cases of Castleman's disease were also examined with primers to both ORF 74 and ORF 75. No additional KSHV-positive cases of Castleman's disease were identified with the additional primer sets (ORF 75; Fig. 4 (6K)). Furthermore, to confirm the integrity of the results of the ORF 26-negative cases, random cases (each extracted from frozen and paraffin embedded tissue) were also examined with an additional set of primers. No PCR amplification product was identified in any of these cases, confirming the previously obtained negative results.
PCR Sequence Analysis of KSHV in the Benign Lymphoid Proliferations
In 3 of the 4 KSHV-positive cases (Cases CD1, CD2, and CD11) for which sufficient DNA was available, direct sequencing using primers to ORF 26 was performed. Approximately 200 base pairs from each case were sequenced. A high degree of conservation (≥99%) was found in all 3 cases, and only rare base-pair discrepancies (2 cases with 1 discrepancy and 1 case with 2 discrepancies) were identified, compared with published sequence analysis of ORF 26 from the prototypic case of primary effusion lymphoma, from which the BC-1 cell line was derived.1
PCR Analysis for EBV in the KSHV-Positive Cases
Because KSHV and EBV are both present in most primary effusion lymphomas and because EBV has been identified in cases of multicentric Castleman's disease, 9 of the 11 cases of Castleman's disease (including all 3 KSHV-positive cases) and the 1 KSHV-positive paracortical hyperplasia were examined for the presence and subtype of EBV. No evidence of EBV was identified by PCR amplification in any of these cases using primers to the EBNA 2, EBNA 3C, or EBER regions (data not shown).
PCR Analysis for HIV in the KSHV-Positive Cases
In 3 of the 4 KSHV-positive cases (Cases CD1, CD2, and CD11), sufficient DNA was available for PCR analysis for the presence of HIV. In the 3 cases tested, HIV was not identified by PCR amplification (data not shown).
Correlation of Clinical Findings and KSHV Positivity
Of the 4 KSHV-positive cases, 3 were from patients with Castleman's disease; however, only patients with multicentric Castleman's disease were KSHV positive. No cases of localized Castleman's disease were KSHV positive. The patients with KSHV-positive lesions were older on average than those patients with KSHV-negative lesions (Table 2). All the patients with KSHV-positive lesions were older than 65 years (age 69, 79, and 83 years), whereas patients with KSHV-negative lesions ranged in age from 11 to 83 years, including 4 younger than 40 years. In addition, it appeared that KSHV in Castleman's disease occurred more frequently in men; 2 of the 4 men (50%) had KSHV-positive lesions, compared with only 1 of 7 women (14%). Furthermore, only 1 of the 8 patients (Case CD4) with KSHV-negative Castleman's disease had aggressive disease. This patient, a male age 39 years, had splenomegaly, diffuse lymphadenopathy, effusions, and systemic symptoms, and required multiagent chemotherapy for resolution of his disease. However, this patient was alive without evidence of Castleman's disease 4 years after initial diagnosis. A second patient who appeared to have had multicentric Castleman's disease was also KSHV negative. This patient, a female age 67 years, presented with megaloblastic anemia, neutropenia, hypothryoidism, diffuse lymphadenopathy, and monoclonal gammopathy. After surgical excision of a mediastinal mass that showed the histologic features of hyaline-vascular Castleman's disease, she died of complications of severe cardiac disease within 1 year. In contrast, all three patients with KSHV-positive multicentric Castleman's disease had widespread aggressive disease. These patients had diffuse lymphadenopathy, splenomegaly (2), renal failure (1), pleural effusion (1), and systemic symptoms (3); in addition, 1 (Case CD1) had Kaposi's sarcoma. The patients for whom clinical follow-up data was available both died of disease within 2 weeks of diagnosis.
Table 2. Comparison of the Clinical Features of KSHV Negative and Positive Cases of Castleman's Disease
No. of patients
Mean age (yrs)
Aggressive clinical course
KSHV: Kaposi's sarcoma-associated herpesvirus.
The 1 patient with aggressive clinical disease was alive and disease free 4 years after diagnosis and treatment.
The one remaining KSHV-positive patient, who had a lymph node that showed paracortical hyperplasia suggestive of a viral etiology, was a female age 27 years with systemic lupus erythematosus. Unfortunately, clinical follow-up of this patient was not available.
In this study we examined a wide variety of benign lymphoid lesions occurring in HIV-negative individuals, using PCR technology to determine the prevalence of KSHV in these tissues. We found KSHV in only 4 of 108 cases (4%). This suggests that KSHV does not occur in a wide range of lymphoid proliferations outside of HIV infection. Thus, these results further support the contention that KSHV is preferentially associated with Kaposi's sarcoma and primary effusion lymphoma.
Three of the four lymphoid lesions positive for KSHV in this study occurred in patients with multicentric Castleman's disease. KSHV has been previously found to be present in cases of Castleman's disease occurring in both HIV-positive and HIV-negative patients.14-16 However, in those studies, as in this one, only the cases that were described as multicentric Castleman's disease were KSHV positive14-17; no case described as localized Castleman's disease was KSHV positive.15 Although the incidence of KSHV positivity is greater than 90% in multicentric Castleman's disease occurring in HIV-positive patients,13-16 the reported incidence is much less in HIV-negative individuals, varying from 0% to 41%15-17--somewhat less than our incidence of 60% (3 of 5 HIV-negative patients with multicentric Castleman's disease).
The clinical course of HIV-negative patients with KSHV-positive multicentric Castleman's disease has not been well described. Soulier et al.16 analyzed 17 HIV-negative individuals with multicentric Castleman's disease for the presence of KSHV and found 7 positive cases (41%). Four of these seven patients had an associated disease, such as Kaposi's sarcoma, B-cell non-Hodgkin's lymphoma, or POEMS syndrome. However, 5 of the 10 KSHV-negative, HIV-negative patients had an associated disease as well, suggesting that KSHV positivity may not affect the clinical course of the patient with multicentric Castleman's disease. In this study, we found that the patients with KSHV-positive multicentric Castleman's disease had an aggressive clinical course; of the patients died of their disease within 2 weeks of diagnosis. Clinical follow-up was not available in the third case, but this patient also had Kaposi's sarcoma as an associated disease, which may or may not account for the KSHV positivity of his multicentric Castleman's disease lesions. However, two patients with multicentric Castleman's disease who were KSHV negative also had relatively aggressive disease. One patient, who subsequently died of cardiac disease, presented with systemic symptoms and clinical evidence of immune dysfunction (such as megaloblastic anemia, hypothyroidism, and monoclonal gammopathy) characteristic of other patients with multicentric Castleman's disease. The other patient, a male age 39 years, presented with clinically aggressive multicentric Castleman's disease and required chemotherapy for remission of his symptoms. This patient, however, was alive and disease free 4 years later. Thus, whether KSHV positivity in patients with multicentric Castleman's disease significantly influences their clinical course has yet to be determined.
The presence of KSHV in multicentric Castleman's disease is noteworthy with respect to the association of KSHV with immune dysfunction, including the relatively high rate of KSHV infection in the HIV-positive patient population.18, 20, 21, 30 Patients with multicentric Castleman's disease are often older and are therefore thought to be somewhat immunodeficient because of their age.31-34 Furthermore, in addition to diffuse lymphadenopathy and frequent splenomegaly, the patients often exhibit clinical findings associated with dysfunctional immune systems, such as autoimmune-related cytopenias, hypergammaglobulinemia, and elevated erythrocyte sedimentation rates.31-34 Histologically, the lymph nodes from many of the patients with multicentric Castleman's disease exhibit histologic findings similar to those of lymph nodes from patients with other immune deficiencies, such as HIV infection or congenital immunodeficiency.35-41 In fact, the follicular involution pattern of HIV-related lymphadenopathy is virtually identical morphologically to that of hyaline-vascular Castleman's disease or the "burned-out" phase of multicentric Castleman's disease.38-41 Multicentric Castleman's disease is thought to be a possible result of dysregulation of B cell proliferation, as evidenced by abnormal follicles and the accumulation of plasma cells.32-34, 42 This may occur as a result of abnormal cellular immune function. This is supported by studies done in mice as well as by immunophenotypic analyses of multicentric Castleman's disease. The mantle zone lymphocytes in multicentric Castleman's disease are CD5 positive and KiB3 negative,42, 43 a phenotype similar to that of Ly-1 sister B cells in mice.44 Antibody production and cell proliferation of these mantle cells, which are thought to be the source of natural autoantibodies, is thought to be more tightly controlled by T cells than other B-cell populations.44-46 Thus, it can be hypothesized that abnormal cellular immune function, as evidenced by abnormal numbers of CD4 and CD8 cells within primary follicles and mantle cell zones of secondary follicles,47 could result in increased autoantibody production, leading to at least some of the clinicopathologic findings of multicentric Castleman's disease. In addition, it has been found that the germinal centers in Castleman's disease have increased levels of IL-6, a cytokine that induces B-cell proliferation and differentiation to plasma cells, which could result in the morphologic features of Castleman's disease.48, 49
Angioimmunoblastic lymphadenopathy (AILD) is another rare, benign, lymphoproliferative disorder associated with immune dysfunction.32 Seven cases of AILD were examined for the presence of KSHV in this study. All 7 cases were found to be negative by PCR analysis. This is in contrast to the study by Luppi et al., in which the authors identified KSHV sequences in 3 of 15 cases (20%).17 These findings, including the discrepancy between our two studies, is not unexpected. Some investigators believe that some of the morphologic features of lymph nodes classified as AILD are similar to those of both multicentric Castleman's disease and HIV-related follicular involution, specifically the presence of increased vascularity and a polymorphic cell population, which they feel suggests a similar underlying pathogenetic mechanism.32 Furthermore, not all cases of either multicentric Castleman's disease or HIV-related follicular involution are KSHV positive (personal observation).14-18 However, clinical follow-up with respect to outcome of the KSHV-positive AILD cases would be noteworthy, in particular to determine whether the KSHV-positive patients had a more aggressive course than those with KSHV-negative lesions. Luppi et al. suggested that the KSHV sequences may merely represent passenger viruses as a result of either reactivation or primary infection induced by the immunosuppression associated with angioimmunoblastic lymphadenopathy.17
The remaining KSHV-positive case identified in this study was a woman age 26 years with systemic lupus erythematosus. Unfortunately, no clinical follow-up was available; however, one may theorize that because she was somewhat immunodeficient secondary to lupus and/or therapy for lupus, she was more susceptible to KSHV infection. Moreover, the histologic features of this case were distinctly different from multicentric Castleman's disease, AILD, and HIV-related lymphadenopathy. Therefore, this case, which histologically showed the features of a virally stimulated lymph node (i.e., paracortical hyperplasia with prominent immunoblasts), may represent, as suggested by Bigoni et al., a privileged site of HHV-8 (KSHV) latency in an individual not infected with HIV.18
KSHV has been previously identified in approximately 7% of reactive lymph nodes from HIV-negative individuals who reportedly do not have an underlying immunodeficiency state.1, 16-18 In five of the reported cases, the lymph nodes showed features of follicular hyperplasia.16, 17 In these patients, the lymph node enlargement appeared to be localized, and in only one case was there evidence of immune dysfunction, as demonstrated by a subsequent diagnosis of idiopathic thrombocytopenia purpura.17 The long term clinical impact of KSHV positivity on these patients has yet to be determined. However, studies have suggested that immunocompetent patients infected with KSHV have a lower viral load than patients who are immunosuppressed.18 Furthermore, serologic studies of KSHV antibody titers suggest that the virus may be under tight immunologic control in immunocompetent patients.30 In addition, serologic studies of HIV-negative individuals without KS from the United States, Italy, and Uganda showed seropositivity rates of 0%, 4% and 50%, respectively, suggesting that KSHV infection may be more prevalent than previously thought, at least in some parts of the world.30 Therefore, although KSHV infection appears to be rare in the general population of the United States, as shown by serologic assays and by the results of this study, the possibility that KSHV cannot be identified in some patients because of the presence of very small copy numbers that cannot be detected with current technology cannot be entirely ruled out. This possibility was strongly suggested by a study by Lennette et al., in which a three-step serologic assay, rather than the standard two-step immunofluorescence serologic test, was used to detect the presence of HHV-8 antibodies.50 In this study, the authors found that approximately 25% of the HIV-negative healthy adults tested were HHV-8 seropositive. However, the mean HHV-8 antibody titers for healthy adults were approximately 15% and 7% of that of HIV-positive patients with KS or generalized lymphadenopathy, respectively.50 These findings raise the possibility that KSHV infection is relatively frequent in the general population but is only present in copy numbers sufficient for detection using DNA-based techniques in patients with abnormal immune systems. It is in this setting of relative immune deficiency that KSHV may act as a opportunitistic infectious organism, resulting in lymph node changes and systemic symptoms in some individuals (as observed in cases of multicentric Castleman's disease) and resulting in the development of either KS or PEL in others. Therefore, these findings strengthen the contention that KSHV is preferentially associated with a specific subset of diseases, namely KS, PEL, and multicentric Castleman's disease.