To the Editor:
In a recent issue of BLOOD Said et al1 reported the presence of human herpesvirus 8 (HHV8) DNA in bone marrow stromal dendritic cells of patients with multiple myeloma and acquired immunodeficiency syndrome patients with plasmocytosis. Using polymerase chain reaction (PCR) and in situ hybridization, the authors corroborate their previous report indicating the presence of HHV8 in stromal cell cultures of patients with multiple myeloma and monoclonal gammopathy of unknown significance (MGUS).2
We recently conducted a similar study analyzing bone marrow biopsies of patients with multiple myeloma using a highly sensitive nested PCR assay with oligonucleotide primers derived from the open reading frame 26.3 HHV8 DNA was detected in only 1 (5.9%) of 17 biopsies tested, although the presence of appropriate DNA was confirmed by amplifying the human β-globin control gene in all samples (Table 1). We further analyzed sera of patients with multiple myeloma and MGUS using an enzyme-linked immunosorbent assay (ELISA) based on the recombinant HHV8 protein 65.2 (kindly provided by T.F. Schulz, Liverpool).4 Anti-HHV8 antibodies were present in 15 (88.2%) of 17 of patients with Kaposi's sarcoma but was absent in 19 sera of patients with multiple myeloma and detected in only 1 (5.6%) of 18 patients with MGUS (Table 1).
Our results are in contrast to the data reported by Said et al. We have previously shown that the PCR assay used in this study can detect HHV8 DNA in all samples of formalin-fixed, paraffin-embedded Kaposi's sarcoma tissue.3 However, in our myeloma samples, a sampling error cannot be completely excluded because the presence of HHV8 has been reported to be restricted to the stromal cells but absent in tumor cells.2 The absence of antibodies to HHV8 in all patients with multiple myeloma and in most patients with MGUS is, however, surprising. Indeed, a “serological footprint” of a previous HHV8 infection was found in 80% to 90% of patients with Kaposi's sarcoma and a similar “footprint” would be expected in multiple myeloma patients. Others have also reported a lack of anti-HHV8 antibodies in a significant proportion of patients with multiple myeloma.5 6 Epidemiological differences may result in these discrepancies. However, because only a single HHV8 antigen has been used in our ELISA, one can speculate that a different mode of infection and pathogenesis may be involved in the development of Kaposi's sarcoma and multiple myeloma, associated with a distinct immune response. Further data are urgently needed to define the role of HHV8 infection in patients with multiple myeloma and MGUS.
Cathomas et al1-1 report detecting human herpesvirus 8 (HHV-8) using a polymerase chain reaction (PCR)-based assay with open reading frame (ORF) 26 primers in only one of 17 bone marrow biopsies obtained from myeloma patients. In addition, there was a lack of serological response to HHV-8 using an enzyme-linked immunosorbent assay technique. These data contrast with our recent report showing HHV-8 in most bone marrow biopsies obtained from myeloma patients using both PCR-based and in situ hybridization techniques.1-2 Recently our findings have been confirmed by Broussett et al1-3 who found this virus in 18 of 20 paraffin-embedded bone marrow biopsies from myeloma patients, whereas bone marrow biopsies from patients with lymphoma (n = 15) or reactive processes (n = 5) did not show viral presence.1-3 Because these patients were from France, epidemilogical differences are unlikely to explain the lack of HHV-8 in the Swiss study.
We have sequenced the KS330233 fragment amplified from myeloma bone marrow DNA and noted significant interpatient variation.1-4 However, samples obtained from several different tissue sources within the same patient (bone marrow biopsy, dendritic cells derived from long-term culture of bone marrow aspirate, and peripheral blood enriched for dendritic cells) showed differences of 0 to 1 base pair. Sequencing of other KSHV ORFs (65 and 72) in myeloma patients also showed differences between patients. These findings make PCR contamination unlikely as an explanation for our results. Interestingly, viral sequences show much more homology between different myeloma patients than when these sequences are compared with HHV-8 from Kaposi's sarcoma. Because the authors used a unique set of primers for initial amplification of the KS330233fragment,1-5 this primer pair may be less specific for viral sequences derived from myeloma patients.
Similar sequence variation may also explain the serological findings. Some of these sequence differences lead to changes in amino acids and frameshifts which will result in alterations of the virally encoded protein products. These findings may explain the lack of a serological response in myeloma patients1-1,1-6,1-7 who may contain variants of HHV encoding viral proteins with different structures. Use of a single HHV-8 antigen may therefore be problematic, as the authors point out. In fact, Alsina et al1-8 have recently reported a serological response in most myeloma patients using antibodies derived against a different viral antigen. Furthermore, lack of detectable antibody may result from the markedly decreased humoral response in these patients. As a result, although a serological response to HHV-8 may exist in myeloma patients, it may be weak and therefore undetectable using these HHV-8 serological assays. Finally, HHV-8 localization within the bone marrow-based dendritic cells may prevent the maturation of an antigenic response due to B-cell tolerance of antigens presented in the bone marrow.1-9
Thus, our recent report1-2 with these additional findings and the reports from Broussett et al1-3 and Alsina et al1-8 make it likely that HHV-8 is frequently present in the bone marrow of myeloma patients. Its possible role in the pathogenesis of this B-cell malignancy remains to be elucidated.
