Rebuttal to Tarte, Chang, and Klein

TARTE ET AL REFER to three studies, two letters to the editor and one recent publication from their group, that have failed to show HHV-8 in multiple myeloma (MM) bone marrow stroma cells (BMSCs).1-3 Importantly, the types of BMSCs produced by long-term marrow culture are highly heterogeneous, and the cell populations that were evaluated for the presence of HHV-8 were not fully characterized in these studies. In addition, no positive controls (Kaposi’s sarcoma [KS] patients) were included in the analysis; thus, the sensitivity of the polymerase chain reaction (PCR) assay could not be assessed.1-3 Anderson’s group4has recently confirmed the presence of HHV-8 in MM BMSCs with a similar immunophenotype to our own.5 Although the Tisdale group was able to amplify ORF26 but not ORF72 products in most MM BM samples,6 we and Anderson’s group have easily amplified product in these patients using primers from both of these ORFs in addition to other HHV-8 ORFs. The inability to find HHV-8 in BM aspirates from myeloma patients is consistent with our own experience.5 In contrast, we and others have detected HHV-8 in the BM biopsies from most MM patients.7,8 We have used primer pairs from many HHV-8 ORFs and have found consistent detection in fresh BM biopsies from MM patients but not in normal subjects or in patients with other malignancies. Importantly, the conditions necessary to amplify product are optimized for each of our primer pairs before use. In addition, the integrity of the sample DNA is assured by serially diluting it until only one copy of it is present and then performing PCR amplification with actin primers. The lack of viral interleukin-6 (vIL-6) expression in the BM biopsies3 is consistent with our recent report showing the infrequent expression of this viral homologue in fresh BM biopsies from these patients.9 

Tarte et al also showed lack of HHV-8 in 30 BM aspirates from MM patients collected after a second high-dose therapy procedure.10 This is consistent with our previous results on BM aspirates posttransplant.5 Unfortunately, studies using fresh BM biopsies have not been reported by this group, except in a single case in which the clinical status of the patient is not recorded.3 In addition, the lack of HHV-8 in BM aspirates obtained from the 4 patients who relapsed after high-dose therapy is consistent with our ability to detect HHV-8 in aspirates from less than 10% of similarly treated patients.

It is intriguing that MM patients have weak11 or no detectable antibodies12,13 against HHV-8. This finding may reflect the relative general immunodeficiency that is the hallmark of MM or a specific immune defect that does not allow generation of antibodies against HHV-8. A more interesting possibility is that the type of HHV-8 present in MM patients is different than that present in KS and body cavity lymphoma (BCL) patients. In support of this, we have identified in MM patients consistent changes in the HHV-8 sequence present in regions of the virus that are largely responsible for the immune response.9 These differences may help explain the inability to identify antibodies that were developed using KS tissues. Recently, specific viral strains of another gammaherpesvirus, Epstein-Barr, have been associated with the development of lymphoid malignancies.14 Similarly, HHV-8 derived from KS tissues show distinct genetic differences from BCL-derived viruses, and these different viral isolates show different biological effects.15 Moreover, recent data from KS patients show that certain subtypes of HHV-8 may even predict more aggressive clinical characteristics of the tumor.16 

The investigators suggest that the HHV-8–infected BMSCs in myeloma patients are not actually dendritic cells (DCs), because they have not been demonstrated to provide antigen-presenting cell function. Quite to the contrary, Anderson’s group has shown that these virally infected DCs are, in fact, functional.17 There is also marked heterogeneity in the type of DCs generated in these different culture systems.18 To assume that the DCs analyzed for viral presence by other groups were identical to those shown to contain virus in our and Anderson’s long-term marrow cultures is a leap of faith. The studies from Tarte et al19 and Yi et al20characterized their DCs as CD1a and CD4-expressing cells, in contrast to the HHV-8–containing DCs in our and Anderson’s studies that lacked expression of both of these markers.4,5 In addition, the DCs in Yi et al’s study also lacked CD83, which was found on the HHV-8–containing DCs. Consistent with Tarte et al19 and others21 and the absence of CD34 on the BMSCs infected with HHV-8 from our and Anderson’s groups,4,5 we also have rarely found HHV-8 in CD34-enriched autograft material.22We certainly believe that it may be possible to generate functional DCs from CD34-selected and other autograft material for clinical use based on our results and these other studies as well as the recent report from Raje et al.17 

Thus, our studies support the presence of HHV-8 in the vast majority of bone marrow and peripheral blood samples from MM patients. It also suggests that the viral strain may be unique in these patients and may help explain the weak or lack of a serological response in these individuals. It remains to be determined the part that this virus plays in the development of this B-cell malignancy, but its uniqueness among these patients suggests that it has an important role in the disease pathogenesis.