- Top of page
- Patients and methods
- Conflicts of interest
Classical Hodgkin lymphoma (HL) exhibits a bi-modal age distribution that suggests an infectious aetiology. However, most cases of nodular sclerosis HL (NSHL) are Epstein–Barr virus (EBV) negative (60–90%). Previous studies regarding human herpesvirus 6 (HHV-6) positivity of HL have led to conflicting results. In order to clarify this situation, we examined NSHL biopsies for the presence and distribution of HHV-6 by immunohistochemistry (IHC), polymerase chain reaction (PCR), and fluorescence in situ hybridization (FISH). PCR identified HHV-6 DNA in 86% of NSHL cases. As HHV-6 DNA was also identified in most cases of reactive lymphoid hyperplasia, we sought to localize the virus to specific cells by IHC, which detected HHV-6 in Reed–Sternberg (RS) cells of nearly half (48%) of NSHL cases. Dual CD30/HHV-6 immunostaining confirmed HHV-6 immunoreactivity in CD30+ RS cells, and HHV-6 PCR positivity was confirmed in laser capture microdissection-isolated CD30+ RS cells. FISH demonstrated multiple copies of HHV-6 genome in scattered cells. In contrast, EBV+ RS cells were identified in only 24% of the cases. HHV-6+ cases trended toward a younger age than EBV+ cases. These results conclusively demonstrate that RS cells in many cases of NSHL are HHV-6 positive, and suggest that HHV-6 may play a role in NSHL pathogenesis, particularly in younger patients with EBV-negative disease.
Classical Hodgkin lymphoma (cHL) is a malignant neoplasm composed of scattered malignant Reed–Sternberg (RS) cells admixed within a reactive inflammatory infiltrate. The RS cells of cHL are hyperdiploid germinal centre-derived B cells that carry numerous immunoglobulin gene mutations and cytogenetic defects (Bräuninger et al, 2006). While the aetiology of cHL remains unknown, the peculiar bi-modal age distribution of the disease, with a young adult peak (15–34 years) and a second peak in the elderly (>55 years), suggests that some cases, particularly those in young adults, may be due to an infectious agent (Grufferman & Delzell, 1984). The discovery that a variable proportion of cases of cHL contain clonal Epstein–Barr virus (EBV) infected RS cells partially confirmed this suggestion (Weiss et al, 1989). However, it is the nodular sclerosis subtype of HL (NSHL), the most common subtype of cHL, which is least often EBV-associated, with EBV positivity limited to 10–40% of cases (Hummel et al, 1992).
While attempts to identify alternate infectious agents in EBV-negative cHL have largely been unsuccessful, there are reports of an association between human herpesvirus 6 (HHV-6) and cHL (Marasca et al, 1990; Torelli et al, 1991; Maeda et al, 1993; Krueger et al, 1994, 2001; Rojo et al, 1994; Trovato et al, 1994; Valente et al, 1996; Luppi et al, 1998; Schmidt et al, 2000; Lacroix et al, 2007). To further clarify the relationship of HHV-6 infection with NSHL, we examined lymphoid tissue from a randomly selected cohort of 21 NSHL cases for the presence of HHV-6 by immunohistochemistry (IHC), polymerase chain reaction (PCR, followed by sequencing and Southern blotting), and fluorescence in situ hybridization (FISH) in order to (i) identify virus by three modalities, (ii) localize virus to specific cell types, and (iii) compare the cellular distribution of HHV-6 with EBV in HL tissue.
HHV-6 infection is very common in the first 2 years of life, often presenting as roseola infantum, an acute febrile illness accompanied by diffuse rash (Yamanishi et al, 1988). In adults, primary infection often presents as heterophile-negative mononucleosis. By adulthood, more than 90% of adults are seropositive (Saxinger et al, 1988; Levy et al, 1990). Like all herpesvirus infections, primary HHV-6 infection is usually followed by lifelong asymptomatic latent persistence. Under conditions of stress or immune compromise, symptomatic recurrence due to lytic viral reactivation can occur. In immunocompromised adults, HHV-6 reactivation may be associated with fever, pneumonitis, hepatitis, neurological dysfunction, and marrow failure (De Bolle et al, 2005). Under normal circumstances, HHV-6 DNA can be detected in situ in a number of organs and tissues, including lymphoid tissue (Chen & Hudnall, 2006). In these sites the virus is capable of infecting a broad range of cell types via the ubiquitous CD46 receptor (Isegawa et al, 1999). HHV-6 is a beta-herpesvirus with close sequence similarity to HHV-7 and cytomegalovirus (Berneman et al, 1992). Two HHV-6 variants, HHV-6A and HHV-6B, differ both in genetic sequence (90% similarity) and biological features (Ablashi et al, 1991; Schirmer et al, 1991; Dominguez et al, 1999; Santoro et al, 1999). While the majority of clinical infections in immune competent hosts are apparently due to the HHV-6B variant, disease in immunocompromised patients and patients with neurological presentations are sometimes due to the HHV-6A variant (Hall et al, 1998).
- Top of page
- Patients and methods
- Conflicts of interest
The bimodal age distribution of cHL has been offered as evidence that the disease may be related to an infectious agent. It is now clear that approximately 40% of all cases of cHL are associated with infection by EBV, a human herpesvirus that is also associated with a number of other tumours, including endemic Burkitt lymphoma and nasopharyngeal carcinoma. EBV has been shown to be an oncogenic virus, capable of transforming B-lymphocytes in vitro due to the concerted actions of latent genes EBNA 1-6, LMP1, and EBER1-2. In EBV-positive cases of cHL the virus can be demonstrated in the malignant RS (RS) cells by LMP-1 immunostaining and EBER ISH. Despite the strong association of cHL with EBV, a relatively large number of cHL cases are EBV-negative. In particular, most cases of the most common subtype of cHL, the nodular sclerosis subtype (NSHL), are EBV-negative. Some previous studies have identified an association of HHV-6 with cHL using a variety of techniques, including IHC and PCR.
Increased HHV-6 antibody titres have been described in patients with HL (Levine et al, 1992; Alexander et al, 1995). In addition, several studies have reported HHV-6 positivity in HL tissue by PCR, IHC, or ISH (Marasca et al, 1990; Torelli et al, 1991; Maeda et al, 1993; Krueger et al, 1994, 2001; Rojo et al, 1994; Trovato et al, 1994; Valente et al, 1996; Luppi et al, 1998; Schmidt et al, 2000; Lacroix et al, 2007). In these studies, HHV-6 PCR positivity ranged widely, from 5% to 79% (Torelli et al, 1991; Trovato et al, 1994; Schmidt et al, 2000; Lacroix et al, 2007). In contrast, in two other studies, HHV-6 was not detected in HL tissues by PCR or Southern blot (Armstrong et al, 1998; Shiramizu et al, 2001). While the positive IHC studies consistently described HHV-6 positivity in reactive lymphocytes and histiocytes (Maeda et al, 1993; Krueger et al, 1994, 2001; Rojo et al, 1994; Valente et al, 1996; Luppi et al, 1998), there is no consensus regarding RS cell positivity. In some studies (Rojo et al, 1994; Kashanchi et al, 1997; Luppi et al, 1998), scattered HHV-6 IHC positive RS cells were seen, while in other studies, HHV-6 positive RS cells were not seen (Maeda et al, 1993; Valente et al, 1996).
As outlined above, previous studies did not reach consensus regarding an association of HHV-6 with HL. Some differences in HHV-6 detection in HL may be due to real biological differences in sample cohorts, including ethnic composition, age distribution (paediatric vs. adult), and HL subtype composition (e.g. NSHL vs. other cHL). Alternatively, discrepant PCR results may be due to differences in tissue quality, sample DNA preparation, Taq DNA polymerase source, amount of loaded DNA, primer design, PCR cycling conditions, gel electrophoresis, and choice of controls. Lastly, discrepant results may stem from reliance upon a single technique (PCR or IHC) for detection of HHV-6 in HL tissue. In our experience, all of these factors may contribute to discrepant HHV-6 PCR results, requiring changes in reagents and methods, use of multiple detection techniques, and confirmation by Southern blotting or sequencing of PCR products.
Our current results conclusively demonstrate the presence of HHV-6 in NSHL tissue using a variety of techniques. Positive PCR results for several different HHV-6 genes were corroborated by Southern blot and/or sequencing. In 10 cases with interpretable U86 PCR results, both HHV-6A and HHV-6B isolates were identified. Both HHV-6A and HHV-6 were identified in three cases. In these cases it is not known if the two subtypes are shared by both RS cells and reactive leucocytes or whether the subtypes segregate between reactive and neoplastic cells. In contrast to our findings, Lacroix et al (2007) reported that nearly all their isolates in HL were HHV-6B, and did not report dual positivity. While the genomes of HHV-6A and HHV-6B subtypes overall share 90% sequence identity, they nevertheless differ considerably in the right end of the genome, spanning a region encoding 15 genes (Dominguez et al, 1999). For discrimination between HHV-6A and HHV-6B, Lacroix et al (2007) utilized a PCR assay based upon a subtype-specific sequence difference in the U31 gene (Collot et al, 2002), while we have used an assay based upon a HHV-6A specific deletion in the U86 gene (Isegawa et al, 1999). Unfortunately, the Collot assay (Collot et al, 2002) is unlikely to be useful to better define the HHV-6 subtype of our cases because DNA obtained from FFPE tissue seldom exceeds 300–350 bp in length. Nevertheless, we plan in the future to develop an alternative PCR assay to confirm the HHV-6 subtype in our cases.
HHV-6 specific immunohistochemistry was performed to localize virus to specific cell types. In those cases with HHV-6 positive RS cells by IHC, DNA obtained from micro-dissected RS cells was subjected to HHV-6 PCR to confirm the presence of HHV-6 in malignant RS cells. Findings of scattered positivity for late antigens p98 and p41, and multiple copies of virus detected by FISH are suggestive of viral replication in some RS cells.
While not reaching statistical significance given the small sample size, HHV6+/EBV- cases trended toward younger age than EBV+ cases, results in agreement with those of a previous study (Lacroix et al, 2007). Our findings suggest that HHV-6 may contribute to NSHL pathogenesis, particularly in the younger patient population with EBV-negative disease. A role for HHV-6 in dual EBV/HHV-6 positive cases is also possible because HHV-6 has been shown to induce EBV replication in dual-infected cells in vitro (Flamand et al, 1993). As HHV-6 has not previously been conclusively associated with cancer, very little is known regarding the potential viral oncogenicity. However, in vitro studies indicate that the HHV-6 DR7 gene product is capable of inducing cellular transformation via inhibition of p53 and activation of NFκB, and thus in some settings may act as an oncogene (Kashanchi et al, 1997; Lacroix et al, 2010).
In order to delineate the exact role played by HHV-6 in the pathogenesis of HL, more detailed analysis of the interaction of HHV-6 with RS cells will be required. As has been previously shown for EBV (Weiss et al, 1989), detection of clonal HHV-6 in the RS cells of individual HL cases would provide strong circumstantial evidence that the virus is not merely an innocent bystander, but instead is probably an aetiollogical agent in HL. Assuming that HHV-6 plays an aetiollogical role in HL, it will be important to delineate the contribution of HHV-6 viral genes with transforming ability (such as DR7) to the malignant phenotype of RS cells. It will also be important to analyse the association of HHV-6 infection of RS cells in a larger cohort of HL cases to determine whether HHV-6 positivity is, in fact, more common in young patients with NSHL.