Association between simian virus 40 and non-Hodgkin lymphoma

Regis A Vilchez, Charles R Madden, Claudia A Kozinetz, Steven J Halvorson, Zoe S White, Jeffrey L Jorgensen, Chris J Finch, Janet S Butel

 

Departments of Medicine (R A Vilchez MD), Molecular Virology and Microbiology (R A Vilchez, C R Madden PhD, S J Halvorson BS, Z S White BS, J S Butel PhD), Pediatrics (C A Kozinetz PhD), and Pathology (J L Jorgensen MD, C J Finch MD), and Baylor Center for AIDS Research (R A Vilchez, C A Kozinetz, J L Jorgensen, C J Finch, J S Butel), Baylor College of Medicine, Houston, TX, USA

 

Background Non-Hodgkin lymphoma has increased in frequency over the past 30 years, and is a common cancer in HIV-1-infected patients. Although no definite risk factors have emerged, a viral cause has been postulated. Polyomaviruses are known to infect human beings and to induce tumours in laboratory animals. We aimed to identify which one of the three polyomaviruses able to infect human beings (simian virus 40 [SV40], JC virus, and BK virus) was associated with non-Hodgkin lymphoma.

Methods We analysed systemic non-Hodgkin lymphoma from 76 HIV-1-infected and 78 HIV-1-uninfected patients, and non-malignant lymphoid samples from 79 HIV-1-positive and 107 HIV-1-negative patients without tumours; 54 colon and breast carcinoma samples served as cancer controls. We used PCR followed by Southern blot hybridisation and DNA sequence analysis to detect DNAs of polyomaviruses and herpesviruses.

Findings Polyomavirus T antigen sequences, all of which were SV40-specific, were detected in 64 (42%) of 154 non-Hodgkin lymphomas, none of 186 non-malignant lymphoid samples, and none of 54 control cancers. This difference was similar for HIV-1-infected patients and HIV-1-uninfected patients alike. Few tumours were positive for both SV40 and Epstein-Barr virus. Human herpesvirus type 8 was not detected. SV40 sequences were found most frequently in diffuse large B-cell and follicular-type lymphomas.

Interpretation SV40 is significantly associated with some types of non-Hodgkin lymphoma. These results add lymphomas to the types of human cancers associated with SV40.

Lancet 2002; 359: 817-23

 

Introduction

Non-Hodgkin lymphoma comprises a biologically diverse group of haematological malignancies with clinical courses ranging from indolent to highly aggressive. During the past 30 years, the reported incidence and death rate of the disease have increased strikingly, nearly doubling since 1970.¹ About 55 000 new cases of non-Hodgkin lymphoma are estimated to be diagnosed annually in the USA,¹ and deaths related to the disorder are ranked fourth and fifth among all cancer deaths in women and men, respectively. Although the reasons for the increase in incidence are not fully understood, a substantial number of cases of non-Hodgkin lymphoma are linked to the HIV-1 epidemic. Indeed, non-Hodgkin lymphoma is a common malignancy in HIV-1-infected patients and the incidence can be up to 300 times higher than in HIV-1-negative individuals.¹

No obvious risk factors have emerged for non-Hodgkin lymphoma in the general population, but a viral cause has been postulated.² Some cases of non-Hodgkin lymphoma in HIV-1-infected patients have been attributed to deficient immune surveillance of oncogenic herpesviruses, such as Epstein-Barr virus (EBV) and human herpesvirus 8 (HHV-8), or perhaps to chronic antigenic stimulation and defective immune regulation.³ EBV is suspected of having a major role in primary central-nervous-system non-Hodgkin lymphoma in HIV-1-infected patients, since most of those tumours contain EBV DNA, but it is detected less frequently (<40%) in systemic non-Hodgkin lymphoma in HIV-1-infected patients.^1-3 EBV is found even less commonly in non-Hodgkin lymphoma from HIV-1-negative patients. HHV-8 is specifically associated with multicentric Castleman's disease and primary effusion lymphoma, which often occurs in a setting of profound immunosuppression.^2,4

Because EBV and HHV-8 are absent from many cases of non-Hodgkin lymphoma, other viral agents should be considered as possible causes. The small DNA-containing polyomaviruses (simian virus 40 [SV40], JC virus, and BK virus) are known to infect human beings, to have oncogenic potential, and to be associated with some human cancers.^2,5,6 SV40 DNA sequences have been found repeatedly in some brain and bone cancers and mesotheliomas.⁵ Polyomaviruses typically establish subclinical and persisting infections in their natural host, with persistence or latency in several organs, including kidney, brain, and spleen.² Studies have identified SV40, JC virus, and BK virus DNA sequences in B lymphocytes from HIV-1-infected and HIV-1-uninfected patients, suggesting that polyomaviruses are lymphotropic in man.^7-9 Polyomaviruses are known to induce tumour formation in animals, including the production of B-cell lymphomas by SV40.¹⁰ The major types of tumours induced by SV40 in laboratory animals are the same as the human cancers found to contain SV40 DNA, with the exception of lymphomas.⁵ In animals, oncogenesis is mediated by the polyomavirus large tumour (T) antigen.^2,5,6 The large T antigen is a multifunctional protein that stimulates host cells to enter S phase and is required for initiation of viral DNA synthesis. Fundamental to the effects of T antigen on host cells is binding to cellular tumour-suppressor proteins p53 and members of the pRB family.^2,5,6

Studies have reported the detection of SV40 DNA sequences in non-Hodgkin lymphoma from HIV-1-infected and HIV-1-uninfected patients,^9,11,12 and the amplification of JC virus DNA sequences from systemic non-Hodgkin lymphoma of HIV-1-infected children.¹³ These findings suggest a possible role for polyomaviruses in lymphoproliferative disorders, but the small size of the study populations, the lack of screens for other tumour viruses, and the limited confirmation of identity of the viral sequences detected made conclusion of whether polyomaviruses were definitely associated with non-Hodgkin lymphoma difficult. We aimed to determine the frequency of detection of polyomavirus T antigen DNA sequences in non-Hodgkin lymphoma among HIV-1-infected and HIV-1-uninfected patients, to identify which one of the three polyomaviruses able to infect humans (SV40, JC virus, and BK virus) was associated with non-Hodgkin lymphoma in adult patients, and to establish clinical correlations between the presence of viral sequences and non-Hodgkin lymphoma among HIV-1-infected and HIV-1-uninfected patients. The HIV-1-infected population was included in this study because of their high incidence of non-Hodgkin lymphoma and because immunocompromised individuals are known to be at risk of development of virus-mediated neoplasms.²
 

Patients and methods

 

Patients

We studied 28 adult patients with HIV-1 infection and 35 HIV-1-uninfected patients who were diagnosed with systemic non-Hodgkin lymphoma between January, 1996, and August, 2001, at the Harris County Hospital District, the Veterans Administration Medical Center, and the Methodist Hospital, all of which are affiliated with Baylor College of Medicine, Houston, TX, USA. Additionally, the AIDS and Cancer Specimen Bank of the US National Cancer Institute, through collaboration with the Baylor Center for AIDS Research, provided non-Hodgkin lymphoma samples and clinical data from 48 HIV-1-positive and 43 HIV-1-negative adult patients diagnosed between November, 1987, and May, 2000, at different medical centres in the USA. The histological types of non-Hodgkin lymphoma among HIV-1-infected and HIV-1-uninfected patients were categorised according to WHO Classification for Neoplastic Diseases of the Lymphoid Tissues.¹⁴ No lymphomas of the central nervous system were included in this study.

Two types of control sample were analysed. Peripheral-blood leucocytes and hyperplastic lymph nodes from 79 HIV-1-positive and 107 HIV-1-negative patients without non-Hodgkin lymphoma or any type of cancer from the Harris County Hospital District, the Methodist Hospital, and the AIDS and Cancer Specimen Bank served as the non-malignant lymphoid control samples. 26 samples of colon carcinoma and 28 of breast carcinoma from patients diagnosed with these malignancies between January and August, 2001, at the Methodist Hospital served as the cancer control group. A preliminary analysis of some non-Hodgkin lymphoma specimens was included in an earlier report.³ Institutional Review Board approval was obtained for this study.

Procedures

All sample processing was done in a laminar flow hood within a biosafety level 3 facility free from viruses and plasmids at the Department of Molecular Virology and Microbiology, Baylor College of Medicine. Total cellular DNA from non-Hodgkin lymphoma and control samples was extracted as previously described.¹⁵

All PCR assays were set up in the PCR Clean Rooms core facility of the Department of Molecular Virology and Microbiology at Baylor College of Medicine to avoid contamination of reaction mixtures. As a further precaution, positive-displacement pipetters and barrier-tip pipettes were used. Oligonucleotide primers used for PCR and DNA sequence analysis have been described previously.^15-19 All DNA samples were tested for suitability for amplification with primers specific for a fragment of the human ß-haemoglobin gene (primers PC03/KM38). Only specimens from which cellular ß-globin gene sequences could be amplified were then examined for viral sequences by PCR amplification with primer sets specific for a region of the large T antigen gene (PYVfor/PYVrev) conserved among all three polyomaviruses capable of infecting humans (SV40, JC virus, and BK virus), for the EBV latent membrane protein 2a (LMP-2a) gene (TP1Q5/TP1Q3), or for a region of the HHV-8 capsid gene (KS1/KS2).^18,19 Primers were obtained from Integrated DNA Technologies (Coralville, IA, USA).

Positive control plasmids were added to the control PCR reactions outside the core facility after tubes containing negative controls and test DNA were closed. The positive controls for polyomavirus PCR reactions were plasmid DNAs containing cloned SV40 (pSVSph21-N), JC virus (pBRJC-MAD-1), or BK virus (pBRBKV-Dunlop) genomes. The SV40 control genome contains an engineered restriction site that distinguishes it from natural isolates.¹⁵ The positive control for EBV reactions was DNA extracted from the EBV-positive Burkitt's lymphoma cell line Namalwa; that for HHV-8 was a plasmid DNA containing cloned viral sequences of the capsid antigen gene. Negative controls for PCR assays were reactions without added DNA template. PCR amplifications (45 cycles) were done with a GeneAmp PCR system 2400 thermocycler (Perkin-Elmer, Norwalk, CT, USA). High-stringency annealing temperatures specific for each primer set have been described elsewhere.^8,15,17-19 PCR amplification products were analysed by agarose gel electrophoresis.¹⁵

Probes specific for each virus were used to discriminate among amplified N-terminus T antigen polyomavirus sequences.^11,16 These specific oligoprobes are 39-labelled with a tail of dUTP-fluorescein by terminal transferase. Electrophoresed polyomavirus PCR products were transferred to a nylon membrane and the DNA was cross-linked to filters by ultraviolet irradiation for 2 min. The fluorescent hybrid was detected with an anti-fluorescein horseradish-peroxidase-conjugated antibody. Autoradiography was done at room temperature for 15 min. Additionally, representative polyomavirus PCR products were cloned into a TA cloning vector (Invitrogen, Carlsbad, CA, USA); multiple clones were screened by PCR and then sequenced with the Thermo Sequenase Radiolabeled Terminator Cycle Sequencing Kit (USB, Cleveland, OH, USA) to confirm the identity of polyomavirus-specific DNA from the tumours.

Statistical analysis

The necessary sample size for the study was calculated a priori from previously published reports. The presence of polyomavirus SV40 neutralising antibodies has been reported in 16% of HIV-1-infected and 11% of HIV-1-uninfected patients.²⁰ Published estimates of detection of SV40 DNA sequences in non-Hodgkin lymphoma range from 10% to 20%.^9,11,12 Therefore, we assumed a conservative rate of 15% for the detection of polyomavirus large T antigen sequences in non-Hodgkin lymphoma in HIV-1-uninfected patients. The rate of polyomavirus T antigen DNA in non-Hodgkin lymphoma was expected to be at least 30% among HIV-1-infected patients. On those assumptions, 120 individuals in each group (HIV-1-infected and HIV-1-uninfected patients) would be necessary to observe a difference of that magnitude assuming a power of 80%, a two-sided test, and a test significance level of 0·05. Post-hoc estimates, however, indicated that, with 75 participants, our study had more than 99% power to identify SV40 detection in non-Hodgkin lymphoma compared with control.

Statistical methods were used to address the third objective of this research. ² analysis was used to compare the distribution of viral sequences in non-Hodgkin lymphoma between HIV-1-infected and HIV-1-uninfected patients. The t test was used to compare the mean age of patients with SV40-positive non-Hodgkin lymphoma between the two groups, and non-parametric analysis of variance was used to compare the difference in CD4 cell count among HIV-1-infected patients with systemic non-Hodgkin lymphoma. Statistical analysis was done with the SAS/PC statistical software package.

Role of the funding source

The funding sources had no role in study design; in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.
 

Results

Table 1 shows the demographic characteristics of the 154 HIV-1-infected and HIV-1-uninfected patients with systemic non-Hodgkin lymphoma. The distribution of histological types of non-Hodgkin lymphoma analysed among HIV-1-infected and HIV-1-uninfected individuals was indicative of the frequency of non-Hodgkin lymphoma in these two populations of patients in general.¹⁴ Diffuse large B-cell lymphoma was the most common histological type of non-Hodgkin lymphoma in HIV-1-infected and HIV-1-uninfected patients. The mean CD4 cell count of HIV-1-infected patients at the time of diagnosis of systemic non-Hodgkin lymphoma was 165/µL (SD 185, range 2-901).

 

	HIV-1-infected patients	HIV-1-uninfected
	(n=76)	patients (n=78)
Demographics
Mean (SD, range) age (years)	40 (7, 28-58)	57 (15, 12-90)
Men/women	68/8	46/32
B-cell neoplasms
Precursor B-cell lymphoblastic	0	2
leukaemia/lymphoma
Mantle cell	0	1
Follicular	1	25
Diffuse large B-cell	58	40
Burkitt's*	13	7
Plasmacytoma	2	0
T-cell neoplasms
Peripheral T-cell, unspecified	2	2
Systemic anaplastic large cell	0	1
*Including six cases of variant Burkitt's lymphoma with atypical cytological features.
Table 1: Demographic characteristics and histological type of non-Hodgkin lymphoma in HIV-1-infected and HIV-1-uninfected patients

 

Polyomavirus large T antigen PCR products were generated from 64 of 154 (42%) samples of non-Hodgkin lymphoma, including from 25 (33%) HIV-1-infected patients and 39 (50%) HIV-1-negative patients (table 2, figure 1). Polyomavirus sequences were not detected in the non-cancer controls (peripheral-blood leucocytes and lymph-node samples from HIV-1-infected and HIV-1-uninfected patients without non-Hodgkin lymphoma) or cancer controls (colon and breast carcinomas). EBV DNA was detected in 30 (39%) of 76 non-Hodgkin lymphoma samples from HIV-1-infected patients and in 12 (15%) of 78 samples in the HIV-1-negative group. HHV-8 sequences were not detected in any of the non-Hodgkin lymphoma samples from either group of patients. Only 11 (7%) of 154 non-Hodgkin lymphoma samples (seven HIV-1-infected and four HIV-1-uninfected patients) were positive for both EBV and polyomavirus sequences.

 

	Polyomavirus	EBV	EBV and	HHV-8
			polyomavirus
Non-Hodgkin lymphoma
All cases (n=154)	64 (42%)*	42 (27%)	11 (7%)	0
HIV-1-positive (n=76)	25 (33%)*	30 (39%)	7 (9%)	0
HIV-1-negative (n=78)	39 (50%)*	12 (15%)	4 (5%)	0
Non-cancer controls
Lymph nodes
HIV-1-positive (n=7)	0	4 (57%)	0	0
HIV-1-negative (n=7)	0	3 (43%)	0	0
Peripheral-blood leucocytes
HIV-1-positive (n=72)	0	NT	..	NT
HIV-1-negative (n=100)†	0	NT	..	NT
Cancer controls
Colon cancer (n=26)‡	0	NT	..	NT
Breast cancer (n=28)‡	0	NT	..	NT
EBV=Epstein-Barr virus. HHV-8=human herpesvirus 8. NT=not tested. *All polyomavirus-positive specimens contained SV40-specific sequences. †Healthy adult volunteers. ‡HIV status was not assessed.
Table 2: Presence of polyomavirus and herpesvirus sequences in non-Hodgkin lymphoma and control samples from HIV-1-infected and HIV-1-uninfected patients

 

Figure 1: Agarose gel electrophoresis and staining with ethidium bromide of PCR-amplified polyomavirus sequences (upper panel), and Southern blotting (lower panels) with probes for individual polyomaviruses M=molecular-weight markers. SV40+, JCV+, and BKV+ are positive controls for SV40, JC virus, and BK virus, respectively. NC=negative controls (no DNA template). CA=cancer controls (colon and breast carcinoma samples).

The polyomavirus amplified products obtained from non-Hodgkin lymphoma samples were analysed by Southern blot hybridisation with specific probes for each virus (SV40, JC virus, and BK virus). The products were identified as polyomavirus SV40 in all cases (figure 1). The detection rate of SV40 T antigen DNA was significantly higher in samples of non-Hodgkin lymphoma than in non-malignant lymphoid samples from HIV-1-infected patients (25 of 76 [33%] vs 0 of 79, p<0·0001) or HIV-1-uninfected patients (39 of 78 [50%] vs 0 of 107, p<0·0001). The rate was also significantly higher in non-Hodgkin lymphoma from HIV-1-uninfected patients than in cancer control samples (39 of 78 [50%] vs 0 of 54, p<0·0001).

To confirm the presence of SV40 T antigen sequences, we further analysed samples of non-Hodgkin lymphoma in which SV40 DNA was detected. Sequence analysis of amplified products obtained from ten samples of non-Hodgkin lymphoma (six HIV-1-infected and four HIV-1-uninfected patients) showed the DNA sequences to be identical to those of the SV40 T antigen gene. The sequences associated with non-Hodgkin lymphoma lacked a 9 bp insert found in both JC virus and BK virus, proving that the sequences were not derived from either of these polyomaviruses (figure 2). Additionally, primers specific for the carboxy (C)-terminal region of the SV40 T antigen gene (TA1/TA2) yielded PCR amplification products of the expected size from 29 non-Hodgkin lymphoma samples. Sequence analysis of PCR products from five samples of non-Hodgkin lymphoma (three from HIV-1-infected and two from HIV-1-uninfected patients) confirmed that their origin was SV40. We then compared these five lymphoma-associated T antigen C sequences with a catalogue of SV40 sequences⁵ (GenBank). One C sequence was similar to that of SV40 strain CPC/MEN, previously detected in several primary human brain cancers,⁵ one was different from any previously reported SV40 sequence, and three (one from an HIV-1-infected patient and two from HIV-1-uninfected patients) were similar to that of SV40 strain MC-028846B--a virus first detected in a sample from a contaminated poliovaccine from 1955.²¹ These results substantiate our belief that the T antigen gene of SV40 was present in the non-Hodgkin lymphoma specimens tested, and that detection of SV40 sequences was not the result of laboratory contamination.

Figure 2: DNA sequence of PCR product from N-terminus of polyomavirus T antigen gene from systemic non-Hodgkin lymphoma Sequence of non-Hodgkin lymphoma (NHL) specimen is identical to that of SV40 and lacks the 9 bp insert found in JC virus and BK virus DNAs.

We saw no significant differences in the mean age of patients with SV40-positive and SV40-negative non-Hodgkin lymphoma within the HIV-1-infected group (42 years [7·0] vs 39 years [6·6], p=0·07) or the HIV-1-uninfected group (59 years [12·9] vs 54 years [17·3], p=0·2) group. Five of the patients with SV40-positive non-Hodgkin lymphoma were born after 1963, the last year that SV40-contaminated poliovirus vaccine was used in the USA.⁵ Among HIV-1-infected patients with systemic non-Hodgkin lymphoma, there was no significant difference in the mean CD4 cell count between patients with EBV-positive and SV40-positive systemic non-Hodgkin lymphoma (190/µL [257] vs 112/µL [93], p=0·3). Additionally, the CD4 cell count did not differ between virus-positive (EBV and SV40) and virus-negative patients with systemic non-Hodgkin lymphoma (166/µL [200] vs 163/µL [166], p=0·9). Non-Hodgkin lymphoma samples were more frequently EBV-positive in HIV-1-infected than HIV-1-uninfected patients (30 of 76 [39%] vs 12 of 78 [15%], p=0·001), whereas non-Hodgkin lymphoma samples were more frequently positive for SV40 T antigen in HIV-1-uninfected than HIV-1-infected patients (39 of 78 [50%] vs 25 of 76 [33%], p=0·03).

The histological types of non-Hodgkin lymphoma positive for SV40 and EBV sequences are presented in table 3. Diffuse large B-cell lymphoma was the most frequent type of non-Hodgkin lymphoma positive for viral sequences. The rate of EBV detection in diffuse large B-cell lymphomas was not significantly different between HIV-1-infected and HIV-1-uninfected patients (p=0·1). However, the detection of SV40 T antigen sequences was significantly more common in diffuse large B-cell non-Hodgkin lymphoma from HIV-1-uninfected than from HIV-1-infected patients (p=0·003). SV40 sequences were also found frequently in follicular tumours from HIV-1-uninfected patients. Virus detection rates did not differ among tumours obtained from different sources (Houston or National Cancer Institute). Of the seven Burkitt's lymphomas found to contain SV40 DNA, six were regarded to be atypical variants since they displayed atypical cytological features.²² The histological types of the 11 non-Hodgkin lymphoma samples positive for both EBV and SV40 sequences were diffuse large B-cell lymphomas (n=8) and Burkitt's lymphomas (n=3). Neither EBV nor SV40 sequences were detected in the five T-cell neoplasms tested.

 

	HIV-1-infected patients			HIV-1-uninfected patients			All cases
	Total	DNA positive		Total	DNA positive		Total	DNA positive
	tested	SV40	EBV	tested	SV40	EBV	tested	SV40	EBV
Diffuse large cell	58	19	20	40	25	8	98	44	28
Follicular	1	1	0	25	11	3	26	12	3
Burkitt's*	13	5	8	7	2	0	20	7	8
Other†	2	0	2	3	1	1	5	1	3
Total	74	25	30	75	39	12	149	64	42
T-cell neoplasms were negative for presence of Epstein-Barr virus (EBV) and SV40 sequences. *Including six specimens of variant Burkitt's lymphoma (see footnote to table 1). †Including other B-cell neoplasms listed in table 1.
Table 3: Presence of SV40 and Epstein-Barr virus DNA sequences by histological type of non-Hodgkin lymphoma from HIV-1-infected and HIV-1-uninfected patients

 

 

Discussion

This investigation showed that polyomavirus SV40 T antigen DNA sequences are significantly associated with non-Hodgkin lymphoma in HIV-1-infected and HIV-1-uninfected patients. This finding sheds new light on the possible genesis of an important group of malignant disorders. The SV40 sequences do not seem to be present simply because non-Hodgkin lymphoma cells are readily susceptible to viral infection; in that case, EBV and SV40 should be found in similar frequencies in non-Hodgkin lymphoma of HIV-1-infected and HIV-1-uninfected patients. The results also suggest that polyomavirus SV40 is not merely an opportunistic superinfection; if so, one would expect similar frequencies of SV40 detection in EBV-positive and EBV-negative non-Hodgkin lymphoma, and in other cancer samples (colon and breast) if those cell types were permissive to SV40 replication. The observation of minimal instances of coinfection with SV40 and EBV and the lack of detection of SV40 in non-malignant lymphoid samples and epithelial cancer control specimens suggest that SV40 might contribute to the development of those lymphomas in which it is present.

Overall, 42% of non-Hodgkin lymphomas tested here contained SV40 DNA sequences--a frequency similar to that found in an independent study (43%).²³ This frequency is higher than reported in previous studies,^9,11,12 and might be a consequence of characteristics of the specific populations of patients from whom specimens were obtained, the histological types of tumours tested, or variations in DNA extraction methods or PCR assay conditions. By contrast with our working hypothesis, the SV40 positivity rate detected here was significantly higher in non-Hodgkin lymphoma from HIV-1-negative patients than in those from HIV-1-infected individuals. We do not yet understand the role of HIV-1 infection in SV40 pathogenesis. Our observed non-Hodgkin lymphoma positivity rates could be a result of the particular sets of tumour specimens we obtained for this study, as well as of the fact that more non-Hodgkin lymphomas are EBV-positive in HIV-1-infected patients than in HIV-1-negative individuals. Our observations do indicate, however, that the development of SV40-positive non-Hodgkin lymphoma is not dependent on pronounced immunodeficiency in the host.

We found EBV associated with 39% of systemic non-Hodgkin lymphoma from HIV-1-infected patients and with 15% from the HIV-1-negative group, similar to rates reported previously.³ We did not detect HHV-8 sequences in non-Hodgkin lymphoma from either group of patients, in agreement with recent studies that showed lack of association between HHV-8 and non-Hodgkin lymphoma in HIV-1-infected and HIV-1-uninfected patients.²⁴ SV40 T antigen sequences were detected frequently in diffuse large B-cell lymphomas in both groups of patients and in follicular lymphoma in HIV-1-uninfected patients. This particular association might be important, since these are the two most common histological types of lymphomas from mature B cells and account for about 50-60% of all cases of non-Hodgkin lymphoma.¹⁴ It also suggests that mature B cells could be more susceptible than precursor cells to the transforming potential of SV40.

The SV40 sequences associated with non-Hodgkin lymphoma identified here were different from those of known laboratory strains, and several examples of the C-terminal T antigen gene sequence were similar to that of an SV40 strain detected in a sample of contaminated poliovaccine from 1955.²¹ We know that several strains of SV40 exist,⁵ but whether the strains detected in non-Hodgkin lymphomas are more lymphomagenic than other strains remains to be determined.

The oncogenic potential of polyomavirus SV40 has been established in laboratory animals.^2,5 In studies in which hamsters were inoculated intravenously with SV40, lymphomas developed among 72% of the animals in the inoculated group and none of the control group.¹⁰ The histological type was consistent with diffuse large cells, and the lymphomas were shown to be of B-cell origin because they expressed cell-surface antigen.²⁵ More recently, a study confirmed the lymphomagenic capacity of the virus and that lymphomas represent a common malignancy induced by SV40.²⁶ After intravenous inoculation, about a third of the animals developed more than one histological type of malignant neoplasm, with osteogenic sarcomas being most common after lymphomas.¹⁰ After intracardiac injection, malignant mesotheliomas and osteosarcomas developed in addition to lymphomas.²⁶ These studies supported a causative role for the virus in lymphomagenesis because SV40 T antigen was expressed in all tumour cells, animals with tumours developed antibody against SV40 T antigen, and neutralisation of SV40 with specific antibody before virus inoculation prevented lymphoma development. Knowledge of these animal studies prompted us to consider a role for SV40 in human lymphomagenesis.

Polyomavirus SV40 has been associated with specific types of solid cancers in human beings, including brain tumours, osteosarcomas, and malignant mesotheliomas.^5,6 These are the types of malignant disorders caused by the virus in laboratory animals--a finding that emphasises the predictive value of the animal studies. Recent reports provide persuasive evidence that the presence of polyomavirus SV40 is meaningful in the development of those human cancers. Immunohistochemical assays have detected the expression of T antigen in tumour cells,^11,16,27 T-antigen protein complexed with p53 has been extracted from some cancer specimens,^27,28 and microdissection of malignant mesothelioma samples followed by PCR assays detected SV40 DNA in tumour cells and not in adjacent non-malignant cells.²⁹ When an antisense SV40 T antigen construct was introduced into SV40-DNA-positive malignant mesothelioma cell lines, the expression of T antigen was abrogated and growth was inhibited.³⁰

The polyomaviruses JC virus and BK virus also have the ability to induce tumour formation in laboratory animals;⁶ they have been associated with some human solid tumours, in particular brain cancers,⁶ but much less frequently than SV40. This observation suggests that SV40 is more oncogenic in humans than are JC virus and BK virus. Up to 80% of the adult population worldwide is seropositive for these viruses,⁶ and JC virus is recognised as the causative agent of progressive multifocal leucoencephalopathy--a subacute opportunistic disease in HIV-1-infected patients. We did not detect JC virus or BK virus DNA sequences in any of the non-Hodgkin lymphoma specimens tested in this study, by contrast with a previous report of JC-virus-positive non-Hodgkin lymphoma involving HIV-1-infected children.¹³ These differences could reflect the age or geographic origin of the patients or a difference in oncogenic capacity among the polyomaviruses. Models of the oncogenicity of JC virus and BK virus do not indicate the development of lymphomas.⁶

The major source of known human exposure to polyomavirus SV40 was immunisation with SV40-contaminated poliovaccines. Inactivated and live, attenuated forms of the poliovaccine were prepared in primary rhesus monkey kidney cells, some of which were from animals naturally infected with SV40--a virus that was unknown at the time. Studies showed that residual infectious SV40 survived the vaccine inactivation treatments, and millions of people were inadvertently exposed to live SV40 from 1955 until early 1963.^5,31 In the USA, vaccine lots received by about 20 states are estimated to have contained 0·75-0·97 mL contaminated vaccine per child, lots from about 15 states were thought to have contained 0·01-0·74 mL contaminated vaccine per child, and about 15 states were believed to have received lots that were free from SV40.³¹ Perhaps this distribution of contaminated vaccines influenced the differences in the rate of SV40-positive non-Hodgkin lymphoma that have been seen in recent studies. Seroepidemiological studies have shown the presence of SV40 neutralising antibodies in 16% of HIV-1-infected patients and 11% of HIV-1-uninfected individuals, some of whom were born after 1963 and could not have been exposed to SV40-contaminated poliovaccines.²⁰ Our study found that five patients with SV40-positive non-Hodgkin lymphoma were born after 1963--a finding similar to previous studies involving brain and bone cancers in which some patients with SV40-positive tumours had been born in recent decades.^5,16 These observations suggest that polyomavirus SV40 might be causing infections in human beings long after the use of the contaminated vaccines. However, how SV40 is transmitted among humans, and the prevalence of infection, remain to be established.

In summary, our study suggests that polyomavirus SV40 is significantly associated with non-Hodgkin lymphoma in HIV-1-infected and HIV-1-uninfected patients and might have a role in the development of these haematological malignancies. Definition of a viral cofactor in the pathogenesis of these tumours could lead to new diagnostic, therapeutic, and preventative approaches.

Contributors

R A Vilchez participated in conception and study design; collection, assembly, and analysis of data; statistical analysis and interpretation of the data; and preparation of the paper. C R Madden helped with analysis of the data. C A Kozinetz assisted with conception, study design, and statistical analysis. S J Halvorson and Z S White helped with collection and assembly of data. J L Jorgensen and C J Finch assisted with collection and histopathological diagnosis of specimens. J S Butel directed the study and was involved in study design, data analysis and interpretation, and preparation of the paper. All authors reviewed and approved the final paper.

Conflict of interest statement

None declared.

Acknowledgments

We thank the AIDS and Cancer Specimen Bank sponsored by the National Cancer Institute for providing specimens of non-Hodgkin lymphoma and lymph nodes for this study, and the staff members of the Pathology Departments at the Methodist Hospital and the Harris County Hospital District for their assistance.

This study was supported in part by the Baylor Center for AIDS Research Core Support Grant Number AI36211 from the National Institute of Allergy and Infectious Diseases, and by Cooperative Agreement NCC 9-58 with the National Space Biomedical Research Institute funded by the National Aeronautics and Space Administration. R A Vilchez is the recipient of the Junior Faculty Development Award from GlaxoSmithKline.

SOURCE: The Lancet

            Volume 359 Issue 9309 Page 817