Human B-lymphotropic virus (human herpesvirus-6)

Journal of Virological Methods, 21 (1988) 29-48 Elsevier

29

JVM 00759

Human B-lymphotropic virus (human herpesvirus-6)

D.V. Ablashi’, S.F. Josephs’, A. Buchbinder2, K. Hellman2, S. Nakamura*, T. Llana*, P. Lusso*, M. Kaplan3, J. Dahlberg4, S. Memon4,

F. Imam4, K.L. Ablashi4, P.D. Markham5, B. Kramarsky6, G.R.F. Krueger7, P. Biberfeld8, F. Wong-Staal*, S.Z. Salahuddin* and R.C.

Gallo*

‘Laboratory of Cellular and Molecular Biology, National Cancer Institute, Bethesda, MD, U.S.A., ‘Laboratory of Tumor Cell Biology, National Cancer Institute, Bethesda, MD, U.S.A., ‘North Shore

University Hospital, Long Island, NY, U.S.A., “Pan-Data Systems. Inc., Rockville, MD, U.S.A., 5Bionetics Research, Inc., Kensington, MD, U.S.A., ‘Electra-Nucleonics, Inc., Silver Spring, MD,

U.S.A., 71nstitute of Pathology, University of Cologne, Cologne, F. R. G. and ‘Department of Pathology, Karolinska Institute, Stockholm, Sweden

(Accepted 18 May 1988)

Summary

Human B-lymphotropic virus (HBLV), also known as human herpesvirus- (HHV-6) was first isolated in 1986 from AIDS patients and patients with other lymphoproliferative disorders. HBLV is distinct from known human herpesvi- ruses, biologically, immunologically and by molecular analysis. HBLV can infect and replicate in fresh and established lines of hemopoietic cells and cells of neural origin, suggesting wide tropism. The prevalence of HBLV antibody in the normal population was 26% though clear differences between different populations were observed. The prevalence of HBLV antibody an elevated antibody titer was higher in sera from certain malignancies, Sjogren’s syndrome and sarcoidosis. Antibody to HBLV was also elevated in AIDS patients and patients with chronic fatigue syndrome. HBLV-DNA was detected in some B-cell lymphomas. The broad in vi- tro tropism, combined with immunological and molecular evidence of HBLV in- fection in individuals raise the question of the pathogenicity of this virus in some diseases. Because in vitro co-infection of CD, cells by HBLV and HIV leads to enhanced degeneration, this raises the possibility that infection in AIDS patients by both viruses can aggravate the HIV-induced immunodeficiency. Specific re-

Correspondence lo: D.V. Ablashi, Laboratory of Cellular and Molecular Biology, National Cancer In- stitute, Bethesda, MD, U.S.A.

30

agents and immunological and molecular assays are currently being investigated, which will aid in virus detection in cells from patients, and in elucidating the pos- sible pathogenesis of HBLV.

HBLVIHHV-6; Biology; Immunology; Molecular hybridization

Introduction

As part of the long-term objectives to understand the mechanisms regulating human hematopoiesis and defects leading to dysfunction, deficiencies and/or ma- lignancy, we have recently focused on diseases frequently associated with infection

by HIV-l. In addition to AIDS, B-cell lymphomas and other lymphoproliferative disorders were investigated. These studies utilized methods for activation and the long-term cultivation of different types of fresh human leukocytes. For these stud- ies, one main source of cells was fresh peripheral mononuclear cells from various patients with AIDS-associated lymphoproliferative disorders.

A small number of shortlived, large, refractile cells in cultures of mononuclear cells from patients with AIDS and other lymphoproliferative disorders were oc- casionally observed after stimulation with PHA (Fig. 1). These large cells were different from PHA blasts, generally mononucleated or binucleated, and fre- quently contained intranuclear and or intracytoplasmic inclusion bodies (Fig. 2). When the primary cells were cocultivated with activated mononuclear cells from healthy controls, e.g., human cord blood leukocytes, large cells also appeared. The

cells died within two weeks. Testing for known viruses showed supernatant fluids from such cell cultures lacked transforming or nontransforming EBV, but herpes- virus-like particles were demonstrated by electron microscopic examination (Fig. 3). In contrast to other human herpesviruses, the majority of these extracellular virus particles were complete virions, and other features also suggested a new her- pesvirus isolate. Further evidence to that was substantiated from the immunolog- ical and molecular characteristics of this virus, now known as human B-lympho- tropic virus (HBLV) or human herpesvirus- (HHV-6). Recently we have standardized procedures for serological determination of antibody to this virus and are in the process of determining its prevalence in the general population. Meth- ods have also been developed to test for the presence of genomic DNA in tissues

and from tumors.

Immunologic identification of infected cells and characterization of cell virus inter- action

The large refractile cells in the cord blood mononuclear culture usually ap- peared within 3-4 days post-infection, with supernatants obtained from cultures of patient lymphocytes exhibiting herpesvirus-like particles. These cells were tested for HBLV antigen by indirect immunofluorescence assay (IFA), after fixation with acetone, using positive patients’ sera (Salahuddin, 1986). The IFA was later mod-

Fig. 1. In vitro cultured mononuclear cells from a patient’s peripheral blood showing large refractile Cdk.

Fig. 2. Cytospin preparation of large refractile. mono- or binucleated cells. These cells were observed in the PHA-stimulated cells in culture. as shown in Fig. I (Giemsa stain).

HUMAN B-LYMPHOTROPIC VIRUS IHBLV)

Fig. 3. Electron micrograph of HBLV particles from the cultured mononuclear cells. Insert shows the

detailed structure of a single virion.

ified by using cells from an infected cell line which increased specificity of the as- say (Ablashi, unpublished data). A characteristic granular nuclear and cytoplasmic immunofluorescence staining was often observed in positive cells where the other cells in the background were always free of detectable HBLV antigen (Fig. 4a). Live HBLV-infected cells exhibited patchy surface immunofluorescence (Fig. 4b). IFA results were confirmed by immune electron microscopy using an indirect im- munoferritin staining according to procedures reported by Biberfeld (1987) (Fig. 5). From such studies it was clear that antibody in patients’ serum bound specifi- cally to the virions and not to the cell membrane.

HBLV-infected cells obtained from patients were initially found to express B- cell surface markers (Salahuddin, 1986). In later studies of the virus tropism, pe- ripheral blood lymphocytes, leukocytes from bone marrow, spleen, thymus and tonsils, were fractionated by Ficoll Hypaque and were infected in vitro with HBLV. In such experiments, infected cells were found predominantly by IFA, and using cell-specific monoclonal antibodies and by radio-immunoprecipitation to possess T-cell associated antigens (CD,, CD,, CD,, CD? and to a lesser extent, CD,) (Ta- ble 1) (Lusso, 1987). A number of established cell lines of T- and B-lymphocytes and other cell types were also found susceptible to infection by HBLV (Ablashi, 1987, Ablashi, 1988a). In addition to CD, cell lines, a number of B-cell lines were infectable with HBLV. We were not able to infect EBV genome negative B-cell

34

Fig. 4. Indirect immunofluorescence of HBLV-infected cells. (a) HBLV-infected acetone-fixed cells.

showing granular, nuclear and cytoplasmic stained cells with a patient’s serum. (b) HBLV-infected live

cell showing membrane fluorescence at the cytoplasm.

lines (Bjab, Louks, Ramos, SKD) until these cell lines were infected by EBV. This suggests that EBV may be inducing a receptor for HBLV (Ablashi, unpublished data). Interestingly, no EBV genome negative cell lines could be infected with HIV-

Fig. 5. Immunocytochem~str~ of an HBLV-infected cell showing virions at the cell margin coated with the electron opaque product of the ~mmunopero~idas~ reaction. The eel1 membranes are not coated

with the reaction product. The bar represents 1.0 pm.

TABLE 1

B- and T-lymphoblastoid and megakaryocyte cells infected with HBLV and HIV

Cell type Percent of cells infected with:

HIV HBLV

Peripheral blood <40 c.55 Mononuclear ceils,

cefts from bone marrow and thymus T-&t tines

hD5, MOLT-3, H9, JM2.7 30-65 15-80 Immature T-cells

HSB> 13.5 290 B-cell lines

(EBV-gename positive) ET-62, LDV-7, IM-9, Craig, NAB-2 10-35 3 lf_l-so

EBV-converted Ramos Bjab, and Luuks IS-20 240

Megakaryocyte cell line (HEL) z=40 230

” As determined on live cells by TFA, using monoclonai antibodies.

Predominant phenotype”

CD.&&. CD,, CD,

CD,, CDT, CDS

CD,, OKT-I#, CD,

CD,,, C&O

CD,,, C&O PDGF receptor

36

1, but EBV converted cell lines were infectable (Salahuddin, 1985). This finding raises the possibility that the receptors induced by EBV are common to HBLV and HIV. This is also supported by the finding that all cell lines infectable with HBLV were infectable with HIV (Table 1). These findings suggest that HBLV has a broader cell tropism than was originally found (Lusso, 1987). Moreover, recent experiments demonstrated that CD, peripheral blood cells can be simultaneously infected by HIV-l and HBLV (Lusso, unpublished data). Similar results were ob- tained with CD, positive T-cell line (Salahuddin, unpublished data). Doubly in- fected cell cultures exhibited positivity for HBLV antigens and for HIV-l P24 and P19, by IFA (Table 1). The co-infected cell cultures showed significantly enhanced

Fig. 6. Thin section electron micrographs of the morphogenesis of HBLV: (A) nucleocapsids forming

in the nucleus of an infected cell (co, nucleoprotein core; ca. capsid); (B) nucleocapsids at the margin

of the nucleus of infected cells showing the beginning of tegument formation (t); (C) enveloped virion in the cisterna of the double nuclear membrane of an infected cell (e, viral envelope); (D) enveloped virions in the cisternae of the rough endoplasmic reticulum of an infected cell; (E) tegument-coated

unenveloped nucleocapsids free in the cytoplasm of an infected cell (t. tegument); (F) extracellular en- veloped HBLV virions associated with infected cells (s, surface spikes; e, envelope; ca, capsid: co, nu- cleoprotein core; t, tegument). The bar represents 0.1 pm. Preparation as previously described (Bi-

berfeld, 1987).

37

degeneration when compared to cultures infected separately with HBLV or HIV- 1. Furthermore, a higher virus-related, extracellular reverse transcriptase activity and higher levels of HIV-l antigens were observed when primary mononuclear cell cultures from AIDS patients were infected in vitro with HBLV (Salahuddin, un- published data). While no definitive disease has been found to be associated with HBLV, these observations raise the possibility that it may act as a co-factor with HIV-l in the development of AIDS.

HBLV (HHV-6) morphology Ultrastructural studies showed that HBLV is an enveloped virion with an ico-

sahedral capsid with 162 capsomers. The diameter of the enveloped particles was estimated to be about 200 nm. The composite of the electron microscopic figure shows in detail various aspects of the morphogenesis of HBLV (Fig. 6A-F). Inside

Fig. 7. High resolution electron micrograph of an HBLV virion. Preparation was negatively stained

with phosphotungstate. The stain has penetrated the envelope revealing internal structures: S, surface

spikes; E, envelope; C, capsid; D, DNA coil; CM, central mass. The DNA coil has apparently loos-

ened and the central mass has become realigned parallel to the DNA coil, instead of perpendicular to it. The bar represents 0.1 pm.

38

the capsid, the DNA is coiled around a cylindrical mass (Fig. 7). This feature de- scribed by Roizman in 1973 in other herpesviruses, appears to be made of subunits arranged in a symmetrical array. The virus particles mature in the nucleus, and are

initially released from the cell by exocytosis, until cytolysis occurs. Since the initial studies reported by us (Salahuddin, 1986), HBLV-like isolates

have been described by others (Downing, 1987; Tedder, 1987), from HIV-l-an- tibody positive African AIDS patients. These isolates morphologically resemble HBLV and hybridize to HBLV DNA. Moreover, we have also isolated HBLV from two patients with chronic fatigue syndrome (CFS) (Lusso and Ablashi, unpub- lished data). The characterization of these isolates is in progress.

Production of HBLV in established cell lines Earlier difficulties in producing large amounts of HBLV for biologic. molecular

and immunologic investigations were recently overcome when HBLV was found to infect and replicate in T- (HSB,, JM2.7, MOLT-3. 6D5) and B- (IM-9, ET-62, LDV-7, P3HR-1) cell lines and in a megakaryocyte cell line (HEL) (Table 1). The amount of virus produced by these cell lines varied considerably. The most effec- tive cell lines for the production of HBLV were T-cell lines HSB,, JM2.7, 6D5 and an EBV genome positive B-cell line, ET-62 (Table 1). Infectivity titers >2.&4.0 logs/ml were generally detected in the cell-free supernatants. Interestingly, a glioblastoma cell line HTB14 could be infected by HBLV. The T-cell lines, and HEL and HTB14 cells lacked EBV, HCMV, HSV and VZV, by Southern blot (Ablashi, 1988a). No EBV or viral antigens (VCAIEA) were detectable after HBLV infection of EBV-genome positive ET-62, LDV-7 and Craig cells. Infec- tion of P3HR-1, an EB virus producer cell line, showed degeneration of EBV VCA and EA positive cells, suggesting that both viruses may be infecting the same cell (Ablashi, unpublished data). Since HSBz and other T-cells are free of other hu- man herpesviruses and produce considerable amounts of HBLV, they can be used for antibody testing to HBLV, by IFA as well as for production of large quantities of virus. At present, HBLV-infected HSB, cells are regularly used by us and oth- ers for serological studies of the prevalence of HBLV antibody by IFA.

Development of immunologic assays other than IFA for analysis of HBLV antigens and the detection of antibody

Besides IFA, which has been the most widely used assay for detecting HBLV antibodies and antibodies to other herpesviruses (Miller, 1985; Pearson, 1986), other immunological assays have also been developed. Radio-immunoprecipita- tion (RIP), Western blots, and ELISA procedures have mainly been used in the identification of various viral proteins of herpesvirus, mainly because of the avail- ability of purified viral proteins and monoclonal antibodies (Pearson, 1985, 1986; Dahlberg, 1985). By IFA it was observed that some sera gave non-specific reac- tions on HBLV-infected human cord blood mononuclear cells as well as HSB, cells. Most of these sera were from Africa or from patients with auto-immune disorders. The RIP assay was found to be useful in detecting the HBLV antibody in these sera. Reactivity to a 120 kDa protein band was consistently found in IFA positive

39

A B 1 2 3 4 1 2 3

kDa

- 200

- 95

- 66

- 45

- 29

Fig. 8. Radio-immunoprecipitation (RIP). Immunoprecipitation of HBLV proteins. (A) HBLV-in-

fected HSB, cells were starved for 1 h in MEM without methionine, then labeled for 3 h in the same

medium, containing 100 Ci/ml [?S]methionine, 5% normal RPM1 1640, and 5% dialysed fetal bovine

serum. The cells were lysed and aliquots were processed for immunoprecipitation, as described earlier

(Dahlberg, 1985). Lane 1, a human serum negative by IFA; Lane 2, a hyperimmune rabbit serum im-

munized with purified HBLV: Lane 3, an extremely high titered serum (by IFA); Lane 4, a typical

IFA-positive human serum. (B) Reactivity of a high titered human serum toward extracts of radiola-

beled HBLV (Lane l), HBLV-infected HSBz cells labeled overnight (Lane 2), and control-uninfected cells labeled overnight.

control sera. For the RIP assay, virus infected and uninfected HSB, cells were la- beled with [35S]methionine and cysteine, and were processed according to the pro- cedure described by Dahlberg (1985). Panel A of Fig. 8 illustrates the reactivity of human and a hyperimmune rabbit serum (Lane 2) toward virus infected HSB, cells labeled for three hours. The serum used for lane 3 had an exceptionally high titer by IFA (>6000). Note that the rabbit has only modest reactivity toward the ~120 and extra reactivities against proteins with approximate molecular weights of 200 000, 180 000 and about 55 000. Panel B of Fig. 8 illustrates the proteins pre- cipitated by the same serum with extracts of labeled virus (l), infected cell extract (2), and uninfected cell extract (3). As can be seen in Lane 1 of panel A of Fig. 8, a serum found to be negative by IFA for HBLV antibody reacted in the HBLV RIP assay, suggesting that the RIP was more sensitive than IFA. The high IFA positive sera showed a good correlation when tested by RIP. Obviously it is dif- ficult to use RIP as the screening test or the test for sero-epidemiological studies, but it should be useful in evaluating those sera which can not be tested by IFA because of non-specific reactivity.

40

Immunoblotting was also performed, by the blotting of electrophoretically re- solved samples of either banded, purified HBLV, or infected cell lysates to nitro-

cellulose membrane followed by specific detection, using the blotting system de- scribed by Towbin (1979) and Gershoni (1982) (Fig. 9). With infected cell lysates, very little reactivity could be detected, perhaps because infected cells are fragile and tend to lyse. thus releasing virus and viral antigens into culture supernatants. Fig. 9 compares the reactivity of three human sera from Sjogren’s patients on blots of purified virus (panel A) and uninfected HSB2 cell extracts (panel B). Lanes 1 and 2 of panel A (Fig. 9) illustrate the activity of two IF positive sera, while Lane 3 utilized a negative serum obtained from Sjogren’s syndrome patients. Although a number of virus-specific bands can be identified by immunoblotting, particularly those of approximately 88, 72, 68 and 58 kDa, reactivity toward the 120 kDa pro- tein seen by RIP is conspicuously reduced or missing. This may reflect the very

A

2%

123

B

123

Fig. 9. Western blot analysis of HBLV proteins (Immunoblot). Immunoblot analysis of IFA positive

and negative sera. (A) Three Sjogren’s syndrome patients’ sera were tested for reactivity on identical strips containing 5 wg of purified HBLV. Lane 1. an IFA positive serum containing detectable reac-

tivity to the ~120 kDa protein in addition to several smaller proteins ranging in size from about 40 to

100 kDa. Lane 2, a different IFA positive serum. lacking reactivity to ~120 in this assay, but positive

against other viral antigens. Lane 3, a negative serum. (B) The same three sera reacted with strips con- taining 5 pg each, of uninfected HSB, antigen. Although one of the IFA positive sera reacted to sev-

eral control antigens, the second (lane 2) was largely non-reactive. In any case, the pattern of reactivity was quite different, and it appears that background bands, with rare exceptions, are less than 60 kDa, so that reactivity of the sera below 60 kDa appear to be difficult to assess because of reactivity with

uninfected cells.

41

low level of this protein seen in Coomassie stained gels of purified virus, suggest- ing that the ~120 may be a protein preferentially lost during virus purification. Since it is a dominant reactivity in the RIP assay, the Western blot may be of limited usefulness in diagnosing ambivalent sera unless a means is found of enriching virus preparations in the 120 kDa protein. The development and use of HBLV ELISA and further significance of Western blot will be discussed in detail by Carl Saxinger in this issue (Saxinger et al., 1988).

The preliminary data provided by these assays so far indicate their usefulness for the identification of the proteins which may be of clinical importance. With the availability of purified virus, and development of monoclonal antibodies, a rapid advance in HBLV immunology similar to that of EBV is anticipated.

Immunological and molecular characterization of HBLV Previous data using monoclonal and polyclonal sera containing antibody to hu-

man and non-human primate herpesviruses showed no reactivity to HBLV (Sa- lahuddin, 1986). To distinguish HBLV from other human and non-human viruses (Salahuddin, 1986), monoclonal and polyclonal antibodies to known human and animal herpesviruses derived from cattle, pigs, horses, cats, dogs, guinea pigs, deer, chickens and turkeys were tested for reactivity against HBLV. These reagents failed to react with HBLV in IFA, immunoblot and Western blot assays (Table 2). The specificity of antibody to HBLV in human serum was demonstrated by absorbing

TABLE 2

Analysis of immunologic reactivities of various animal herpesvirus sera and swine fever virus sera to

HBLV by IFA and dot-blot assay”

Animal serum description

Deer (contain antibody to deer herpesvirus)

Swine (anti-pseudo rabies) Swine fever (sera from infected cases)

Equine (anti-equine HV) Bovine (naturally infected with bovine HV)

Anti-bovine viral diarrhea

Anti-infectious bovine rhinotracheitis

Feline (naturally infected with feline HV)

Anti-G. pig HV

Anti-mouse CMV

Owl monkey (anti-HV type-l)

Rhesus monkey (anti-B virus) Chicken (naturally infected with MDHV)

Monoclonal to MDHV (Marek’s dis. herpesvirus)

Rabbit anti-human CMV Human NPC reactiveh

Dilution tested

10, 20 5, 10

5, 10. 20

5. 10 5, 10

5, 10

5, 10

5. 10

5, 10

5, 10

5, 10

5, 10

5, 10 5, 10

5, 10, 20

5. 10. 20. 40

Incidence of positivity:

IFA dot-blot

0124 0124 oi 1 O/l

018 015 Oil o/ 1

012 Oil

012 012

012 012 Oil N.D. Oil N.D.

Oil N.D.

o/2 Oil

013 013 117 (1:5) weak 015 015 015

Oil Oil

l/l l/l

,’ All sera were classified and heat inactivated at 56°C for 112 h to remove non-specificity. Each serum

was also tested on HBLV uninfected cells in IFA. ” This serum was reactive against HBLV infected HSB, cells at 1:40 dilution.

42

the HBLV antibody-positive serum with EBV, HCMV, VZV and HSV. After ab- sorption, no loss of HBLV antibody titer was detected.

Molecular analysis of HBLV HBLV contains a large molecular weight, double-stranded DNA genome of ap-

proximately 170 kb, which is consistent with its morphological classification as a herpesvirus. A 9.0 kb molecular clone designated PZVHl4 was obtained from HBLV dp DNA (Josephs, 1986). Another clone of 23.0 kb is also being investi- gated. Molecular analysis demonstrated that HBLV is a new herpesvirus (Josephs. 3986) thus supporting the immunologic, biologic and morphologic observations. The PZVH14 clone, used as a probe, detected DNA and RNA sequences of HBLV in experimentally infected cells from human cord blood, and not in uninfected cells in the same culture (Fig. 10). This same probe was used extensively in searching for HBLV genomes in tumor tissues. Dot-blot analysis of genomic DNA of human herpesviruses (EBV, HCMV, HSV and VZV) did not reveal cross-hybridization with other herpesviruses, using HBLV probe ZVH14. Conversely, while they hy- bridized to the homologous genomic DNA. none of the other viral probes hybrid- ized to the lanes containing HBLV-DNA (Fig. 11). Also HBLV-DNA failed to cross-hybridize with Her~es~~~r~ saimiri of squirrel monkeys (Josephs. 1986). Thus. no sequence similarities were found with other herpesviruses. A clone of human

Fig. 10. In situ hybridization of HBLV-infected human cord blood mononuclear cells with PZVHl4 clone of HBLV (GS strain).

43

Genomic DNA

CMV ’ Labeled 0.1

Probe

Fig. 11. DNA dot blot hybridization of HBLV DNA probes to DNA of EBV, CMV, HSV and VZV. 1 unit = 25 wg DNA.

CMV was recently found to cross-hybridize, to some extent, with HBLV (Efsta- thiou, 1987). However, the same CMV clone did not hybridize with HBLV clone PZVH14. Sequence similarity was also found between the BamB fragment of HBLV and the BamA fragment of Marek’s disease virus (Kishi, 1988). Prelimi- nary data show that there appears to be a pleomorphism between HBLV isolates (Josephs, unpublished data). These isolates are also being compared for biological and immunological differences.

Prevalence of HBLV antibody in ~y~phopro~~erative diseases and in healthy sub- jects

IFA of sera from 1095 healthy donors, obtained from blood banks and normal laboratory donors in the U.S.A.: Canada and Europe, revealed that 26.0% con- tained HBLV-TgG antibody when tested at 1:20 dilution. The mean titer of anti- body was 31:20, and 95% of these sera had a titer of 1:40. A higher prevalence of HBLV antibody was found in sera collected from West Africa (52%). Since the lowest antibody distribution was found in sera from healthy Malaysian donors (9%), this suggested major differences between regionally different donor populations (Ablashi, 1988b).

Sera obtained from patients with lymphoid malignancies (Hodgkin’s disease, African BL and acute lymphocytic leukemia) had a higher prevalence and titer of

44

HBLV antibody. The highest prevalence and titer of HBLV antibody were found in African patients with BL (86.7%) and Hodgkin’s disease (77.2%). The African BL sera showed that antibody titers of HBLV and EBV were almost identical (>640-2560). In contrast, HBLV prevalence was normal in NPC and infectious mononucleosis, both of which are EBV-associated diseases. Patients with other malignancies such as chronic lymphocytic leukemia, hairy cell leukemia and car- cinomas, did not show an increased prevalence of HBLV antibody titers (Ablashi, 1988b). The higher prevalence to HBLV antibody titers was obtained in patients with acute sarcoidosis (52%) and SjGgren’s syndrome (>55%). Sera from lupus and rheumatoid arthritis showed normal distributions. It was difficult to test Sjo- gren’s sera by IFA, because >XO% possessed anti-nuclear activity. Therefore, these sera were tested by RIP and by Western blot analysis, as described earlier.

The prevalence of HBLV antibody in people infected with the AIDS retrovirus (HIV-l antibody positive) was also significantly higher (70%) than in uninfected populations. Among the asymptomatic HIV-l positive individuals, 76% had HBLV antibody, with 90% of them having a titer of <1:80. Even though 46% of sera from HIV-l antibody negative homosexual men contained HBLV antibody; their titers were <1:40. Titers of all groups of symptomatic HIV-l antibody positive individ- uals ranged from l:SO->1:640 (Ablashi, 1988b). These data seem to indicate an increased prevalence of HBLV infections in all stages of HIV-1 disease develop- ment. Since HBLV can apparently infect both fresh and established T-cells, pos- sessing CD4 and other receptors, HBLV may further impair immune function in HIV-l-infected people. Secondly, T-cell cultures infected by HBLV and HIV-l show enhanced degeneration of cells. Thirdly, except for a few isolations of HBLV from CFS patients, all other HBLV isolates were reported from AIDS. The effects of both HBLV and HIV-l infections in the same patient are under investigation.

The epidemiology of a cluster of 134 patients from Lake Tahoe, Nevada, with CFS suggested an epidemic disease with a possible viral etiology. CFS is also known as chronic EBV, because of the presence of elevated antibody titers to EBV-VCA, EA and EBNA-2 in some patients exhibiting fatigue, exhaustion, low grade fever, parethesias, headache, joint and muscle ache, and many with mental changes (confusion, lack of concentration). However, a causative link between EBV and CFS has not yet been found (Holmes, 1987). In 1987, we began to explore the possibility that HBLV could be a cofactor in CFS. The first group of patients tested for HBLV were those being studied by Dr. Komaroff from Harvard Medical School and Dr. Paul Cheney of Lake Tahoe Clinic. Besides the CFS cases from Lake Tahoe, sporadic incidences of CFS have been reported from other parts of the U.S.A. and Europe (Krueger, 1987a.b). The aggregate seroprevalence of HBLV in 332 CFS patients from Lake Tahoe, Boston, Bethesda and other parts of the U.S.A. was >80%. HBLV antibody in control populations from the above-men- tioned areas, ranged between 3545%. There was a marked difference in the range of antibody titers in CFS sera (1:8&>2560). The antibody titers from the majority of controls ranged between 1:20-(1: 160. We were successful in isolating HBLV from two HBLV seropositive CFS patients. The peripheral blood lymphocytes from these patients, after PHA stimulation, exhibited large cells and numbered be-

45

tween l&20%. The cell-free supernatant from infected cells induced a similar

population of large cells in PHA stimulated cord blood mononuclear cells. These cells were IFA positive with HBLV antibody, and electron microscopic examina- tion showed herpesvirus particles. Further immunovirological analyses of these isolates are currently under investigation.

While elevated anti-HBLV titers are not unique to CFS patients. the potential for latent infection, the cytopathic nature of the virus and the diverse cell tropism suggest that HBLV has a possible pathogenic role in at least some patients with persistent post-viral syndrome.

Detection of HBLV-DNA in human tumors As part of the investigations into the possible association of HBLV with disease,

we examined DNA from a variety of malignant and benign tumors (Table 3), in- cluding B- and T-cell lymphomas, Kaposi’s sarcomas and sarcoidosis for HBLV

sequences, by Southern blot hybridization, and by in situ hybridization using

HBLV-DNA probe ZVH14. BamV fragment of EBV, which was kindly supplied by Dr. E. Kieff, of the Harvard University Medical School, was also used as a probe for the detection of EBV-DNA. HBLV sequences were found in an EBV-DNA positive BL from a five-year old African male from Ghana (Table 3). The restric-

tion patterns with EcoRl, Hind111 and BamH were identical to the prototype HBLV (GS) from which the probe was derived (Josephs, 1988). The EBV se- quences in this tumor were approximately 15 times more prevalent than HBLV sequences. The HBLV antibody titer in this patient was >1:640, and a similar titer to EBV-VCA antibody was detected (>l: 1280). Seven out of ten African BL tu-

TABLE 3

Detection of HBLV-genome in human tumors” and tissues obtained from scarcoidosis patients

Tumors

Large follicular lymphoma” Sjogren’s Syndrome B-cellh’

Lymphoma (small cleaved cell)

Burkitt’s Lymphoma’ Kaposi Sarcoma endothelial cells from AIDS

CML

T-cell lymphoma

Southern blot In situ

hybridization

114 HBLV antibody N.D.

218 l/l 117 7110d

017 N.D.

015 N.D.

O/l 1 N.D. Sarcoidosi9 N.D. 113

*’ For hybridization, PZVH14 was used.

h No EBV-DNA was detected. ‘ Both EBV and HBLV-DNA were detected, EBV-DNA was 10 times more than HBLV-DNA.

d Sarcoidosis tissues positive by in situ for HBLV-DNA contained <5 cells per field. * Formaline fixed sections of tumor tissue contained clusters of cells (2-10) positive for HBLV, how-

ever, the ratio of positive to negative cells was 1 to >lOOO. EBV-DNA was also present. The BL tumor, found to possess HBLV-DNA by Southern blot hybridization, was also positive by in situ.

46

mors positive for EBV-DNA also showed HBLV-DNA by in situ hybridization us- ing PZVH14 (Table 3). The sera from these patients contained high titers of HBLV (>1:440) and EBV-VCA antibody (>1:640). The DNA hyb~dization for EBV and HBLV together with elevated antibody to both herpesviruses suggest that HBLV may be another important co-factor for BL. The other two tumors found positive for HBLV by Southern blot hybridization were a B-cell lymphoma from a 75year- old white female with Sjogren’s syndrome and a large B-cell follicular lymphoma from a 55-year-old white male. Neither of these B-cell lymphomas hybridized to EBV BumV fragment. DNA extracted from other tissues of the patient with Sjo- gren’s syndrome (liver, parotid glands, peripheral blood) failed to show any HBLV- DNA (Josephs, 1988). The sera from both of these patients had elevated HBLV antibody (>1:160). Dr. R. Jarrett also found HBLV-DNA in a B-cell lymphoma from a Sjogren’s syndrome patient. Dr. Peter Biberfeld also detected, by in situ hybridization, HBLV-DNA in tissues from an HBLV antibody positive sarcoidosis patient. Two other sarcoidosis tissues from HBLV antibody negative patients, were found to be free of HBLV-DNA (Biberfeld et al., 1988). The role of HBLV in these tumors is unclear.

Concluding remarks The present findings of elevated HBLV antibody in certain non-malignant dis-

eases and the detection of HBLV-DNA in some B-cell lymphomas as well as im- pressive in vitro tropism of HBLV for B- and T-cells, megakaryocytes and glio- blastoma cell lines and its synergistic cytopathic effect on HIV infected CD4 cells, raise the possibility of a role for this virus in some immune-suppressive disorders such as AIDS and in lymphoproliferative disorders.

Although HBLV had been found to be a lytic virus in vitro, the possibility should be considered that, under the right circumstances, HBLV infection could result in a selective clonal proliferation of cells resistant to the cytopathic effects of the vi- rus and/or infection with replication of defective virions. If the lytic phase is pre- vented, for example, by virus mutation of delection of part of the genome, a direct growth-promoting effect may occur. Thus, some chain of events could give rise to in vivo malignancy.

We thank Mrs. Kristine L. Abiashi for her assistance in the critical review of this manuscript.

References

Ablashi, D.V., Salahuddin, S.Z., Joseph% S.F., Imam, F., Lusso, P., Gallo, R.C., Hung, CL... Lemp. J. and Markham, P.D. (1987) HBLV (or HHV-6) in human cells. Nature (London) 329, 207.

Ablashi, D.C. t Lusso, P., Hung, C.L., Salahuddin, S.Z., Josephs, S.F., Llana, T., Kramarsky, B., Biberfeld. P.. Markham, P.D. and Gallo, R.C. (1988a) Utilization of human hematopoietic cell lines

37

for the propagation, detection and characterization of HBLV (Human Herpesvirus-6). Int. J. Can-

cer (in press). Ablashi, D.V., Kaplan, M., Saxinger, C.. Salahuddin, S.Z.. Levine, P.H., Streicher, H.Z., Krueger,

G.R.F., Josephs. S.F.. Dalgleish, A.G., Imam, F.. Biberfeld, P. and Gallo. R.C. (1988b) Anti- body to HBLV (Human Herpesvirus-6) in lymphoproliferativ~ diseases and in healthy subjects: a preliminary survey. Science, submitted.

Biberfeld, P., Kramarsky, B., Salahuddin, S.Z. and Gailo, R.C. (1987) Ultrastructure characterization of a new B-lymphotropic DNA virus (HBLV) isolated from patients with lymphoproliferative dis-

ease. J. Natl. Cancer Inst. 79, 933-941.

Biberfeld, P., Petren, A.-L., Ehlund, A., Lindemalm, A., Barkhem, T., Ekman, M., Ablashi, D. and

Salahuddin, Z. (1988) Human herpesvirus (HHV-6, HBLV) in sarcoidosis and lymphoprolife~ative

disorders. J. Viral. Methods 21, 49-59.

Dahlberg, J.E., Zintz, E. and Ablashi. D.V. (1985) Analysis of Herpesvirus saintin‘ structural proteins

and monoclonal antibodies. J. Virol. 53. 279-286.

Downing, R.C., Sweankambo. N., Serwadda, D., Honess, R., Crawford, I)., Jarrett, R. and Griffin.

B. (1987) Isolation of human lymphotropic herpesvirus from Uganda. Lancet II. 390.

Efstathiou, S.. Gompels, U.A., Craxton, M.A.. Honess, R.S. and Ward, K. (1988) DNA homology between a novel human herpesvirus (HHV-6) and human cyto-megalovirus. Lancet I, 8575-8576.

Gershoni, J.M. and Plade, G.E. (1982) Electrophoretic transfer of proteins from SRS-polyacrylamide

gels to a positivity charged membrane filter. Anal. Biochem. 124, 396405.

Holmes, G.P., Kaplan. J.E., Steward, J.A., Hunt, B., Pinsky. R.F. and Schonbergcr, L.B. (1987) A

cluster of patients with a chronic mononucleosis syndrome: Epstein-Barr Virus the cause? J. Am.

Med. Assoc. 257. 2297-2302.

Josephs, S.F., Salahuddin, S.Z., Ablashi, D.V., Schachter, F., Wong-Staal. F. and Gallo. R.C. (1986) Genomic analysis of the human B-lymphotropic virus. Science 234, 601-603.

Josephs. S.F., Buchbinder, A., Streicher, H.. Ablashi. D.V., Salahuddin, S.Z., Guo. H.G.. Wong-

Staal, F., Cossman, J., Rafeld, M.. Sundeen. J., Levine. P.H.. Biggar. B., Krueger, G.R.F.. Fox.

R. and Gallo. R.C. (1988) Detection of human B-lymphotropic virus (Human Herpesvirus 6) se-

quence in B-cell lymphoma tissue of three patients. Leukemia 2. 132-135.

Kishi, M., Harada, H.. Takahashi, M., Tanaka, A., Hayashi, M., Nonoyama, M., Josephs, S.. Buch-

binder, A., Schachter, F., Ablashi, D.V., Wong-Staal, F., Salahuddin, S.Z. and Gallo. R.C. (1988)

A repeat sequence GGGTTA is shared by DNA of Human B-lymphotropic virus (HI-IV-6 or HBLV)

and Marek’s Disease Virus (MDV). J. Virol. (in press).

Krueger, G.R.F., Koch, B. and Abiashi. D.V. (1987) Persistent fatigue and depression in patients with antibody to human B-iymphotropic virus. Lancet II, 36.

Krueger, G.R.F. and Ablashi, D.V. (1987) Human B-lymphotropic virus in Germany. Lancet II, 694.

Lusso, P., Salahuddin, S.Z.. Ablashi, D.V., Gallo. R.C., Veronese. F.D. and Markham, P.D. (1987) Diverse tropism of HBLV (human herpesvirus-6). Lancet II. 743-744.

Miller, G.,(1985) Biology of Epstein-Barr virus. In: B.N. Fields, D.M. Knipe, J.S. Melnick. R.M.

Chanock, B. Roizman and R.E. Shape (Eds.). Virology, pp. 497-526. Raven Press, New York.

Pearson. G.R., Luka, J. and Docherty, J. (1986) Monoclonal antibodies in the diagnosis of EBV-in-

duced disease. In: C. Lopez and B. Roizman (Eds.), Human Herpesvirus Infections, pp. 21 l-220.

Raven Press, New York.

Roizman, B. and Spear. P.G. (1973) Herpesviruses. In: Ultrastructure of Animal Viruses and Bacte-

riophages. Salahuddin, S.Z., Ablashi, D.V., Markham, P.D.. Josephs, S.F.. Sturzenegger, S., Kaplan, M., Hal-

ligan, G., Biberfeld, P.. Wong-Staal, F., Kramarsky, B. and Gallo, R.C. (1986) Isolation of a new virus, HBLV, in patients with lymphoproliferative disorders. Science 234, 596-601.

Salahuddin, S.Z., Ablashi, D.V., Hunter, E.A.. Gonda. M.A., Sturzenegger. S., Markham, P.D. and

Gallo, R.C. (1987) HTLV-III infection of EBV genome-positive B-lymphoblastoid cells with or without detectable TJ antigen. Int. J. Cancer. 39. 198-202.

Saxinger, C., Poleski, H., Eloy, N., Grufferman, S., Murphy, R.. Tegtmeir. G., Parekh, V., Memon, S. and Hung, C. (1988) Antibody reactivity with HBLV (HHV-6) in U.S. populations. J. Virol.

Methods 21, 199-208.

4x

Tedder. R.S.. Cammeron. C.H.. Iltoness. R.. Robertson. D. and Whittle. H. (1987) A novel Iym- photropic virus. Lancet II. 39&393.

Towbin. H., Staehelin, T.. Gordon. J. (1979) Electrophoretic transfer of proteins from polyacrylamide

gel to microcellulose sheets. Proc. Natl. Acad. Sci. USA 76. 435&4354.