Human Immunodeficiency Virus Activates c-Myc and Epstein-Barr Virus in Human B Lymphocytes SUSAN M. ASTRIN AND JEFFREY LAURENCE aThe Institute for Cancer Research Fox Chase Cancer Center 7101 Burholme Avenue Philadelphia, Pennsylvania 19111 bDepartment of Medicine Cornell University Medical College 41 1 East 69th Street New York. New York 10021
INTRODUCIION AIDS-associated malignancies have become increasingly prevalent among HIV- 1 infected individuals, with B-cell lymphoma' and Kaposi's sarcoma2causing significant numbers of deaths in this population. Although published data do not indicate a direct role for virus in the transformation of B cells,)-' lymphadenopathy, polyclonal B-cell proliferation, and even lymphoma may precede overt compromise of T-cell i m m ~ n i t y . ~ . ~ It thus seems possible that these conditions might result from direct infection of a subset of B cells by HIV. We therefore sought to investigate more closely the interaction between primary B lymphocytes and HIV-1.
MATERIAL AND METHODS Cells
Peripheral blood mononuclear cells were isolated from heparinized blood by FicollHypaque density gradient centrifugation. Monocytes were depleted by adherence to plastic; B cells were purified by rosetting with neuraminidase-treated sheep erythrocytes for 45 min at 4 "C, followed by density gradient centrifugation and collection of the erythrocyte-rosette negative fraction. Using this procedure the B-lymphocyte pool typically contains < 1% CD3' T cells, 95% surface Ig+ cells. Immunojluorescence Assays for Membrane Antigens
Expression of viral and cellular surface antigens was assayed by indirect immunofluorescence and cytofluorimetry using an EPICS V cell sorter. Monoclonal antibodies (mAb) used included anti-CD3 (mAb 454.3), anti-LMP (generously supplied by D.A. Thorley-Lawson), and anti-CD2 1 (B2, Epstein-Barr virus (EBV) receptor). Counterstaining was performed using fluorescein-conjugated F(ab), fragments of goat anti422
ASTRIN & LAURENCE: c-M&
423
mouse IgG (for CD3 and LMP) or IgM (for CD21). Surface immunoglobulin was directly detected with fluoresceinated goat anti-human IgG, IgM, and IgA. Detection of EB V
In situ hybridization for EBV nucleic acid was performed using a cosmid clone, pM966/20,I0 containing approximately 25 kb of EBV sequence, encompassing 5' leader sequences, and the gene for latent membrane protein (LMP). It was derivitized with digoxigenin-11-dUTP. Procedures for DNA probe labeling by the random primer method, paraformaldehyde cell fixation, hybridization, and detection have been described elsewhere." Detection of c-Myc
In situ hybridization was performed with a 1.5 kb DNA fragment containing the third exon of c-myc, I 2 labeled as described above for the EBV probe. c-Myc protein was detected using indirect immunofluorescence with rabbit polyclonal antibody to human c-myc and rhodamine-conjugated goat anti-rabbit immunoglobulin. c-Myc RNA was analyzed using Quick-Blots (Schleicher and Schuell, and N. H. Keene),'3 in which mRNA is immobilized onto nitrocellulose and hybridized with digoxigeninlabeled c-myc exon 3 probe, as described elsewhere.I4Controls for amount of poly-A containing RNA bound to filters were performed by rehybridizing blots, from which the c-myc probe had been eluted, with 3ZP-labeledoligo-dT. Assays for Malignant Transformation Growth in soft agar was assayed by embedding 5 X lo3 cells in 0.5 mL of RPMI1640 plus 20% fetal calf serum and 0.3% agarose. Duplicate plates were scored for colonies 12 days after seeding. Tumorigenicity assays in SCID mice were performed by subcutaneous injection of 10 million washed cells contained in 0.2 mL phosphate-buffered saline into 8- 10-weekold female CB-17 SCID mice. Animals were kept in an isolator for 12 weeks and examined daily for the appearance of tumors. All tumor masses were removed, fixed in Bouin's solution, stained with hematoxylin and eosin, and examined in blinded fashion by a single pathologist. Using this protocol, Burkitt lymphoma cell lines will elicit fully malignant tumors at the site of injection in 17-75% of SCID rni~e.'~-~' By contrast, whereas EBV-immortalized B-cell lines may produce tumors in SCID mice, these tumors, with one exception,'*do not fulfill the histologic, immunophenotypic, or in vitro clonigenic criteria for malignancy.I6*l7
RESULTS Immortalization of B Lymphocytes by Infection with HIV Ten million B cells were purified19 from peripheral blood leukocytes of an HIV-1seronegative, EBV-seropositive adult and exposed to 5 X 104 50% tissue-culture infectious doses of the HTLV-IIIB strain of HIV-1 for 2 hours at 37 "C, followed by washing
424
ANNALS NEW YORK ACADEMY OF SCIENCES
and resuspension in fresh culture medium. A second set of B cells from the same individual was infected with 1 X lo5 transforming units of EBV from supernatants of the B95-8 marmoset lymphoma cell line. A third set was exposed to buffer alone. Cells exposed to HIV- 1 exhibited polyclonal proliferation followed by outgrowth of an immortalized cell population after 21 days; this line was termed B-HIV. Similarly, an immortalized population emerged from the EBV-infected cells after 28 days; this line was termed B-EBV. The lines were negative for T-cell antigen CD3 and were positive for membrane IgM, indicating that they were both B-cell lines. Cells exposed to buffer alone failed to proliferate and died.
HIV Activation of EBV The production of a cell line from HIV-infected B lymphocytes was somewhat surprising; however, HIV has been reported to be infectious for certain B-cell lines and to activate endogenous EBV in those lines.*' Both B-HIV and B-EBV expressed the EBV-encoded gene product LMP and the EBV receptor (CD21). LMP expression is associated with immortalization of B lymphocytes. We thus postulated that HIV had entered a cell harboring EBV (the frequency of such cells in peripheral blood B cells is estimated to be 21-23), and that EBV had been activated and had spread to to other cells causing outgrowth of an immortalized cell population. Indeed, all B-HIV cells contained EBV nucleic acid, as evidenced by in situ hybridization using a probe containing 25 kb of EBV sequence, including the region encoding LMP (FIG. 1A). In addition, the levels of EBV RNA and DNA were significantly enhanced in B-HIV cells relative to levels in B-EBV cells (FIG.IB). As a further test of the ability of HIV-1 to activate EBV, B-EBV cells were infected with HIV-1 and levels of EBV nucleic acid assayed by in situ hybridization (FIG. 1C). Levels of EBV DNA and RNA were significantly upregulated in every cell of the HIV-infected population. Exposure of B-EBV cells to heat-inactivated HIV-1 did not result in upregulation of EBV.
Frequency of Immortalization
How frequently does exposure of primary B cells to HIV-1 result in an immortalized cell line? B cells from five different donors were exposed to HIV-1 as described above. All samples showed polyclonal proliferation for several weeks; an immortalized cell population emerged in two cases, including that described above. Several factors probably limit the production of a cell line, including the ability of HIV to enter a sufficient fraction of the cell population such that an EBV-harboring cell becomes infected. Further experimentation will be required to determine the mechanism and efficiency of entry of HIV into primary B lymphocytes. Analysis by polymerase chain reaction for the presence of HIV-1 genomes in the two immortalized cell populations revealed that one population contained, on average, one proviral copy per ce11,I9 whereas HIV was undetectable in the other population. Presumably the second population had been immortalized through activation and spread of EBV, but HIV-1-containing cells had not been selected for in the resultant cell line. Subsequent experiments were all performed with the cell line harboring both HIV and EBV.
ASTRIN & LAURENCE: C-MYC
425
FIGURE 1. Detection of EBV nucleic acids by in siru hybridization. B-HIV cells (A), B-EBV cells (B), and B-EBV cells exposed to HIV-I followed by incubation in fresh media for 3 days (C) were hybridized with an EBV DNA probe of 25 kb encompassing the region encoding LMP. Hybridization is detected by the formation of blue-black precipitates. The sensitivity of this method is such that the lower threshold for detection of transcripts or genomes is on the order of 100 per cell. Data from ref. 19.
Growth Parameters of B-HZV Assuming that the presence of the HIV genome might confer a special growth advantage on the B-HIV cell line, we tested the line for several parameters associated with malignant transformation: growth in 1% serum and the ability to clone in soft agar. After maintaining the cell line in 10% serum for 3 months, cells were grown in 1% fetal bovine serum (FBS) for 4 weeks, with repetitive Ficoll-Hypaque density gradient centrifugations to remove nonviable cells. A subline emerged, termed B-HIV 1; this subline was characterized further. B-HIV1 grew logarithmically in media containing 1% FBS, with kinetics similar to those of the EBV+ Burkitt lymphoma cell line, Raji; B-EBV cells did not exhibit any significant growth under these conditions
426
ANNALS NEW YORK ACADEMY OF SCIENCES
FIGURE 2. Growth of B-HIV and B-EBV cell lines in media containing 1 % fetal bovine serum (FBS). B-HIV1 cells (open circles) or B-EBV cells (closed circles) were suspended at a concentration of 5 x 105/mLin RPMI containing 1% FBS, and cells were counted daily for 5 days without replenishing the media. Data from ref. 19.
-
0
1
2
3
4
5
FIG.^). The ability to grow in low serum is one of the hallmarks of malignantly transformed cells; B-HIV1 cells have retained this property and have been in continuous culture in 1% FBS for more than seven months. In addition, the cells exhibit a second property associated with malignant transformation: the ability to form colonies in soft agar (FIG.3). Indeed, the cloning efficiency (ca. 1%) was similar to that of some Burkitt lymphoma cell lines.24B-EBV cells, on the other hand, formed no colonies in soft agar (FIG. 3). HIV Activation of c-Myc The malignant phenotype in human B cells is often associated with deregulated c-myc expression. Therefore, m y RNA and protein levels were assayed in B-HIV1. Slot blot filter hybridization for c-myc transcripts and immunohistochemical analysis for c-myc protein indicated that both myc RNA and protein levels were elevated by 10- to 20-fold compared to those in EBV-immortalized cells (FIG. 4, panels A, B, and C). In fact, myc RNA levels were very similar to those observed for Raji. Was elevated expression of myc a direct effect of the HIV genome, or was another mechanism responsible, such as chromosomal translocation? In order to shed light on this question, we analyzed the karyotype of B-HIV 1 by examining 75 metaphases. No translocations involving chromosome 8 were observed; the karyotype appeared completely normal. It thus seemed possible that c-myc overexpression was a direct result of HIV-1 infection. Several lines of evidence indicated that the population still harbored HIV genomes: analysis by polymerase chain reaction with env and pol primer pairs indicated the presence of HIV at one copy per cell,I9 electron microscopy indicated the presence of budding virions,25and virus could be induced by treatment with phorbol ester as assayed by particulate reverse transcriptase or p24 Gag protein.25 We thus chose to examine myc expression in B-EBV cells shortly after infection with HIV. Raji cells were chosen for these experiments because, in our hands, they are readily infectable by HIV; they have an elevated basal level of myc expression, making detection of elevated levels of transcripts possible by in situ hybridization; and, despite having a chromosomal translocation, myc promoter elements and > 1 kb of 5’ flanking sequences are normal in structure.26Three days after infection with HIV, 100% of Raji cells showed detectable increases in myc RNA expression as assayed by in situ
ASTRIN & LAURENCE C-MYC
421
X 106 B-EBV, B-HIVI, or Raji cells; serial fivefold dilutions were made; and mRNA was immobilized onto nitrocellulose using the Quick-Blot technique. Filters were probed with a digoxigenin-labeled c-myc DNA probe of 1.5 kb encompassing exon 3 (top lanes), then washed and probed with 32P-labeledoligo-dT to control for the amount of poly A containing mRNA in the samples. Panels B and C c-Myc protein was assayed by indirect immunofluorescence. B-HIV cells (B) show intense nuclear staining; B-EBV cells (C) show little staining.
FIGURE 4. c-Myc RNA and protein in B-HIV cells. Panel A Nucleic acids were extracted from 2.0
B-EBV B-HIVI Raji
B-EBV B-HIVI Raji
t 00
ASTRIN & LAURENCE: c-M~v
429
FIGURE 5. Detection of c - m y transcripts by in situ hybridization in Raji cells exposed to HIV. Raji cells were exposed to buffer (top) or to HIV-1 (bottom) for 2 hours at 37 "C, washed with phosphate-buffered saline, and incubated in fresh culture medium at 37 "C. In situ hybridization using a digoxigenin-labeled c-myc exon 3 probe was performed on day 3 after HIV exposure. Hybridization is detected by the formation of blue-black precipitates. The lower threshold for the detection of transcripts in this assay is on the order of several hundred transcripts per cell.
hybridization (FIG.5 , top and bottom); seven days after infection, an 8- to 16-fold enhancement in message levels was observed by filter hybridization (data not shown). These findings are consistent with the interpretation that HIV itself is able to upregulate c-myc. Tumorigenicity of B-HIVl Cells
Because 3-HIV1 cells possess two hallmarks of malignantly transformed cells, growth in 1% serum and cloning in soft agar, and, in addition, overexpress the c-myc proto-oncogene, we decided to test their malignant potential in SCID (severe combined immunodeficient) mice. Ten million B-HIV 1 cells were inoculated subcutaneously into each of four SCID mice; ten million B-EBV cells were inoculated into each of nine
430
ANNALS NEW YORK ACADEMY OF SCIENCES
FIGURE 6. Histology of a tumor induced by B-HIV cells in a SCID mouse. Excised tumor tissue was fixed in Bouin’s solution, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Examination under 100 X (top) and 430 x (bottom) initial magnifications revealed a malignant, invasive lymphoma of immunoblastic phenotype.
control SCID mice. Mice were observed for tumor formation over a period of 6 weeks. Two of four mice inoculated with B-HIV1 cells developed masses > 2 cm in diameter by 4 weeks; histology revealed a highly malignant invasive lymphoma of immunoblastic phenotype (FIG.6). None of the mice inoculated with B-EBV cells developed tumors within the observation period; however, between 8 and 12 weeks, 7 of the 9 mice developed masses, which upon histologic examination were classified as noninvasive plasmacytomas.
SUMMARY We have established a line of malignantly transformed human B cells by infecting purified primary B lymphocytes with human immunodeficiency virus type 1 (HIV-1).
ASTRIN & LAURENCE C-MYC
431
This line, termed B-HIVl, may serve as a model system for a subset of AIDS-related B-cell lymphomas in which the transformed phenotype may be initiated and/or maintained through an HIV-1 gene product. The B-HIV1 line contains both Epstein-Barr virus (EBV) and HIV-1 genomes. In addition, the c-myc gene is expressed at levels 10 to 20 times those in normal B cells. Similarly, EBV sequences, including those for the latent membrane protein (LMP), are expressed at greatly enhanced levels relative to expression in normal, EBV-immortalized B cells. The upregulation of c-myc and of EBV gene expression can both be produced by infection of susceptible B cells (not already harboring HIV) with exogenous HIV-1. The B-HIV1 line exhibits properties of malignantly transformed cells, in that it grows logarithmically in 1% serum, clones in soft agar, and produces invasive, malignant B-cell lymphomas in severe combined immunodeficient (SCID) mice. We have shown that HIV-1 has the ability to infect primary human B cells and to activate expression of EBV and c-myc. HIV activation of EBV has been documented previously in certain cell lines;20here we note that such activation can occur in primary B cells and, under certain conditions, can result in outgrowth of immortalized cell lines. This phenomenon may contribute to the clinical manifestation of lymphadenopathy early after infection with HIV. In addition, we have demonstrated that HIV infection of primary B cells in v i m can result in appearance of a fully malignant phenotype. This phenotype is likely to be due, at least in part, to the activation of c-myc by HIV. Preliminary experiments indicate that Tat,27the gene product of the transactivator of viral gene transcription tat, can upregulate c-myc transcription after addition to the culture media of certain B-cell lines. This raises the possibility that Tat can bind to target sequences in cellular RNA and enhance transcription as it does for HIV. It also raises the possibility that a necessary, but perhaps not sufficient, component for the expression of the malignant phenotype in certain AIDS-related lymphomas may be the presence and expression of the Tat gene product of HIV. Although published data would appear to rule out the presence of intact HIV genomes in the majority of AIDS-related the possibility of the presence of highly defective viral genomes has not been addressed.
REFERENCES
1. LEVINE,A. M. 1990. Semin. Oncol. 17: 104-121. 2. RABKIN,C. S., R. J. BIGGAR & J. W. HORM.1991. Int. J. Cancer 47: 692-695. 3. PELICCI,P. G., D. M. KNOWLES11, Z. ARLIN,R. WIECZOREK, P. LUCIW,D . DINA,C. BASILICO & R. DALLA-FAVERA. 1986. J. Exp. Med. 164: 2049-2076. 4. GROOPMAN, J. E., J. C. SULLIVAN, C. MULDER,D. GINSBURG, S. H. ORKIN,C. J. O'HARA, K. FALCHUK, F. WONG-STAAL & R. C. GALLO.1986. Blood 67: 612-615. 5. RECHARI, G., I. BEN-BASSAT,M. BERKOWICZ, U. MARTINOWITZ, F. BROK-SIMONI, Y. NEUMANN, A. VANSOVER, T. GOTLIEB-STEMATSKY & B. RAMOT.1987. Blood 70: 17131717. 6. SHIBATA, D., R. K. BRYNES,B. NATHWANI, S.KWOK,J. SHINSKY & N . ARNHEIM. 1989. Am. J. Pathol. 135: 697-702. 7. SELLERI, L., R. MARASCA, M. LUPPI,M. MONTOSSI & L. CECCHERINI-NELLI. 1990. AIDS Res. Hum. Retroviruses 6 149-150. 8. NERI,A., F. BARRIGA, D. M. KNOWLES,I. T. MCGRATH& R. DALLA-FAVERA. 1988. Proc. Natl. Acad. Sci. USA 8 5 2748-2752. 9. BECKHARDT, R. N., N. FARADY,M. MAY,R. A. TORRES& J. A. STRAUCHEN. 1988. Mt. Sinai J. Med. 55: 383. 10. POLACK,A,, G. HARTL,U. ZIMBER,U-K FREESE,G. LAWN,K. TAKABI,B. HOHN, L. GISSMAN& G. W. BORNKAUM. 1984. Gene 27: 279-288.
432
ANNALS NEW YORK ACADEMY OF SCIENCES
11. LAURENCE, J., H. CWKE & S. K. SIKDER.1990. Blood 7 5 696-703. 12. ROTHBERG, P. G., M. D. ERISMAN, R. E. DIEHL,U. G. ROVIGATTI & S. M. ASTRIN.1984. Mol. Cell. Biol. 4 1096-1103. 13. BRESSER, J., H. HUBBELL & D. GILLESPIE.1983. Proc. Natl. Acad. Sci. USA 8 0 65236527. 14. MUHLEGGER, K.,H-G BETZ, S. BOHM,H. V. D. ELTZ, H-J HOLTZKE& C. KESSLER. 1989. Nucleosides & Nucleotides 8 1161-1 166. 15. GHETIE,M-A, J. RICHARDSON, T. TUCKER,D. JONES,J. W. UHR& E. S. V I T E ~ A 1990. . Int. J. Cancer 4 5 481-485. 16. OKANO, M., Y. TAGUCHI, H. NAKAMIRE, S. J. PIRRUCCELLO, J. R. DAVIS,K. W. BEISEL, W. G. SANGER,R. R. FORDYCE & D. T.PURTILO.1990. Am. J. K. L. KLEVELAND, Pathol. 137: 517-522. 17. ROWE,M., L. S. YOUNG,J. CROCKER, H. STOKES,S. HENDERSON & A. B. RICKINSON. 1991. J. Exp. Med. 173: 147-158. 18. MOSIER,D. E., R. J. GULIZIA, S. M. BAIRD,D. D. RICHMAN, D. B. WILSON,R.I. Fox & T. J. KIPPS. 1989. Blood 74 (Suppl. 1): 52A. 19. LAURENCE, J. & S. M. ASTRIN.1991. Proc. Natl. Acad. Sci. USA. 88: 7635-7639. 20. LAI,P., X. LI & D. VOLSKY. 1989. Int. J. Cancer 43: 1104. 21. ANDIMAN, W. A., K. MARTIN,A. RUBENSTEIN, S. PAHWA,R. EASTMAN, B. 2.KATZ,J. P i n & G. MILLER.1985. Lancet 2 1390-1393. 22. BIRX,D. L., R. R. REDFIELD& G. TOSATO.1986. N. Engl. J. Med. 314 874-879. R., R. R. REDFIELD& S. BRODER.1986. J. Clin. Invest. 7 8 439-447. 23. YARCHOAN, 24. LOMBARDI, L., E. W. NEWCOMB & R. DALLA-FAVERA. 1987. Cell 49: 161-170. 25. LAURENCE, J., B. GRIMISON & S. M. ASTRIN.Submitted for publication. 26. RABBITTS, T. H., P. H. HAMLYN & R. BAER.1983. Nature 306 760-765. 27. FRANKEL, A. D., S. BIANCALANA & D. HUDSON.1989. Proc. Natl. Acad. Sci. USA 8 6 7397-7401.
INTRODUCIION AIDS-associated malignancies have become increasingly prevalent among HIV- 1 infected individuals, with B-cell lymphoma' and Kaposi's sarcoma2causing significant numbers of deaths in this population. Although published data do not indicate a direct role for virus in the transformation of B cells,)-' lymphadenopathy, polyclonal B-cell proliferation, and even lymphoma may precede overt compromise of T-cell i m m ~ n i t y . ~ . ~ It thus seems possible that these conditions might result from direct infection of a subset of B cells by HIV. We therefore sought to investigate more closely the interaction between primary B lymphocytes and HIV-1.
MATERIAL AND METHODS Cells
Peripheral blood mononuclear cells were isolated from heparinized blood by FicollHypaque density gradient centrifugation. Monocytes were depleted by adherence to plastic; B cells were purified by rosetting with neuraminidase-treated sheep erythrocytes for 45 min at 4 "C, followed by density gradient centrifugation and collection of the erythrocyte-rosette negative fraction. Using this procedure the B-lymphocyte pool typically contains < 1% CD3' T cells, 95% surface Ig+ cells. Immunojluorescence Assays for Membrane Antigens
Expression of viral and cellular surface antigens was assayed by indirect immunofluorescence and cytofluorimetry using an EPICS V cell sorter. Monoclonal antibodies (mAb) used included anti-CD3 (mAb 454.3), anti-LMP (generously supplied by D.A. Thorley-Lawson), and anti-CD2 1 (B2, Epstein-Barr virus (EBV) receptor). Counterstaining was performed using fluorescein-conjugated F(ab), fragments of goat anti422
ASTRIN & LAURENCE: c-M&
423
mouse IgG (for CD3 and LMP) or IgM (for CD21). Surface immunoglobulin was directly detected with fluoresceinated goat anti-human IgG, IgM, and IgA. Detection of EB V
In situ hybridization for EBV nucleic acid was performed using a cosmid clone, pM966/20,I0 containing approximately 25 kb of EBV sequence, encompassing 5' leader sequences, and the gene for latent membrane protein (LMP). It was derivitized with digoxigenin-11-dUTP. Procedures for DNA probe labeling by the random primer method, paraformaldehyde cell fixation, hybridization, and detection have been described elsewhere." Detection of c-Myc
In situ hybridization was performed with a 1.5 kb DNA fragment containing the third exon of c-myc, I 2 labeled as described above for the EBV probe. c-Myc protein was detected using indirect immunofluorescence with rabbit polyclonal antibody to human c-myc and rhodamine-conjugated goat anti-rabbit immunoglobulin. c-Myc RNA was analyzed using Quick-Blots (Schleicher and Schuell, and N. H. Keene),'3 in which mRNA is immobilized onto nitrocellulose and hybridized with digoxigeninlabeled c-myc exon 3 probe, as described elsewhere.I4Controls for amount of poly-A containing RNA bound to filters were performed by rehybridizing blots, from which the c-myc probe had been eluted, with 3ZP-labeledoligo-dT. Assays for Malignant Transformation Growth in soft agar was assayed by embedding 5 X lo3 cells in 0.5 mL of RPMI1640 plus 20% fetal calf serum and 0.3% agarose. Duplicate plates were scored for colonies 12 days after seeding. Tumorigenicity assays in SCID mice were performed by subcutaneous injection of 10 million washed cells contained in 0.2 mL phosphate-buffered saline into 8- 10-weekold female CB-17 SCID mice. Animals were kept in an isolator for 12 weeks and examined daily for the appearance of tumors. All tumor masses were removed, fixed in Bouin's solution, stained with hematoxylin and eosin, and examined in blinded fashion by a single pathologist. Using this protocol, Burkitt lymphoma cell lines will elicit fully malignant tumors at the site of injection in 17-75% of SCID rni~e.'~-~' By contrast, whereas EBV-immortalized B-cell lines may produce tumors in SCID mice, these tumors, with one exception,'*do not fulfill the histologic, immunophenotypic, or in vitro clonigenic criteria for malignancy.I6*l7
RESULTS Immortalization of B Lymphocytes by Infection with HIV Ten million B cells were purified19 from peripheral blood leukocytes of an HIV-1seronegative, EBV-seropositive adult and exposed to 5 X 104 50% tissue-culture infectious doses of the HTLV-IIIB strain of HIV-1 for 2 hours at 37 "C, followed by washing
424
ANNALS NEW YORK ACADEMY OF SCIENCES
and resuspension in fresh culture medium. A second set of B cells from the same individual was infected with 1 X lo5 transforming units of EBV from supernatants of the B95-8 marmoset lymphoma cell line. A third set was exposed to buffer alone. Cells exposed to HIV- 1 exhibited polyclonal proliferation followed by outgrowth of an immortalized cell population after 21 days; this line was termed B-HIV. Similarly, an immortalized population emerged from the EBV-infected cells after 28 days; this line was termed B-EBV. The lines were negative for T-cell antigen CD3 and were positive for membrane IgM, indicating that they were both B-cell lines. Cells exposed to buffer alone failed to proliferate and died.
HIV Activation of EBV The production of a cell line from HIV-infected B lymphocytes was somewhat surprising; however, HIV has been reported to be infectious for certain B-cell lines and to activate endogenous EBV in those lines.*' Both B-HIV and B-EBV expressed the EBV-encoded gene product LMP and the EBV receptor (CD21). LMP expression is associated with immortalization of B lymphocytes. We thus postulated that HIV had entered a cell harboring EBV (the frequency of such cells in peripheral blood B cells is estimated to be 21-23), and that EBV had been activated and had spread to to other cells causing outgrowth of an immortalized cell population. Indeed, all B-HIV cells contained EBV nucleic acid, as evidenced by in situ hybridization using a probe containing 25 kb of EBV sequence, including the region encoding LMP (FIG. 1A). In addition, the levels of EBV RNA and DNA were significantly enhanced in B-HIV cells relative to levels in B-EBV cells (FIG.IB). As a further test of the ability of HIV-1 to activate EBV, B-EBV cells were infected with HIV-1 and levels of EBV nucleic acid assayed by in situ hybridization (FIG. 1C). Levels of EBV DNA and RNA were significantly upregulated in every cell of the HIV-infected population. Exposure of B-EBV cells to heat-inactivated HIV-1 did not result in upregulation of EBV.
Frequency of Immortalization
How frequently does exposure of primary B cells to HIV-1 result in an immortalized cell line? B cells from five different donors were exposed to HIV-1 as described above. All samples showed polyclonal proliferation for several weeks; an immortalized cell population emerged in two cases, including that described above. Several factors probably limit the production of a cell line, including the ability of HIV to enter a sufficient fraction of the cell population such that an EBV-harboring cell becomes infected. Further experimentation will be required to determine the mechanism and efficiency of entry of HIV into primary B lymphocytes. Analysis by polymerase chain reaction for the presence of HIV-1 genomes in the two immortalized cell populations revealed that one population contained, on average, one proviral copy per ce11,I9 whereas HIV was undetectable in the other population. Presumably the second population had been immortalized through activation and spread of EBV, but HIV-1-containing cells had not been selected for in the resultant cell line. Subsequent experiments were all performed with the cell line harboring both HIV and EBV.
ASTRIN & LAURENCE: C-MYC
425
FIGURE 1. Detection of EBV nucleic acids by in siru hybridization. B-HIV cells (A), B-EBV cells (B), and B-EBV cells exposed to HIV-I followed by incubation in fresh media for 3 days (C) were hybridized with an EBV DNA probe of 25 kb encompassing the region encoding LMP. Hybridization is detected by the formation of blue-black precipitates. The sensitivity of this method is such that the lower threshold for detection of transcripts or genomes is on the order of 100 per cell. Data from ref. 19.
Growth Parameters of B-HZV Assuming that the presence of the HIV genome might confer a special growth advantage on the B-HIV cell line, we tested the line for several parameters associated with malignant transformation: growth in 1% serum and the ability to clone in soft agar. After maintaining the cell line in 10% serum for 3 months, cells were grown in 1% fetal bovine serum (FBS) for 4 weeks, with repetitive Ficoll-Hypaque density gradient centrifugations to remove nonviable cells. A subline emerged, termed B-HIV 1; this subline was characterized further. B-HIV1 grew logarithmically in media containing 1% FBS, with kinetics similar to those of the EBV+ Burkitt lymphoma cell line, Raji; B-EBV cells did not exhibit any significant growth under these conditions
426
ANNALS NEW YORK ACADEMY OF SCIENCES
FIGURE 2. Growth of B-HIV and B-EBV cell lines in media containing 1 % fetal bovine serum (FBS). B-HIV1 cells (open circles) or B-EBV cells (closed circles) were suspended at a concentration of 5 x 105/mLin RPMI containing 1% FBS, and cells were counted daily for 5 days without replenishing the media. Data from ref. 19.
-
0
1
2
3
4
5
FIG.^). The ability to grow in low serum is one of the hallmarks of malignantly transformed cells; B-HIV1 cells have retained this property and have been in continuous culture in 1% FBS for more than seven months. In addition, the cells exhibit a second property associated with malignant transformation: the ability to form colonies in soft agar (FIG.3). Indeed, the cloning efficiency (ca. 1%) was similar to that of some Burkitt lymphoma cell lines.24B-EBV cells, on the other hand, formed no colonies in soft agar (FIG. 3). HIV Activation of c-Myc The malignant phenotype in human B cells is often associated with deregulated c-myc expression. Therefore, m y RNA and protein levels were assayed in B-HIV1. Slot blot filter hybridization for c-myc transcripts and immunohistochemical analysis for c-myc protein indicated that both myc RNA and protein levels were elevated by 10- to 20-fold compared to those in EBV-immortalized cells (FIG. 4, panels A, B, and C). In fact, myc RNA levels were very similar to those observed for Raji. Was elevated expression of myc a direct effect of the HIV genome, or was another mechanism responsible, such as chromosomal translocation? In order to shed light on this question, we analyzed the karyotype of B-HIV 1 by examining 75 metaphases. No translocations involving chromosome 8 were observed; the karyotype appeared completely normal. It thus seemed possible that c-myc overexpression was a direct result of HIV-1 infection. Several lines of evidence indicated that the population still harbored HIV genomes: analysis by polymerase chain reaction with env and pol primer pairs indicated the presence of HIV at one copy per cell,I9 electron microscopy indicated the presence of budding virions,25and virus could be induced by treatment with phorbol ester as assayed by particulate reverse transcriptase or p24 Gag protein.25 We thus chose to examine myc expression in B-EBV cells shortly after infection with HIV. Raji cells were chosen for these experiments because, in our hands, they are readily infectable by HIV; they have an elevated basal level of myc expression, making detection of elevated levels of transcripts possible by in situ hybridization; and, despite having a chromosomal translocation, myc promoter elements and > 1 kb of 5’ flanking sequences are normal in structure.26Three days after infection with HIV, 100% of Raji cells showed detectable increases in myc RNA expression as assayed by in situ
ASTRIN & LAURENCE C-MYC
421
X 106 B-EBV, B-HIVI, or Raji cells; serial fivefold dilutions were made; and mRNA was immobilized onto nitrocellulose using the Quick-Blot technique. Filters were probed with a digoxigenin-labeled c-myc DNA probe of 1.5 kb encompassing exon 3 (top lanes), then washed and probed with 32P-labeledoligo-dT to control for the amount of poly A containing mRNA in the samples. Panels B and C c-Myc protein was assayed by indirect immunofluorescence. B-HIV cells (B) show intense nuclear staining; B-EBV cells (C) show little staining.
FIGURE 4. c-Myc RNA and protein in B-HIV cells. Panel A Nucleic acids were extracted from 2.0
B-EBV B-HIVI Raji
B-EBV B-HIVI Raji
t 00
ASTRIN & LAURENCE: c-M~v
429
FIGURE 5. Detection of c - m y transcripts by in situ hybridization in Raji cells exposed to HIV. Raji cells were exposed to buffer (top) or to HIV-1 (bottom) for 2 hours at 37 "C, washed with phosphate-buffered saline, and incubated in fresh culture medium at 37 "C. In situ hybridization using a digoxigenin-labeled c-myc exon 3 probe was performed on day 3 after HIV exposure. Hybridization is detected by the formation of blue-black precipitates. The lower threshold for the detection of transcripts in this assay is on the order of several hundred transcripts per cell.
hybridization (FIG.5 , top and bottom); seven days after infection, an 8- to 16-fold enhancement in message levels was observed by filter hybridization (data not shown). These findings are consistent with the interpretation that HIV itself is able to upregulate c-myc. Tumorigenicity of B-HIVl Cells
Because 3-HIV1 cells possess two hallmarks of malignantly transformed cells, growth in 1% serum and cloning in soft agar, and, in addition, overexpress the c-myc proto-oncogene, we decided to test their malignant potential in SCID (severe combined immunodeficient) mice. Ten million B-HIV 1 cells were inoculated subcutaneously into each of four SCID mice; ten million B-EBV cells were inoculated into each of nine
430
ANNALS NEW YORK ACADEMY OF SCIENCES
FIGURE 6. Histology of a tumor induced by B-HIV cells in a SCID mouse. Excised tumor tissue was fixed in Bouin’s solution, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Examination under 100 X (top) and 430 x (bottom) initial magnifications revealed a malignant, invasive lymphoma of immunoblastic phenotype.
control SCID mice. Mice were observed for tumor formation over a period of 6 weeks. Two of four mice inoculated with B-HIV1 cells developed masses > 2 cm in diameter by 4 weeks; histology revealed a highly malignant invasive lymphoma of immunoblastic phenotype (FIG.6). None of the mice inoculated with B-EBV cells developed tumors within the observation period; however, between 8 and 12 weeks, 7 of the 9 mice developed masses, which upon histologic examination were classified as noninvasive plasmacytomas.
SUMMARY We have established a line of malignantly transformed human B cells by infecting purified primary B lymphocytes with human immunodeficiency virus type 1 (HIV-1).
ASTRIN & LAURENCE C-MYC
431
This line, termed B-HIVl, may serve as a model system for a subset of AIDS-related B-cell lymphomas in which the transformed phenotype may be initiated and/or maintained through an HIV-1 gene product. The B-HIV1 line contains both Epstein-Barr virus (EBV) and HIV-1 genomes. In addition, the c-myc gene is expressed at levels 10 to 20 times those in normal B cells. Similarly, EBV sequences, including those for the latent membrane protein (LMP), are expressed at greatly enhanced levels relative to expression in normal, EBV-immortalized B cells. The upregulation of c-myc and of EBV gene expression can both be produced by infection of susceptible B cells (not already harboring HIV) with exogenous HIV-1. The B-HIV1 line exhibits properties of malignantly transformed cells, in that it grows logarithmically in 1% serum, clones in soft agar, and produces invasive, malignant B-cell lymphomas in severe combined immunodeficient (SCID) mice. We have shown that HIV-1 has the ability to infect primary human B cells and to activate expression of EBV and c-myc. HIV activation of EBV has been documented previously in certain cell lines;20here we note that such activation can occur in primary B cells and, under certain conditions, can result in outgrowth of immortalized cell lines. This phenomenon may contribute to the clinical manifestation of lymphadenopathy early after infection with HIV. In addition, we have demonstrated that HIV infection of primary B cells in v i m can result in appearance of a fully malignant phenotype. This phenotype is likely to be due, at least in part, to the activation of c-myc by HIV. Preliminary experiments indicate that Tat,27the gene product of the transactivator of viral gene transcription tat, can upregulate c-myc transcription after addition to the culture media of certain B-cell lines. This raises the possibility that Tat can bind to target sequences in cellular RNA and enhance transcription as it does for HIV. It also raises the possibility that a necessary, but perhaps not sufficient, component for the expression of the malignant phenotype in certain AIDS-related lymphomas may be the presence and expression of the Tat gene product of HIV. Although published data would appear to rule out the presence of intact HIV genomes in the majority of AIDS-related the possibility of the presence of highly defective viral genomes has not been addressed.
REFERENCES
1. LEVINE,A. M. 1990. Semin. Oncol. 17: 104-121. 2. RABKIN,C. S., R. J. BIGGAR & J. W. HORM.1991. Int. J. Cancer 47: 692-695. 3. PELICCI,P. G., D. M. KNOWLES11, Z. ARLIN,R. WIECZOREK, P. LUCIW,D . DINA,C. BASILICO & R. DALLA-FAVERA. 1986. J. Exp. Med. 164: 2049-2076. 4. GROOPMAN, J. E., J. C. SULLIVAN, C. MULDER,D. GINSBURG, S. H. ORKIN,C. J. O'HARA, K. FALCHUK, F. WONG-STAAL & R. C. GALLO.1986. Blood 67: 612-615. 5. RECHARI, G., I. BEN-BASSAT,M. BERKOWICZ, U. MARTINOWITZ, F. BROK-SIMONI, Y. NEUMANN, A. VANSOVER, T. GOTLIEB-STEMATSKY & B. RAMOT.1987. Blood 70: 17131717. 6. SHIBATA, D., R. K. BRYNES,B. NATHWANI, S.KWOK,J. SHINSKY & N . ARNHEIM. 1989. Am. J. Pathol. 135: 697-702. 7. SELLERI, L., R. MARASCA, M. LUPPI,M. MONTOSSI & L. CECCHERINI-NELLI. 1990. AIDS Res. Hum. Retroviruses 6 149-150. 8. NERI,A., F. BARRIGA, D. M. KNOWLES,I. T. MCGRATH& R. DALLA-FAVERA. 1988. Proc. Natl. Acad. Sci. USA 8 5 2748-2752. 9. BECKHARDT, R. N., N. FARADY,M. MAY,R. A. TORRES& J. A. STRAUCHEN. 1988. Mt. Sinai J. Med. 55: 383. 10. POLACK,A,, G. HARTL,U. ZIMBER,U-K FREESE,G. LAWN,K. TAKABI,B. HOHN, L. GISSMAN& G. W. BORNKAUM. 1984. Gene 27: 279-288.
432
ANNALS NEW YORK ACADEMY OF SCIENCES
11. LAURENCE, J., H. CWKE & S. K. SIKDER.1990. Blood 7 5 696-703. 12. ROTHBERG, P. G., M. D. ERISMAN, R. E. DIEHL,U. G. ROVIGATTI & S. M. ASTRIN.1984. Mol. Cell. Biol. 4 1096-1103. 13. BRESSER, J., H. HUBBELL & D. GILLESPIE.1983. Proc. Natl. Acad. Sci. USA 8 0 65236527. 14. MUHLEGGER, K.,H-G BETZ, S. BOHM,H. V. D. ELTZ, H-J HOLTZKE& C. KESSLER. 1989. Nucleosides & Nucleotides 8 1161-1 166. 15. GHETIE,M-A, J. RICHARDSON, T. TUCKER,D. JONES,J. W. UHR& E. S. V I T E ~ A 1990. . Int. J. Cancer 4 5 481-485. 16. OKANO, M., Y. TAGUCHI, H. NAKAMIRE, S. J. PIRRUCCELLO, J. R. DAVIS,K. W. BEISEL, W. G. SANGER,R. R. FORDYCE & D. T.PURTILO.1990. Am. J. K. L. KLEVELAND, Pathol. 137: 517-522. 17. ROWE,M., L. S. YOUNG,J. CROCKER, H. STOKES,S. HENDERSON & A. B. RICKINSON. 1991. J. Exp. Med. 173: 147-158. 18. MOSIER,D. E., R. J. GULIZIA, S. M. BAIRD,D. D. RICHMAN, D. B. WILSON,R.I. Fox & T. J. KIPPS. 1989. Blood 74 (Suppl. 1): 52A. 19. LAURENCE, J. & S. M. ASTRIN.1991. Proc. Natl. Acad. Sci. USA. 88: 7635-7639. 20. LAI,P., X. LI & D. VOLSKY. 1989. Int. J. Cancer 43: 1104. 21. ANDIMAN, W. A., K. MARTIN,A. RUBENSTEIN, S. PAHWA,R. EASTMAN, B. 2.KATZ,J. P i n & G. MILLER.1985. Lancet 2 1390-1393. 22. BIRX,D. L., R. R. REDFIELD& G. TOSATO.1986. N. Engl. J. Med. 314 874-879. R., R. R. REDFIELD& S. BRODER.1986. J. Clin. Invest. 7 8 439-447. 23. YARCHOAN, 24. LOMBARDI, L., E. W. NEWCOMB & R. DALLA-FAVERA. 1987. Cell 49: 161-170. 25. LAURENCE, J., B. GRIMISON & S. M. ASTRIN.Submitted for publication. 26. RABBITTS, T. H., P. H. HAMLYN & R. BAER.1983. Nature 306 760-765. 27. FRANKEL, A. D., S. BIANCALANA & D. HUDSON.1989. Proc. Natl. Acad. Sci. USA 8 6 7397-7401.
