- Email: [email protected]
Immunoglobulin G abnormalities in HIV-1 infected individuals with lymphoma
Immunotechnology 4 (1998) 29 – 36
Immunoglobulin G abnormalities in HIV-1 infected individuals with lymphoma Peter Lundholm a,b,*, Erik Lucht b, Erik Svedmyr d, Ulla Rude´n c, Annika Linde c, Britta Wahren a,c a
Swedish Institute for Infectious Disease Control, Microbiology and Tumorbiology Center, Karolinska Institute, S-105 21 Stockholm, Sweden b Department of Immunology, Microbiology, Pathology and Infectious Diseases, Huddinge Uni6ersity Hospital, Karolinska Institute, Huddinge, Sweden c Department of Virology, Swedish Institute for Infectious Disease Control, Stockholm, Sweden d Department of Oncology, Karolinska Hospital, Stockholm, Sweden Received 1 September 1997; received in revised form 22 January 1998; accepted 13 February 1998
Abstract Background: Polyclonal B-cell activation precedes the occurrence of malignant B-cell clones. Several recent reports suggest a perturbed cytokine regulation in HIV-related lymphomagenesis and Epstein-Barr virus (EBV) involvement in approximately half of the cases with generalized lymphoma. Objecti6es: We investigated whether altered immunoglobulin properties would be detected by fine analysis of the immunoglobulin G (IgG) subclass patterns against HIV and EBV epitopes. Study Design: HIV-1 infected patients in early stage, late stage and with lymphoma were analyzed by ELISA for anti HIV and EBV IgG class and subclass antibodies. Avidity and affinity of the antibodies were studied. The lymphoma patients were also studied by PCR for EBV DNA in serum. Results: The total IgG reactivity to several HIV antigens was similar in the three patient groups. However, lymphoma patients had a more restricted subclass pattern with significantly lower IgG1 and IgG3 anti gp120 titers compared to other HIV-infected patients but good and persistent total IgG and IgG1 (excluding the gp120 antigen) reactivities in contradiction to their low CD4 counts. IgG4 reactivity was sparse, detectable to significant levels in the symptomatic group only. The observed relative affinity of the HIV-specific IgG and IgG1 of lymphoma patients was similar to that of asymptomatic and symptomatic patients. The subclass reactivity to the EBV peptide was similar in all groups but lymphoma patients with EBV DNA in serum exhibited significantly lower anti EBV peptide titers than those who were EBV DNA negative. Conclusion: These findings indicate that subclass analysis to defined viral antigens may be a means to detect immune dysregulation in tumor development. © 1998 Elsevier Science B.V. All rights reserved. Keywords: HIV-1 infection; HIV-1 related lymphoma; IgG subclasses; EBV DNA; HIV-1 serology; EBV serology
* Corresponding author. 1380-2933/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S1380-2933(98)00003-7
30
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
1. Introduction In the setting of immunodeficiency, be it congenital, acquired or iatrogenic, malignancies develop at a higher rate than expected in the immunocompetent host. Lymphoma associated with HIV-1 infection is most often a relatively late manifestation of the disease [1], although in approximately one third of the cases the malignancy is the first sign of HIV-1 infection without a previous history of AIDS or ARC [2,3]. The lymphomas often carry the EBV genome and consist of high-grade, B-cell tumors, with predilection for extranodal involvement, and short survival. In contrast to classical, or EBV negative, lymphomatous malignancies, HIV-1 associated lymphomas often have an aggressive clinical course and a rapidly fatal outcome [4,5]. Antibody response in the course of HIV-1 infection has been extensively studied to elucidate the role of the humoral arm in a strategy to eliminate free virus. Antiviral IgG subclass reactivity consists mainly of IgG1 and IgG3 [6,7]. IgG is the dominant anti HIV-1 immunoglobulin isotype and its restriction in subclass patterns was described [8–11]. In the case of reactivity to whole HIV-1 there is an IgG1 and IgG3 restriction followed by a restriction to synthesis of IgG1 during late progression of the disease [12]. Immunological deprivation is considered a reason for development of monoclonal B-cell lymphomatous proliferation in HIV infection analogous to the situation following bone marrow transplantation. Recent work [13] suggests a role for EBV specific antibodies, purified from EBV seropositive donors, in the prevention of EBV-related tumor development in SCID mice repopulated with human PBMCs. This might be related to antibody interactions with the EBV nuclear antigens (EBNAs) and/or EBV latent membrane proteins (LMPs) which occur at the tumor cell surface since these antigens are implicated in B-cell transformation [14]. IL-10, produced in normal monocytes following in vitro gp120 treatment, could exert an inhibitory effect on CD4 + and CD8 + proliferation [15]. In addition to its growth-promoting and differen-
tiating activities on human B cells [16], IL-10 is reported to act as a switch factor for IgG1 and IgG3. The present study was undertaken to analyze possible immunological imbalances in HIVinfected individuals with lymphoma regarding the IgG subclass distribution to previously defined epitopes of HIV-1 and EBV proteins. We defined the IgG subclass distribution in HIV-1 infected individuals with HIV-1 associated lymphoma in comparison to various clinical stages. Our findings point toward a similar IgG subclass restriction in the lymphoma patient group as in the asymptomatic patients. Patients with low CD4 cell counts exhibited significantly lower titers of IgG3 to the HIV-1 envelope and patients with active lymphoma had low anti EBV titers.
2. Materials and methods
2.1. Patients Serum samples from 12 HIV-1 infected patients diagnosed with lymphoma, 15 CDC group A (asymptomatic) patients and 11 patients in CDC groups B and C (symptomatic) were collected and stored frozen at − 70°C prior to examination. Nine were highly malignant lymphomas, two were widespread subcutaneous lymphomas and one was of Hodgkin’s mixed type. One was located intracerebrally, one rectally, two located in peritoneal lymph nodes. The remaining were generalized or located in peripheral lymphnodes. CD4+ T-lymphocyte counts in peripheral blood are presented in Fig. 1. HIV-1 or EBV negative laboratory workers were used as controls.
2.2. Antigens HIV-1 LAI derived recombinant glycoprotein gp160 (MicroGeneSys, Meriden, CT) was expressed in a baculovirus system. Linear peptides had a length of 13–20 amino acids (a a) [17]. HIV-1 sequences were derived from the MNstrain for V3/A12 (a a 309–321; RKRIHIGPGRAFY) and the LAI isolate for gp120/C90 (a a 494–508; VKIEPLGVAPTKAKR). The im-
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
munodominant peptide gp41/JB-4 (a a 594 – 613; GIWGCSGKLICTTAVPWNAS) and EBV peptide p107, representing the glycine-alanine repeat of the EBV nuclear antigen 7 (AGAGGGAGGAGAGGGAGGAG), were purified to greater than 95% purity.
2.3. Enzyme-linked immunosorbent assay 2.3.1. Total IgG Total IgG was performed as previously described [11]. Briefly, microplates (Maxisorp, Nunc, Odense, Denmark) were coated with 1 mg peptide or 0.1 mg recombinant protein per well in 100 ml of 0.05 mol/l sodium carbonate buffer (pH 9.6). The plates were washed five times with 0.9% NaCl/0.05% Tween 20 before 100 ml of serum diluted in phosphate-buffered saline with 0.5% bovine serum albumin/0.05% Tween 20/0.01% merthiolate/2% goat serum was added to the well and incubated for 1 h in 37°C. After an additional wash, goat anti-human IgG conjugated with alkaline phosphatase (ALP; Sigma, St. Louis, MO) diluted 1:3000 was incubated for 1 h in 37°C. The ALP substrate p-nitrophenyl phosphate (Sigma) was added after and incubated for 30 min in 37°C. The absorbance was measured at 405 nm after terminating the reaction with 100 ml 1 mol/l NaOH.
Fig. 1. Box plot illustrating CD4 + cell counts (106 cells/l) 10th, 25th, 50th (median), 75th and 90th percentiles in lymphoma (Ly, median 125, mean 155), asymptomatic (Asy, median 390, mean 398) and symptomatic (Sy, median 260, mean 247) patients. Values below the 10th or above the 90th percentile are shown as circles.
31
2.3.2. Enzyme linked immunosorbent assay (ELISA) for IgG subclasses Enzyme linked immunosorbent assay (ELISA) for IgG subclasses was performed as previously described [10]. Briefly, sera diluted in buffer A (0.5% bovine serum albumin (BSA) and 0.05% Tween 20 in phosphate-buffered saline (PBS)) were incubated on plates coated as described above for 60 min at 37°C. After four washes mouse anti-human IgG monoclonal antibodies (Mabs) diluted in buffer A with 20% fetal calf serum were added for 105 min at 37°C. The Mabs specific for human IgG were NL16 for IgG1 (Oxoid, Hampshire, UK), G0M1 for IgG2, clone ZG4 for IgG3 and clone RJ4 for IgG4. After washing four times, conjugated horseradish peroxidase (HRP) labeled rabbit anti-mouse Ig (Dakopatts, Copenhagen, Denmark) was added and incubated for 105 min at 37°C. After washes, the HRP substrate ortophenylene diamine (Sigma) was added and incubated for 30 min at room temperature. The reaction was terminated with 2.5 mol/l H2SO4 and the absorbance was read at 492 nm (OD492). End-point titers were determined at a cut off of mean + 3× S.D. of negative controls. 2.4. A6idity and inhibition ELISA Urea treatment of antigen-antibody complexes allows approximate determination of immunoglobulin avidities [18]. In the urea wash ELISA, peptide was coated as described. After serum incubation for 60 min at 37°C the wells were washed three times for 5 min with 8 M urea. Avidity index [18] was obtained by dividing the endpoint titer with and without urea wash. In the inhibition ELISA, the peptide was coated as described above. Sera were added in a predefined dilution together with increasing amounts of the corresponding peptide. The molar inhibitory concentrations required for 50% inhibition of the antibody binding without inhibitor were determined [19]. The values obtained represent an average anti-peptide antibody affinity in the polyclonal serum.
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
32
Table 1 Anti HIV and EBV antibody distribution pattern Group
Lymphoma Asymptomatic Symptomatic
3.2. gp41 /JB-4 peptide
Frequency of reactivity to antigen gp160
gp41
V3
gp120
p107
12/12 15/15 11/11
12/12 15/15 11/11
11/12 15/15 11/11
8/12 12/15 11/11
12/12 9/15a 6/11a
Total IgG response frequencies for the different patient groups to indicated antigens. a Indicating significant frequence differences. Comparing lymphoma and asymptomatic patients; P= 0.02. Comparing lymphoma and symptomatic patients; P= 0.014. Fisher’s exact test.
2.5. EBV DNA nested polymerase chain reaction The nPCR was performed as described [20]. EBV DNA was extracted from serum with the QIA amp blood kit (Qiagen, Hilden, Germany) with minor alterations of the manufacturers’ instructions. Statistical analysis was performed with MannWhitney U-test and Fisher’s exact test as indicated in the text and tables.
3. Results
3.1. Whole rgp160 The patients with lymphoma were compared with asymptomatic and symptomatic HIV-infected individuals (Fig. 1). The serological results were also related to the CD4 cell levels. All patients had detectable IgG antibody reactivity to the precursor envelope glycoprotein gp160 (Table 1). The subclass response was mainly IgG1 and IgG3 restricted (Fig. 2). Patients with lymphoma and asymptomatic patients exhibited similar IgG1 and IgG3 titers to rgp160. The symptomatic patient group had additional IgG4 reactivities to rgp160 with significantly higher titers of IgG4 (P=0.05 Mann-Whitney U-test, Fig. 2).
The gp41 conserved immunodominant epitope was previously reported to be IgG1 and IgG2 restricted [21]. In our study, all IgG subclasses were noted, in a frequency of IgG1\ G2\ G3\ G4. This antigen was the only one to which IgG2 responses occurred in detectable titers (Fig. 2). The IgG3 titer was significantly lower (P= 0.01 Mann-Whitney U-test) in the lymphoma group compared to the symptomatic group.
3.3. V3 /A12 peptide Antibodies reactive to this region have been correlated to the neutralization of HIV-1 infectivity and probably interfering with coreceptor binding [22–24]. For the lymphoma patients, subclass reactivity to the V3 peptide was restricted to subclasses IgG1 and IgG3. In comparison with symptomatic patients, the lymphoma patients exhibited significantly lower titers of IgG3 (P=0.005 Mann-Whitney U-test).
3.4. gp120 /C90 peptide The peptide C90 of gp120 has been suggested to represent an antibody-dependent cellular cytotoxicity epitope [25] and to contribute to neutralization [23]. Only the symptomatic group had 100% reactivity of total IgG to this peptide (Table 1). The pattern of subclass reactivity was dominated by IgG1 and IgG3, the lymphoma group of patients having the lowest reactivities. Differences in titers (Fig. 2) could be observed comparing lymphoma and symptomatic patients for IgG1 (P =0.001 Mann-Whitney U-test) and IgG3 (P= 0.05 Mann-Whitney U-test).
3.5. EBV peptide p107 EBV is associated with approximately 50% of the lymphomas in HIV-disease and those with demonstrable EBV DNA in serum have a poorly controlled lymphoma. Antibody subclass distribution to EBV p107 was studied in order to
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
33
Fig. 2. Mean titers + 1 S.E.M. for indicated suclasses and antigens in the different patient categories. Significant mean differences indicated by P-values (Mann-Whitney U-test).
compare immune reactivities to another longstanding latent infection. All patients were EBV antibody positive (Table 1). All patients had both IgG1 and IgG3 responses to p107. We observed higher IgG1 titers (P =0.05 Mann-Whitney Utest) in the lymphoma group compared to the symptomatic one (Fig. 2). IgG4 responses did not occur to the EBV peptide (Fig. 2). EBV DNA PCR in serum was performed in the lymphoma patients. The EBV DNA positive patients (n=3/12) exhibited lower IgG1 and total IgG titers (Fig. 3) to the EBV peptide p107 (P = 0.05 for both IgG1 and total IgG, Mann-Whitney U-test) although the reactivities to HIV-1 derived antigens and the CD4 values were comparable to EBV DNA negative lymphoma patients (not shown).
3.6. IgG subclass reacti6ity, CD4 cell counts and affinity The lymphoma and asymptomatic patient groups had comparable frequencies of overall antibody responses. IgG subclass restriction in patients with lymphoma exhibited a pattern of responses to gp160 where concurrent IgG1 and IgG3 were present in 75% of them. With respect to IgG3, the lymphoma patients had significantly lower titers to three of the HIV antigens. The anti-EBV p107 was low in patients with demonstrable EBV DNA (Figs. 2 and 3). IgG3 plays a role in neutralization and complement fixation. We analyzed the relation of IgG3 occurrence to immunological competence. Patients with low CD4 cells (50% of the patient
34
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
material had CD4 values below 270 CD4 cells/ml) had significantly lower IgG3 titers to the envelope protein gp160 (P=0.005 Mann-Whitney U-test), as well as low titers of total IgG to the V3 peptide (P =0.005 Mann-Whitney U-test). These individuals were predominantly from the lymphoma group. The IgG3 reactivity to all antigens compared together was significantly lowered in the patients with CD4 cells below 270 (P = 0.02 Mann-Whitney U-test). This is further illustrated by the observation that 3/95 of the IgG3 assays in patients with low CD4 cells had IgG3 titers above 1000 compared to 19/95 assays for the individuals with high levels of CD4 cells (\ 270). The symptomatic patient group had higher frequencies and an extended subclass distribution including IgG4. IgG3 and IgG4 titers to all antigens were significantly higher (P =0.02 for IgG3 and P =0.002 for IgG4, Mann-Whitney U-test) in the symptomatic group (n= 55 tests) compared to the lymphoma patients (n =60 tests). Avidity properties of the HIV-1 specific IgG and IgG1 were found to be similar in the different patient groups. No decreased binding capacities were observed with low CD4 cell counts (Fig. 4). Binding avidity of the isolated IgG3 subclass was
Fig. 3. Mean titers +1 S.E.M. in EBV DNA positive patients (n=3) and EBV DNA negative patients (n= 9) exhibiting different reactivities to the EBV peptide p107.
not possible to measure due to difficulties in complete separation of subclass IgG3.
4. Discussion The HIV-infected patients with lymphoma had an IgG1 subclass pattern directed to HIV antigens that was similar to that of the other HIV-infected patients except for a defective reactivity to one immunodominant epitope of the C-terminal of gp120. IgG3 reactivities of the lymphoma patients were lower against all antigens and also IgG4 reactivity was sparse. The lymphoma group had the lowest CD4 cell values, a property that may decrease the CD4 help, particularly to IgG3 responses. With EBV p107 antibody, an inverse relationship is observed. Higher titers and frequencies could be observed in the lymphoma group although low levels of antibodies were found in patients where EBV DNA was present in the sera. The latter finding is in concordance with previous findings where low levels of p107 antibodies in EBV DNA positive subjects reflect an impaired EBV T-cell control [20]. This impairment may consist of an altered IL-10 expression with inhibitory effects of IL-10 or vIL-10 on T-cells. IgG subclass disorders may develop as a manifestation of an impairment of the B-cell function due to the malignancy per se, but more likely as a result of functional imbalance in immune regulation developing with declining CD4 cells. Previous findings point to a loss of HIV-1 specific minor IgG subclasses during disease progression [12]. We interpret the observed loss of IgG subclass reactivities in the lymphoma group as a sign of progression of disease since the mean CD4 count of these patients is below that of even the symptomatic group. IgG1 is the dominating subclass in anti-HIV antibody responses and is elicited in response to protein antigens together with IgG3. Among the anti-HIV IgG subclasses of sera from HIV-1 subjects, IgG1 was dominant regardless of the clinical stage of infection. The patients with lymphoma retained their responses to immunodominant antigens despite their generally low CD4 values. IgG1
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
35
Fig. 4. Inhibitory ELISA for gp41/JB4. Sera were added together with increasing amounts of the corresponding peptide. Six representative patient sera are shown. Solid symbols represent patients with low CD4 cell counts: 13× 106/l – 130 ×106/l. Open symbols represent patients with high CD4 counts: 430 × 106/l – 440 ×106/l.
responses to gp160 were for example slightly higher in lymphoma patients compared to the symptomatic patients. IgG2 is the predominant subclass in response to bacterial polysaccharides. It was present only in response to the immunodominant gp41 peptide. In general, IgG3 was significantly decreased in individuals with low CD4 counts and lymphoma patients often had lower IgG3 compared to the symptomatic patients. IgG4 typically follows repeated antigenic stimulation. HIV-specific IgG4 reactivity was seen almost exclusively in the symptomatic group compared to the asymptomatic and lymphoma groups, and thus might represent a late broadened subclass pattern to a high antigenic burden. The lymphoma patients, compared to the symptomatic patients, had decreased responses of IgG1 and IgG3 to gp120 peptide and IgG3 to gp41 and V3 peptides (Fig. 2). These findings could have implications in the development of lymphoma since gp120 and gp41 recently have been reported to stimulate B-cell growth and dif-
ferentiation through IL-10, possibly towards a development of malignant B-cell clones [26–28]. Since the presence of lymphomas related to HIV-infection was evident [29], several mechanisms have been suggested contributing to HIV1 related lymphomagenisis. Direct polyclonal activation of B lymphocytes [30], and the influence of Epstein-Barr virus infection are likely to have significant roles in the pathogenesis [28,31– 34]. Recent work points towards EBV negative tumors having a less aggressive clinical course [35]. The lymphoma patients with poor EBV p107 reactivities had detectable serum EBV DNA, which has previously been implicated as a marker for decreased T-cell control of EBVtransformed cells [20]. However, EBV specific IgG also appears to be important in inhibiting outgrowth of transformed EBV-carrying cells [13]. The present small material indicates that a retro- and prospective study of IgG3 changes during lymphoma development with concurrent analysis of HIV antigen reactivities should, indeed, be performed.
36
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
Acknowledgements This work was supported by funds from the Faculty of Odontology, Karolinska Institute, Swedish Medical Research Council and funds from the Stockholm County Council.
References [1] Levine AM. Blood 1992;80:8–20. [2] Beral V, Peterman T, Berkelman R, Jaffe H. Lancet 1991;937:805 – 9. [3] Ziegler JL, McGrath MS. In: Magrath IT, editor. The Non-Hodgkin’s Lymphomas: Lymphomas in HIV-Positive Individuals. London: Edward Arnold, 1990:155–159. [4] Knowles DM, Chamulak GA, Subar M, Burke JS, Dugan M, Wernz J, Slywotzky C, Pelicci PG, Dalla-Favera R, Raphael B. Ann Intern Med 1988;108:744. [5] Levine AM. Ann Oncol 1994;5(Suppl 2):29–35. [6] Sundqvist VA, Linde A, Wahren B. J Clin Microbiol 1984;20:94 – 8. [7] Linde GA. J Clin Microbiol 1985;21:117–21. [8] Khalife J, Guy B, Capron M, Kieny MP, Ameisen JC, Montagnier L, Lecocq JP, Capron A. AIDS Res Hum Retroviruses 1988;4:3–9. [9] Ljunggren K, Broliden PA, Morfeldt-Manson L, Jondal M, Wahren B. Clin Exp Immunol 1988;73:343–7. [10] Broliden PA, Morfeldt-Ma˚nsson L, Rosen J, Jondal M, Wahren B. Clin Exp Immunol 1989;76:216–21. [11] Jansson M, Wahren B, Scarlatti G, Principi N, Lombardi V, Livadiotti S, Elia L, Plebani A, Wigzell H, Rossi P. AIDS 1992;6:365 – 71. [12] Sundqvist V-A, Linde A, Kurth R, Werner A, Helm EB, Popovic M, Gallo RC, Wahren B. J Infect Dis 1986;153:970. [13] Abedi MR, Linde A, Christensson B, Mackett M, Hammarstrom L, Smith CI. Int J Cancer 1997;71:624–9. [14] Kieff E, Liebowitz D. In: Fields B, Knipe D, editors. Virology: Epstein-Barr Virus and Its Replication. New York: Raven Press, 1990:1889–1920. [15] Schols D, De Clercq E. J Virol 1996;70:4953–60.
.
[16] Briere F, Servet-Delprat C, Bridon JM, Saint-Remy JM, Banchereau J. J Exp Med 1994;179:757 – 62. [17] Sa¨llberg M, Rude´n U, Magnius LO, Norrby E, Wahren B. Immunol Lett 1991;30:59 – 68. [18] Hedman K, Seppa¨la¨ I. J Clin Immunol 1988;8:214 – 21. [19] Steward MW, Stanley CM, Dimarchi R, Mulcahy G, Doel TR. Immunology 1991;72:99 – 103. [20] Barkholt LM, Dahl H, Enbom M, Linde A. Transplant Int 1996;9:439 – 45. [21] Chiodi F, Mathiesen T, Albert J, Parks E, Norrby E, Wahren B. J Immunol 1989;142:3809– 14. [22] Javaherian K, Langlois AJ, LaRosa GJ, Profy AT, Bolognesi DP, Herlihy WC, Putney SD, Matthews TJ. Science 1990;250:1590– 3. [23] Broliden PA, VonGegerfelt A, Clapham P, Rosen J, Fenyo¨ EM, Wahren B, Broliden K. Proc Natl Acad Sci USA 1992;89:461 – 5. [24] Cocchi F, DeVico AL, Garzino-Demo A, Cara A, Gallo RC, Lusso P. Nat Med 1996;2:1244 – 7. [25] Palker TJ, Clark ME, Langlois A, Matthews TJ, Weinhold KJ, Randall RR, Bolognesi DI, Haynes BF. Proc Natl Acad Sci USA 1988;85:1932 – 6. [26] Edelman L, Deveau C, Raphael M, Monchatre E, Gabarre J, Deville-Chabrol A, Pialoux G, Emilie D, Joab I, Galanaud P. Eur Cytokine Netw 1996;7:785 – 91. [27] Miyazaki I, Cheung RK, Dosch HM. J Exp Med 1993;178:439 – 47. [28] Stuart AD, Stewart JP, Arrand JR, Mackett M. Oncogene 1995;11:1711 – 9. [29] Ziegler JL, Beckstead JA, Volberding PA, Abrams DI, Levine AM, Lukes RJ, Gill PS, Burkes RL, Meyer PR, Metroka CE, et al. N Engl J Med 1984;311:565 – 70. [30] Schnittman SM, Lane HC, Higgins SE, Folks T, Fauci AS. Science 1986;233:1084– 6. [31] Herndier BG, Kaplan LD, McGrath MS. AIDS 1994;8:1025 – 49. [32] Pelicci PG, Knowles DMd, Arlin ZA, Wieczorek R, Luciw P, Dina D, Basilico C, Dalla-Favera R. J Exp Med 1986;164:2049– 60. [33] Subar M, Neri A, Inghirami G, Knowles DM, Dalla-Faverta R. Blood 1988;72:667 – 71. [34] Muller JR, Janz S, Goedert JJ, Potter M, Rabkin CS. Proc Natl Acad Sci USA 1995;92:6577 – 81. [35] Kaplan LD, Shiramizu B, Herndier B, Hahn J, Meeker TC, Ng V, Volberding PA, McGrath MS. Blood 1995;85:1727 – 35.
Immunoglobulin G abnormalities in HIV-1 infected individuals with lymphoma Peter Lundholm a,b,*, Erik Lucht b, Erik Svedmyr d, Ulla Rude´n c, Annika Linde c, Britta Wahren a,c a
Swedish Institute for Infectious Disease Control, Microbiology and Tumorbiology Center, Karolinska Institute, S-105 21 Stockholm, Sweden b Department of Immunology, Microbiology, Pathology and Infectious Diseases, Huddinge Uni6ersity Hospital, Karolinska Institute, Huddinge, Sweden c Department of Virology, Swedish Institute for Infectious Disease Control, Stockholm, Sweden d Department of Oncology, Karolinska Hospital, Stockholm, Sweden Received 1 September 1997; received in revised form 22 January 1998; accepted 13 February 1998
Abstract Background: Polyclonal B-cell activation precedes the occurrence of malignant B-cell clones. Several recent reports suggest a perturbed cytokine regulation in HIV-related lymphomagenesis and Epstein-Barr virus (EBV) involvement in approximately half of the cases with generalized lymphoma. Objecti6es: We investigated whether altered immunoglobulin properties would be detected by fine analysis of the immunoglobulin G (IgG) subclass patterns against HIV and EBV epitopes. Study Design: HIV-1 infected patients in early stage, late stage and with lymphoma were analyzed by ELISA for anti HIV and EBV IgG class and subclass antibodies. Avidity and affinity of the antibodies were studied. The lymphoma patients were also studied by PCR for EBV DNA in serum. Results: The total IgG reactivity to several HIV antigens was similar in the three patient groups. However, lymphoma patients had a more restricted subclass pattern with significantly lower IgG1 and IgG3 anti gp120 titers compared to other HIV-infected patients but good and persistent total IgG and IgG1 (excluding the gp120 antigen) reactivities in contradiction to their low CD4 counts. IgG4 reactivity was sparse, detectable to significant levels in the symptomatic group only. The observed relative affinity of the HIV-specific IgG and IgG1 of lymphoma patients was similar to that of asymptomatic and symptomatic patients. The subclass reactivity to the EBV peptide was similar in all groups but lymphoma patients with EBV DNA in serum exhibited significantly lower anti EBV peptide titers than those who were EBV DNA negative. Conclusion: These findings indicate that subclass analysis to defined viral antigens may be a means to detect immune dysregulation in tumor development. © 1998 Elsevier Science B.V. All rights reserved. Keywords: HIV-1 infection; HIV-1 related lymphoma; IgG subclasses; EBV DNA; HIV-1 serology; EBV serology
* Corresponding author. 1380-2933/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII S1380-2933(98)00003-7
30
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
1. Introduction In the setting of immunodeficiency, be it congenital, acquired or iatrogenic, malignancies develop at a higher rate than expected in the immunocompetent host. Lymphoma associated with HIV-1 infection is most often a relatively late manifestation of the disease [1], although in approximately one third of the cases the malignancy is the first sign of HIV-1 infection without a previous history of AIDS or ARC [2,3]. The lymphomas often carry the EBV genome and consist of high-grade, B-cell tumors, with predilection for extranodal involvement, and short survival. In contrast to classical, or EBV negative, lymphomatous malignancies, HIV-1 associated lymphomas often have an aggressive clinical course and a rapidly fatal outcome [4,5]. Antibody response in the course of HIV-1 infection has been extensively studied to elucidate the role of the humoral arm in a strategy to eliminate free virus. Antiviral IgG subclass reactivity consists mainly of IgG1 and IgG3 [6,7]. IgG is the dominant anti HIV-1 immunoglobulin isotype and its restriction in subclass patterns was described [8–11]. In the case of reactivity to whole HIV-1 there is an IgG1 and IgG3 restriction followed by a restriction to synthesis of IgG1 during late progression of the disease [12]. Immunological deprivation is considered a reason for development of monoclonal B-cell lymphomatous proliferation in HIV infection analogous to the situation following bone marrow transplantation. Recent work [13] suggests a role for EBV specific antibodies, purified from EBV seropositive donors, in the prevention of EBV-related tumor development in SCID mice repopulated with human PBMCs. This might be related to antibody interactions with the EBV nuclear antigens (EBNAs) and/or EBV latent membrane proteins (LMPs) which occur at the tumor cell surface since these antigens are implicated in B-cell transformation [14]. IL-10, produced in normal monocytes following in vitro gp120 treatment, could exert an inhibitory effect on CD4 + and CD8 + proliferation [15]. In addition to its growth-promoting and differen-
tiating activities on human B cells [16], IL-10 is reported to act as a switch factor for IgG1 and IgG3. The present study was undertaken to analyze possible immunological imbalances in HIVinfected individuals with lymphoma regarding the IgG subclass distribution to previously defined epitopes of HIV-1 and EBV proteins. We defined the IgG subclass distribution in HIV-1 infected individuals with HIV-1 associated lymphoma in comparison to various clinical stages. Our findings point toward a similar IgG subclass restriction in the lymphoma patient group as in the asymptomatic patients. Patients with low CD4 cell counts exhibited significantly lower titers of IgG3 to the HIV-1 envelope and patients with active lymphoma had low anti EBV titers.
2. Materials and methods
2.1. Patients Serum samples from 12 HIV-1 infected patients diagnosed with lymphoma, 15 CDC group A (asymptomatic) patients and 11 patients in CDC groups B and C (symptomatic) were collected and stored frozen at − 70°C prior to examination. Nine were highly malignant lymphomas, two were widespread subcutaneous lymphomas and one was of Hodgkin’s mixed type. One was located intracerebrally, one rectally, two located in peritoneal lymph nodes. The remaining were generalized or located in peripheral lymphnodes. CD4+ T-lymphocyte counts in peripheral blood are presented in Fig. 1. HIV-1 or EBV negative laboratory workers were used as controls.
2.2. Antigens HIV-1 LAI derived recombinant glycoprotein gp160 (MicroGeneSys, Meriden, CT) was expressed in a baculovirus system. Linear peptides had a length of 13–20 amino acids (a a) [17]. HIV-1 sequences were derived from the MNstrain for V3/A12 (a a 309–321; RKRIHIGPGRAFY) and the LAI isolate for gp120/C90 (a a 494–508; VKIEPLGVAPTKAKR). The im-
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
munodominant peptide gp41/JB-4 (a a 594 – 613; GIWGCSGKLICTTAVPWNAS) and EBV peptide p107, representing the glycine-alanine repeat of the EBV nuclear antigen 7 (AGAGGGAGGAGAGGGAGGAG), were purified to greater than 95% purity.
2.3. Enzyme-linked immunosorbent assay 2.3.1. Total IgG Total IgG was performed as previously described [11]. Briefly, microplates (Maxisorp, Nunc, Odense, Denmark) were coated with 1 mg peptide or 0.1 mg recombinant protein per well in 100 ml of 0.05 mol/l sodium carbonate buffer (pH 9.6). The plates were washed five times with 0.9% NaCl/0.05% Tween 20 before 100 ml of serum diluted in phosphate-buffered saline with 0.5% bovine serum albumin/0.05% Tween 20/0.01% merthiolate/2% goat serum was added to the well and incubated for 1 h in 37°C. After an additional wash, goat anti-human IgG conjugated with alkaline phosphatase (ALP; Sigma, St. Louis, MO) diluted 1:3000 was incubated for 1 h in 37°C. The ALP substrate p-nitrophenyl phosphate (Sigma) was added after and incubated for 30 min in 37°C. The absorbance was measured at 405 nm after terminating the reaction with 100 ml 1 mol/l NaOH.
Fig. 1. Box plot illustrating CD4 + cell counts (106 cells/l) 10th, 25th, 50th (median), 75th and 90th percentiles in lymphoma (Ly, median 125, mean 155), asymptomatic (Asy, median 390, mean 398) and symptomatic (Sy, median 260, mean 247) patients. Values below the 10th or above the 90th percentile are shown as circles.
31
2.3.2. Enzyme linked immunosorbent assay (ELISA) for IgG subclasses Enzyme linked immunosorbent assay (ELISA) for IgG subclasses was performed as previously described [10]. Briefly, sera diluted in buffer A (0.5% bovine serum albumin (BSA) and 0.05% Tween 20 in phosphate-buffered saline (PBS)) were incubated on plates coated as described above for 60 min at 37°C. After four washes mouse anti-human IgG monoclonal antibodies (Mabs) diluted in buffer A with 20% fetal calf serum were added for 105 min at 37°C. The Mabs specific for human IgG were NL16 for IgG1 (Oxoid, Hampshire, UK), G0M1 for IgG2, clone ZG4 for IgG3 and clone RJ4 for IgG4. After washing four times, conjugated horseradish peroxidase (HRP) labeled rabbit anti-mouse Ig (Dakopatts, Copenhagen, Denmark) was added and incubated for 105 min at 37°C. After washes, the HRP substrate ortophenylene diamine (Sigma) was added and incubated for 30 min at room temperature. The reaction was terminated with 2.5 mol/l H2SO4 and the absorbance was read at 492 nm (OD492). End-point titers were determined at a cut off of mean + 3× S.D. of negative controls. 2.4. A6idity and inhibition ELISA Urea treatment of antigen-antibody complexes allows approximate determination of immunoglobulin avidities [18]. In the urea wash ELISA, peptide was coated as described. After serum incubation for 60 min at 37°C the wells were washed three times for 5 min with 8 M urea. Avidity index [18] was obtained by dividing the endpoint titer with and without urea wash. In the inhibition ELISA, the peptide was coated as described above. Sera were added in a predefined dilution together with increasing amounts of the corresponding peptide. The molar inhibitory concentrations required for 50% inhibition of the antibody binding without inhibitor were determined [19]. The values obtained represent an average anti-peptide antibody affinity in the polyclonal serum.
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
32
Table 1 Anti HIV and EBV antibody distribution pattern Group
Lymphoma Asymptomatic Symptomatic
3.2. gp41 /JB-4 peptide
Frequency of reactivity to antigen gp160
gp41
V3
gp120
p107
12/12 15/15 11/11
12/12 15/15 11/11
11/12 15/15 11/11
8/12 12/15 11/11
12/12 9/15a 6/11a
Total IgG response frequencies for the different patient groups to indicated antigens. a Indicating significant frequence differences. Comparing lymphoma and asymptomatic patients; P= 0.02. Comparing lymphoma and symptomatic patients; P= 0.014. Fisher’s exact test.
2.5. EBV DNA nested polymerase chain reaction The nPCR was performed as described [20]. EBV DNA was extracted from serum with the QIA amp blood kit (Qiagen, Hilden, Germany) with minor alterations of the manufacturers’ instructions. Statistical analysis was performed with MannWhitney U-test and Fisher’s exact test as indicated in the text and tables.
3. Results
3.1. Whole rgp160 The patients with lymphoma were compared with asymptomatic and symptomatic HIV-infected individuals (Fig. 1). The serological results were also related to the CD4 cell levels. All patients had detectable IgG antibody reactivity to the precursor envelope glycoprotein gp160 (Table 1). The subclass response was mainly IgG1 and IgG3 restricted (Fig. 2). Patients with lymphoma and asymptomatic patients exhibited similar IgG1 and IgG3 titers to rgp160. The symptomatic patient group had additional IgG4 reactivities to rgp160 with significantly higher titers of IgG4 (P=0.05 Mann-Whitney U-test, Fig. 2).
The gp41 conserved immunodominant epitope was previously reported to be IgG1 and IgG2 restricted [21]. In our study, all IgG subclasses were noted, in a frequency of IgG1\ G2\ G3\ G4. This antigen was the only one to which IgG2 responses occurred in detectable titers (Fig. 2). The IgG3 titer was significantly lower (P= 0.01 Mann-Whitney U-test) in the lymphoma group compared to the symptomatic group.
3.3. V3 /A12 peptide Antibodies reactive to this region have been correlated to the neutralization of HIV-1 infectivity and probably interfering with coreceptor binding [22–24]. For the lymphoma patients, subclass reactivity to the V3 peptide was restricted to subclasses IgG1 and IgG3. In comparison with symptomatic patients, the lymphoma patients exhibited significantly lower titers of IgG3 (P=0.005 Mann-Whitney U-test).
3.4. gp120 /C90 peptide The peptide C90 of gp120 has been suggested to represent an antibody-dependent cellular cytotoxicity epitope [25] and to contribute to neutralization [23]. Only the symptomatic group had 100% reactivity of total IgG to this peptide (Table 1). The pattern of subclass reactivity was dominated by IgG1 and IgG3, the lymphoma group of patients having the lowest reactivities. Differences in titers (Fig. 2) could be observed comparing lymphoma and symptomatic patients for IgG1 (P =0.001 Mann-Whitney U-test) and IgG3 (P= 0.05 Mann-Whitney U-test).
3.5. EBV peptide p107 EBV is associated with approximately 50% of the lymphomas in HIV-disease and those with demonstrable EBV DNA in serum have a poorly controlled lymphoma. Antibody subclass distribution to EBV p107 was studied in order to
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
33
Fig. 2. Mean titers + 1 S.E.M. for indicated suclasses and antigens in the different patient categories. Significant mean differences indicated by P-values (Mann-Whitney U-test).
compare immune reactivities to another longstanding latent infection. All patients were EBV antibody positive (Table 1). All patients had both IgG1 and IgG3 responses to p107. We observed higher IgG1 titers (P =0.05 Mann-Whitney Utest) in the lymphoma group compared to the symptomatic one (Fig. 2). IgG4 responses did not occur to the EBV peptide (Fig. 2). EBV DNA PCR in serum was performed in the lymphoma patients. The EBV DNA positive patients (n=3/12) exhibited lower IgG1 and total IgG titers (Fig. 3) to the EBV peptide p107 (P = 0.05 for both IgG1 and total IgG, Mann-Whitney U-test) although the reactivities to HIV-1 derived antigens and the CD4 values were comparable to EBV DNA negative lymphoma patients (not shown).
3.6. IgG subclass reacti6ity, CD4 cell counts and affinity The lymphoma and asymptomatic patient groups had comparable frequencies of overall antibody responses. IgG subclass restriction in patients with lymphoma exhibited a pattern of responses to gp160 where concurrent IgG1 and IgG3 were present in 75% of them. With respect to IgG3, the lymphoma patients had significantly lower titers to three of the HIV antigens. The anti-EBV p107 was low in patients with demonstrable EBV DNA (Figs. 2 and 3). IgG3 plays a role in neutralization and complement fixation. We analyzed the relation of IgG3 occurrence to immunological competence. Patients with low CD4 cells (50% of the patient
34
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
material had CD4 values below 270 CD4 cells/ml) had significantly lower IgG3 titers to the envelope protein gp160 (P=0.005 Mann-Whitney U-test), as well as low titers of total IgG to the V3 peptide (P =0.005 Mann-Whitney U-test). These individuals were predominantly from the lymphoma group. The IgG3 reactivity to all antigens compared together was significantly lowered in the patients with CD4 cells below 270 (P = 0.02 Mann-Whitney U-test). This is further illustrated by the observation that 3/95 of the IgG3 assays in patients with low CD4 cells had IgG3 titers above 1000 compared to 19/95 assays for the individuals with high levels of CD4 cells (\ 270). The symptomatic patient group had higher frequencies and an extended subclass distribution including IgG4. IgG3 and IgG4 titers to all antigens were significantly higher (P =0.02 for IgG3 and P =0.002 for IgG4, Mann-Whitney U-test) in the symptomatic group (n= 55 tests) compared to the lymphoma patients (n =60 tests). Avidity properties of the HIV-1 specific IgG and IgG1 were found to be similar in the different patient groups. No decreased binding capacities were observed with low CD4 cell counts (Fig. 4). Binding avidity of the isolated IgG3 subclass was
Fig. 3. Mean titers +1 S.E.M. in EBV DNA positive patients (n=3) and EBV DNA negative patients (n= 9) exhibiting different reactivities to the EBV peptide p107.
not possible to measure due to difficulties in complete separation of subclass IgG3.
4. Discussion The HIV-infected patients with lymphoma had an IgG1 subclass pattern directed to HIV antigens that was similar to that of the other HIV-infected patients except for a defective reactivity to one immunodominant epitope of the C-terminal of gp120. IgG3 reactivities of the lymphoma patients were lower against all antigens and also IgG4 reactivity was sparse. The lymphoma group had the lowest CD4 cell values, a property that may decrease the CD4 help, particularly to IgG3 responses. With EBV p107 antibody, an inverse relationship is observed. Higher titers and frequencies could be observed in the lymphoma group although low levels of antibodies were found in patients where EBV DNA was present in the sera. The latter finding is in concordance with previous findings where low levels of p107 antibodies in EBV DNA positive subjects reflect an impaired EBV T-cell control [20]. This impairment may consist of an altered IL-10 expression with inhibitory effects of IL-10 or vIL-10 on T-cells. IgG subclass disorders may develop as a manifestation of an impairment of the B-cell function due to the malignancy per se, but more likely as a result of functional imbalance in immune regulation developing with declining CD4 cells. Previous findings point to a loss of HIV-1 specific minor IgG subclasses during disease progression [12]. We interpret the observed loss of IgG subclass reactivities in the lymphoma group as a sign of progression of disease since the mean CD4 count of these patients is below that of even the symptomatic group. IgG1 is the dominating subclass in anti-HIV antibody responses and is elicited in response to protein antigens together with IgG3. Among the anti-HIV IgG subclasses of sera from HIV-1 subjects, IgG1 was dominant regardless of the clinical stage of infection. The patients with lymphoma retained their responses to immunodominant antigens despite their generally low CD4 values. IgG1
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
35
Fig. 4. Inhibitory ELISA for gp41/JB4. Sera were added together with increasing amounts of the corresponding peptide. Six representative patient sera are shown. Solid symbols represent patients with low CD4 cell counts: 13× 106/l – 130 ×106/l. Open symbols represent patients with high CD4 counts: 430 × 106/l – 440 ×106/l.
responses to gp160 were for example slightly higher in lymphoma patients compared to the symptomatic patients. IgG2 is the predominant subclass in response to bacterial polysaccharides. It was present only in response to the immunodominant gp41 peptide. In general, IgG3 was significantly decreased in individuals with low CD4 counts and lymphoma patients often had lower IgG3 compared to the symptomatic patients. IgG4 typically follows repeated antigenic stimulation. HIV-specific IgG4 reactivity was seen almost exclusively in the symptomatic group compared to the asymptomatic and lymphoma groups, and thus might represent a late broadened subclass pattern to a high antigenic burden. The lymphoma patients, compared to the symptomatic patients, had decreased responses of IgG1 and IgG3 to gp120 peptide and IgG3 to gp41 and V3 peptides (Fig. 2). These findings could have implications in the development of lymphoma since gp120 and gp41 recently have been reported to stimulate B-cell growth and dif-
ferentiation through IL-10, possibly towards a development of malignant B-cell clones [26–28]. Since the presence of lymphomas related to HIV-infection was evident [29], several mechanisms have been suggested contributing to HIV1 related lymphomagenisis. Direct polyclonal activation of B lymphocytes [30], and the influence of Epstein-Barr virus infection are likely to have significant roles in the pathogenesis [28,31– 34]. Recent work points towards EBV negative tumors having a less aggressive clinical course [35]. The lymphoma patients with poor EBV p107 reactivities had detectable serum EBV DNA, which has previously been implicated as a marker for decreased T-cell control of EBVtransformed cells [20]. However, EBV specific IgG also appears to be important in inhibiting outgrowth of transformed EBV-carrying cells [13]. The present small material indicates that a retro- and prospective study of IgG3 changes during lymphoma development with concurrent analysis of HIV antigen reactivities should, indeed, be performed.
36
P. Lundholm et al. / Immunotechnology 4 (1998) 29–36
Acknowledgements This work was supported by funds from the Faculty of Odontology, Karolinska Institute, Swedish Medical Research Council and funds from the Stockholm County Council.
References [1] Levine AM. Blood 1992;80:8–20. [2] Beral V, Peterman T, Berkelman R, Jaffe H. Lancet 1991;937:805 – 9. [3] Ziegler JL, McGrath MS. In: Magrath IT, editor. The Non-Hodgkin’s Lymphomas: Lymphomas in HIV-Positive Individuals. London: Edward Arnold, 1990:155–159. [4] Knowles DM, Chamulak GA, Subar M, Burke JS, Dugan M, Wernz J, Slywotzky C, Pelicci PG, Dalla-Favera R, Raphael B. Ann Intern Med 1988;108:744. [5] Levine AM. Ann Oncol 1994;5(Suppl 2):29–35. [6] Sundqvist VA, Linde A, Wahren B. J Clin Microbiol 1984;20:94 – 8. [7] Linde GA. J Clin Microbiol 1985;21:117–21. [8] Khalife J, Guy B, Capron M, Kieny MP, Ameisen JC, Montagnier L, Lecocq JP, Capron A. AIDS Res Hum Retroviruses 1988;4:3–9. [9] Ljunggren K, Broliden PA, Morfeldt-Manson L, Jondal M, Wahren B. Clin Exp Immunol 1988;73:343–7. [10] Broliden PA, Morfeldt-Ma˚nsson L, Rosen J, Jondal M, Wahren B. Clin Exp Immunol 1989;76:216–21. [11] Jansson M, Wahren B, Scarlatti G, Principi N, Lombardi V, Livadiotti S, Elia L, Plebani A, Wigzell H, Rossi P. AIDS 1992;6:365 – 71. [12] Sundqvist V-A, Linde A, Kurth R, Werner A, Helm EB, Popovic M, Gallo RC, Wahren B. J Infect Dis 1986;153:970. [13] Abedi MR, Linde A, Christensson B, Mackett M, Hammarstrom L, Smith CI. Int J Cancer 1997;71:624–9. [14] Kieff E, Liebowitz D. In: Fields B, Knipe D, editors. Virology: Epstein-Barr Virus and Its Replication. New York: Raven Press, 1990:1889–1920. [15] Schols D, De Clercq E. J Virol 1996;70:4953–60.
.
[16] Briere F, Servet-Delprat C, Bridon JM, Saint-Remy JM, Banchereau J. J Exp Med 1994;179:757 – 62. [17] Sa¨llberg M, Rude´n U, Magnius LO, Norrby E, Wahren B. Immunol Lett 1991;30:59 – 68. [18] Hedman K, Seppa¨la¨ I. J Clin Immunol 1988;8:214 – 21. [19] Steward MW, Stanley CM, Dimarchi R, Mulcahy G, Doel TR. Immunology 1991;72:99 – 103. [20] Barkholt LM, Dahl H, Enbom M, Linde A. Transplant Int 1996;9:439 – 45. [21] Chiodi F, Mathiesen T, Albert J, Parks E, Norrby E, Wahren B. J Immunol 1989;142:3809– 14. [22] Javaherian K, Langlois AJ, LaRosa GJ, Profy AT, Bolognesi DP, Herlihy WC, Putney SD, Matthews TJ. Science 1990;250:1590– 3. [23] Broliden PA, VonGegerfelt A, Clapham P, Rosen J, Fenyo¨ EM, Wahren B, Broliden K. Proc Natl Acad Sci USA 1992;89:461 – 5. [24] Cocchi F, DeVico AL, Garzino-Demo A, Cara A, Gallo RC, Lusso P. Nat Med 1996;2:1244 – 7. [25] Palker TJ, Clark ME, Langlois A, Matthews TJ, Weinhold KJ, Randall RR, Bolognesi DI, Haynes BF. Proc Natl Acad Sci USA 1988;85:1932 – 6. [26] Edelman L, Deveau C, Raphael M, Monchatre E, Gabarre J, Deville-Chabrol A, Pialoux G, Emilie D, Joab I, Galanaud P. Eur Cytokine Netw 1996;7:785 – 91. [27] Miyazaki I, Cheung RK, Dosch HM. J Exp Med 1993;178:439 – 47. [28] Stuart AD, Stewart JP, Arrand JR, Mackett M. Oncogene 1995;11:1711 – 9. [29] Ziegler JL, Beckstead JA, Volberding PA, Abrams DI, Levine AM, Lukes RJ, Gill PS, Burkes RL, Meyer PR, Metroka CE, et al. N Engl J Med 1984;311:565 – 70. [30] Schnittman SM, Lane HC, Higgins SE, Folks T, Fauci AS. Science 1986;233:1084– 6. [31] Herndier BG, Kaplan LD, McGrath MS. AIDS 1994;8:1025 – 49. [32] Pelicci PG, Knowles DMd, Arlin ZA, Wieczorek R, Luciw P, Dina D, Basilico C, Dalla-Favera R. J Exp Med 1986;164:2049– 60. [33] Subar M, Neri A, Inghirami G, Knowles DM, Dalla-Faverta R. Blood 1988;72:667 – 71. [34] Muller JR, Janz S, Goedert JJ, Potter M, Rabkin CS. Proc Natl Acad Sci USA 1995;92:6577 – 81. [35] Kaplan LD, Shiramizu B, Herndier B, Hahn J, Meeker TC, Ng V, Volberding PA, McGrath MS. Blood 1995;85:1727 – 35.