JOURNAL OF VIROLOGY, Oct. 1996, p. 7327–7330 0022-538X/96/$04.0010 Copyright q 1996, American Society for Microbiology
Vol. 70, No. 10
Lymphoproliferative Responses after Infection with Human Parvovirus B19 ANDREAS
POBLOTZKI,* CHRISTIAN GERDES, UDO REISCHL, HANS WOLF, AND SUSANNE MODROW
VON
Institut fu ¨r Medizinische Mikrobiologie und Hygiene der Universita ¨t Regensburg, Universita ¨t Regensburg, D-93053 Regensburg, Germany Received 8 April 1996/Accepted 18 July 1996
Immunity after infection with the parvovirus B19 is assumed to be conferred by a humoral immune response with development of neutralizing antibody. In contrast, little is known about the nature of T-cell-mediated responses to parvovirus B19 infection in humans. We used recombinant proteins VP1, VP2, and NS1, as well as a recombinant VP1-specific amino-terminal sequence, to test the proliferative responses of peripheral blood mononuclear cells after infection of otherwise healthy individuals with parvovirus B19. These proteins were used as antigens for the stimulation of freshly isolated cells. The results show that a B19 virus-specific cellular immunity develops that is directed against the capsid proteins VP1 and VP2. We also demonstrate that viral determinants are presented to CD41 T cells by HLA class II molecules. responses directed against parvovirus B19 appears to be important for understanding viral pathogenesis. To investigate the presence of B19 virus-specific T cells after infection, we used B19 virus proteins as antigens to stimulate freshly isolated peripheral blood mononuclear cells (PBMCs) from these patients. Since the supply of native viral antigens was limited by the lack of a cell culture system that allows propagation of parvovirus B19, we generated recombinant viral proteins in bacteria. Construction of the expression vectors and purification of the proteins VP1, VP2, and NS1 have been described before (19). Briefly, the coding parts of the viral genome were amplified by PCR and inserted in frame with an amino-terminal histidine tag. After induction of expression in Escherichia coli, the proteins were purified by affinity chromatography and subsequent preparative sodium dodecyl sulfatepolyacrylamide gel electrophoresis. The purification of the VP1 region encompassing amino acids 1 to 227 of the VP1 protein was done similarly. These proteins were used to stimulate PBMCs from 16 voluntary blood donors. In 10 individuals, specific IgG was detected in a commercial parvovirus B19 enzyme-linked immunosorbent assay (ELISA) (Virotech GmbH, Ru ¨sselsheim, Germany), indicating previous viral infection. None of these persons had detectable amounts of virus-specific IgM in their sera; thus, acute infection could be excluded. Serum samples from the six remaining individuals showed no reactivity in the B19 virus-specific IgM and IgG (ELISAs), nor were there any clinical symptoms related to parvovirus B19 infection. The PBMCs were isolated by Ficoll-density centrifugation with Histopaque-1077 (Sigma Chemicals, Deisenhofen, Germany) and washed three times with phosphate-buffered saline (PBS) prior to use in proliferation assays. Cells (2 3 105 per well) were incubated at a concentration of 106 cells per ml in 96-well round-bottom microtiter plates without antigen to determine the baseline proliferation (referred to as blank values in the text) or with 1 mg of the recombinant B19 virus antigens per ml. PBMCs stimulated with tetanus toxoid antigen and phytohemagglutinin (PHA) (1 mg/ml) served as positive controls. RPMI 1640 medium containing 10% human serum, 2 mM glutamine, and 100 mg of gentamicin per ml was used in all assays. The purified tetanus toxoid was kindly provided by Klaus-Dieter Hungerer, Behringwerke AG, Marburg, Ger-
Infection with the human parvovirus B19 usually results in a mild disease known as erythema infectiosum, or fifth disease (2). Frequently, the patients develop arthralgias, which may become long lasting and can proceed to severe chronic arthritis resembling rheumatoid arthritis (4, 15, 20, 21). Complications arise in patients with underlying hemolytic diseases, such as patients with sickle cell disease in which parvovirus B19 is the main cause of aplastic crises (13, 17), AIDS patients, and otherwise immunocompromised people (5, 6, 8, 9). When infection is transmitted to the fetus during pregnancy, hydrops fetalis and fetal loss may occur (1, 16). The antibody response to parvovirus B19 is well characterized, whereas attempts to identify cellular immune reactions after infection have been unsuccessful (9). Several lines of evidence led us to examine the cellular response generated during B19 virus infection. First, recent investigations indicated that the persistence of the virus was the cause of complications in individuals suffering from B19 virus-associated arthritis or severe cytopenias. In these patients, viral DNA was detectable in various tissues despite the presence of neutralizing antibodies. In other patients, the antibody response appeared to be disturbed, manifested as a recurrent capsid protein-specific immunoglobulin M (IgM) response without the establishment of a stable IgG response (8, 14, 20). Recent data indicated that chronic anemia in AIDS patients also may be due to B19 virus infection (5). These individuals fail to produce stable B19 virus-specific antibody titers and, in turn, are unable to eliminate the virus after infection. This is most likely due to the destruction of CD41 T cells by human immunodeficiency virus type 1 leading to the absence of the T-cell help that is required for B-cell-mediated antibody production. These observations suggest that T cells are important to overcome B19 virus infection and to establish long-term immunity. Moreover, immune-mediated pathogenesis is suspected as a cause of several B19 virus-related clinical symptoms, including arthralgia (4, 12, 21). Therefore, the investigation of cellular immune
* Corresponding author. Present address: SmithKline Beecham Pharma GmbH, Leopoldstrasse 175, D-80804 Mu ¨nchen, Germany. Phone: 49-941-9446454. Fax: 49-941-9446402. Electronic mail address:
[email protected]. 7327
7328
NOTES
J. VIROL.
TABLE 1. Individual responses for all patients testeda SI (cpmwith antigen/cpmblank)
Patient VP1
VP2
DV
NS1
IgG positive 1 2 3 4 5 6 7 8 9 10
2.32 3.69 17.17 3.98 1.67 8.49 12.78 3.84 11.19 1.95
4.03 6.13 13.98 6.77 5.82 6.66 17.92 0.93 2.87 5.12
2.16 2.36 19.58 1.49 1.71 7.01 10.31 3.69 3.6 1.47
1.87 0.52 2.1 1.07 1.07 3.92 1.39 0.87 1.61 0.88
IgG and IgM negative 11 12 13 14 15 16
1.63 1.55 0.75 0.9 0.74 0.98
1.47 1.23 1.2 0.67 0.58 1.15
1.93 1.91 0.9 1.88 0.57 0.7
1.11 0.46 0.88 0.54 1.46 0.97
a PBMCs were isolated, and 2 3 105 cells per well were incubated without any antigen (blank) or with 1 mg of the corresponding parvovirus B19 antigen per ml. After 6 days, the cells were labeled with [3H]thymidine for 14 h and harvested on filter plates, and the amount of incorporated radioactivity was measured as described in the text. Because the baseline proliferation among different individuals was variable (480 up to 4,080 cpm), proliferation was represented as SIs to facilitate comparisons. The data are based on three independent experiments. Positive responses with SIs of more than 2.0 were observed only in persons who were seropositive for parvovirus B19-specific IgG.
many. The human serum was positive for B19 virus-specific IgG but not IgM. The presence of antibodies may aid antigen presentation through antibody-mediated antigen uptake by Fcreceptor-positive cells. After 6 days at 378C, 6% CO2, and 95% humidity, the cells were labeled with 1 mCi of [3H]thymidine (2 Ci/mmol [Amersham, Braunschweig, Germany]) per well for 14 h and then harvested on filter plates (GF/C Unifilter; Canberra Packard, Frankfurt, Germany) with an automated cell harvester. The filter plates were dried, and 30 ml of scintillation fluid (Microscint-O; Canberra Packard) was added per well before counting with a microplate scintillation counter (TopCount; Canberra Packard). All samples, including blank values, were run as triplicates. For comparison, the results from different patients were calculated as stimulatory indices (SIs) (cpmwith antigen/cpmblank). The responses to the two capsid proteins VP1 and VP2, the VP1-unique sequence (DV), and the NS1 protein were investigated for all 16 participants. The general proliferative capacity of the cells was assessed by stimulation with PHA or purified tetanus toxoid, since all donors were vaccinated against tetanus. SIs in these controls were usually greater than 100 (data not shown). Statistical analysis comparing the blank and antigen stimulation values for each capsid antigen within the groups showed significant differences only in the seropositive group (Student’s t test; P , 0.05). Moreover, we found that an efficient stimulation of the cells with an SI of .2.0 to any of the antigens was found only in the seropositive group (Table 1 and Fig. 1). According to previous results, we defined values above this threshold as positive (10, 11, 18). Individuals in the seropositive group responded to at least one of the capsid antigens. Nine of 10 donors displayed VP2-specific reactions, 8 of 10 samples contained cells stimulated by VP1, and 6 of 10 persons appeared to recognize the VP1-unique sequence. There were
only two examples of reactivity to VP2 and there was only one VP1-unique-specific response seen by stimulation by both VP1 and the VP1-unique sequence, but not by VP2. The sera which tested positive for B19 virus IgG contained antibodies directed against at least one of the three capsid proteins tested (Table 1). No statistically significant differences between seropositive and seronegative persons were observed for the NS1 protein, but two individuals had SIs .2.0, indicating a specific response to this protein. Both individuals had low titers of NS1 IgG antibody, a situation that has been described only for persistently infected persons. Since viral DNA was not found in the sera nor were any clinical symptoms related to B19 virus infection observed, these persons do not seem to be persistently infected. Both had worked with recombinant NS1 protein in the laboratory, and, therefore, they may have been exposed repeatedly to this antigen. Such exposure may have generated the specific immune responses we detected. To define the specific immune cells responding to particular antigens, we investigated the influence of HLA-specific monoclonal antibodies on the reactions of four patients. The monoclonal antibodies W6/32 and L243 recognize constant regions of HLA class I and class II molecules, respectively. They have been shown to block specific proliferation responses induced through antigen presentation by HLA molecules (3, 11). These monoclonal antibodies were added to the culture at setup at a final concentration of 1 mg/ml. After 6 days, the cultures were labeled with [3H]thymidine and harvested 14 h later. The
FIG. 1. Proliferation of isolated PBMCs from four patients (1, 3, 6, and 7) after B19 virus infection. Ficoll-gradient purified PBMCs (2 3 105) from these patients were stimulated with the parvoviral proteins purified from E. coli or with tetanus toxoid (TT) as control. Blank values give the proliferation with medium only. All individuals displayed a specific response upon exposure to at least one of the capsid proteins VP1 or VP2 or to the VP1-unique sequence (DV). Little or no proliferation was observed after stimulation with the nonstructural protein NS1. Values for the two patients (3 and 6) with NS1-specific lymphocytes among the 10 individuals investigated are shown. All values are given as counts per minute of [3H]thymidine incorporated and are the means of three experiments. Error bars indicate 1 standard deviation.
VOL. 70, 1996
NOTES
7329
Depletion of all CD31 T lymphocytes or the CD41 fraction of cells removed the antigen-responsive population of the PBMCs. It did not, however, impair the general proliferative capacity or viability of the cells, as demonstrated by the strong incorporation of radioactivity after stimulation with PHA (Fig. 3). In contrast, PBMCs depleted of CD81 T lymphocytes were still able to respond to the parvoviral antigens. However, the overall responses in these experiments were diminished slightly compared with the control responses (Fig. 3). In summary, our results suggest that infection with parvovirus B19 activates not only a humoral response but also a cellular immune response. The virus does not behave immunologically differently from other viruses such as poliovirus or rubella virus (7, 10, 11, 18). The inability of other groups to detect B19 virus-specific cellular responses may reflect difficulty in producing sufficient amounts of native virus for stimulation experiments in vitro (9). The use of recombinantly produced proteins allowed us to produce enough antigen not only for detection of memory cells but also for analysis of responses to all of the viral antigens encoded by parvovirus B19. The CD41 T lymphocytes appear to make up the major population of reactive cells. These lymphocytes function mainly as T-helper cells for B cells leading to the maturation of the antibody response with resultant long-term immunity. They can provide support for specific B cells producing antibodies to
FIG. 2. Impact of HLA class-specific antibodies on the cellular response. PBMCs of four patients (1, 3, 6, and 7) were tested. PBMCs (2 3 105) were incubated with medium only (blank) or 1 mg of the capsid proteins per ml for 6 days before being labeled with radioactivity. The presence of antibody W6/32, specific for HLA class I molecules, does not significantly reduce the amount of [3H]thymidine incorporated. In contrast, antibody L243, specific for class II molecules, abrogates the proliferation upon antigen exposure. Values are the means of three independent experiments.
amount of incorporated radioactivity was determined as described above. The presence of W6/32 had little or no effect on the incorporation of radioactive nucleotides by the proliferating cells (Fig. 2). In contrast, the class II-specific L243 antibody strongly reduced or fully abrogated the responses to the structural proteins in all cases investigated. This strongly indicated that a major component of the cellular immune response was restricted by HLA class II molecules. This would indicate involvement of CD41 cells as an important effector cell population. We depleted all T cells by using a CD3-specific monoclonal antibody and magnetic separation or selectively depleted CD41 or CD81 T lymphocytes. The antibodies used for selection were derived from clones UHCT-1 (CD3), UHCT-4 (CD8), (Biozol, Munich, Germany) and T151 (CD4; Boehringer Mannheim, Mannheim, Germany). After isolation, the cells were incubated with the appropriate monoclonal antibody for 30 min at 48C. After threefold washing with PBS–5% fetal calf serum, magnetic beads coated with mouse IgG-specific secondary antibodies were added (Dynal GmbH, Hamburg, Germany), and the incubation continued for another 30 min at 48C. Bound cells were removed by magnetic separation. The separation was repeated, and the remaining cell fraction was washed with PBS–5% fetal calf serum, concentrated by centrifugation, adjusted to 2 3 106 cells per ml in RPMI 1640 containing the necessary supplements, and used in the stimulation assays as described above.
FIG. 3. Lymphoproliferative response of patient 7 after selective depletion of T-cell subsets. PBMCs from patient 7 were used without depletion of T cells (control) or were deprived either of all T cells by a CD3-specific monoclonal antibody and magnetic separation (CD3) or of the CD41 or CD81 populations (CD4 and CD8, respectively). A total of 2 3 105 of the remaining cells were exposed to the viral antigens VP1 and VP2 as described in the text. The removal of all T lymphocytes or the CD41 fraction abolishes the ability to respond to the capsid proteins or tetanus toxoid (TT). The viability of the cells is not impaired, as demonstrated by proliferation after incubation with PHA. Values are the means of three experiments.
7330
NOTES
J. VIROL.
VP1, VP2, and the VP1-unique sequence even in persons in whom cellular responses only against VP2 or VP1 were detected. The absence of Ig class switch and affinity maturation in patients with severe clinical symptoms due to persistent B19 virus infection strongly hints to a T-cell defect in these cases (9, 14, 19). The finding that some of these persons show unaltered responses to other antigens, such as tetanus toxoid, and are not more prone to infection than the general population favors a model that postulates the absence of specific helper functions rather than an inability to process and present antigens to T lymphocytes. The presence of proliferation deficiencies in response to some, but not all, antigens strengthens this hypothesis (8). Whether this is caused by an altered T-cell-receptor repertoire or is linked to certain HLA haplotypes remains to be investigated. Further experiments defining the major cellular epitopes recognized by the CD41 T cells may help to clarify this question. We believe that the importance of the cellular response for immunity to parvovirus B19 should be considered in the development of a B19 virus-specific vaccine.
6. 7. 8.
9. 10. 11. 12. 13. 14.
The monoclonal antibodies W6/32 and L243 were kindly provided by Christoph Bogedain, Institut fu ¨r Medizinische Mikrobiologie und Hygiene, Universita¨t Regensburg. We also thank Rudi Wank and Dolores Schendel, Institute for Immunology, Ludwig-MaximiliansUniversita¨t Mu ¨nchen, and Wolfgang Jilg, Institut fu ¨r Medizinische Mikrobiologie und Hygiene, for initial help with the proliferation experiments. REFERENCES 1. Anderson, L. J. 1990. Human parvoviruses. J. Infect. Dis. 161:603–608. 2. Anderson, M. J., S. E. Jones, S. P. Fisher-Hoch, E. Lewis, S. M. Hall, C. L. R. Bartlett, B. J. Cohen, P. P. Mortimer, and M. S. Pereira. 1983. Human parvovirus: the cause of erythema infectiosum (fifth disease)? Lancet i:1378. 3. Bogedain, C., H. Wolf, S. Modrow, G. Stuber, and W. Jilg. 1995. Specific cytotoxic T lymphocytes recognize the immediate-early transactivator Zta of Epstein-Barr virus. J. Virol. 69:4872–4879. 4. Cohen, B. J., M. M. Buckley, J. P. Clewley, V. E. Jones, A. H. Puttick, and R. K. Jacoby. 1986. Human parvovirus infection in early rheumatoid and inflammatory arthritis. Ann. Rheum. Dis. 45:832–838. 5. Frickhofen, N., J. L. Abkowitz, M. Safford, J. M. Berry, J. Antunez de Mayolo, A. Astrow, R. Cohen, I. Halperin, L. King, and D. Mintzer. 1990. Persistent B19 parvovirus infection in patients infected with human immu-
15. 16. 17.
18.
19.
20.
21.
nodeficiency virus type 1 (HIV-1): a treatable cause of anemia in AIDS. Ann. Intern. Med. 113:926–933. Frickhofen, N., and N. S. Young. 1989. Persistent parvovirus B19 infections in humans. Microb. Pathog. 7:319–327. Graham, S., E. C. Y. Wang, O. Jenkins, and L. K. Borysiewicz. 1993. Analysis of the human T-cell response to picornaviruses: identification of T-cell epitopes close to B-cell epitopes in poliovirus. J. Virol. 67:1627–1637. Kurtzman, G., N. Frickhofen, J. Kimball, D. W. Jenkins, A. W. Nienhuis, and N. S. Young. 1989. Pure red-cell aplasia of 10 years’ duration due to persistent parvovirus B19 infection and its cure with immunoglobulin therapy. N. Engl. J. Med. 321:519–523. Kurtzman, G. J., B. J. Cohen, A. M. Field, R. Oseas, R. M. Blaese, and N. S. Young. 1989. Immune response to B19 parvovirus and an antibody defect in persistent viral infection. J. Clin. Invest. 84:1114–1123. Lovett, A. E., M. McCarthy, and J. S. Wolinsky. 1993. Mapping cell-mediated immunodominant domains of the rubella virus structural proteins using recombinant proteins and synthetic peptides. J. Gen. Virol. 74:445–452. McCarthy, M., A. Lovett, R. H. Kerman, A. Overstreet, and J. S. Wolinsky. 1993. Immunodominant T-cell epitopes of rubella virus structural proteins defined by synthetic peptides. J. Virol. 67:673–681. Naides, S. J., L. L. Scharosch, F. Foto, and E. J. Howard. 1990. Rheumatologic manifestations of human parvovirus B19 infection in adults. Initial two-year clinical experience. Arthritis Rheum. 33:1297–1309. Pattison, J. R., S. E. Jones, and J. Hodgson. 1981. Parvovirus infections and hypoplastic crises in sickle cell anaemia. Lancet i:664. Pont, J., E. Puchhammer Stockl, A. Chott, T. Popow Kraupp, H. Kienzer, G. Postner, and N. Honetz. 1992. Recurrent granulocytic aplasia as clinical presentation of a persistent parvovirus B19 infection. Br. J. Haematol. 80: 160–165. Reid, D. M., T. M. Reid, T. Brown, J. A. Rennie, and C. J. Eastmond. 1985. Human parvovirus-associated arthritis: a clinical and laboratory description. Lancet i:422–425. Schwarz, T. F., M. Roggendorf, B. Hottentra ¨ger, F. Deinhardt, G. Enders, K.-P. H. Gloning, T. Schramm, and M. Hansmann. 1988. Human parvovirus infection in pregnancy. Lancet ii:566–567. Serjeant, G. R., B. E. Serjeant, P. W. Thomas, M. J. Anderson, G. Patou, and J. R. Pattison. 1993. Human parvovirus infection in homozygous sickle cell disease. Lancet 341:1237–1240. Simons, J., M. Kutubuddin, and M. Chow. 1993. Characterization of poliovirus-specific T lymphocytes in the peripheral blood of Sabin-vaccinated humans. J. Virol. 67:1262–1268. von Poblotzki, A., A. Gigler, B. Lang, H. Wolf, and S. Modrow. 1995. Antibodies to parvovirus B19 NS-1 protein in infected individuals. J. Gen. Virol. 76:519–527. von Poblotzki, A., A. Hemauer, A. Gigler, E. Puchhammer-Sto¨ckl, F.-X. Heinz, J. Pont, K. Laczika, H. Wolf, and S. Modrow. 1995. Antibodies to the non-structural protein of parvovirus B19 in persistently infected patients: implications for pathogenesis. J. Infect. Dis. 172:1356–1359. White, D. G., A. D. Woolf, P. P. Mortimer, B. J. Cohen, D. R. Blake, and P. A. Bacon. 1985. Human parvovirus arthropathy. Lancet i:419–421.
