- Email: [email protected]
FIV vaccine studies. I. Immune response to recombinant FIV env gene products and outcome after challenge infection
Veterinary Immunology and Immunopathology 46(1995)103-113
ELSEVIER
Veterinary immunology and immunopathology
FIV vaccine studies. I. Immune response to recombinant FIV env gene products and outcome after challenge infection -I. Lutz”,” , R. Hofmann-Lehmann”, K. Bauer-Pham”, E. Holznagel”, F. Tozzinib, M. Bendinellib, G. Reubel’, A. Aubertd, D. Davis”, D. Coxe, E. Younge aDepartment of Internal Veterinary Medicine, University of Zurich, Zurich, Switzerland “Department of Animal Pathology, University of Pisa, Pisa, Italy ‘Department of Veterinary Medicine, University of California, Davis, CA, USA *Virbac Laboratories Inc. Nice. France ‘Cambridge Biotech Inc.. Worcester, MA, USA
Abstract We have vaccinated five groups of cats (n = 25) four times with five preparations of recombinant feline immunodeficiency virus (FIV) env gene products; one group (n = 7) served as control. The vaccine formulations were as follows: ( 1) envelope glycoprotein of FIV Zurich 2 (FIV Z2) expressed in a Baculovirus system and isolated by gel electroelution (denatured form) ; (2) insect cells expressing FIV 22 glycoprotein; (3) envelope glycoprotein of a Boston strain (FIV Bangston) expressed in insect cells and isolated by gel electroelution (denatured form); (4) glycosylated Bangston envelope protein made in insect cells and isolated in a native form; (5) non-glycosylated Bangston envelope protein made in Escherichia cofi. All cats were challenged with 20 50% cat infective doses (CIDS,) of FIV 22 previously titrated in cats. All vaccinated cats developed high enzyme-linked immunosorbent assay ( ELISA) antibodies to the homologous antigen; crossreactivity to heterologous antigens was seen at a lower level. Virus neutralizing antibodies (tested with Petaluma virus) reached titers up to 32. After challenge, all cats seroconverted (as judged by anti gag antibodies in Western blot) and became infected (as judged by virus isolation and/or polymerase chain reaction) between 4 and 11 weeks with the exception of one cat. It is concluded that it is relatively easy to induce high ELISA antibody liters using recombinant enu gene products. ELISA antibody titers do not correlate with virus neutralization or with protection.
* Corresponding author. 016%2427/95/$09.50
0 1995 Elsevier Science B.V. All rights reserved
SSDIO165-2427(94)07010-5
104
H. Lutz et al. / Veterinary Immunology and Immunopathology 46 (1995) 103-l 13
Abbreviations CID,,,, 50% cat infective dose; ELBA, enzyme-linked immunosorbent assay; FeLV, feline leukemia virus; FIV, feline immunodeficiency virus; PCR, polymerase chain reaction; SIV, Simian immunodeficiency virus; SPF, specified pathogen free; VN, virus neutralizing (antibodies).
1. Introduction The feline immunodeficiency virus (FIV) is a lentivirus infecting cats leading to immunosuppression (Pedersen et al., 1987) by gradual destruction of the CD4+ lymphocytes (Torten et al., 1991) . The infection is widespread in domestic cats and occurs worldwide. Up to now no etiologic treatment and no vaccine have been available to control FIV infection under field conditions. Until a few years ago it was generally believed that vaccines protecting against retrovirus infections would be almost impossible to develop because during replication retroviruses become integrated into the host genome in the form of the provirus which may persist and be re-activated after long periods. Only recently have effective vaccines against feline leukemia virus (FeLV) been introduced into the field to control FeLV infection. FeLV and FIV and the host’s immune responses to these agents are different in many aspects. Development of an FIV vaccine may be more complicated than was the case with the FeLV vaccine. As FIV is judged to be a good model for HIV and because FIV infection is an important feline disease, several groups have begun work on developing a vaccine against FIV infection. Very promising data have been reported by Yamamoto’s group (Yamamoto et al., 1991, 1993). In their experiments inactivated FIV-infected cells and inactivated and purified FIV were used as vaccine antigens. Other approaches using purified viral proteins were less succesful (for review see Hosie, 1994). Based on the positive experience we had with a recombinant FeLV SU vaccine (Lehmann et al., 1991)) we investigated whether cats could be protected against FIV infection by vaccination with recombinant FIV SU expressed in a Baculovirus system and in Escherichia coli. The use of recombinant envelope components as immunogens appeared to be especially attractive because of the relative ease with which these antigens can be produced in defined purity and the fact that seroconversion induced by natural infection can clearly be differentiated from vaccine-induced antibodies by demonstration of anti gag antibodies.
2. Materials and methods
2.1. Study design Em gene products derived from two FIV isolates were used to vaccinate five groups of cats (n = 25); seven cats served as non-vaccinated controls. The isolates included the strains Zurich 2 (FIV 22; Morikawa et al. 1990) and Bangston isolated from a cat in the Boston area (Young E., personal communication, 1994). The SU components of the two isolates differ in 15% of their amino acid composition (Young E.,
H. Lutz et al. /Veterinary
Immunology
and Immunopathoiogy
105
46 (1995) 103-l 13
personal communication, 1994). By challenging the vaccinated animals with the FIV 22 strain, homolgous and/or heterologous protection was to be examined. 2.2. Cats and sample collection Specified pathogen free (SPF) cats of both sexes were obtained at the age of 15 weeks from Ciba-Geigy, Werk Stein, Stein, Switzerland. They were housed under climatized conditions in a specialized facility at the University of Zurich. Males were castrated and randomly assigned with females to the different vaccine groups. Prior to the first vaccination, the animals were adapted to the new housing conditions for a period of 4 weeks. All cats were clinically examined once per week. Blood and serum samples were collected under slight sedation under routine conditions. 2.3. Vaccine antigens and vaccination The vaccine antigens were expressed in the Baculovirus system (FIV 22 and Bangston) and in E. coti (FIV Bangston) and applied in highly purified denatured or native form (Table 1; Young, E, et al., in preparation). Vaccine antigens were applied at Weeks 0,2,4 Table 1 Vaccine antigens used to vaccinate
the different groups of cats
No. of cats
FIV isolate
Expressed in
Nature of antigen
Amount per injection
I
s
2 3 4 5 6
s 5 5 5 7
Zurich 2 Zurich 2 Bangston Bangston Bangston Ovalbumin
Insect cells Insect cells Insect cells Insect cells E. coli NA
Highly purified, denatured Native Highly purified, denatured Native Highly purified, denatured NA
‘00 CLg Corr. lo* cells
Group
NA, not applicable.
Corr.. corresponding
100 !J,g Corr. 10’ cells 100 CLg 100 pg
to.
Table 2 Levels of antibodies to the three antigens used in the study determined by ELISA Group no.
Antigen used in vaccine
FIV 22 Baculovirus FIV Bangston Baculovirus FIV Bangston
Nature of antigen
Purified Native Purified Native Purified
Antigen used for testing FIV 22 Baculovirus
FIV Bangston Baculovirus
FIV Bangston
100% 62% 11% 74% 26%
28% 14% 100% 39% 108%
4% 4% 78% 21% 100%
0%
0%
0%
E. coli
E. coli
Control The homologous
Ovalbumin
reaction was defined as 100%
H. LUIZet al. /Veterinary Immunology and Immunopathology 46 (1995) 103-113
106
cats/g1 70%
~.
60%
;
50% 4 40%
j
30%
1
20%
.I
.; t
a
(8 group
1
E&
gro”p
2
group
4
F_Z,
group
5
Fig. 1. VN antibodies were determined in serum samples collected considered to be induced by the respective vaccine preparations.
titer
32
16
.....I m
group
3
group
6 (control)
on tbe day of challenge.
Titers > 8 were
and 8 by intramuscular injection in conjunction with 2 mg of aluminium hydroxide and 20 pg of QS2 1, a non-toxic fraction prepared from Quilluju saponaria ( Kensil et al., 199 1) , as adjuvants.
2.4. Challenge infection
Challenge virus infection was done in Week 10 by intraperitoneal injection of 20 50% cat infectious doses (CID& of the FIV 22 strain previously titrated in ten cats 15 weeks of age.
2.5. Antibodies determined by ELlSA and Western blots
Antibody levels to the three purified Baculovirus and E. coli expression products were determined by enzyme-linked immunosorbent assay (ELISA) under conditions described previously (Lehmann et al., 1991). Antigens were coated in amounts of 100 ng per well, sera were diluted 1:5000, and the conjugate (goat anti cat IgG horseradish peroxidase (Milan, Duedingen, Switzerland) was used at a dilution of 1:lOOO. The absorbance values were determined in a Dynatech reader M580 (Dynatech, Embraport, Switzerland) and the homologous reaction was defined as 100%. Western blots were performed under conditions described (Lutz et al., 1988a) using 500 ng of gradient purified FIV 22.
H. Lutz et al. /Veterinary Immunology and Immunopathology 46 (1995) 103-I 13
107
challenge c
101112
week o
14
16
18
2021
cat no 405 group
1
416 425 434 449 411 421
group
2
441 454 428 1
3
,“,“;
406 420
group
456
group
5
411) 419 432 435 444 : 4oi 415 414
group
6
4’33 439 453 461
n FlVpr~,t,ie
I
~~ ~.___-~__~
up24
qLI.E
3FI‘., ___
ne&31&
Fig. 2. Western blot results obtained from all cats before (three samples) and after (seven samples) infection. For clarity, reactions with the gag components p24 and p I7 are summarized in this figure.
2.6. Virus neutralizing
challenge
(VN) antibodies
Sterile serum samples were tested by a procedure described elsewhere using the Petaluma strain (Tozzini et al., 1992). 2.7. Virus isolation Virus isolation was performed as described 2.8. Polymerase
(Lutz et al., 1988b).
chain reaction (PCR)
Peripheral blood mononuclear cells were purified by Ficoll/Hypaque gradients (Pharmacia, Duebendorf, Switzerland), DNA extracted from lo6 cells was used for nested PCR under conditions described (Reubel et al., 1994). Briefly, a gag sequence was amplified covering nucleotides 929-l 394 (465 base pairs (bp) ) in the first round and 1236-1394
108
H. Lutz et al. /Veterinary Immunologyand Immunopathology46 (1995)103-113 group
1
group
2
group
3
group
4
group
5
group
6 0
1
2
3 no positive
4 of
5
6
7
cats
J
negative
Fig. 3. Virus isolation assays performed with blood samples collected 10 weeks after challenge infection.
(158 bp) in the second round. With every PCR cycle and every gel run, a DNA sample collected from a known infected and a known non-infected SPF cat were included.
3. Results 3.1. Serological
response to vaccination,
antibodies determined by ELISA
Serum samples collected on the day of challenge infection were tested for antibodies to the different SU components by ELISA (Table 2). From Table 2 it can be concluded that the homologous reaction was considerably stronger than the heterologous reaction. Glycosylation which was the main difference between the Baculovirus and E. coli expression products had a less pronounced effect on the antibody reaction. Serum samples collected from cats immunized with the Bangston E. coEi expression product contained antibodies that displayed a stronger reaction with the Baculovirus expression product than with the protein expressed in E. coli. 3.2. VN antibodies Serum samples collected under sterile conditions on the day of challenge were analysed for VN activity (Fig. 1). The highest titers observed were 32. In Group 3 (Bangston Baculovirus expression product) four of five cats had VN antibodies of I6 and higher. In Group 1 (Zurich Baculovirus expression product) and in Group 5 (Bangston E. coli) one cat each had a titer of 32. All other cats including those of the control group (Group 6) had titers of 8 or below. 3.3. Western blot results throughout the course of challenge infection Serum samples collected Week 10 (day of challenge)
at Week 0 (first vaccination), Week 4 (third vaccination), and on seven occasions throughout the challenge observation
H. Lutz et al. /Veterinary Immunology and Immunopathology 46 (19951 103-l 13
109
Table 3
WOUD CM
week12
week16
week46
period ( up to Week 2 1) were submitted to Western blot analysis (Fig. 2). With the exception of one cat (Cat 406 in Group 3) all cats eventually seroconverted as judged by the development of antibodies specific for p24 and ~15. Cats of Groups 2 and 5 appeared to seroconvert as a group somewhat earlier than cats of Groups 1, 3,4 and 6. 3.4. Virus isolation Peripheral blood mononuclear cells were purified from blood samples collected on the day of challenge and 10 weeks after challenge. The cells were co-cultivated with mononuclear cells collected from SPF cats for 4 weeks. The supernatants were tested for the presence of FIV p24 in Weeks 2,3 and 4 after establishing the cultures. Virus isolation was considered positive when one of the culture supernatants showed a positive FIV p24 test (Fig. 3). Blood samples tested on the day of challenge were all negative (data not shown). Ten weeks after challenge infection, in Groups 1 and 4 all cats were virus isolation positive. In Groups 2, 3 and 5, FIV was isolated from four of five cats. In the control group FIV was isolated from five of seven animals. 3.5. PCR results Mononuclear were subjected
cells collected 10 weeks before and 2, 12, 16 and 48 weeks after challenge to PCR analysis (Table 3). With the exception of one cat (Cat 406) all
110
H. Lutz et al. / Veterinary Immunology and Immunopathology 46 (1995) 103-l 13
animals were positive by PCR at least once until Week 16 after challenge. Cat 424 was positive by PCR 10 weeks before challenge; this result is considered to be due to contamination of the PCR reaction tube. At Week 48 some cats had become PCR negative. Cat 406, which had remained PCR-negative until Week 16 after challenge, was PCR-positive in Week 48.
4. Discussion In our experiments we have vaccinated five groups of cats with different preparations of recombinant FIV SU. After challenge infection with 20 ClD,,,, all control cats and all vaccinated cats became seropositive with the exception of one vaccinated animal in Group 3. From the Western blot results and the virus isolation and PCR experiments, it was concluded that there was no complete protection against homologous and heterologous challenge infection. Protection against FIV infection was recently demonstrated using inactivated virus infected cells or inactivated cell-free FIV vaccines (Yamamoto et al., 1991, 1993). Protection was correlated with high titers of VN antibodies; cats with lower anti SU and VN antibody titers were not protected (Yamamoto et al., 1993). Passive immunization using sera from FIV vaccinated cats or from cats with FIV infection also resulted in protection. However, VN antibodies alone may not always be sufficient for protection. This was concluded from experiments of the Glasgow group in which it was found that some kittens with maternal antibodies became infected (Callanan et al., 1991). Obviously, VN antibodies are important for protection against FIV infection at least to some degree as a first line of defense. For the prevention or reduction of cell-to-cell transmission of FIV, a strong cellular immune response will certainly also be required. The immunization of our cats induced high levels of antibodies detectable by ELISA. Immunologic differences between the two isolates were readily revealed by ELISA. Reactivity in the homologous assay (i.e. cats vaccinated with FIV 22 protein assayed on FIV 22 protein coated plates) was stronger than the heterologous immune response. However, there was also some cross-reaction between the isolates. This cross-reactivity may be explained in part by antibodies directed against the carbohydrate moiety of the glycoproteins. This was concluded from the observation that cats vaccinated, for example, with the FIV 22 glycoprotein showed a much stronger reaction with the Bangston SU component expressed in Baculovirus than with the Bangston E. coli expression product which is not glycosylated. That part of the antibodies were directed to the glycoproteins was also concluded from the observation that the animals immunized with the Bangston SU component expressed in the Baculovirus system had higher antibody levels against the Bangston Baculovirus expression product than against the SU expressed in E. coEi. It is less clear why cats immunized with the E. coli expression product made higher antibody levels against the Baculovirus-expressed material than with the SU synthesized by E. coli. One explanation could be that the Baculovirus expression product is at least to some degree adsorbed to the surface of the ELISA well via the carbohydrate part, rendering more epitopes accessible to antibodies than when the E. coli expression product is adsorbed to the plastic surface or that the glycosylated protein binds more easily to the plate. The cats immunized with the native antigens had somewhat lower antibody levels than the animals which received the purified
H. Lutz et al. /Veterinary Immunology and Immunopathology 46 (1995) 103-I 13
III
immunogens. This could be explained by the possibly lower concentration of the SU component in this antigen preparation. The SU concentration could not be determined precisely because of the other proteins present in the insect cells. VN antibodies with titers of 16 and 32 were found on the day of challenge in some cats immunized with purified immunogens (Groups 1, 3 and 5). These titers were considered to be induced by the vaccine antigens as the control cats had titers of 8 or lower. These VN titers were much lower than those found in cats with natural FIV infection (Tozzini et al., 1992) and in cats vaccinated with an inactivated FIV-infected cell vaccine (Yamamoto et al., 1991) or an inactivated cell-free FIV preparation (Yamamoto et al., 1993). The relatively low VN titers found in the present study probably were not sufficient for complete neutralization of the challenge virus and therefore inadequate for protection. Although the virus strain used for the determination of VN antibodies differed from the vaccine strains, this probably does not explain the low titers. Yamamoto et al. (1993) were able to demonstrate in a similarly heterologous system high VN titers. The discrepancy between high ELISA and relatively low VN titers may be explained by the fact that the purified immunogens were denatured and therefore did not have the three-dimensional conformation of the natural glycoprotein. As we were aware of this possibility, cats of Groups 2 and 4 were vaccinated with native SU present in the insect cells. The SU in the native form was unable to induce VN antibodies although high levels of ELISA antibodies were found. It is difficult to understand why cell-free virus and inactivatedcells but not purified or native recombinant SU can induce VN antibodies. The nature of the glycosylation by the Baculovirus may explain some but not all of the difference. By analogy to the FeLV system, where VN antibodies were readily induced by a recombinant FeLV SU (Lehmann et al. 1991), we had expected that VN antibodies would also be found in the FIV experiments described here. In the Simian immunodeficiency virus (SIV) system, protection was found in experiments in which the vaccine virus was grown in the same cells (C8166) as those in which the challenge virus was prepared (Stott et al., 1991). Some degree of protection could also be induced by vaccination of macaques with uninfected C8166 cells (Stott et al., 1991). Protection by antibodies directed against cell constituents can be explained by the fact that retroviruses incorporate cellular components into their membranes during virus assembly and budding from the cell (Orenta and Hildreth, 1993). It can be postulated that a combination of anti cellular and anti viral antibodies are especially efficacious in inducing VN antibodies. In the cat, cellular components were shown to be less important for protection against FIV than in the SIV/macaquqe model (Hohdatsu et al., 1993; Hosie, 1994). The fact that monkeys vaccinated by recombinant gp 160 of SIV were protected against challenge infection (Hu et al., 1992) is suggestive that anti viral immune reaction can be achieved and will be efficacious in preventing lentivirus infections. Whether or not the cats of this study may have been partially protected resulting in a reduced virus load is unclear. Quantitative PCR and and serial virus isolation experiments were not performed. However, visual interpretation of Fig. 2 leads to the impression that the animals of Groups 1,3 and 4 seroconverted to p24 and ~15 somewhat later than the cats of the control group and of Group 2. In addition, the negative PCR results of the three cats of Group 1 available in Week 48 also suggested that in this group the virus load may have been smaller than in the control group. Cat 406 in Group 3 turned positive by PCR in Week
112
H. Lurz et al. /Veterinary Immunology and Immunopathology 46 (1995) 103-113
48. Whether it became truly infected by contact with the other cats that were FIV-positive or whether this positive result is the consequence of a laboratory contamination was not determined. Hosie et al. ( 1992) reported that cats immunized with different vaccine preparations including recombinant p24 showed enhancement of infection after challenge infection. In our study, based on Western blot results, there seemed to be no faster seroconversion in vaccinated cats over control animals, with the possible exception of Group 2 where complete seroconversion to p 24 and ~15 was observed in four of five cats already in Week 16. In conclusion, it will be important in future experiments to improve induction of VN antibodies and to quantitate virus load througout the course of challenge infection. It has to be expected that partial protection can be induced which may lead to lower virus load and therefore to improved clinical outcome. Use of recombinant SU components thereby allows the differentiation of vaccine-induced seroconversion and natural infection.
Acknowledgments This project was supported by a grant from the Swiss Banking Corporation on behalf of a customer, by the European concerted action on FIV vaccination and by the Rassekatzenvereinigung Ostschweiz. Cat food was kindly donated by Effems, Zug, Switzerland. The expert technical assistance of Zihu Zhang and Peter Fidler is acknowledged.
References Callanan, J.J., Hosie, M.J. and Jarret, O., 1991. Transmission of feline immunodeficiency virus from mother to kitten. Vet. Rec., 128: 322-323. Hohdatsu, T., Pu, R., Torres, B.A., Trujillo, S., Gardner, M.B. and Yamamoto, J.K., 1993. Passive antibody protection of cats against feline immunodeficiency virus infection. J. Virol., 67: 2344-2348. Hosie, M.J., 1994. The development of a vaccine against feline immunodeticiency virus. Br. Vet. J., 150: 25-39. Hosie, M.J., Osborne, R.. Reid, G., Neil, J. and Jarrett, 0.. 1992. Enhancement after feline immunodeficiency virus vaccination. Vet. Immunol. Jrnmunopathol., 35: 191-197. Hu, S.L.. Abrams, K., Barber, G.N., Moran, P., Zarling, J.M., Langlois. A.J., Kuller, L., Morton. W.R. and Benveniste, R.E., 1992. Protection of macaques against SIV infection by subunit vaccines of SIV envelope glycoprotein gp 160. Science, 255: 456-459. Kensil, C.R., Patel, U.. Lennick. M. and Marciani, D., 1991. Separation and characterization of saponins with adjuvant activi’y from QuiZlajasaponaria Molina cortex. J. Immunol., 146: 431437. Lehmann, R., Franchini, M., Aubert. A., Wolfensberger, C., Cronier, J. and Lutz, H.. 1991. Vaccination of cats experimentally infected with feline immunodehciency virus with a recombinant feline leukemia virus vaccine. J. Am. Vet. Med. Assoc., 199: 1446-1452. Lutz. H., Arnold, P., Htibscher, U.. Egberink, H., Pedersen, N.C. and Horzinek, M.C., 1988a. Specificity assessment of feline T-lymphotropic Jentivirus serology. Zbl. Vet. Med. B, 35: 773-778. Lutz, H.. Egberink, H. Arnold, P., Winkler, G., Wolfensberger, C., Jarrett, 0.. Parodi, A.L., Pedersen, N.P. and Horzinek, M.C., 1988b. Felines T-Jymphotropes Lentivirus (FTLV): Experimentelle lnfektion und Vorkommen in einigen Landem Europas. Kleintier Prax.. 33: 455461. Morikawa, S., Lutz, H., Albert, A. and Bishop, D.H.L., 1991. Identification of conserved and variable regions in the envelope glycoprotein sequences of two feline imnmnodeficiency viruses isolated in Zurich, Switzerland. Virus Res., 21: 53-63.
H. Lutz et al. / Veterinary Immunology and Immunopathology
46 (1995) 103-I 13
113
Orenta, R.J. and Hildreth, J.E., 1993. Association of host cell surface adhesion receptors and other membrane proteins with HIV and SILV. AIDS Res. Hum. Retrovituses, 9: 1157-I 165. Pedersen. N.C., Ho, E.W., Brown, M.L. and Yamamoto, J.K., 1987. Isolation of a T-Iymphotropic virus from domestic cats with an immunodeficiency-like syndrome. Science, 235: 790-793. Reubel. G.H.. George, J.W., Higgins, J. and Pedersen, N.C., 1994. Effect of chronic feline immunodeficiency virus infection on experimental feline calicivirus-induced disease. Vet. Microbial., 39: 335-35 1. Stott, E.J., Kitchen, P.A., Page M., Flanagan B., Taffs, L.F. Chan, W.L.. Mills, K.H.G., Silvera. P. and Rodgers, A., 1991. Anti-cell antibody in macaques. Nature, 353: 393. Torten, M.. Franchini, M., Barlough, J., George, J., Mazes, E., Lutz, H. and Pedersen, N.C., 1991. Progressive immune dysfunction in cats experimentally infected with feline immunodeficiency virus, J. Viral., 65: 22252230. Tozzini, F., Matteucci, D., Bandecchi, P., Baldinotti, F., Poli, A., Pistello, M., Siebelink, K.H., Ceccherini-Nelli, L. and Bendinelli. M., 1992. Simple in vitro methods for titrating feline immunodeficiency virus (FIV) and FIV neutralizing antibodies. J. Viral. Methods, 37: 241-252. Yamamoto. J.K.. Okuda, T., Ackley, C., Louie, H., Zochlinski, E., Pembroke, E. and Gardner, M.B., 1991. Experimental vaccine protection against feline immunodeficiency virus, AIDS Res. Hum. Retroviruses, 7: 91 l-922. Yamamoto, J.K.. Hohdatsu, T., Olmsted, R.A.. Pu, R., Louie, H., Zochlinski, H.A., Acevedo, V.. Johnson, H.M., Soulds, G.A. and Gardner. M.S., 1993. Experimental vaccine protection against homologous and heterologous strains of feline immunodeliciency virus. J. Viral., 67: 601605.
ELSEVIER
Veterinary immunology and immunopathology
FIV vaccine studies. I. Immune response to recombinant FIV env gene products and outcome after challenge infection -I. Lutz”,” , R. Hofmann-Lehmann”, K. Bauer-Pham”, E. Holznagel”, F. Tozzinib, M. Bendinellib, G. Reubel’, A. Aubertd, D. Davis”, D. Coxe, E. Younge aDepartment of Internal Veterinary Medicine, University of Zurich, Zurich, Switzerland “Department of Animal Pathology, University of Pisa, Pisa, Italy ‘Department of Veterinary Medicine, University of California, Davis, CA, USA *Virbac Laboratories Inc. Nice. France ‘Cambridge Biotech Inc.. Worcester, MA, USA
Abstract We have vaccinated five groups of cats (n = 25) four times with five preparations of recombinant feline immunodeficiency virus (FIV) env gene products; one group (n = 7) served as control. The vaccine formulations were as follows: ( 1) envelope glycoprotein of FIV Zurich 2 (FIV Z2) expressed in a Baculovirus system and isolated by gel electroelution (denatured form) ; (2) insect cells expressing FIV 22 glycoprotein; (3) envelope glycoprotein of a Boston strain (FIV Bangston) expressed in insect cells and isolated by gel electroelution (denatured form); (4) glycosylated Bangston envelope protein made in insect cells and isolated in a native form; (5) non-glycosylated Bangston envelope protein made in Escherichia cofi. All cats were challenged with 20 50% cat infective doses (CIDS,) of FIV 22 previously titrated in cats. All vaccinated cats developed high enzyme-linked immunosorbent assay ( ELISA) antibodies to the homologous antigen; crossreactivity to heterologous antigens was seen at a lower level. Virus neutralizing antibodies (tested with Petaluma virus) reached titers up to 32. After challenge, all cats seroconverted (as judged by anti gag antibodies in Western blot) and became infected (as judged by virus isolation and/or polymerase chain reaction) between 4 and 11 weeks with the exception of one cat. It is concluded that it is relatively easy to induce high ELISA antibody liters using recombinant enu gene products. ELISA antibody titers do not correlate with virus neutralization or with protection.
* Corresponding author. 016%2427/95/$09.50
0 1995 Elsevier Science B.V. All rights reserved
SSDIO165-2427(94)07010-5
104
H. Lutz et al. / Veterinary Immunology and Immunopathology 46 (1995) 103-l 13
Abbreviations CID,,,, 50% cat infective dose; ELBA, enzyme-linked immunosorbent assay; FeLV, feline leukemia virus; FIV, feline immunodeficiency virus; PCR, polymerase chain reaction; SIV, Simian immunodeficiency virus; SPF, specified pathogen free; VN, virus neutralizing (antibodies).
1. Introduction The feline immunodeficiency virus (FIV) is a lentivirus infecting cats leading to immunosuppression (Pedersen et al., 1987) by gradual destruction of the CD4+ lymphocytes (Torten et al., 1991) . The infection is widespread in domestic cats and occurs worldwide. Up to now no etiologic treatment and no vaccine have been available to control FIV infection under field conditions. Until a few years ago it was generally believed that vaccines protecting against retrovirus infections would be almost impossible to develop because during replication retroviruses become integrated into the host genome in the form of the provirus which may persist and be re-activated after long periods. Only recently have effective vaccines against feline leukemia virus (FeLV) been introduced into the field to control FeLV infection. FeLV and FIV and the host’s immune responses to these agents are different in many aspects. Development of an FIV vaccine may be more complicated than was the case with the FeLV vaccine. As FIV is judged to be a good model for HIV and because FIV infection is an important feline disease, several groups have begun work on developing a vaccine against FIV infection. Very promising data have been reported by Yamamoto’s group (Yamamoto et al., 1991, 1993). In their experiments inactivated FIV-infected cells and inactivated and purified FIV were used as vaccine antigens. Other approaches using purified viral proteins were less succesful (for review see Hosie, 1994). Based on the positive experience we had with a recombinant FeLV SU vaccine (Lehmann et al., 1991)) we investigated whether cats could be protected against FIV infection by vaccination with recombinant FIV SU expressed in a Baculovirus system and in Escherichia coli. The use of recombinant envelope components as immunogens appeared to be especially attractive because of the relative ease with which these antigens can be produced in defined purity and the fact that seroconversion induced by natural infection can clearly be differentiated from vaccine-induced antibodies by demonstration of anti gag antibodies.
2. Materials and methods
2.1. Study design Em gene products derived from two FIV isolates were used to vaccinate five groups of cats (n = 25); seven cats served as non-vaccinated controls. The isolates included the strains Zurich 2 (FIV 22; Morikawa et al. 1990) and Bangston isolated from a cat in the Boston area (Young E., personal communication, 1994). The SU components of the two isolates differ in 15% of their amino acid composition (Young E.,
H. Lutz et al. /Veterinary
Immunology
and Immunopathoiogy
105
46 (1995) 103-l 13
personal communication, 1994). By challenging the vaccinated animals with the FIV 22 strain, homolgous and/or heterologous protection was to be examined. 2.2. Cats and sample collection Specified pathogen free (SPF) cats of both sexes were obtained at the age of 15 weeks from Ciba-Geigy, Werk Stein, Stein, Switzerland. They were housed under climatized conditions in a specialized facility at the University of Zurich. Males were castrated and randomly assigned with females to the different vaccine groups. Prior to the first vaccination, the animals were adapted to the new housing conditions for a period of 4 weeks. All cats were clinically examined once per week. Blood and serum samples were collected under slight sedation under routine conditions. 2.3. Vaccine antigens and vaccination The vaccine antigens were expressed in the Baculovirus system (FIV 22 and Bangston) and in E. coti (FIV Bangston) and applied in highly purified denatured or native form (Table 1; Young, E, et al., in preparation). Vaccine antigens were applied at Weeks 0,2,4 Table 1 Vaccine antigens used to vaccinate
the different groups of cats
No. of cats
FIV isolate
Expressed in
Nature of antigen
Amount per injection
I
s
2 3 4 5 6
s 5 5 5 7
Zurich 2 Zurich 2 Bangston Bangston Bangston Ovalbumin
Insect cells Insect cells Insect cells Insect cells E. coli NA
Highly purified, denatured Native Highly purified, denatured Native Highly purified, denatured NA
‘00 CLg Corr. lo* cells
Group
NA, not applicable.
Corr.. corresponding
100 !J,g Corr. 10’ cells 100 CLg 100 pg
to.
Table 2 Levels of antibodies to the three antigens used in the study determined by ELISA Group no.
Antigen used in vaccine
FIV 22 Baculovirus FIV Bangston Baculovirus FIV Bangston
Nature of antigen
Purified Native Purified Native Purified
Antigen used for testing FIV 22 Baculovirus
FIV Bangston Baculovirus
FIV Bangston
100% 62% 11% 74% 26%
28% 14% 100% 39% 108%
4% 4% 78% 21% 100%
0%
0%
0%
E. coli
E. coli
Control The homologous
Ovalbumin
reaction was defined as 100%
H. LUIZet al. /Veterinary Immunology and Immunopathology 46 (1995) 103-113
106
cats/g1 70%
~.
60%
;
50% 4 40%
j
30%
1
20%
.I
.; t
a
(8 group
1
E&
gro”p
2
group
4
F_Z,
group
5
Fig. 1. VN antibodies were determined in serum samples collected considered to be induced by the respective vaccine preparations.
titer
32
16
.....I m
group
3
group
6 (control)
on tbe day of challenge.
Titers > 8 were
and 8 by intramuscular injection in conjunction with 2 mg of aluminium hydroxide and 20 pg of QS2 1, a non-toxic fraction prepared from Quilluju saponaria ( Kensil et al., 199 1) , as adjuvants.
2.4. Challenge infection
Challenge virus infection was done in Week 10 by intraperitoneal injection of 20 50% cat infectious doses (CID& of the FIV 22 strain previously titrated in ten cats 15 weeks of age.
2.5. Antibodies determined by ELlSA and Western blots
Antibody levels to the three purified Baculovirus and E. coli expression products were determined by enzyme-linked immunosorbent assay (ELISA) under conditions described previously (Lehmann et al., 1991). Antigens were coated in amounts of 100 ng per well, sera were diluted 1:5000, and the conjugate (goat anti cat IgG horseradish peroxidase (Milan, Duedingen, Switzerland) was used at a dilution of 1:lOOO. The absorbance values were determined in a Dynatech reader M580 (Dynatech, Embraport, Switzerland) and the homologous reaction was defined as 100%. Western blots were performed under conditions described (Lutz et al., 1988a) using 500 ng of gradient purified FIV 22.
H. Lutz et al. /Veterinary Immunology and Immunopathology 46 (1995) 103-I 13
107
challenge c
101112
week o
14
16
18
2021
cat no 405 group
1
416 425 434 449 411 421
group
2
441 454 428 1
3
,“,“;
406 420
group
456
group
5
411) 419 432 435 444 : 4oi 415 414
group
6
4’33 439 453 461
n FlVpr~,t,ie
I
~~ ~.___-~__~
up24
qLI.E
3FI‘., ___
ne&31&
Fig. 2. Western blot results obtained from all cats before (three samples) and after (seven samples) infection. For clarity, reactions with the gag components p24 and p I7 are summarized in this figure.
2.6. Virus neutralizing
challenge
(VN) antibodies
Sterile serum samples were tested by a procedure described elsewhere using the Petaluma strain (Tozzini et al., 1992). 2.7. Virus isolation Virus isolation was performed as described 2.8. Polymerase
(Lutz et al., 1988b).
chain reaction (PCR)
Peripheral blood mononuclear cells were purified by Ficoll/Hypaque gradients (Pharmacia, Duebendorf, Switzerland), DNA extracted from lo6 cells was used for nested PCR under conditions described (Reubel et al., 1994). Briefly, a gag sequence was amplified covering nucleotides 929-l 394 (465 base pairs (bp) ) in the first round and 1236-1394
108
H. Lutz et al. /Veterinary Immunologyand Immunopathology46 (1995)103-113 group
1
group
2
group
3
group
4
group
5
group
6 0
1
2
3 no positive
4 of
5
6
7
cats
J
negative
Fig. 3. Virus isolation assays performed with blood samples collected 10 weeks after challenge infection.
(158 bp) in the second round. With every PCR cycle and every gel run, a DNA sample collected from a known infected and a known non-infected SPF cat were included.
3. Results 3.1. Serological
response to vaccination,
antibodies determined by ELISA
Serum samples collected on the day of challenge infection were tested for antibodies to the different SU components by ELISA (Table 2). From Table 2 it can be concluded that the homologous reaction was considerably stronger than the heterologous reaction. Glycosylation which was the main difference between the Baculovirus and E. coli expression products had a less pronounced effect on the antibody reaction. Serum samples collected from cats immunized with the Bangston E. coEi expression product contained antibodies that displayed a stronger reaction with the Baculovirus expression product than with the protein expressed in E. coli. 3.2. VN antibodies Serum samples collected under sterile conditions on the day of challenge were analysed for VN activity (Fig. 1). The highest titers observed were 32. In Group 3 (Bangston Baculovirus expression product) four of five cats had VN antibodies of I6 and higher. In Group 1 (Zurich Baculovirus expression product) and in Group 5 (Bangston E. coli) one cat each had a titer of 32. All other cats including those of the control group (Group 6) had titers of 8 or below. 3.3. Western blot results throughout the course of challenge infection Serum samples collected Week 10 (day of challenge)
at Week 0 (first vaccination), Week 4 (third vaccination), and on seven occasions throughout the challenge observation
H. Lutz et al. /Veterinary Immunology and Immunopathology 46 (19951 103-l 13
109
Table 3
WOUD CM
week12
week16
week46
period ( up to Week 2 1) were submitted to Western blot analysis (Fig. 2). With the exception of one cat (Cat 406 in Group 3) all cats eventually seroconverted as judged by the development of antibodies specific for p24 and ~15. Cats of Groups 2 and 5 appeared to seroconvert as a group somewhat earlier than cats of Groups 1, 3,4 and 6. 3.4. Virus isolation Peripheral blood mononuclear cells were purified from blood samples collected on the day of challenge and 10 weeks after challenge. The cells were co-cultivated with mononuclear cells collected from SPF cats for 4 weeks. The supernatants were tested for the presence of FIV p24 in Weeks 2,3 and 4 after establishing the cultures. Virus isolation was considered positive when one of the culture supernatants showed a positive FIV p24 test (Fig. 3). Blood samples tested on the day of challenge were all negative (data not shown). Ten weeks after challenge infection, in Groups 1 and 4 all cats were virus isolation positive. In Groups 2, 3 and 5, FIV was isolated from four of five cats. In the control group FIV was isolated from five of seven animals. 3.5. PCR results Mononuclear were subjected
cells collected 10 weeks before and 2, 12, 16 and 48 weeks after challenge to PCR analysis (Table 3). With the exception of one cat (Cat 406) all
110
H. Lutz et al. / Veterinary Immunology and Immunopathology 46 (1995) 103-l 13
animals were positive by PCR at least once until Week 16 after challenge. Cat 424 was positive by PCR 10 weeks before challenge; this result is considered to be due to contamination of the PCR reaction tube. At Week 48 some cats had become PCR negative. Cat 406, which had remained PCR-negative until Week 16 after challenge, was PCR-positive in Week 48.
4. Discussion In our experiments we have vaccinated five groups of cats with different preparations of recombinant FIV SU. After challenge infection with 20 ClD,,,, all control cats and all vaccinated cats became seropositive with the exception of one vaccinated animal in Group 3. From the Western blot results and the virus isolation and PCR experiments, it was concluded that there was no complete protection against homologous and heterologous challenge infection. Protection against FIV infection was recently demonstrated using inactivated virus infected cells or inactivated cell-free FIV vaccines (Yamamoto et al., 1991, 1993). Protection was correlated with high titers of VN antibodies; cats with lower anti SU and VN antibody titers were not protected (Yamamoto et al., 1993). Passive immunization using sera from FIV vaccinated cats or from cats with FIV infection also resulted in protection. However, VN antibodies alone may not always be sufficient for protection. This was concluded from experiments of the Glasgow group in which it was found that some kittens with maternal antibodies became infected (Callanan et al., 1991). Obviously, VN antibodies are important for protection against FIV infection at least to some degree as a first line of defense. For the prevention or reduction of cell-to-cell transmission of FIV, a strong cellular immune response will certainly also be required. The immunization of our cats induced high levels of antibodies detectable by ELISA. Immunologic differences between the two isolates were readily revealed by ELISA. Reactivity in the homologous assay (i.e. cats vaccinated with FIV 22 protein assayed on FIV 22 protein coated plates) was stronger than the heterologous immune response. However, there was also some cross-reaction between the isolates. This cross-reactivity may be explained in part by antibodies directed against the carbohydrate moiety of the glycoproteins. This was concluded from the observation that cats vaccinated, for example, with the FIV 22 glycoprotein showed a much stronger reaction with the Bangston SU component expressed in Baculovirus than with the Bangston E. coli expression product which is not glycosylated. That part of the antibodies were directed to the glycoproteins was also concluded from the observation that the animals immunized with the Bangston SU component expressed in the Baculovirus system had higher antibody levels against the Bangston Baculovirus expression product than against the SU expressed in E. coEi. It is less clear why cats immunized with the E. coli expression product made higher antibody levels against the Baculovirus-expressed material than with the SU synthesized by E. coli. One explanation could be that the Baculovirus expression product is at least to some degree adsorbed to the surface of the ELISA well via the carbohydrate part, rendering more epitopes accessible to antibodies than when the E. coli expression product is adsorbed to the plastic surface or that the glycosylated protein binds more easily to the plate. The cats immunized with the native antigens had somewhat lower antibody levels than the animals which received the purified
H. Lutz et al. /Veterinary Immunology and Immunopathology 46 (1995) 103-I 13
III
immunogens. This could be explained by the possibly lower concentration of the SU component in this antigen preparation. The SU concentration could not be determined precisely because of the other proteins present in the insect cells. VN antibodies with titers of 16 and 32 were found on the day of challenge in some cats immunized with purified immunogens (Groups 1, 3 and 5). These titers were considered to be induced by the vaccine antigens as the control cats had titers of 8 or lower. These VN titers were much lower than those found in cats with natural FIV infection (Tozzini et al., 1992) and in cats vaccinated with an inactivated FIV-infected cell vaccine (Yamamoto et al., 1991) or an inactivated cell-free FIV preparation (Yamamoto et al., 1993). The relatively low VN titers found in the present study probably were not sufficient for complete neutralization of the challenge virus and therefore inadequate for protection. Although the virus strain used for the determination of VN antibodies differed from the vaccine strains, this probably does not explain the low titers. Yamamoto et al. (1993) were able to demonstrate in a similarly heterologous system high VN titers. The discrepancy between high ELISA and relatively low VN titers may be explained by the fact that the purified immunogens were denatured and therefore did not have the three-dimensional conformation of the natural glycoprotein. As we were aware of this possibility, cats of Groups 2 and 4 were vaccinated with native SU present in the insect cells. The SU in the native form was unable to induce VN antibodies although high levels of ELISA antibodies were found. It is difficult to understand why cell-free virus and inactivatedcells but not purified or native recombinant SU can induce VN antibodies. The nature of the glycosylation by the Baculovirus may explain some but not all of the difference. By analogy to the FeLV system, where VN antibodies were readily induced by a recombinant FeLV SU (Lehmann et al. 1991), we had expected that VN antibodies would also be found in the FIV experiments described here. In the Simian immunodeficiency virus (SIV) system, protection was found in experiments in which the vaccine virus was grown in the same cells (C8166) as those in which the challenge virus was prepared (Stott et al., 1991). Some degree of protection could also be induced by vaccination of macaques with uninfected C8166 cells (Stott et al., 1991). Protection by antibodies directed against cell constituents can be explained by the fact that retroviruses incorporate cellular components into their membranes during virus assembly and budding from the cell (Orenta and Hildreth, 1993). It can be postulated that a combination of anti cellular and anti viral antibodies are especially efficacious in inducing VN antibodies. In the cat, cellular components were shown to be less important for protection against FIV than in the SIV/macaquqe model (Hohdatsu et al., 1993; Hosie, 1994). The fact that monkeys vaccinated by recombinant gp 160 of SIV were protected against challenge infection (Hu et al., 1992) is suggestive that anti viral immune reaction can be achieved and will be efficacious in preventing lentivirus infections. Whether or not the cats of this study may have been partially protected resulting in a reduced virus load is unclear. Quantitative PCR and and serial virus isolation experiments were not performed. However, visual interpretation of Fig. 2 leads to the impression that the animals of Groups 1,3 and 4 seroconverted to p24 and ~15 somewhat later than the cats of the control group and of Group 2. In addition, the negative PCR results of the three cats of Group 1 available in Week 48 also suggested that in this group the virus load may have been smaller than in the control group. Cat 406 in Group 3 turned positive by PCR in Week
112
H. Lurz et al. /Veterinary Immunology and Immunopathology 46 (1995) 103-113
48. Whether it became truly infected by contact with the other cats that were FIV-positive or whether this positive result is the consequence of a laboratory contamination was not determined. Hosie et al. ( 1992) reported that cats immunized with different vaccine preparations including recombinant p24 showed enhancement of infection after challenge infection. In our study, based on Western blot results, there seemed to be no faster seroconversion in vaccinated cats over control animals, with the possible exception of Group 2 where complete seroconversion to p 24 and ~15 was observed in four of five cats already in Week 16. In conclusion, it will be important in future experiments to improve induction of VN antibodies and to quantitate virus load througout the course of challenge infection. It has to be expected that partial protection can be induced which may lead to lower virus load and therefore to improved clinical outcome. Use of recombinant SU components thereby allows the differentiation of vaccine-induced seroconversion and natural infection.
Acknowledgments This project was supported by a grant from the Swiss Banking Corporation on behalf of a customer, by the European concerted action on FIV vaccination and by the Rassekatzenvereinigung Ostschweiz. Cat food was kindly donated by Effems, Zug, Switzerland. The expert technical assistance of Zihu Zhang and Peter Fidler is acknowledged.
References Callanan, J.J., Hosie, M.J. and Jarret, O., 1991. Transmission of feline immunodeficiency virus from mother to kitten. Vet. Rec., 128: 322-323. Hohdatsu, T., Pu, R., Torres, B.A., Trujillo, S., Gardner, M.B. and Yamamoto, J.K., 1993. Passive antibody protection of cats against feline immunodeficiency virus infection. J. Virol., 67: 2344-2348. Hosie, M.J., 1994. The development of a vaccine against feline immunodeticiency virus. Br. Vet. J., 150: 25-39. Hosie, M.J., Osborne, R.. Reid, G., Neil, J. and Jarrett, 0.. 1992. Enhancement after feline immunodeficiency virus vaccination. Vet. Immunol. Jrnmunopathol., 35: 191-197. Hu, S.L.. Abrams, K., Barber, G.N., Moran, P., Zarling, J.M., Langlois. A.J., Kuller, L., Morton. W.R. and Benveniste, R.E., 1992. Protection of macaques against SIV infection by subunit vaccines of SIV envelope glycoprotein gp 160. Science, 255: 456-459. Kensil, C.R., Patel, U.. Lennick. M. and Marciani, D., 1991. Separation and characterization of saponins with adjuvant activi’y from QuiZlajasaponaria Molina cortex. J. Immunol., 146: 431437. Lehmann, R., Franchini, M., Aubert. A., Wolfensberger, C., Cronier, J. and Lutz, H.. 1991. Vaccination of cats experimentally infected with feline immunodehciency virus with a recombinant feline leukemia virus vaccine. J. Am. Vet. Med. Assoc., 199: 1446-1452. Lutz. H., Arnold, P., Htibscher, U.. Egberink, H., Pedersen, N.C. and Horzinek, M.C., 1988a. Specificity assessment of feline T-lymphotropic Jentivirus serology. Zbl. Vet. Med. B, 35: 773-778. Lutz, H.. Egberink, H. Arnold, P., Winkler, G., Wolfensberger, C., Jarrett, 0.. Parodi, A.L., Pedersen, N.P. and Horzinek, M.C., 1988b. Felines T-Jymphotropes Lentivirus (FTLV): Experimentelle lnfektion und Vorkommen in einigen Landem Europas. Kleintier Prax.. 33: 455461. Morikawa, S., Lutz, H., Albert, A. and Bishop, D.H.L., 1991. Identification of conserved and variable regions in the envelope glycoprotein sequences of two feline imnmnodeficiency viruses isolated in Zurich, Switzerland. Virus Res., 21: 53-63.
H. Lutz et al. / Veterinary Immunology and Immunopathology
46 (1995) 103-I 13
113
Orenta, R.J. and Hildreth, J.E., 1993. Association of host cell surface adhesion receptors and other membrane proteins with HIV and SILV. AIDS Res. Hum. Retrovituses, 9: 1157-I 165. Pedersen. N.C., Ho, E.W., Brown, M.L. and Yamamoto, J.K., 1987. Isolation of a T-Iymphotropic virus from domestic cats with an immunodeficiency-like syndrome. Science, 235: 790-793. Reubel. G.H.. George, J.W., Higgins, J. and Pedersen, N.C., 1994. Effect of chronic feline immunodeficiency virus infection on experimental feline calicivirus-induced disease. Vet. Microbial., 39: 335-35 1. Stott, E.J., Kitchen, P.A., Page M., Flanagan B., Taffs, L.F. Chan, W.L.. Mills, K.H.G., Silvera. P. and Rodgers, A., 1991. Anti-cell antibody in macaques. Nature, 353: 393. Torten, M.. Franchini, M., Barlough, J., George, J., Mazes, E., Lutz, H. and Pedersen, N.C., 1991. Progressive immune dysfunction in cats experimentally infected with feline immunodeficiency virus, J. Viral., 65: 22252230. Tozzini, F., Matteucci, D., Bandecchi, P., Baldinotti, F., Poli, A., Pistello, M., Siebelink, K.H., Ceccherini-Nelli, L. and Bendinelli. M., 1992. Simple in vitro methods for titrating feline immunodeficiency virus (FIV) and FIV neutralizing antibodies. J. Viral. Methods, 37: 241-252. Yamamoto. J.K.. Okuda, T., Ackley, C., Louie, H., Zochlinski, E., Pembroke, E. and Gardner, M.B., 1991. Experimental vaccine protection against feline immunodeficiency virus, AIDS Res. Hum. Retroviruses, 7: 91 l-922. Yamamoto, J.K.. Hohdatsu, T., Olmsted, R.A.. Pu, R., Louie, H., Zochlinski, H.A., Acevedo, V.. Johnson, H.M., Soulds, G.A. and Gardner. M.S., 1993. Experimental vaccine protection against homologous and heterologous strains of feline immunodeliciency virus. J. Viral., 67: 601605.