Materials and Methods

Transcription

1 734 Memory Cytotoxic T Lymphocyte Responses in Human Immunodeficiency Virus Type 1 (HIV-l)-Negative Volunteers Immunized with a Recombinant Canarypox Expressing gp160 of HIV-l and Boosted with a Recombinant gp160 Beatrice Fleury, Genevieve Janvier, Gilles Pialoux, Florence Buseyne, Michael N. Robertson,* James Tartaglia, Enzo Paoletti, Marie Paule Kieny, Jean Louis Excler, and Yves Riviere Unite de Virologie et Immunologie Cellula ire, Institut Pasteur, and Hopital de L'Institut Pasteur, Paris, Transgene, Strasbourg, and Pasteur Merieux Serums et Vaccins, Marnes-la-Coquette, France; Virogenetics Corp., Troy, New York A vaccine against human immunodeficiency virus (HIV) should induce virus-specific cytotoxic T lymphocyte (CTL) activity. Immunization of uninfected volunteers with a canarypox virus expressing HIV envelope was carried out in a phase I trial. Two injections of canarypox expressing HIVl M N gp160 (months 0 and 1) were followed by two boosts of recombinant envelope protein (months 3 and 6). HIV envelope-specific CTL were detected in peripheral blood mononuclear cells stimulated with autologous HIV-l-infected blast cells. T cell lines were obtained from 18 of 20 donors: CTL were detected at least once following immunization in 7 (39%)of these 18. This activity was mediated by major histocompatibility complex class I-restricted CD3+CD8+ T cells. For two subjects, this activity was still present 2 years after the initial immunization. The CTL responses with this primeboost regimen are the best observed with any HIV vaccine tested in humans. Vaccination against the human immunodeficiency virus type I (HIV-I) is a key strategy for the eventual control of the AIDS pandemic. Since the discovery of the etiologic agent of AIDS, different candidate vaccines have been designed and developed, including whole killed virus, live attenuated virus, virus-like particles including pseudovirions, protein subunits based on HIV structural genes, peptides based on selected immunoreactive parts ofthe viral proteins, live recombinant viral or bacterial vectors, and DNA-based immunogens encoding one or more HIV proteins [1-3]. Live recombinant vaccinia viruses have been used successfully to induce protective immunity against animal infectious diseases such as rabies [4], but because the use ofvaccinia virus for human vaccination against smallpox has been shown to induce rare but serious complications [5], other recombinant poxviruses such as avian poxviruses have been developed [6, 7]. As was the case for most live expression vectors in AIDS vaccine studies, initial efforts have focused on products based on the HIV-I envelope protein, since several epitopes of env have been described to induce neutralizing antibodies and cell-mediated immune responses, including cytotoxic T lymphocytes (CTL). In HIV-seronegative volunteers at low risk ofhiv infection, the safety and immuno- Received 31 January 1996; revised 21 May Grant support: Institut Pasteur, Agence Nationale de Recherches sur Ie SIDA. YR. is an Elisabeth Glaser Scientist supported by the Pediatic AIDS Foundation. Reprints or correspondence: Dr. Yves Riviere, Unite de Virologie et Immunologie Cellulaire, Institut Pasteur, 28 rue du Docteur Roux, Paris Cedex 15, France. * Present affiliation: Department of Medicine, Division of Infectious Diseases, University of Washington Medical Center, Seattle. The Journal of Infectious Diseases 1996; 174: by The University of Chicago. All rights reserved /96/ $01.00 genicity of a combined vaccine regimen consisting of priming with a recombinant canarypox virus expressing the HIV-1MN env gene (ALVAC-HIV, vcp125) followed by booster immunizations with a soluble recombinant envelope glycoprotein gp 160MN/LAI has been reported [8]. The characterization of the CD8-mediated, major histocompatibility complex (MHC) class I-restricted cytolytic response elicited in 39% ofthe vaccinees is presented here. Materials and Methods Study Population Twenty volunteers, 10 men and 10 women, were enrolled into the study as described [8]. All were healthy HIV-I-seronegative adults with low risk for HIV-1 exposure. Vaccine Products and Protocol Design The recombinant canarypox expressing the HIV-1 gp160 gene from the MN isolate (ALVAC-HIV; vcp125; Virogenetics, Troy, NY) and the soluble recombinant hybrid envelope MN/LAI (rgp 160; Transgene, Strasbourg, France) were previously described in detail together with the results obtained on the safety of this phase I, randomized, open-labeled trial [8]. All subjects received 10 6 TCID so of vcp 125 at months 0 and 1 followed by booster injections of 176 f-lg of rgpl60 formulated in alum (n = 10) or in incomplete Freund's adjuvant (Montanide ISA 51; Seppic, Paris) (n = 10) at months 3 and 6 according to random assignment. Cytotoxic Assays Target cells. For each patient, an autologous Epstein-Barr virus (EBV)-transformed B cell line was obtained by infection of peripheral blood mononuclear cells (PBMC) with supernatant from

2 JID 1996; 174 (October) HIV-1 - Specific CTL Induction by a Canarypox 735 the EBV-producing cell line B95-8 (ATCC CRL 1612). EBVtransformed B cells (3 X 10 6 ) were infected with the Copenhagen strain of wild type vaccinia virus (vvwt) or with the various recombinant vaccinia viruses at an MOl of 10 pfu/cell, for 1 h at 37 C, then washed and cultured for 16 h at 37 C in 5% CO 2 before their use in the cytotoxic assays. Recombinant vaccinia viruses vvtg1139 and vvtg5l67, encoding the HIV-I LA 1 and the HIV I MN env proteins, respectively, have been described [9]. Where cited, recombinant vaccinia viruses encoding signal peptide deletion mutants ofhiv-i LA I or HIV-I M N env proteins (vvtg3183 and vvtg61l0, respectively) were used [9, 10]. The human K562 cell line (ATCC CCL243) was used as a target to detect NK activity. Effector cells. PBMC were isolated from heparinized blood by density gradient centrifugation on Ficoll-Paque (Pharmacia, Les VIis, France). PBMC collected at different times after immunization were aliquoted and frozen in liquid nitrogen. CTL lines were generated by two different procedures using either nonspecific or HIV-specific stimulation. Nonspecific stimulation was done after in vitro activation of thawed PBMC with an anti-cd3 monoclonal antibody (OKT3; Ortho Diagnostics System, Roissy, France; 25 ng/ml) in RPMlc: RPMJ 1640 (reference catalog 12-l67B; Whittaker, Gagny, France) supplemented with 5% heat-inactivated human AB serum (ETS, les Ulis, France), L-glutamine (2 rum, reference H; GIBCO, Cergy Pontoise, France), sodium pyruvate (1 mm, reference H; GIBCO), and nonessential amino acids (1 x: reference H; GIBCO). HIV-specific stimulation was done by in vitro activation of thawed PBMC with autologous irradiated phytohemagglutinin (PHA)-stimulated blasts (2 j.lg/ml; Difco, Paris) infected 6 or 7 days previously with HIV-1LAI as stimulator cells (responder/stimulator ratio = 5/1) in RPMIc. After 3 days and for both nonspecific and HIV-specific stimulation protocols, the cells were washed and resuspended (10 6 / ml) in RPMlc supplemented with 10% T cell growth factor (Lymphocult-T-LF; Biotest, Buc, France). Four days later, recombinant human interleukin-2 (l00 IV/mL, gift ofd. Lando, Roussel Uclaf, Romainville, France) was added to the cultures. The cultures were fed every 3-5 days for 2-6 weeks. The cytolytic activity and the characteristics of the T cell lines mediating this activity were evaluated from day 12 to day 35 after in vitro stimulation [9-11]. Chromium release assays. Cytotoxic assays were conventional 4-h 51Crrelease assays as described [12]. Briefly, 10 6 target cells were labeled for 1 h with 3.7 MBq of [51Cr]Na2Cr04 (reference 62014; ICN, Orsay, France), washed three times, and distributed at 5 X 10 3 cells/well in 0.1 ml of medium in round-bottom 96 well microtiter plates (Costar; OSI, Maurepas, France). Various concentrations of effector cells in 0.1 ml (in triplicate or quadruplicate) were added. The plates were incubated at 37 C in 5% CO 2 for 4 h, then centrifuged at 150 g for 5 min. Fifty microliters of supernatant from each well was transferred together with 100 j.ll of scintillation liquid (Optiphase Hisafe 3; EGG, Evry, France) into a 96-well harvesting plate (reference ; EGG), and the radioactivity was measured in a beta counter (Microbeta 1450; EGG). The percentage of specific lysis was calculated as 100 X [(experimental release - spontaneous release)/(maximal release spontaneous release)]. Maximal release was obtained from targets lysed by 5% Triton X-IOO with 1% SDS. Spontaneous release was obtained from targets incubated for 4 h in medium alone. Assays in which spontaneous 51Cr release exceeded 35% were discarded. The CTL responses against HIV antigens were considered positive Table 1. Longitudinal analysis of envelope-specific CTL activity in volunteers primed with canarypox virus expressing HIV envelope and boosted with recombinant gp160. Envelope-specific lysis at month* Group, subject no Alum adjuvant ND 102 NG ND 106 NG ND NG NG + ND 110 ND 112 NG NG NG ND 116 ND NG + ND NG ND ND Incomplete Freund's adjuvant ND + ND ND 108 NG ND 109 NG NG ND 111 NG NG NG ND 113 ND ND + ND ND 117 NG ND 119 NG NG ND 120 ND ND ND ND ,_._-- NOTE. Subjects received 10 6 TCTD so of vcp 125 at months 0 and 1, followed by booster injections of 176 f.1g of rgp160 formulated in alum or in incomplete Freund's adjuvant (n = loin each group) at months 3 and 6 according to random assignment. Boldface indicates subjects with env-specific lysis. -, no specific lysis; +, specific lysis (results were derived from cytotoxic assays in which 3 consecutive effector-to-target ratios, routinely 60/1, 20/1, and 7/1, were used); ND, not done; NG, no growth. * After first immunization. if they exceeded the mean of vvwt-specific lysis by 3 SO and by 10%. Lymphocyte surface marker studies. Cell surface markers were detected by direct labeling with the following antibodies: fluorescein-conjugated anti - Leu-2a (CD8), fluorescein-conjugated anti Leu-4a (CD3), phycoerythrin-conjugated anti-vleu-ll c (CD16) (Becton Dickinson, Le Pont de Claix, France), and phycoerythrinconjugated OKT4 (Ortho). Samples were analyzed with a FACScan flow cytometer (Becton Dickinson). Results Envelope-specific cytolytic activity is detected after immunization. T cell lines were obtained from 18 of 20 donors; for 2 donors, it was not possible to grow the PBMC after in vitro stimulation. Nonspecific stimulation with anti-cd3 monoclonal antibodies did not provide detectable env-specific CTL activity (not shown). On the contrary, HIV-I envelope-vspecific cytolytic activity was only observed when the T cell lines were derived from PBMC specifically stimulated with autologous HIV-1 infected blasts (table 1). Seven (39%) of these 18 subjects presented envelope-specific cytolytic activity at least once after immunization (months 2, 4, 7, 12, or 24 after primary

3 736 Fleury et al. JID 1996; 174 (October) 40--r , 35 T 30 /If J.: o e- ---A Target/Effector Figure 1. Cytotoxic actrvrty is envelope-specific and is detected after immunization. Targets: Epstein-Barr virus-transformed B cells infected with vvwt (wild type vaccinia virus) or vv5167 (recombinant vaccinia virus expressing env gene ofhiv-1mn)' Effectors: cell lines derived from peripheral blood mononuclear cells from subject 113, collected at months (M) 0 and 7. CTL activity was tested 21 days after initiation ofculture after in vitro stimulation with autologous HIV-1-infected blasts. immunization). The cytolytic activity was tested for from days 13 to 35 after the initiation of the culture and was found from days 21 to 35. For each postimmunization time, the results generally agree: for example, donors 102, 106, los, 110, and 117 were tested three or four different times after immunization and were negative each time. For 6 of the 7 donors who had an env-specific CTL response, this activity was detectable at two different times following immunization. For 3 subjects (101, 105, and 107), the cytolytic activity was detected at month 2, I month after the second AL VAC-HIV administration. The cytolytic activity was still present at months 7 or 12, 1-6 months after the second immunization with rgpl60. Four other subjects were positive: at months 4 and 7 for subject 112, at months 7 and 24 for subjects 113 and 116, and at month 12 for subject lis. No significant difference was observed between the alum versus incomplete Freund's adjuvant groups. As shown in figure 1 and table 1, envelopespecific cytolytic activity was detected from PBMC collected from donor 113 at month 7, but no envelope-specific cytolytic activity was detected from the T cell lines derived from preimmune PBMC. Envelope-specific cytolytic activity is mediated by CD3+CD8+ MHC class Ir-restricted T cells. Most ofthe cell lines mediating env-specific cytolytic activity were CDS T cells (64%-90%) with <4% CD4 cells as shown by cell cytometric analysis. As an example, envelope-specific effector cells from subject 107, measured 2S days after HIV-specific in vitro stimulation of the PBMC collected at month 12, were 97% CDS T cells, and the cytolytic activity was directed against the envelope of both MN and LAI isolates (table 2). In this experiment, the target cells were infected with vvtg 31S3 and vvtg 6110, which encode the signal peptide deletion mutant ofhiv-i env. Target cells infected with these recombinant vaccinia viruses are not susceptible to antibody-mediated cellular cytotoxicity, Table 2. T cells. Effector-to-target ratio vvwt/mo Targets 60/1 20/1 6.6/1 vvwt/m7 vvwt 18 (1.4) 10 (l.8) 4 (3.4) vv5167/mo vvtg (2.2) 32.7 (3.2) 14 (2.6) --ft-- vv5167/m7 vvtg (2.6) 29 (6.5) 15 (3.1) NOTE. Effectors were derived from peripheral blood mononuclear cells of volunteer 107, at month 12, and stimulated in vitro with autologous HIV-1-infected blasts. Targets were autologous Epstein-Barr virus-transformed B cells infected 16 h before with wild type vaccinia virus (vvwt) or with recombinant vaccinia virus encoding for signal peptide-deleted env gene of HIV-I LA 1 (vvtg3183) or HIV-I M N (vvtg61 10). Data are mean % specific lysis oftriplicate (SO). Positive results are indicated in boldface type. as shown previously [9, 10]. For two cell lines, the activity was shown to be restricted by the MHC class I. As shown in figure 2, envelope-specific cytolytic activity was observed against autologous EBV-transformed B target cells (but not against MHC class I-mismatched EBV-transformed B target cells) from subject 105, as measured 29 days after HIV-Ispecific in vitro stimulation of the PBMC collected at month 2. The cell line consisted of 65% CD3 +CDS+ T cells and 31% CD 16 cells. In this case, it is likely that the presence of CD 16 cells was responsible for the background lysis against the vvwt-infected target cells described previously [13]. Long-term memory CTL activity. Measurable HIV-l envspecific CTL activity was observed I month after the second ALVAC-HIV administration (month 2), 1 month after the first and second rgp160 boosts (months 4 and 7), and 6 months 40 HIV envelope-specific cytotoxicity is mediated by CD8 35 T. 30 T 1 Target r.j)... T vvwt Auto u 1,'.l ' A--- vvwthetero (]) T /',-1 " P vv6110 Auto r.j) 10 --i::j:-- vv6110 Hetero 5 =","-Ii Figure 2. Envelope-specific cytotoxic activity is major histocompatibility complex class I-restricted. Targets: autologous (auto) or heterologous (hetero) Epstein-Barr virus - transformed B cells infected with vvwt (wild type vaccinia virus) or vv61 10 (recombinant vaccinia virus synthesizing the env protein of HIV-1MN)' Effectors: cell line derived from peripheral blood mononuclear cells from subject 105, collected at month 2. CTL activity was tested 29 days after initiation of in vitro stimulation with autologous HIV-1-infected blasts.

4 JID 1996; 174 (October) HIV-1-Specific CTL Induction by a Canarypox 737 after the last rgp 160 boost (month 12). Three subjects who had env-specific CTL activity at one or more of these times (112, 113, and 116) were tested IS months after the last rgp 160 boost, 2 years after the first ALVAC-HIV immunization. Three weeks after in vitro stimulation with autologous HIV-l-infected blasts, 2 of the 3 cultures had env-specific cytolytic activity (table 1). The results of a positive T cell line from subject 116 are shown in figure 3. At day 21 after in vitro stimulation with autologous HIV-1-infected blasts, there was no detectable env-specific CTL activity with PBMC collected at month 0, but env-specific activity was observed with the PBMC collected at months 7 and 24. The T cell line was further restimulated with PHA (2 i-lg/ml), and the cells mediating the env-specific cytolytic activity were found to be 94% CD3 +CDS+ by cell cytometric analysis. Further, the activity was shown to be MHC class I-restricted. A panel ofallogeneic target cells matched at one or more HLA class I alleles infected with a recombinant vaccinia virus encoding HIV-l M N env (vvenv) were used as target for this CTL line. in addition to the autologous vv-env-infected targets, one EBV target (target 1) was also lysed when infected with vv-env (table 3). This target and autologous target share HLA A32 and C7. Target 4, sharing HLA A29 and C7, was not lysed; the HLA-restricting antigen was thus demonstrated to be A32. Discussion This report shows that immunization with a live recombinant canarypox virus vector (AL VAC) expressing the HIV-1MN env gene (vcp125) induced HIV-1 envelope-specific CTL activity in 7 (39%) of the 18 HIV-l-seronegative human volunteers tested. For 6 of 7 donors, cytolytic activity was detected at Target/Effector vvwf/mo vvwf/m / CJ) --A:-- rob 40- IS':./,/...., vvwf/m24 'u 30- T/ vv5167/mo <lj 0.. CJ) vv5167/m ,-0 -, vv5167/m Figure 3. Envelope-specific cytotoxic activity is present 2 years after first canarypox injection. Targets: autologous Epstein-Barr virus-transformed B cells infected with vvwt (wild type vaccinia virus) or vv5167 (recombinant vaccinia virus expressing env gene of HIV-1 MN)' Effectors: cell lines derived from peripheral blood mononuclear cells from subject I 16, collected at month (M) 0, 7, or 24. CTL activity was tested 21 days after initiation of in vitro stimulation with autologous HIV-1-infected blasts. Table 3. CTL line 116 is restricted by HLA A32. Effector-to-target ratio HLA alleles,* targets 30/1 10/1 3/1 A29,32 B7,17 C7 (autologous) vvwt 4.5 (3.5) 1.1 (0.9) 2.1 (2.1) vvtg (2.2) 15.9 (1.2) 8.1 (1.5) A32,2 B8,51 Cl,7 (target 1) vvwt 1.2 (1.6) 0.4 (1.6) 0.6 (2.4) vvtg6l10 23 (2.3) 11.1 (2.4) 8.8 (1.5) A28,30 B7,55 C2,6 (target 2) vvwt 35.8 (6.2) 17.2 (5.8) 4.8 (1.6) vvtg (2.7) 10.1 (7.1) 4 (3.6) A30,68 B53 C4 (target 3) vvwt 0.7 (2.9) -0.9 (2.7) -1.7 (1.1) vvtg (1) 0.5 (1.2) 1.4 (0.9) Al,29 B8,44 C5,7 (target 4) vvwt 1.9 (0.8) 0(3.8) 1.5 (1.8) vvtg61l0 1.8 (2.4) 0(2) 1.1 (4.2) NOTE. Effectors were CTL line from month 24 peripheral blood mononuclear cells from volunteer 116, stimulated with autologous HIV -1 infected blasts at day 0 and restimulated with phytohemagglutinin (2 Ilg/mL) at day 23; at day 13 of restimulation, cell line was 94% CD3 CD8 as shown by cell cytometry. Chromium release assay was done 18 days after initiation of restimulation. vvwt, wild type vaccinia virus; vvtg611o, recombinant vaccinia virus encoding for signal peptide-deleted env gene of HIV-I M N Data are mean % specific lysis oftriplicate (SO). Positive results are indicated in boldface type. * HLA alleles are in boldface type when they match with effector cells. various times after immunization (subjects 101, 105, 107, 112, 113, and 116; table 1). In some cases, the cytotoxic response was transient (table 1). For example, the cytotoxic response of donor 112 was present at months 4 and 7 but not at months 12 and 24; this might be due to the loss of memory T cells in this donor. Consistent with this hypothesis, 9 of the 11 donors tested for CTL activity at month 12 were negative. For donor 101, the cytotoxic response was present at months 2 and 7 but not at months 4 and 12. in this case, the absence of detectable CTL activity at month 4 might be due to fluctuations in the kinetics ofthe CTL response in vitro, thus reflecting the tremendous difficulty in assessing the capacity of vaccine candidates to induce CTL. These results reinforce the need to establish a consensus for techniques to detect CTL in vaccinees. PBMC from 3 of 10 subjects tested at month 2 were found to contain envelope-specific cytolytic activity, showing that the recombinant canarypox per se was sufficient for CTL induction, despite the absence of detectable HIV-specific antibodies by ELISA or Western blot [S]. Five of the 11 subjects tested at month 7 and 4 of the 13 tested at month 12 or 24 were positive for envelope cytolytic activity. There were no major differences in the frequency of donors with a positive CTL response after rgp 160 boost, but there was a significant increase, for all subjects, in anti-gp 160 and anti-v3 antibodies after each rgp160 boost [S]. The response with the preimmune PBMC was negative. No significant difference could be attributed to the specific adjuvant used; in fact, the overall frequency of responses was 4 of 9 for subjects who had received

5 738 Fleury et al, no 1996; 174 (October) alum-adjuvanted rgp160 and 3 of 9 for those who had received the subunit in incomplete Freund's adjuvant. PBMC from 3 subjects were assessed for HlV-specific CTL activity 2 years after the first ALVAC-HIV immunization. These 3 subjects were previously positive for env-specific cytotoxic activity at month 7. For 2 of them, CTL activity was still present in in vitro-stimulated PBMC cultures at this 2-year time point. This result indicates the persistence of memory CTL activity 18 months after the last immunization. There has been much interest in understanding how T cell memory is maintained, the dogma being that long-term T cell memory is dependent on persistent antigenic stimulation and presumably that the chronic stimulus originates from antigen persistence on follicular dendritic cells [14]. Recent experiments, however, describing data from three murine models of viral infection have challenged the current dogma and suggested that memory T cells generated in these experimental models did not require specific antigen for their persistence [15-17]. Because avian pox viruses undergo abortive replicative cycles in mammalian cells, it is unlikely that there exists a persistent reservoir of immunogen in the host. Thus, our data support models in which long-term memory CD8 CTL persist in the absence of the priming antigen, indicating that T cell memory might be independent of continuous antigenic exposure. Induction of envelope-specific CTL activity has been reported from PBMC of HIV-l-seronegative human subjects immunized with HIV-l gp160-based candidate vaccine [1]. The combination consisting ofpriming with a live recombinant vaccinia virus expressing HIV -1 envelope and boosting with a recombinant soluble gp 160 envelope glycoprotein has been shown to enhance humoral antibodies and cytolytic responses compared with those induced by the rgp 160 or recombinant vaccinia alone [18]. Recently, induction of HIV -1 CTL responses in seronegative adults by the same canarypox construct has been shown by Egan et a1. [19]. These investigators described envelope-specific CTL detection in 3 (25%) of 12 subjects immunized with two or three injections of vcp125. The overall frequency ofctl responses reported herein was higher (7/18, 39%). This difference may be linked to the different stimulation protocols used (autologous EBV-transformed B cells infected with a recombinant vaccinia virus encoding the HIV-1 env gene, treated with psoralen and UV -light inactivated [19], vs. autologous irradiated HIV-1-infected PBMC in our experiments) or a unique assay at day 9 after in vitro stimulation for CTL detection [19] versus at least two assays between day 13 and 35 after in vitro stimulation in our experiments. The CTL responses reported with this vaccine approach using a nonreplicating recombinant live vector and a boost with rgp 160 are the best observed for any AIDS vaccine regimen tested, to date, in humans [1-3, 18, 19]. Acknowledgments We thank Jean Fleury, Carole Cluchat, Caroline Renaudie, and Christine Marie for their assistance in obtaining volunteers' specimens. We are grateful to Michael McChesney and Michele Fevrier for helpful discussions. References 1. Burke OS. Human trials of experimental HIV vaccines. AIDS 1995; 9(suppl A):S Dolin R. Human studies in the development of human immunodeficiency virus vaccines. J Infect Dis 1995; 172: Graham BS, Wright PF. Candidate AIDS vaccines. N Engl J Med 1995; 333: Brochier B, Kieny MP, Costy F, et al. Large-scale eradication of rabies using recombinant vaccinia-rabies vaccine. Nature 1991;354: Fenner F. Poxviruses. In: Fields BN, Knipe DM, Chanock RM, et al. Fields virology. New York: Raven Press, 1990: Taylor J, Weinberg R, Kawaoka Y, Webster RG, Paoletti E. Protective immunity against avian influenza induced by a fowlpox virus recombinant. Vaccine 1988; 6: Wild F, Giraudon P, Spehner D, Drillien R, Lecocq JP. Fowlpox virus recombinant encoding the measles virus fusion protein: protection of mice against fatal measles encephalitis. Vaccine 1990; 8: Pialoux G, Excler JL, Riviere Y, et al. A prime-boost approach to HIV preventive vaccine using a recombinant canarypox virus expressing glycoprotein 160 (MN) followed by a recombinant glycoprotein 160 (MNILAI). AIDS Res Hum Retroviruses 1995; 11: Robertson MN, Buseyne F, Kieny MP, Riviere Y. Strain specificity of cellmediated cytotoxic responses specific for the human immunodeficiency virus type 1 (HIV -I) envelope protein in seropositive donors: HIV-1LAI is more commonly recognized than HIV -1MN. J Infect Dis 1994; 169: McChesney M, Tanneau F, Regnault A, et al. Detection of primary cytotoxic T lymphocytes specific for the envelope glycoprotein of HIV-l by deletion of the env amino terminal signal sequence. Eur J Immunol 1990; 20: Buseyne F, McChesney M, Porrot F, Kovarik S, Guy B, Riviere Y. Gag specific cytotoxic T lymphocytes from HIV-l infected individuals: gag epitopes are clustered in three regions of the p24 gag protein. J Virol 1993;67: Riviere Y, Tanneau-Salvadori F, Regnault A, et al. HIV-specific cytotoxic responses of seropositive individuals: distinct types of effector cells mediate killing of targets expressing gag and env proteins. J Viro11989; 63: Buseyne F, Blanche S, Schmitt D, Griscelli C, Riviere Y. Detection of HIV-specific mediated cytotoxicity in peripheral blood from infected children. J lmmunol 1993; 150: Matzinger P. Memories are made of this? Nature 1994; 369: Mullbacher A. The long-term maintenance of cytotoxic T cell memory does not require persistence of antigen. J Exp Med 1994; 179: Hou S, Hyland L, Ryan KW, Portner A, Doherty Pc. Clonal burst size determines virus specific CD8+ cell memory. Nature 1994; 369: Lau LL, Jamieson BD, Somasundaram T, Ahmed R. Cytotoxic T cell memory without antigen. Nature 1994;369: Cooney EL, McElrath MJ, Corey L, et al. Enhanced immunity to human immunodeficiency virus (HIV) envelope elicited by a combined vaccine regimen consisting of priming with a vaccinia recombinant expressing HlV envelope and boosting with gp 160 protein. Proc Nat! Acad Sci USA 1993;90: Egan MA, Pavlat WA. Tartaglia J, Paoletti E, Clements ML, Siliciano RF. Induction of human immunodeficiency virus type I (HIV-I)-specific cytolytic T lymphocyte responses in seronegative adults by a nonreplieating, host range-restricted canarypox vector (ALVAC) carrying the HIV-I M N env gene. J Infect Dis 1995; 171: