Enhanced activation of human T cells via avipox vector-mediated hyperexpression of a triad of costimulatory molecules in human dendritic cells.

T-cell activation usually requires at least two signals. The first signal is antigen-specific, and the second signal(s) involves the interaction of a T-cell costimulatory molecule(s) on the antigen-presenting cell (APC) with its ligand on the T cell. Dendritic cells (DCs) are the most potent APCs, attributable, in part, to their expression of several T-cell costimulatory molecules. Human DCs generated in vitro, however, will vary in methods of generation and maturation and in terms of expression of different phenotypic markers-including costimulatory molecules-among different donors. We report here that a recombinant avipox (fowlpox, rF) vector has been constructed that can efficiently express the transgenes for three human T-cell costimulatory molecules (B7-1, ICAM-1, and LFA-3) as a result of individual early avipox promoters driving the expression of each transgene. This triad of costimulatory molecules (designated TRICOM) was selected because each has an individual ligand on T cells and each has been shown previously to prime a unique signaling pathway in T cells. We report here that rF-TRICOM can efficiently infect human DCs of different states of maturity and hyperexpress each of the three costimulatory molecules on the DC surface without affecting other DC phenotypic markers. Infection of influenza or human papilloma virus 9-mer peptide-pulsed DCs from different individuals, or at different stages of maturity with rF-TRICOM, resulted in enhanced activation of T cells from peripheral blood mononuclear cells of autologous donors after 24 h of incubation with DCS: This enhanced activation was analyzed by both titrating the peptide and differing the DC:effector cell ratios. No effect was observed using the control wild-type avipox vector. No increase in apoptosis was observed in T cells hyperstimulated with the TRICOM vector, and no decrease in interleukin-12 production was seen in lipopolysaccharide-stimulated DCs infected with rF-TRICOM. Antibody-blocking experiments demonstrated that enhanced T-cell activation by TRICOM was attributed to each of the three costimulatory molecules. Peptide-pulsed, rF-TRICOM-infected DCs were also shown to be more effective than peptide-pulsed DCs in activating T cells to 9-mer peptides derived from two relatively weak "self" immunogens, i.e., human prostate-specific antigen and human carcinoembryonic antigen. These studies thus demonstrate for the first time that a vector that can simultaneously hyperexpress three costimulatory molecules can be used to efficiently infect human DCs, leading to enhanced peptide-specific T-cell activation. The use of this approach for in vitro studies and clinical applications in immunotherapy is discussed.

Granulocyte/macrophage-colony stimulating factor produced by recombinant avian poxviruses enriches the regional lymph nodes with antigen-presenting cells and acts as an immunoadjuvant.

Recombinant avian poxviruses [fowlpox and canarypox (ALVAC)], restricted for replication in nonavian cell substrates and expressing granulocyte/macrophage-colony stimulating factor (avipox-GM-CSF), were evaluated for their ability to enrich an immunization site with antigen-presenting cells (APCs) and, in turn, function as biological vaccine adjuvants. Avipox-GM-CSF administered as a single s.c. injection significantly enhanced the percentage and absolute number of APCs in the regional lymph nodes that drain the injection site. Both the magnitude and duration of the cellular and phenotypic increases within the lymph nodes induced by the avipox-GM-CSF viruses were significantly (P < 0.05) greater than those measured in mice treated with four daily injections of recombinant GM-CSF protein. Temporal studies revealed that the APC enrichment of regional lymph nodes was sustained for 21-28 days after injection of the recombinant avipox virus expressing GM-CSF and, moreover, three injections of the recombinant virus could be given without any appreciable loss of in vivo bioactivity. Mice expressing human carcinoembryonic antigen (CEA) as a transgene (CEA.Tg) developed CEA-specific humoral and cell-mediated immunity after being immunized with avipox-CEA. The coadministration of recombinant avipox viruses expressing CEA and GM-CSF significantly enhanced CEA-specific host immunity with an accompanying immunotherapeutic response in tumor-bearing CEA.Tg mice. The optimal use of avipox-GM-CSF, in terms of dose and dose schedule, especially when used with different immunogens, remains to be determined. Nonetheless, the present findings demonstrate: (a) the effective delivery of GM-CSF to an immunization site using a recombinant avian poxvirus; (b) the compatibility of delivering an antigen and GM-CSF in replication-defective viruses to enhance antigen-specific immunity; and (c) the combined use of recombinant avipox viruses expressing CEA and GM-CSF to generate antitumor immunity directed at a self tumor antigen.

Identification of NY-ESO-1 peptide analogues capable of improved stimulation of tumor-reactive CTL.

Expression of NY-ESO-1 in a high proportion of different human tumors makes this protein a very attractive vaccine target. NY-ESO-1 peptides, recognized by HLA-A2-restricted CTL, have recently been described. However, it remains unclear how efficiently tumors generate these epitopes, and whether peptide analogues can be used for optimal expansion and activation of NY-ESO-1-specific HLA-A2-restricted CTL. By generating unique CTL clones, we demonstrate that NY-ESO-1-positive tumor cells are efficiently killed by HLA-A2-restricted CTL specific for the peptide epitope NY-ESO-1 157-165. Presentation of this epitope is not affected by the presence or absence of the proteasome subunits low molecular proteins 2 and 7 and is not blocked by proteasome inhibitors, while it is impaired in the TAP-deficient cell line LBL 721.174. NY-ESO-1 157-165 peptide analogues were compared for their antigenicity and immunogenicity using PBL from melanoma patients. Three peptides, containing the carboxyl-terminal cysteine substituted for either valine, isoleucine, or leucine, were recognized at least 100 times more efficiently than the wild-type peptide by specific CTL. Peptide analogues were capable of stimulating the expansion of NY-ESO-1-specific CTL from PBL of melanoma patients much more efficiently than wild-type peptide. These findings define the processing requirements for the generation of the NY-ESO-1 157-165 epitope. Identification of highly antigenic NY-ESO-1 peptide analogues may be important for the development of vaccines capable of expanding NY-ESO-1-specific CTL in cancer patients.

A phase I trial of a recombinant vaccinia virus expressing prostate-specific antigen in advanced prostate cancer.

A recombinant vaccinia virus encoding human prostate-specific antigen (rV-PSA) was administered as three consecutive monthly doses to 33 men with rising PSA levels after radical prostatectomy, radiation therapy, both, or metastatic disease at presentation. Dose levels were 2.65 x 10(6), 2.65 x 10(7), and 2.65 x 10(8) plaque forming units. Ten patients who received the highest dose also received 250 microg/m2 granulocyte-macrophage colony-stimulating factor (GM-CSF) as an immunostimulatory adjunct. No patient experienced any virus-related effects beyond grade I cutaneous toxicity. Pustule formation and/or erythema occurred after the first dose in all 27 men who received > or =2.65 x 10(7) plaque forming units. GM-CSF administration was associated with fevers and myalgias of grade 2 or lower in 9 of 10 patients. PSA levels in 14 of 33 men treated with rV-PSA with or without GM-CSF were stable for at least 6 months after primary immunization. Nine patients remained stable for 11-25 months; six of these remain progression free with stable PSA levels. Immunological studies demonstrated a specific T-cell response to PSA-3, a 9-mer peptide derived from PSA. rV-PSA is safe and can elicit clinical and immune responses, and certain patients remain without evidence of clinical progression for up to 21 months or longer.

Neutralizing antibody-independent containment of immunodeficiency virus challenges by DNA priming and recombinant pox virus booster immunizations.

Eight different protocols were compared for their ability to raise protection against immunodeficiency virus challenges in rhesus macaques. The most promising containment of challenge infections was achieved by intradermal DNA priming followed by recombinant fowl pox virus booster immunizations. This containment did not require neutralizing antibody and was active for a series of challenges ending with a highly virulent virus with a primary isolate envelope heterologous to the immunizing strain.

Recombinant vaccinia-PSA (PROSTVAC) can induce a prostate-specific immune response in androgen-modulated human prostate cancer.

OBJECTIVES: Prostate cancer recurrence, evidenced by rising prostate-specific antigen (PSA) levels after radical prostatectomy, is an increasingly prevalent clinical problem in need of new treatment options. Preclinical studies have suggested that for tumors in general, settings of minimal cancer volume may be uniquely suitable for recombinant vaccine therapy targeting tumor-associated antigens. A clinical study was undertaken to evaluate the safety and biologic effects of vaccinia-PSA (PROSTVAC) administered to subjects with postprostatectomy recurrence of prostate cancer and to assess the feasibility of interrupted androgen deprivation as a tool for modulating expression of the vaccine target antigen, as well as detecting vaccine bioactivity in vivo. METHODS: A limited Phase I clinical trial was conducted to evaluate the safety and biologic effects of vaccinia-PSA administered in 6 patients with androgen-modulated recurrence of prostate cancer after radical prostatectomy. End points included toxicity, serum PSA rise related to serum testosterone restoration, and immunologic effects measured by Western blot analysis for anti-PSA antibody induction. RESULTS: Toxicity was minimal, and dose-limiting toxicity was not observed. Noteworthy variability in time required for testosterone restoration (after interruption of androgen deprivation therapy) was observed. One subject showed continued undetectable serum PSA (less than 0.2 ng/mL) for over 8 months after testosterone restoration, an interval longer than those reported in previous androgen deprivation interruption studies. Primary anti-PSA IgG antibody activity was induced after vaccinia-PSA immunization in 1 subject, although such antibodies were detectable in several subjects at baseline. CONCLUSIONS: Interrupted androgen deprivation may be a useful tool for modulating prostate cancer bioactivity in clinical trials developing novel biologic therapies. Immune responses against PSA may be present among some patients with prostate cancer at baseline and may be induced in others through vaccinia-PSA immunization.

Inhibition of allergic encephalomyelitis in marmosets by vaccination with recombinant vaccinia virus encoding for myelin basic protein.

A primary demyelinating form of experimental allergic encephalomyelitis (EAE) resembling human multiple sclerosis (MS) occurs in Callithrix jacchus marmosets following immunization with human white matter. Participation of a T-cell immune response against myelin basic protein (MBP) in this disease model is supported by observations of increased reactivity against MBP in PBMC and of adoptive transfer of an inflammatory form of EAE by MBP-reactive T-cells. To evaluate the effects of ectopic presentation of MBP on marmoset EAE, animals were vaccinated prior to induction of EAE by subcutaneous injection of attenuated strains of vaccinia virus genetically engineered to contain either the entire coding sequence for human MBP (vT15) or the equine herpes virus glycoprotein gH gene (vAbT249). Vaccination with vT15 was followed by transient cytoplasmic and surface membrane expression of MBP in circulating PBMC (15-45 days). The onset of clinical EAE after immunization (pi) was markedly delayed in vT15-vaccinated animals (37-97 days pi, n = 4) compared to vAbT249-vaccinated controls (14-18 days pi, n = 3). Proliferative responses against MBP but not against vaccinia antigens or phytohemagglutinin were suppressed in protected animals. Thus, development of attenuated live viruses carrying genes for myelin antigens could be useful for induction of immunologic tolerance and for modulation of autoimmune demyelination.

Gene gun-based nucleic acid immunization alone or in combination with recombinant vaccinia vectors suppresses virus burden in rhesus macaques challenged with a heterologous SIV.

Gene gun-based DNA immunization alone or in combination with recombinant vaccinia vectors was evaluated for the ability to elicit protective immune responses in rhesus macaques challenged with a pathogenic, heterologous simian immunodeficiency virus (SIV). Six monkeys primed with seven consecutive doses of DNA encoding SIVmac239 gp120 and gp160 (DNA + DNA) were divided into two groups. Three of these animals received another DNA booster immunization and the remaining three received a booster immunization containing a homologous, live recombinant vaccinia virus expressing SIVmac251 gp160 (DNA + VAC). In addition, a group of 15 animals primed with recombinant vaccinia vectors were divided into two groups. One group of six monkeys received another immunization of vaccinia (VAC + VAC) and the other nine animals received a DNA (mac239) booster immunization (VAC + DNA). Geometric mean end-point IgG titres in the DNA + VAC and VAC + DNA groups were substantially higher than the responses seen in the VAC + VAC and DNA + DNA groups, demonstrating a synergistic relationship between DNA-based vaccines and recombinant vaccinia virus-based vaccines. All vaccinates and five naive controls were challenged 19 weeks after the final booster immunization with 10 animal infectious doses of SIVDelta/B670. The vaccines did not prevent infection. However, all vaccine groups showed significant virus load reductions from seven to 56 days post challenge when compared to controls. Although the DNA + DNA group developed the lowest prechallenge antibody responses, the most significant reduction (200-fold) in virus load was associated with this group. In addition, a significant delay in CD4+ T cell loss relative to controls was observed in the DNA + DNA group. These results demonstrate that a gene gun-based DNA vaccine provided some attenuation of infection and CD4+ T cell loss after a heterologous challenge.

Human immunodeficiency virus type 1-specific cytotoxic T lymphocytes (CTL), virus load, and CD4 T cell loss: evidence supporting a protective role for CTL in vivo.

The relationships between primary human immunodeficiency virus type 1 (HIV-1) Gag-specific cytotoxic T lymphocyte (CTL) frequency, virus load, and CD4 T cell loss were evaluated in a group of 46 HIV-1-infected persons with hemophilia. Freshly isolated peripheral blood mononuclear cells in limiting dilution assays were used to measure HIV-1 Gag-specific CTL frequencies. Concurrent measurements of virus load and lymphocyte surface markers were obtained. No correlation between Gag-specific CTL frequency and concurrent CD4 cell count was observed. A significant inverse relationship was observed between HIV-1 Gag-specific CTL frequency and provirus load as measured by polymerase chain reaction. Subjects with higher CTL frequencies were found to have more stable CD4 cell counts over time. These results provide additional evidence to support the concept that the predominant role of this virus-specific cellular immune response is to limit viral replication and CD4 cell loss in HIV-1 infection.

Early HIV-1 envelope-specific cytotoxic T lymphocyte responses in vertically infected infants.

High frequencies of cytotoxic T lymphocyte precursors (CTLp) recognizing HIV-1 laboratory strain gene products have been detected in adults within weeks of primary infection. In contrast, HIV-1-specific CTLp are uncommonly detected in infants younger than 6 mo. To address the hypothesis that the use of target cells expressing laboratory strain env gene products might limit the detection of HIV-1 env-specific CTLp in early infancy, recombinant vaccinia vectors (vv) expressing HIV-1 env genes from early isolates of four vertically infected infants were generated. The frequencies of CTLp recognizing target cells infected with vv-expressing env gene products from early isolates and HIV-1 IIIB were serially measured using limiting dilution followed by in vitro stimulation with mAb to CD3. In one infant, the detection of early isolate env-specific CTLp preceded the detection of IIIB-specific CTLp. CTLp recognizing HIV-1 IIIB and infant isolate env were detected by 6 mo of age in two infants. In a fourth infant, HIV-1 IIIB env and early isolate env-specific CTLp were simultaneously detected at 12 mo of age. These results provide evidence that young infants can generate HIV-1-specific CTL responses and provide support for the concept of neonatal vaccination to prevent HIV-1 transmission. However, the early predominance of type-specific CTL detected in some young infants suggests that the use of vaccines based on laboratory strains of HIV-1 may not protect against vertical infection.

Strategies for the development of recombinant vaccines for the immunotherapy of breast cancer.

The development of recombinant vaccines for specific immunotherapy of carcinoma represents a novel approach for the treatment of breast cancer and other tumor types. This article reviews the various parameters that should be considered in the development of recombinant vaccines. Several breast cancer associated antigens are also discussed which may provide potential target molecules. The human carcinoembryonic antigen (CEA), which is expressed on approximately 50% of breast cancers, represents one such target for immunotherapy. To enhance the immunogenicity of this antigen, a recombinant CEA-vaccinia vaccine, designated rV-CEA, was produced. To study the effects of this vaccine in an animal model, a murine colon carcinoma cell line was transduced with CEA and transplanted into immunocompetent mice for protection and therapy studies. Pre-clinical toxicity studies were also conducted in non-human primates. The results of these studies showed the rV-CEA vaccine to be immunogenic and safe in both rodents and primates, and to elicit good anti-tumor responses in the rodent model. In a Phase I clinical trial in metastatic breast, lung, and colorectal cancer patients involving three immunizations of rV-CEA, at three dose levels, enhancement of T-cell and antibody responses to vaccinia virus proteins were observed with no toxicity. Specific T-cell responses were studied via stimulation of peripheral blood lymphocytes with specific peptide epitopes from the CEA molecule. These studies demonstrated clear cut differences in establishment of T-cell lines pre- versus post-immunization. The T-cell lines were shown to be CD8+ and/or CD4+/CD8+, to lyse EBV transformed B-cells transduced with the CEA gene, and to lyse CEA positive carcinoma cells in a HLA restricted manner. Thus, in a Phase I clinical trial the rV-CEA vaccine has been shown to stimulate a CTL response specific for CEA defined epitopes in cancer patients.

A recombinant vaccinia virus expressing human prostate-specific antigen (PSA): safety and immunogenicity in a non-human primate.

Prostate-specific antigen (PSA) is a serine protease secreted by prostatic epithelial cells and is widely used as a marker for prostate cancer. The tissue specificity of PSA makes it a potential target for active specific immunotherapy, especially in prostate cancer patients who have undergone prostatectomy and in whom the only PSA-expressing tissue in the body resides in metastatic deposits. We report here the cloning, construction and immunological consequences of immunization of rhesus monkeys with a recombinant vaccinia virus expressing human PSA (designated rV-PSA). The prostate gland of the rhesus is structurally and functionally similar to the human prostate. While rodent and other mammalian species do not share homology with human PSA, there is 94% homology between the amino acid sequences of rhesus and human PSA. Immunization of rhesus monkeys with wild-type vaccinia virus or rV-PSA elicited the usual low-grade constitutional symptoms of vaccinia virus infection. There was no evidence of any adverse effects in any immunized monkeys. A short-lived PSA-specific IgM antibody response was noted in all rV-PSA immunized monkeys regardless of dose level. All monkeys receiving the 10(8)pfu dose of rV-PSA demonstrated PSA-specific T-cell responses that were maintained up to 270 days. No differences in anti-PSA immune responses or toxicity were observed in animals that received prostatectomy prior to immunization. Our results thus demonstrate the safety and immunogenicity of rV-PSA in a non-human primate and have implications for potential specific immunotherapy protocols using PSA as a target.

High levels of anti-human immunodeficiency virus type 1 (HIV-1) memory cytotoxic T-lymphocyte activity and low viral load are associated with lack of disease in HIV-1-infected long-term nonprogressors.

Lack of disease in long-term nonprogressors with human immunodeficiency virus type 1 (HIV-1) infection was strongly associated with very low copy numbers of HIV-1 DNA and RNA in peripheral blood mononuclear cells and plasma and the presence of high levels of anti-HIV-1 CD8+ memory cytotoxic T lymphocytes specific for Gag, Pol, and Env, compared with levels present in intermediate and advanced progressors. CD8+ memory cytotoxic T lymphocytes may have an important role in controlling HIV-1 replication and preventing disease in long-term nonprogressors.

Virus burden in long-term survivors of human immunodeficiency virus (HIV) infection is a determinant of anti-HIV CD8+ lymphocyte activity.

Persons infected with human immunodeficiency virus (HIV) for > 8 years were studied to delineate virologic and immunologic attributes of long-term survival. Whereas those with 300-700 CD4+ cells/microL often had circulating cytotoxic T lymphocytes (CTL) against HIV antigens, those with > 1000 CD4+ cells/microL did not. The subjects with > 1000 CD4+ cells/microL had low virus burden, low levels of Gag-specific CTL precursors, and minimal CD8+ cell activation. Overall, elevated levels of CD8+ cells, CD38 antigen expression on CD8+ cells, and anti-HIV functions were correlated with increased virus burden, provirus load, and HIV plasma RNA levels. A factor that suppressed HIV replication was spontaneously secreted from CD8+ cells of most subjects but not from those with high CD4+ cell counts. CD8+ cell activities, therefore, may reflect chronic viral stimulation of the immune system. Long-term survivors with high levels of CD4+ cells maintained control of viral replication but lacked the CD8+ cell activities.

Active immunotherapy of cancer with a nonreplicating recombinant fowlpox virus encoding a model tumor-associated antigen.

Some tumor cells express Ags that are potentially recognizable by T lymphocytes and yet do not elicit significant immune responses. To explore new immunotherapeutic strategies aimed at enhancing the recognition of these tumor-associated Ags (TAA), we developed an experimental mouse model consisting of a lethal clone of the BALB/c tumor line CT26 designated CT26.WT, which was transduced with the lacZ gene encoding beta-galactosidase, to create CT26.CL25. The growth rate and lethality of CT26.CL25 and CT26.WT were virtually identical despite the expression by CT26.CL25 of the model tumor Ag in vivo. A recombinant fowlpox virus (rFPV), which is replication incompetent in mammalian cells, was constructed that expressed the model TAA, beta-galactosidase, under the influence of the 40-kDa vaccinia virus early/late promoter. This recombinant, FPV.bg40k, functioned effectively in vivo as an immunogen, eliciting CD8+ T cells that could effectively lyse CT26.CL25 in vitro. FPV.bg40k protected mice from both subcutaneous and intravenous tumor challenge by CT26.CL25, and most surprisingly, mice bearing established 3-day pulmonary metastasis were found to have significant, Ag-specific decreases in tumor burden and prolonged survival after treatment with the rFPV. These observations constitute the first reported use of rFPV in the prevention and treatment of an experimental cancer and suggest that changing the context in which the immune system encounters a TAA can significantly and therapeutically alter the host immune response against cancer.

HIV-1-specific cytotoxic T lymphocyte responses in the first year of life.

HIV-1-specific CTL responses were prospectively evaluated in infants born to HIV-1-seropositive women to assess the capability of the young infant to generate HIV-1-specific CTL and to examine the potential role of HIV-1-specific CTL in the pathogenesis of vertical infection. Our results indicate that some young infants, and even the fetus, seem to be capable of generating virus-specific CTL responses. The detection of HIV-1-specific CTL responses varied among infants, however, with respect to timing, HIV-1 gene product recognition, and the magnitude of detectable responses; HIV-1-specific CTL responses were uncommonly detected in the first few months of life. The less consistent detection of HIV-specific CTL in early infancy contrasts with reports of the detection of HIV-1-specific CTL soon after primary infection in adults. HIV-1-specific CTL were not detected in any uninfected infants born to HIV-1-seropositive women. This description of HIV-1-specific CTL in infants may have important implications for understanding the pathogenesis of vertical HIV-1 infection and for the development of a vaccine to interrupt vertical infection.

High-titer immune responses elicited by recombinant vaccinia virus priming and particle boosting are ineffective in preventing virulent SIV infection.

Eighteen rhesus monkeys were vaccinated with recombinant vaccinia viruses expressing SIVmac antigens in 3 separate rounds of experiments. Twelve of the monkeys were primed with a trivalent vaccinia virus recombinant expressing Gag, Pol, and Env polypeptides that can assemble into SIV pseudovirion particles and boosted with SIV particles in adjuvant. Four of the monkeys were primed with different vaccinia virus recombinants expressing env or gag+env followed by SIV particle boosts; two received vaccinia virus recombinants alone (env or env+gag). Despite the induction of vigorous immune responses, 17 of 18 rhesus monkeys became infected on challenge with a low dose of virulent SIVmac. The single protected animal was one of three challenged with homologous cloned SIV exactly matched to the clone used for construction of trivalent vaccinia virus recombinant and particles. Vaccination may have diminished SIV burdens and rates of CD4+ cell declines in some of the animals, but vaccinated/challenge/infected animals eventually developed fatal disease similar to control animals. These results highlight the extreme difficulty in achieving vaccine protection against virulent SIVmac infection even under idealized laboratory conditions.

Induction of antigen-specific killer T lymphocyte responses using subunit SIVmac251 gag and env vaccines containing QS-21 saponin adjuvant.

Subunit vaccines based on recombinant proteins have proved useful for inducing antibody responses and they are safe for widespread use because they do not contain any live components. Unfortunately, they do not typically induce the types of cell-mediated immune responses required to control viral pathogens; specifically, they do not induce CD8+ cytotoxic T lymphocyte (CTL) responses. To increase the immunogenicity of recombinant proteins, we have used the QS-21 saponin adjuvant in subunit vaccine formulations. In the current study, experimental subunit vaccine formulations containing recombinant p55gag or gp120env proteins from the mac251 strain of the simian immunodeficiency virus (SIVmac251) and the QS-21 adjuvant were used to immunize rhesus macaques. These formulations induced SIV gag- or env-specific cellular immunity that was detectable in vitro and included killer cell activity. The induction of killer cells required prior vaccination and the responses were antigen specific for the immunogens contained in the vaccine formulations. Autologous target cells were required to detect these responses, suggesting genetic restriction, and effector cells appeared to be present in both the CD4+ and CD8+ T lymphocyte subpopulations. These data suggest that the vaccine-induced killer cell activity that was detected was mediated by both CD4+ and CD8+ lymphocytes. Despite the presence of these killer cells, all of the animals became infected with the SIVmac251 on experimental challenge. These findings demonstrated that antigen-specific killer cell responses could be induced by a subunit vaccine formulated with the QS-21 saponin adjuvant. The characteristics of the responses suggested that the effector cells were T lymphocytes, expressing either CD4 or CD8.(ABSTRACT TRUNCATED AT 250 WORDS)