Safety and Immunogenicity of a Candidate HIV-1 Vaccine in Healthy Adults: Recombinant Glycoprotein (rgp) 120: A Randomized, Double-Blind Trial

Methods

Preparation of the Vaccine Product

Human immunodeficiency virus gp120 was produced in genetically engineered Chinese hamster ovary cells under the control of the cytomegalovirus immediate early-1 promoter. It was purified by ion exchange, hydrophobic interaction, and gel chromatography. The emulsifier MF59 consists of polysorbate 80 (Tween 80, polyoxethylene sorbitan monooleate), sorbitan trioleate (Span 85, Arlacel 85), and squalene. Emulsification by high-pressure homogenization results in a stable droplet size of less than 200 nm. MTP-PE (Ciba-Geigy, Ltd., Basel, Switzerland) is a muramyl tripeptide (N-acetylmuramyl-L-alanyl-D-isoglutaminyl-L-alanine) that is linked covalently with dipalmitoyl phosphatidylethanolamine. The gp120 antigen was supplied in single-dose vials that each contained either 15 µg or 50 µg of antigen in 0.5 mL of 2 times phosphate-buffered saline and were stored at −70°C. The MF59 emulsion was stored at 4 °C. The rgp120 antigen was mixed with MF59 with or without MTP-PE within 4 hours before administration.

Study Participants

All participants were healthy, HIV-seronegative adults at low risk for exposure to HIV-1. Participants in part A were men and women between 18 and 60 years of age; all participants in part B were women. Part B was intended to provide pilot data for anticipated trials in HIV-infected pregnant women and to investigate the role of the vaccination schedule on immune response.

Each volunteer had a physical examination, complete blood count, urinalysis, measurement of alanine aminotransferase (ALT) and creatinine levels, and intradermal skin test (Multitest; Merieux Institute, Inc., Miami, Florida) [6]. Participants were excluded if they used immunosuppressive medications; had a history of anaphylaxis; had a chronic illness, such as tuberculosis, autoimmunity, or hepatitis B; or had received immunoglobulin or commercial vaccines within 2 months, experimental treatments within 30 days, or blood or cryoprecipitate transfusions within 3 months.

All volunteers gave informed consent as approved by local institutional review boards and institutional biosafety committees. The study was done according to the human experimentation guidelines of the U.S. Department of Health and Human Services.

Experimental Design

In this multicenter, randomized, double-blind trial, we evaluated the safety and immunogenicity of a candidate HIV-1 vaccine composed of rgp120 (derived from Chinese hamster ovary cells) formulated in MF59 with or without the biological response modifier MTP-PE. Part A compared two different quantities of antigen used with or without MTP-PE. Thirty-two participants were randomly assigned to one of four groups (Table 1). MTP-PE was omitted from the injection given at 6 months because of local reactions seen at 0 months and at 1 month. In part B, 16 women were randomly assigned to receive either 50 µg of rgp120 in MF59 or 50 µg of MF59 alone (Table 1). One additional woman was recruited because 1 participant dropped out of the study early. Study participants in part A and vaccine recipients in part B received a booster vaccination at 12 to 18 months.

The study was done at five AIDS (acquired immunodeficiency syndrome) Vaccine Evaluation Units supported by the National Institute of Allergy and Infectious Disease. Vanderbilt University, the University of Rochester, and the University of Washington entered participants into part A, and Johns Hopkins University and St. Louis University entered participants into part B. Serologic assays were done at the Central Immunology Laboratory for the AIDS Vaccine Evaluation Group at Duke University Medical Center and at Biocine/Chiron Corp. All lymphoproliferation assays were done on cryopreserved cells at Vanderbilt University. EMMES Corp. served as the central unit for data management and statistical analysis.

The randomization scheme was a permuted block design, blocked according to study site. Assignments were concealed by central allocation, and only the site pharmacists and EMMES Corp. were aware of product allocation codes. Investigators from Biocine/Chiron Corp. aided in protocol design, CD4-binding inhibition assays, and cytotoxic T-lymphocyte evaluation. Data were analyzed at EMMES Corp. and interpreted by investigators from the AIDS Vaccine Evaluation Group.

Clinical and Laboratory Safety Evaluation

Each participant had a physical examination, complete blood count, CD4 and CD8 lymphocyte subset analyses, measurement of serum creatinine and ALT levels, urine pregnancy test, and urinalysis at the time of each vaccination. After each vaccination, participants were observed for 1 hour. Oral temperatures and symptoms were recorded 1, 6, and 12 hours after vaccination. Participants were examined on day 1 or day 2 after vaccination. During the 3-day period after vaccination, they were contacted by telephone on the days on which they were not examined. Each participant kept a written diary of symptoms through day 10. Brief examinations and clinical laboratory screening assays were also done on days 28, 42, 56, 112, 168, 196, 224, and 336. Visits related to the late booster vaccination were made on days 0, 14, 28, 84, and 168.

Laboratory Immunogenicity Evaluation

Serologic Testing

Western blot assays (Du Pont Co., Wilmington, Delaware) and HIV-1 lysate enzyme-linked immunosorbent assays (ELISAs) (Abbott Laboratories, Chicago, Illinois) were done according to the manufacturers' directions. We did HIV-1 envelope-specific ELISAs using homologous SF-2 gp120 derived from Chinese hamster ovary cells, truncated gp120 derived from yeast (Chiron/Biocine Corp., Emeryville, California), LAI gp120 derived from baculovirus (Repligen, Cambridge, Massachusetts), or a V3 loop peptide based on the SF-2 sequence. Results of ELISA from the Central Immunology Laboratory were recorded as the arithmetic mean difference in optical density (±SE) between duplicate antigen-positive and antigen-negative wells for serum dilutions 1:100 for gp120 and 1:50 for the peptides [7]. Neutralization assays against homologous and heterologous laboratory strains of HIV-1 were done as described elsewhere [7], and fusion inhibition was determined by plaque reduction assays as described elsewhere [8]. The MN CD4-binding inhibition titer, measured at the Central Immunology Laboratory, was defined as the serum dilution that reduced binding of ¹²⁵I-radiolabeled MN rgp120 to CD4-positive MOLT-4 cells by 50% [8]. The SF-2 CD4-binding inhibition assays were done at Chiron/Biocine Corp. using a modification of the method in which inhibition of fluorescein isothiocyanate-conjugated recombinant CD4 binding to SF-2 gp120 was the end point. The CD4-blocking titer was the reciprocal of the serum dilution that resulted in CD4 binding of 50% or less. Antibody-dependent cell-mediated cytotoxicity studies were done as described elsewhere [9, 10] using CEM.NKR target cells coated with purified LAI gp120 and SF-2 gp120.

Measurement of Cell-Mediated Immune Response

We evaluated cryopreserved peripheral blood mononuclear cells for lymphoproliferative responses by measuring ³H-thymidine incorporation after exposure to phytohemagglutinin (5 µg/mL) (Sigma, St. Louis, Missouri), Candida antigen (20 µg/mL) (Greer Laboratories, Lenoir, North Carolina), rgp120, env2-3 antigen (1 µg/mL) (Chiron/Biocine Corp.), and control substances, including medium alone, an extract of Chinese hamster ovary cells, and yeast-derived feline leukemia virus protein, as described elsewhere [11]. Results were expressed as the mean stimulation index at each time point, where stimulation index = (mean sample counts per minute)/(mean counts per minute from either the Chinese hamster ovary cell extract, feline leukemia virus protein, or medium alone).

Cytotoxic T-lymphocyte responses were measured on days 175, 182, and 196 in a subset of participants. Fresh peripheral blood mononuclear cells were co-cultivated with autologous vaccinia-env-infected peripheral blood mononuclear cells for 10 to 14 days, after which time cultures were evaluated for MHC class I-restricted, CD8 (+) cytotoxic T-lymphocyte reactivity. Target cells were B-lymphoblastoid cell lines transformed by the Epstein-Barr virus from each participant infected overnight with recombinant vaccinia virus constructs expressing gp160 or irrelevant proteins. Cytolytic activity was determined by measuring specific ⁵¹Cr release from target cells after 4 hours of incubation with effector cells (effector cells:target cells = 100:1 or 50:1) as described elsewhere [9]. Percent specific lysis was determined using the following formula: (sample counts per minute − spontaneous release counts per minute)/(total release counts per minute − spontaneous release counts per minute) × 100.

Statistical Analysis

Repeated measures analysis of variance was used to evaluate differences in local reactions between groups and to assess changes in hematologic, chemistry, and lymphocyte measurements. Standard nonparametric tests (the Kruskal-Wallis test and the Wilcoxon test) were used to assess the frequency and intensity of systemic reactions or local pain or tenderness and the measurements of immunogenicity. Differences in the ELISA results of participants who received 15 µg of gp120 and participants who received 50 µg of gp120 were compared using the Kruskal-Wallis test. Two-way analysis of variance was used to evaluate differences in functional antibody results that may have been due either to an effect of MTP-PE or to an antigen-dose effect. Associations between ELISA titers and neutralization titers or cross-reactivity of neutralization between HIV-1 strains were analyzed using the Spearman rank correlation.

Results

Participants

Thirty-two participants were enrolled in part A, 8 in each of the four groups. Thirty-one received the second vaccination, given at 1 month; 28 received the third vaccination, given at 6 months. One participant was withdrawn from the study after the first vaccination because of self-limited nasal congestion and myalgia, one received a diagnosis of breast cancer and did not receive the 6-month vaccination, one declined further participation after the second vaccination, and one was lost to follow-up after the second vaccination. The late booster injection was offered as an option in a protocol amendment; 22 participants chose to receive it. Sixteen women were initially enrolled in part B. One of the placebo recipients refused to participate after the first vaccination for an unspecified reason, and she was replaced by an additional participant. Of the 8 vaccine recipients in part B, 6 agreed to receive the late booster vaccination. The demographic profiles of the group in part A and the group in part B were similar (Table 2).

Clinical Response to Vaccination

Participants who received vaccine that contained MTP-PE had a statistically significant increase in local and systemic side effects (Table 3). In part A, moderate or severe local pain or tenderness was noted in 14 of 16 MTP-PE recipients and 7 of 16 participants who received only rgp120 in MF59; systemic reactions, including fever, were noted in 13 of 16 recipients of MTP-PE and 0 of 16 participants who received only rgp120 in MF59. Severe symptoms were seen only in participants who received MTP-PE, occurred within 48 hours of vaccination, and resolved within 48 hours of onset. Three participants had temperatures greater than 101 °F; all 3 had received MTP-PE. A severe reaction after the first vaccination did not predict a severe reaction after the second vaccination.

Laboratory Safety Profile

All clinical laboratory measurements made to monitor toxicity were normal. No statistically significant changes in CD4 lymphocyte counts were seen.

Humoral Response to Vaccination

Vaccinated participants consistently produced antibody that was detected by ELISAs done using gp120 antigen derived from Chinese hamster ovary cells or yeast, LAI gp120 derived from baculovirus, or a V3 loop peptide based on the SF-2 sequence. The magnitude of the response to homologous gp120 and the V3 peptide was similar to that seen in a set of randomly selected serum specimens obtained from persons with clinically asymptomatic HIV-1 infection (Figure 1). Adding MTP-PE did not significantly improve binding antibody responses, but a trend toward higher antibody titers was seen for the participants who received the 15-µg antigen dose and MTP-PE (Figure 1).

Functional antibody responses were measured by evaluating neutralization; fusion inhibition; blocking of CD4 binding by gp120; and antibody-dependent, cell-mediated cytotoxicity. All participants who received 50 µg of gp120 with or without MTP-PE developed type-specific neutralizing activity after the third vaccination (Table 4). The neutralizing titer induced by the 50-µg dose was significantly higher than that induced by the 15-µg dose for SF-2 (P < 0.005) and MN (P < 0.05), but the magnitude of neutralizing activity was not significantly affected by the addition of MTP-PE. Neutralization titers were 5-fold to 10-fold lower than those measured in samples obtained from a randomly selected group of persons with HIV-1 infection. Similarly, the 50-µg gp120 dose induced a greater frequency and magnitude of fusion inhibition activity against SF-2 than did the 15-µg dose (P < 0.005) (Table 4). Vaccine-induced antibody inhibited CD4 binding to MN gp120 as well as to SF-2 gp120 and was found in most participants who received 50 µg of rgp120 with MTP-PE. However, the activity was more than 10-fold lower in these participants than in persons infected with HIV-1 (Table 4).

The monthly vaccination schedule used in part B did not substantially boost either binding or neutralizing antibody responses (Figure 1), and it resulted in a shorter duration of serum antibody response than did the schedule used in part A (Figure 1). Although repeated booster injections at short intervals dampened the antibody response of participants in part B, the response after the late booster vaccination was similar for participants in parts A and B. The late booster vaccination also showed cross-neutralizing activity in part B participants. Of six participants who had neutralizing activity against HIV-1/SF-2, six neutralized HIV-1/MN and two neutralized HIV-1/LAI, all at titers similar to those of participants in part A. Low-titered, antibody-dependent, cell-mediated cytotoxicity was shown in only 4 of 36 participants; each of the 4 was in a different group.

Serum specimens obtained from participants 4 weeks after the third vaccination were analyzed to determine the frequency of vaccine-induced reactions that falsely suggested HIV infection. Only 6 of 36 vaccinated participants (16%) had positive HIV ELISA results. Their samples had optical densities just greater than the cutoff value (mean, 0.178 [range, 0.160 to 0.200]; cutoff value, 0.156). Three of the samples that showed positive HIV ELISA results had positive bands at gp120 of 1+ intensity on the Western blot assay. Overall, 21 of 36 vaccinated participants had visible low-intensity bands at gp120 on Western blot assay. At the end of the follow-up period (6 months after the late booster injection) 1 of 28 participants had a positive HIV ELISA result and 9 had detectable envelope bands on Western blot assay.

Cellular Immune Response to Vaccination

Cytotoxic T-cell activity was evaluated in a subset of 11 vaccinated participants. A total of 41 assays were done 7, 14, or 28 days after the third vaccination. In one participant who received 15 µg of rgp120 in MF59 plus MTP-PE, specific cytolytic activity was detected. In this participant, cytotoxic T-lymphocyte activity was also documented 14 and 28 days after the fourth vaccination, with specific lysis of 16% and 11.5%, respectively. Depletion of CD8+ T lymphocytes using immunomagnetic beads eliminated the activity, but depletion of CD4+ T-lymphocytes did not. After this participant received a fifth vaccination, the activity was confirmed by an alternative method: Peripheral blood mononuclear cells were stimulated with peptide pools from the carboxy terminus of gp120 and were found to have cytolytic activity against peptide-labeled target cells. The participant had no history of high-risk exposure and was shown to be HIV-negative by ELISA, Western blot assay, p24 antigen assay, co-culture assay, and polymerase chain reaction.

Lymphoproliferative assays were done on samples obtained from 3 to 7 participants at each time point (Figure 2). Responses to phytohemagglutinin and Candida antigen in all participants remained unchanged throughout the study period. A specific lymphoproliferative response to gp120 was detected in all vaccinated participants. Responses to the rgp120 derived from Chinese hamster ovary cells were fourfold greater than responses to yeast-derived, denatured gp120 antigen: env2-3. In participants from parts A and B, responses were still present after 12 months. However, lymphoproliferative activity was about 10 times less in participants from part B than in those from part A. The booster vaccination increased the magnitude of lymphoproliferative responses, especially in participants who received the 15-µg dose of rgp120, but responses did not exceed the peak achieved after the third vaccination. MTP-PE did not substantially affect the induction of lymphoproliferative activity.

Discussion

Using a subunit gp120 strategy is a logical initial approach to developing a vaccine for HIV. Glycoprotein120 is a surface-expressed glycoprotein responsible for attachment to susceptible cells. It is a target for most neutralizing antibodies and is analogous to the hepatitis B surface antigen, which is the basis for an effective, licensed vaccine product. Previous studies by the AIDS Vaccine Evaluation Group have shown the safety and immunogenicity of rgp160 products [12-14], MN rgp120 [15] LAI rgp120 [16], and yeast-derived envelope protein [17]. In general, the gp120 products have induced higher neutralizing activity although they lack a neutralizing epitope found in the gp41 component of gp160 [18].

The structure and conformation of envelope glycoproteins may be an important determinant of the immunogenicity of these glycoproteins. The SF-2 rgp120 that we used was made in mammalian cells to ensure authentic glycosylation patterns and was purified by methods designed to prevent denaturation. The vaccine was immunogenic, but it induced type-specific functional antibody responses that were 5- to 10-fold lower than those induced by natural infection. Comparing the magnitude and breadth of neutralizing antibody responses with those of responses induced by yeast-derived, denatured SF-2 gp120 formulations [17, 19] suggests that native structure is important for the induction of functional antibody responses; this has been shown elsewhere in animals [20]. The magnitude of antibody response was among the highest seen with any candidate vaccines studied to date and was similar to that induced by an MN rgp120 product formulated with aluminum hydroxide [15]. However, some features of the antibody response were suboptimal. Although glycosylation or conformational attributes improved the immunogenicity of this product, the CD4-binding domain was not as immunogenic as that in previous animal studies [21]. Finding ways to improve the immunogenicity of the CD4-binding domain could improve the neutralizing titer and breadth of response. In addition, serum specimens obtained from participants who had neutralizing activity against laboratory isolates failed to neutralize primary field isolates of HIV-1 grown in peripheral blood mononuclear cells [22]. This may be related either to the magnitude of the response or to the monomeric structure of the vaccine. There may be important conformational neutralization determinants on the oligomer to which antibody response cannot be induced by the monomer [23].

The development of new adjuvants and vaccine delivery systems are as important to vaccine development as the improvement of vaccine antigens. When formulating a vaccine product, one has to consider not only the antigenic target for the immune response but also the composition of the immune effector mechanisms induced [24, 25]. The candidate vaccine evaluated in this study used an oil-in-water, microfluidized emulsion—MF59—that induced both humoral and cellular immune responses, including CD8+ cytotoxic T-lymphocyte activity in one participant. The microfluidization process creates 200-nm oil droplets in the emulsion; these may be antigen carriers and may interact with cell membranes much like liposomes [26, 27]. Studies done in mice and baboons [28] have shown that MF59 can induce MHC class I-restricted CD8+ cytotoxic T-lymphocyte activity in addition to antibody responses [28]. CD8+ cytotoxic T lymphocytes are thought to be important for clearance of HIV-infected cells [29-31]. An early, vigorous cytotoxic T-lymphocyte response after exposure to HIV has the potential to clear infection before dissemination. This mechanism has been suggested by the recent finding of HIV-specific cytotoxic T lymphocyte in persons who remained uninfected despite numerous high-risk exposures [32]. That a gp120-specific CD8+ cytotoxic T-lymphocyte response was induced suggests that antigen escaped the endocytic compartment and entered the MHC class I processing pathway [33]. Presumably, the nature of the interaction between the antigen and the MF59 emulsion allows this to happen in at least some instances. This activity has not been detected in participants who received other formulations of HIV-protein subunit vaccines.

MTP-PE is an immunomodulator analogous to compounds present in bacterial cell walls. In animal studies [21], it has been shown to increase neutralization titers when combined with incomplete Freund's adjuvant. In our study, it caused substantial reactogenicity. A trend toward improved immunogenicity was seen for MTP-PE at the 15-µg dose of rgp120, but the incremental benefit of this compared with MF59 alone did not outweigh the increase in reactogenicity. Similar conclusions were made in studies that evaluated a series of MTP-PE doses in combination with a fixed dose of rgp120 [34] or env2-3 antigen [17].

Defining the optimal vaccination schedule is also important. Monthly injections of rgp120 in MF59 resulted in a diminished antibody response and an attenuated cellular response after the fourth consecutive monthly dose. An interval of 5 months between the second and third vaccinations resulted in higher peak antibody titers and lymphoproliferative activity. After a 6- to 12-month interval, the same antibody titer was found in participants receiving vaccinations at 0, 1, and 6 months and in participants receiving monthly vaccinations. Therefore, for this product, the vaccination schedule should include no more vaccinations than those done at 0, 1, and 2 months before a booster vaccination is given several months later. We did not address the need for additional booster vaccinations. Although these findings may have implications for other vaccine antigens, adjuvants, and delivery systems, each product may have unique properties that determine its optimal vaccination schedule. For example, rgp160 formulated with alum and deoxycholate given on an accelerated schedule did not lead to the attenuation of antibody responses even after five consecutive monthly injections [35].

Although antibody responses were consistently induced, participants rarely had positive results on standard HIV diagnostic tests. Nevertheless, the risk for seropositivity continues to be a major deterrent to volunteer enrollment in phase I studies of HIV. Ongoing attention to vaccine design and development of diagnostic tests is needed to minimize the effect of vaccine-induced seropositivity that may exclude participants from blood product donation and endanger their insurability [36].

Vaccination of healthy adult participants with native rgp120 formulated in MF59 induced 1) neutralizing antibodies against laboratory strains of HIV-1, 2) sustained T-cell memory detected by lymphoproliferation, and 3) CD8+ cytotoxic T-lymphocyte activity in 1 of 11 participants. Vaccination with rgp120 in MF59 was associated with no adverse systemic reactions and only mild local reactions. The product thus appears to be safe after more than 4 years of follow-up and is immunogenic for both the humoral and cellular limbs of the immune system. However, it did not induce antibody responses that could neutralize primary HIV-1 field isolates grown in peripheral blood mononuclear cells, and its neutralizing activity against laboratory strains of HIV-1 was about one tenth that seen in asymptomatic HIV-infected persons (Table 4). Also, vaccination with rgp120 was rarely able to induce CD8+ cytotoxic T-lymphocyte activity, which may be an important mechanism of HIV clearance.

Ultimately, the value of a candidate vaccine depends on its ability to prevent infection or disease progression. One participant who had received this candidate vaccine in another study [37] was infected with HIV-1 acquired during a high-risk sexual exposure, but this person's long-term clinical course is not yet known [37]. Testing in large-scale trials is needed to determine whether neutralizing antibody activity against laboratory strains of HIV-1 combined with memory T cells that can recognize HIV-1 antigens in lymphoproliferation assays and rare induction of cytotoxic T-lymphocyte activity is sufficient to partially protect against HIV-1 infection or disease progression. Recombinant glycoprotein 120 in MF59 is now being evaluated 1) in a phase II study in high-risk persons, 2) as a booster vaccination after primary vaccination with live vector vaccine candidates, 3) with alternative adjuvant products, and 4) in phase I studies at international sites.

Appendix

Glossary

Biological response modifier: Stimulates a nonspecific inflammatory response that can change the composition or magnitude of immune response to a vaccine antigen.

CD8+ cytotoxic T lymphocyte: A T cell that causes lysis of virus-infected cells after recognizing processed peptide epitopes from viral proteins presented in the context of the host's MHC class I molecule on the surface of the infected cell.

Envelope: The product of the env gene of HIV that is the only protein component of the outer virus membrane. It is a 160-kd glycoprotein that is endoproteolytically cleaved to result in a 120-kd external domain, gp120, and a transmembrane domain, gp41.

Env2-3: An antigen that is a denatured, recombinant protein derived from yeast containing SF-2 gp120.

MF59: An adjuvant that consists of polysorbate 80 (Tween 80, polyoxethylene sorbitan monooleate), sorbitan trioleate (Span 85, Arlacel 85), and squalene. Emulsification by high-pressure homogenization results in a stable droplet size of less than 200 nm. Squalene constitutes the oil phase of the oil-in-water emulsion.

MTP-PE: A biological response modifier that is a muramyl tripeptide (N-acetylmuramyl-L-alanyl-D-isoglutaminyl-L-alanine) linked covalently with dipalmitoyl phosphatidylethanolamine.

Neutralization: The reduction of viral infectivity that occurs when antibody is mixed with cell-free virus before inoculation of susceptible cells.

rgp120: A purified recombinant 120-kd glycoprotein derived from Chinese hamster ovary cells containing the gene for SF-2 gp120.

Variable region 3 (V3) loop: The third variable domain of the HIV-1 envelope gene contains a sequence that is a principal neutralizing determinant of the virus. The sequence is in a region that forms a loop because of disulfide bonding between two cysteine molecules.

From Vanderbilt University School of Medicine, Nashville, Tennessee; University of Rochester School of Medicine and Dentistry, Rochester, New York; University of Washington School of Medicine, Seattle, Washington; St. Louis University School of Medicine, St. Louis, Missouri; Johns Hopkins School of Hygiene and Public Health and Johns Hopkins School of Medicine, Baltimore, Maryland; Duke University Medical Center, Durham, North Carolina; EMMES Corp., Potomac, Maryland; Biocine/Chiron Corp., Emeryville, California; and the National Institute of Allergy and Infectious Diseases, Bethesda, Maryland.

Presented in part at the 5th Annual National Cooperative Vaccine Development Groups for AIDS in Chantilly, Virginia, on 30 August-2 September 1992 and at the IXth International Conference on AIDS in Berlin, Germany, on 7-11 June 1993.

Additional authors are Peter F. Wright, MD; Raphael Dolin, MD; Lawrence Corey, MD; Robert B. Belshe, MD; Mary Lou Clements, MD; Dani P. Bolognesi, PhD; Donald M. Stablein, PhD; David Chernoff, MD; Anne Marie Duliege, MD; and Christopher M. Walker, PhD.

Drs. Keefer and Dolin: Infectious Diseases, Box 689, University of Rochester Medical Center, 1601 Elmwood Avenue, Rochester, NY 14642.

Drs. McElrath and Corey: University of Washington, AIDS Vaccine Evaluation Unit, Room 9301, 1200 12th Avenue South, Seattle, WA 98144.

Drs. Gorse and Belshe: St. Louis University School of Medicine, Room FDT-8N, 3635 Vista Avenue, St. Louis, MO 63110.

Drs. Schwartz and Clements: Johns Hopkins University, Center for Immunization Research, Hampton House, 624 North Broad-way, Baltimore, MD 21205.

Drs. Weinhold, Matthews, and Bolognesi: Duke University Medical Center, SORF Building, Room 204, La Salle Street Extension, Durham, NC 27710.

Ms. Esterlitz and Dr. Stablein: EMMES Corporation, 11325 Seven Locks Road, Suite 214, Potomac, MD 20854.

Drs. Sinangil, Chernoff, Duliege, and Walker: Chiron/Biocine Corporation, 4560 Horton Street, Emeryville, CA 94608-2916.

Dr. Fast: Division of AIDS, National Institute of Allergy and Infectious Diseases, Solar Building Room 2A03, 6003 Executive Boulevard, Rockville, MD 20892-7620.

Dr. Wright: D-7226 MCN, Vanderbilt University School of Medicine, Nashville, TN 37232.