Comparison of Immunity Generated by Nucleic Acid-, MF59-, and ISCOM-Formulated Human Immunodeficiency Virus Type 1 Vaccines in Rhesus Macaques: Evidence for Viral Clearance

Abstract

The kinetics of T-helper immune responses generated in 16 mature outbred rhesus monkeys (Macaca mulatta) within a 10-month period by three different human immunodeficiency virus type 1 (HIV-1) vaccine strategies were compared. Immune responses to monomeric recombinant gp120_SF2 (rgp120) when the protein was expressed in vivo by DNA immunization or when it was delivered as a subunit protein vaccine formulated either with the MF59 adjuvant or by incorporation into immune-stimulating complexes (ISCOMs) were compared. Virus-neutralizing antibodies (NA) against HIV-1_SF2 reached similar titers in the two rgp120_SF2 protein-immunized groups, but the responses showed different kinetics, while NA were delayed and their levels were low in the DNA-immunized animals. Antigen-specific gamma interferon (IFN-γ) T-helper (type 1-like) responses were detected in the DNA-immunized group, but only after the fourth immunization, and the rgp120/MF59 group generated both IFN-γ and interleukin-4 (IL-4) (type 2-like) responses that appeared after the third immunization. In contrast, rgp120/ISCOM-immunized animals rapidly developed marked IL-2, IFN-γ (type 1-like), and IL-4 responses that peaked after the second immunization. To determine which type of immune responses correlated with protection from infection, all animals were challenged intravenously with 50 50% infective doses of a rhesus cell-propagated, in vivo-titrated stock of a chimeric simian immunodeficiency virus-HIV_SF13 construct. Protection was observed in the two groups receiving the rgp120 subunit vaccines. Half of the animals in the ISCOM group were completely protected from infection. In other subunit vaccinees there was evidence by multiple assays that virus detected at 2 weeks postchallenge was effectively cleared. Early induction of potent type 1- as well as type 2-like T-helper responses induced the most-effective immunity.

The continued spread of the AIDS epidemic and the dramatic increase in the number of new cases, especially in developing countries, illustrate the urgent need for effective human immunodeficiency virus (HIV) vaccines. The development of vaccines for the prevention of AIDS is hampered not only by the variability of the virus but also by the relatively poor immunogenicity of the HIV type 1 (HIV-1) envelope. It is generally agreed that HIV-1 envelope antigens are a necessary component but that alone or in their current form they may be an insufficient part of a prophylactic HIV-1 vaccine. The specific nature of the immune response required to generate HIV-1-specific immunity by vaccination remains undefined. However, data from a large number of different clinical as well as nonhuman primate model studies are accumulating. Current evidence suggests that to generate effective HIV-1-specific immunity, both neutralizing antibodies (NA) as well as cytotoxic T lymphocytes (CTL) are required (8, 19). Whereas studies suggest that NA are primarily important for blocking infection (23), CTL are most likely critical once an infection has been established, as evidenced by the correlation of high-level CTL activity and the containment of virus loads (25, 38, 46). In order to induce and sustain such humoral and cellular effector responses, potent T-helper immune responses must be generated. Studies with macaques have shown that animals with impaired T-helper responses become infected with chimeric simian-human immunodeficiency virus (SHIV) and develop higher virus loads (3). Similarly, in HIV-1-immunized chimpanzees, the most-vigorous T-helper responses correlate with the highest NA titers and the best protection (7). These observations are supported by the fact that individuals in which viremia is controlled and who survive for up to or more than 18 years with normal CD4⁺-T-cell counts have vigorous HIV-1-specific CD4⁺-T-cell (T-helper) responses (45). Understanding the induction and kinetics of such HIV-1-specific T-helper responses as well as the nature of these responses, i.e., type 1 like (gamma interferon [IFNγ] or interleukin-2 [IL-2]) or type 2 like (IL-4), required to establish protective immunity in outbred primates may be of great value in the design of an effective HIV-1 vaccine.

Efforts to solve the problem of poor immunogenicity of HIV-1 envelope antigens have been focused on adjuvant development, with an emphasis on different means of presenting antigens to the immune system. A number of novel adjuvant formulations, such as oil-in-water emulsions (e.g., MF59), have been found to be safe as well as superior to alum, which is widely used in human vaccine preparations (34, 51). Another approach is to incorporate antigen into immune-stimulating complexes (ISCOMs) to form cage-like structures composed of QuilA derivatives, cholesterol, and phospholipids (35, 36). By changing the components and formulation of ISCOMs, either type 1-like or type 2-like T-helper responses were generated in mice (2, 48, 55).

A promising alternative to adjuvant-associated protein subunit vaccines is nucleic acid immunization (44). Using this method, DNA expression vectors can be administered intramuscularly (i.m.) or epidermally by gene gun delivery. Immune responses elicited by DNA-based vaccines include T-helper type 1 (Th1) or type 2 (Th2) responses in mice, with the nature of the response being influenced by the expression vector, antigen, route of administration, interval between immunizations, number of doses, and animal model used (12, 13, 16, 39, 40, 43). DNA vaccines have been shown to protect small animals against experimental infections with viruses such as rotavirus (24), herpes simplex virus type 2 (33), and influenza virus (52). Moreover, this approach has been effective in protecting newborn chimpanzees against hepatitis B virus infection (41).

For the rational design of HIV-1 vaccines, a combined evaluation of immunogenicity and vaccine efficacy in outbred primates is necessary to determine the nature of the protective immunity provided. An important development in AIDS vaccine research is the successful infection of rhesus macaques with SHIVs consisting of a simian immunodeficiency virus (SIV) genetic backbone into which genes have been replaced by their counterparts from HIV-1 (26, 29, 32). The existence of chimeras containing the HIV-1 envelope genes allows the use of rhesus macaques for combined immunogenicity and efficacy evaluation of HIV-1 vaccine candidates which contain envelope antigens (3, 28, 34). Although the majority of SHIVs constructed to date are nonpathogenic, they are all highly infectious (4) and can be utilized to pose proof-of-principle questions similar to those posed by HIV-1 challenges in chimpanzees yet allow analysis in larger groups of primates.

In this study, we evaluated and compared the kinetics of antibody and T-helper responses to envelope antigens following immunization of 16 (four groups of 4) outbred rhesus monkeys (Macaca mulatta) by three different HIV-1 vaccines strategies, i.e., rgp120/MF59, gp120/DNA, and rgp120/ISCOM, and with control preparations. These three vaccines were chosen because of their potential to induce different types of T-helper responses. As a reference point, we utilized the currently available monomeric recombinant gp120 (rgp120) antigen of HIV-1_SF2, which is widely used in clinical trials. The characteristics of the immune responses in each group were assessed after every three or four immunizations and over a period of 10 months. To determine which type of T-helper immune responses best correlated with protection from infection, animals were challenged intravenously 1 month following the last immunization with an in vivo-titered macaque peripheral blood mononuclear cell (PBMC)-propagated stock of SHIV_SF13.

MATERIALS AND METHODS

Animals.

Captive-outbred, 4- to 5-year-old M. mulatta macaques were housed at the Biomedical Primate Research Center, Rijswijk, The Netherlands. Animals were negative for antibodies to SIV, simian T-cell lymphotropic virus type 1, simian type D retrovirus, and herpes B virus. During the course of the study they were checked twice daily for appetite and behavior. Protocols were approved by the institute’s Animal Care and Use Committee according to international ethical and scientific standards and guidelines.

Preparation of immunogens.

The pUCgp120_SF2 construct used for DNA immunization was based on a modification of pCMV6agp120_SF2 which has been previously described (9). pUCgp120 expresses gp120 of HIV-1_SF2 by using the cytomegalovirus promoter-intron A, tissue plasminogen activator signal sequences, and bovine growth hormone termination sequences; the control plasmid expresses an irrelevant antigen, using the same expression vector. Plasmid DNA was isolated by using plasmid purification columns and endotoxin-free buffers (Qiagen, Chatsworth, Calif.). DNA was bound to 2.6-μm-diameter gold particles to a concentration of 2 μg of DNA/mg of gold. Gene gun cartridges were prepared to a final payload of 1.0 μg of DNA bound to 0.5 mg of gold per target site.

HIV-1_SF2 rgp120 was produced in Chinese hamster ovary cells and has been described previously (18). It was formulated to yield 50 μg of rgp120 (rgp120/MF59) in MF59C-0 adjuvant (5% squalene, 0.5% Tween 80, and 0.5% Span 85 in 10 mM sodium citrate) (51). In contrast, 30 μg of rgp120 was formulated in ISCOMs (rgp120/ISCOM) consisting of a mixture of Quillajasaponin fractions QH-A and QH-C (QH703). This formulation was based on the activities of QH-A and QH-C, to result in enhanced type 1 responses but also a type 2 response as shown by enhanced IL-4 production. ISCOMs were prepared from rgp120 after lipidification of the rgp120 with phosphatidylethanolamine essentially as described by Sjölander et al. (48). ISCOM preparations were characterized by negative-staining electron microscopy and analytical 10 to 50% (wt/wt) sucrose gradient centrifugation (18 h at 200,000 × g and 10°C). The sucrose gradients were fractionated into 16 fractions which were analyzed for protein content (14) and for [¹⁴C]phosphatidylethanolamine by liquid scintillation. The ISCOMs were purified away from nonincorporated rgp120, excess lipid, and QH703 by sedimentation through 30% (wt/wt) sucrose for 18 h at 200,000 × g and 10°C and then resuspended in phosphate-buffered saline. Protein and Quillaja saponin content were determined. Aliquots of ISCOM preparations were stored at −70°C until use.

Schedule of immunizations and challenge.

Sixteen rhesus monkeys were divided into four experimental groups of four. The first group received four immunizations with 8 μg of the DNA expression vector pUCgp120_SF2 (gp120/DNA), the second group was given three immunizations with 50 μg of rgp120 in the adjuvant MF59 (rgp120/MF59), the third group was immunized three times with 30 μg of rgp120 incorporated into ISCOMs (rgp120/ISCOM), and the fourth group consisted of two animals which received three immunizations with 50 μg of an irrelevant antigen in the MF59 adjuvant and two animals that were immunized four times with 8 μg of an irrelevant DNA expression vector (controls).

DNA vaccines were administered epidermally four times, at weeks 0, 12, 24, and 36. Protein subunit immunizations were given at weeks 0, 12, and 36. DNA vaccines were delivered into cells of the epidermis by using a PowderJect model XR gene gun (PowderJect Vaccines, Madison, Wis.). Gene gun inoculations were given over the inguinal lymph nodes and in the lower part of the abdomen, just above the inguinal lymph nodes. Each DNA immunization consisted of eight inoculations with a total of 8 μg of DNA. Protein subunit immunizations were given as i.m. bolus injections at a single site in the posterior part of the right leg. Unique animal numbers per group and vaccine are listed together with the immunization schedules in Table 1. Four weeks after the last immunization at week 40, all animals were intravenously inoculated with 50 50% monkey infective doses of in vivo-titrated SHIV_SF13 that was prepared in rhesus PBMC (4).

TABLE 1.

Vaccine groups and their unique animal constituents, doses, and immunization schedules

Group	Animal	Vaccine	Dose	Routea	Immunization at wkb:
					0	12	24	36
gp120/DNA	I038	pUCgp120	4 × 8 μg	e.d.	×	×	×	×
	Q062
	I044
	X007
rgp120/MF59	I046	rgp120 in MF59	3 × 50 μg	i.m.	×	×		×
	T118
	J041
	Z64
rgp120/ISCOM	T122	rgp120 in ISCOMs	3 × 30 μg	i.m.	×	×		×
	L159
	Q048
	X009
DNA controls	Q045	Control plasmid	4 × 8 μg	e.d.	×	×	×	×
	Q054
Subunit controls	EP4	Control in MF59	3 × 50 μg	i.m.	×	×		×
	WK2

^aVaccine was delivered by gene gun immunization at four sites. e.d., epidermal immunization. 

^b× indicates immunization was performed. 

Determination of gp120-specific antibody responses, virus neutralization, and Chiron RIBA.

Enzyme-linked immunosorbent assays designed to measure HIV-1_SF2 gp120-specific antibody titers were performed on sera by a modification of previously described methods (18). Sera were tested at 1:10 or 1:100 dilutions and then at serial three- or fourfold dilutions thereafter. The titers reported are the reciprocals of the serum dilutions that gave half-maximal optical densities. Assays for serum NA activity against the HIV-1_SF2 laboratory strain were performed as described previously (49). Assays for the neutralization of the SHIV_SF13 challenge strain were performed similarly, except that C8166 cells were employed as the target cells for infection, instead of Hut78 cells, and sera were tested at only three dilutions (1:10, 1:50, and 1:250). NA titers are given as the reciprocals of the dilutions at which 50% inhibition of virus infection was observed. Seroconversion to SIV or HIV-1 positivity post-viral challenge was evaluated by the Chiron HIV-1/HIV-2 radioimmunoblot assay (RIBA) as performed at the Chiron Reference Laboratory (Emeryville, Calif.).

Cell-mediated immune responses.

Freshly isolated PBMC from each monkey were assayed for gp120-specific T-cell responses by using rgp120. Enumeration of antigen-specific cytokine (IFN-γ, IL-2, and IL-4)-secreting cells by ELIspot assay was performed as previously reported (54). Briefly, PBMC were stimulated overnight in triplicate with antigen in plates coated with antibodies specific for either IFN-γ, IL-4, or IL-2. After an incubation period of 18 to 24 h, the cells were removed and cytokine-producing cells were enumerated by using a second labeled antibody with the same specificity but recognizing epitopes different from those recognized by the capture antibody. Results were expressed as the frequency of antigen (rgp120)-specific cytokine-secreting cells per 4 × 10⁵ PBMC. Concanavalin A (5 μg/ml) was used as a positive control.

gp120-specific proliferation of PBMC was measured in triplicate with different concentrations of antigen. Proliferation was determined at the end of the stimulation period (72 h) by the incorporation of [³H]thymidine (0.5 μCi per well) during an 18- to 24-h period. The results are presented as mean counts per minute ± the standard deviation for triplicate cultures and expressed as stimulation indexes, i.e., mean counts per minute of antigen/mean counts per minute of medium alone (RPMI) (54).

Measurement of plasma virus loads and detection of proviral DNA.

The plasma virus load was determined by a quantitative competitive reverse transcription-PCR, using plasma from EDTA-treated blood samples. The lower detection limit of this assay is 100 RNA copies/ml (50). Viral RNA was coamplified with a calibrated amount of internal-standard RNA which was added prior to RNA purification to the sample to be analyzed. As the target sequence, a highly conserved 267-bp region in the SIV gag gene was chosen, with the primer and probe regions being homologous to SIVmac, SIVsm, HIV-2, and chimeric SHIVs. The internal standard was based on the same 267-bp target sequence; however, by PCR, the 26-bp probe region was replaced by a rearranged 26-bp sequence. This fragment was cloned into a transcription vector, and in vitro transcripts were synthesized by using T7 RNA polymerase. The RNA was reverse transcribed and amplified within one reaction protocol by rTth DNA polymerase (Perkin-Elmer), using biotinylated primers. The amplification products were alkaline denatured and were hybridized in six fivefold dilutions to a capture probe that was covalently bound to microwells. The products were detected by a streptavidin-horseradish peroxidase-mediated calorimetric reaction. The amplified internal standard was hybridized to a different capture probe in separate microwells. The amount of RNA in the plasma sample was determined by calculating the ratio of the optical densities of the sample well and the corresponding internal-standard well (a quantitative comparison of the wells detecting the amplified sample with the wells detecting the amplified internal standard). To confirm that animals were free of proviral DNA, a nested PCR for two regions of the chimeric SHIV genome (SIV gag and HIV-1 env) was utilized (4, 34). Nested PCR assays for both regions of the proviral genome as well as quantitative virus isolation assays were performed on PBMC samples at 2-week intervals following challenge to determine if true sterilizing immunity was achieved.

Statistical analysis.

Data were calculated as the means ± standard errors of the means and were analyzed by either the Kruskal-Wallis nonparametric or Wilcoxon statistical test, depending on the comparison being made.

RESULTS

Humoral immune responses following immunization.

The patterns of antibody development following sequential immunizations are shown in Fig. 1A. gp120-specific antibodies were first detected in sera from the animals of the rgp120/ISCOM and rgp120/MF59 groups 2 weeks after the first immunization. The highest antibody titers in these animals were measured 2 weeks after the second immunization; both groups had mean titers of >10,000. The antibody titers then gradually declined in the period between the first (week 12) and second (week 36) protein boosts. A boosting effect was again detectable 2 weeks after the final immunization at week 36 in both protein subunit groups. Curiously, titers remained lower than after the first booster immunization and did not exceed 10,000. Although the mean anti-gp120 titer was highest in the rgp120/ISCOM group throughout the immunization period (P < 0.05 at weeks 14, 26, and 36), at the time of challenge the mean titers developed to approximately the same level in both groups (MF59 titer, 1,823; ISCOM titer, 1,640) (Fig. 1A). Among the DNA-immunized animals, one animal (X007) developed a gp120-specific antibody response at week 26 (titer, 200), while very low antibody titers (<50) were measured in three of four animals on the day of challenge (Table 2).

FIG. 1.

(A) Development of HIV-1_SF2-specific antibody titers to gp120_SF2 in sera of vaccinated rhesus monkeys. Recombinant protein subunit vaccines were given three times, at weeks 0, 12, and 36 (closed arrows), while DNA vaccinees received an additional immunization at week 24 (open arrow). The titer of each group represents the mean of values for four monkeys. ——, rgp120/MF59; –––, rgp120/ISCOM; ⋯⋯, gp120/DNA; ———, control animals immunized with control protein in MF59 adjuvant (EP4 and WK2) or with a control plasmid (Q045 and Q054). (B) Development of antibodies capable of neutralizing the HIV-1_SF2 strain. █, rgp120/ISCOM group; ▩, rgp120/MF59 group; ▧, gp120/DNA-immunized animals. Error bars indicate SEMs.

TABLE 2.

Individual humoral immune responses at challenge

Study group or statistical test	Animal	Antibody titer			Infection status
		gp120	HIV-1SF2a	SHIVSF13a
Study groups
gp120/DNA	I038	12	<10	<10	Infected
	Q062	<10	<10	<10	Infected
	I044	35	10	<10	Infected
	X007	34	<10	<10	Infected
rgp120/MF59	I046	857	67	10	Transient infection
	T118	1,976	150	10	Transient infection
	J041	2,426	450	50	Transient infection
	Z64	2,700	412	50	Transient infection
rgp120/ISCOM	T122	491	142	50	Protected
	L159	1,432	±10	50	Transient infection
	Q048	3,227	56	<10	Transient infection
	X009	3,191	±10	<10	Protected
DNA controls	Q045	<10	NDb	ND	Infected
	Q054	<100	ND	ND	Infected
Subunit controls	EP4	<10	ND	ND	Infected
	WK2	<10	ND	ND	Infected
Statistical tests
Kruskal-Wallisc		11.625	8.283	ND
P value		<0.05	<0.02	ND
ISCOM vs MF59
Wilcoxond		7	1	ND
P value		NSe	0.1 > P > 0.05	ND

^aNA titers. 

^bND, not determined. 

^cStatistical analysis to determine whether there was a significant difference between groups. 

^dStatistical analysis to determine whether there was a significant difference between the two protein subunit groups. 

^eNS, not significant. 

In all three vaccine groups, animals developed antibodies that were able to neutralize HIV-1_SF2 (Fig. 1B; Table 2), although the kinetics and mean titers differed among the groups. In general, development of NA followed a trend similar to that observed with the total anti-gp120 antibody titers. NA were measurable in the animals immunized with either the rgp120/MF59 or rgp120/ISCOM vaccine throughout the immunization period, whereas in those immunized with gp120 DNA, low levels of NA were detectable in one of four animals, and only after four immunizations, on the day of challenge. Differences in the kinetics of NA development also existed between the rgp120/ISCOM and rgp120/MF59 groups. Animals in the rgp120/ISCOM group developed their highest mean NA titers after the second immunization, and these titers gradually declined. In contrast, the mean NA titers of the rgp120/MF59 vaccinees peaked later and remained high and relatively constant after the second immunization until the time of challenge.

In a second neutralization assay, day-of-challenge sera were evaluated for the presence of antibodies that were able to neutralize the challenge virus, SHIV_SF13. Six of 16 monkeys, all four of the rgp120/MF59-immunized animals and two of the four in the rgp120/ISCOM group (T122 and L159), developed heterologous-NA titers against the challenge virus, with titers ranging between 10 and 50 (Table 2). It is important to note that the SHIV_SF13 chimeric used for this study was derived from the envelope gene of a biological variant of HIV-1_SF2 isolated from the patient from which the vaccine strain had been obtained but 5 months later (10, 11).

Cell-mediated immunity.

The kinetics of antigen-specific proliferative responses and the induction of cytokine-secreting cells after immunization are illustrated in Fig. 2. Lymphocyte proliferation in response to gp120 was measured (Fig. 2A), as was enumeration of the number of gp120-specific Th1 cytokine (IFN-γ [Fig. 2B] and IL-2 [Fig. 2C])- and Th2 cytokine (IL-4 [Fig. 2D])-secreting cells during the course of immunization. Immune responses elicited with the rgp120/ISCOM vaccine were characterized by a relatively rapid increase of the gp120-specific proliferative response as well as specific type 1-like (IL-2) and type 2-like (IL-4) responses. Those responses reached maximum values 2 weeks after the second immunization but declined thereafter despite a booster immunization at week 36. The IFN-γ-secreting cells increased in number more slowly than the other cytokine-secreting cell populations and reached peak values at week 26. Again, no boosting effect was detectable after the last immunization. The rgp120/MF59-induced immune responses were characterized by the induction of small numbers of IL-2-secreting cells, while IL-4-secreting cells were exclusively detected at week 14. Only relatively low antigen-specific proliferative and IFN-γ responses could be measured, and then only at later time points, in the rgp120/MF59 group. When comparing the rgp120/ISCOM vaccinees with those receiving the MF59 vaccine, significantly higher numbers of cytokine-producing cells could be found in the ISCOM group at both weeks 2 and 14 (IL-2, P < 0.05 and P < 0.05; IL-4, P < 0.05 and 0.1 > P > 0.05, respectively) or at week 2 (proliferation, P < 0.05) or week 14 (IFN-γ, 0.1 > P > 0.05) only. Two weeks after the final DNA immunization, proliferative responses and numbers of antigen-specific IFN-γ-secreting cells in the DNA vaccinees clearly emerged above background values; however, this was primarily due to the immune responses measured in one of the four animals (X007).

FIG. 2.

Cell-mediated immune responses in rhesus macaques after immunization with DNA and recombinant subunit vaccines. Data are plotted as the means of values for four animals ± the SEM per group. Data for control groups are given as means of values for two animals. Samples were obtained and assayed at the start of the study and 2 weeks after each immunization. (A) Lymphocyte proliferation in response to gp120 is expressed as the stimulation index (antigen-induced proliferation/background proliferation). The numbers of gp120-specific-cytokine-producing (pd) cells per 4 × 10⁵ PBMC are expressed for IFN-γ (B), IL-2 (C), and IL-4 (D). ▨, rgp120/MF59 group; █, rgp120/ISCOM group; ▩, gp120/DNA-immunized animals; ▧, control animals immunized with control protein in MF59 adjuvant (EP4 and WK2); □, control plasmids (Q045 and Q054). nd, not determined.

Assessment of vaccine efficacy.

Four weeks after the last immunization (week 40), animals were challenged intravenously with 50 50% monkey infective doses of SHIV_SF13. At regular time points after challenge, blood samples were taken and analyzed for evidence of the challenge virus. Cell-free plasma virus loads were measured by a quantitative reverse transcription-PCR (Fig. 3). The serological status of the animals was assessed by using the Chiron RIBA assay to detect the presence of antibodies reactive to the challenge virus (Fig. 4). None of the animals which received the rgp120/ISCOM vaccine, nor three of the four rgp120/MF59 vaccinees, had evidence of viral RNA in the plasma (Fig. 3). Also, none of the monkeys immunized with either of the rgp120 subunit protein vaccines showed any evidence of seroconversion to positivity for any of the non-gp120 SHIV-specific antigens (SIV/HIV-2 p27 and HIV-1 gp41) (Fig. 4), indicating the absence of persistent virus replication in these animals. In contrast, all control animals and two of the four animals that received the DNA vaccine became persistently infected, as evidenced by their RNA virus loads 6 to 12 weeks postchallenge, and remained plasma virus positive and seropositive until the end of the study. In contrast to the controls, two DNA-immunized animals (I038 and I044) became and remained plasma virus negative after the primary viremic peak that occurred by week 6, suggesting a positive vaccine-induced effect of DNA immunization. In addition, three of the four DNA-immunized animals showed a gp120 antibody response after challenge, in contrast to the control animals, of which only one animal showed anti-gp120 antibodies 12 weeks postchallenge. This may represent a more rapid (anamnestic) immune response due to effective priming with the nucleic acid vaccine. Animal Q062 of the DNA group, which failed to exhibit a gp120 immune response before and even following challenge (Fig. 4), was the only gp120-immunized animal in which the virus load was not controlled (Fig. 3B).

FIG. 3.

Virus load in plasma of individual rhesus macaques. Virus load is expressed as the number of viral RNA genome equivalents (Eq.) per milliliter of plasma. (A) Animals immunized with control protein in MF59 adjuvant (EP4 and WK2) or a control plasmid (Q045 and Q054); (B) gp120/DNA-immunized animals; (C) rgp120/MF59 group; (D) rgp120/ISCOM group.

FIG. 4.

Seroconversion to SIV and HIV-1 antigen positivity of vaccinated rhesus macaques challenged with SHIV_SF13. Serum specimens were analyzed by using the Chiron HIV-1/HIV-2 RIBA. The location of a given antigen on a strip is indicated to the left of the panels. Level I and level II immunoglobulin G bands are internal controls for moderate and strong antibody reactivity, respectively. Reactivity to gp120 reflects antibody responses to the vaccine. Reactivity to HIV-1 gp41 and/or HIV-1/HIV-2 p26 indicates reactivity to SHIV-specific viral antigens. C, day-of-challenge serum; PC, serum from 12 weeks postchallenge.

To determine whether complete protection from infection actually had been achieved in any of the animals, nested PCR assays for both env and gag regions of the chimeric SHIV genome were performed on PBMC DNA at 2-week intervals postchallenge. In contrast to control animals and gp120/DNA-immunized animals, which were all routinely positive for provirus at all subsequent time points, a PCR signal was detected only at 2 weeks postchallenge in some animals in the other two vaccine groups (Table 3). Evidence of a transient proviral infection was found in three of the four gp120/MF59-immunized animals, with the fourth animal of this group having a transient plasma viral RNA signal. Only two gp120/ISCOM-immunized animals had transient DNA PCR signals in the absence of viral RNA in plasma. All subsequent samples after this 2-week time point, plus the absence of antibodies to non-gp120 SHIV antigens 3 months postchallenge (Fig. 4), confirmed the transient nature of these infections and suggested that the animals had cleared provirus-infected cells after challenge.

TABLE 3.

Postchallenge virological follow-up to determine whether sterilizing immunity was achieved

Group	Animal	Results obtained by PCR and QVI at wka:							Infection statusc postchallenge
		2		4		6		12b
		PCR	QVI	PCR	QVI	PCR	QVI
gp120/DNA	I038	+	27	+	3	+	3	+	Infected
	Q062	+	0	+	81	+	9	+	Infected
	I044	+	9	+	9	+	0	+	Infected
	X007	+	0.5	+	27	+	0	+	Infected
rgp120/MF59	I046	+	0	−	0	−	0	−	Transient infection
	T118	+	0	−	0	−	0	−	Transient infection
	J041	+	0	−	0	−	0	−	Transient infection
	Z64	−d	0	−	0	−	0	−	Transient infection
rgp120/ISCOM	T122	−	0	−	0	−	0	−	Protected
	L159	+	0	−	0	−	0	−	Transient
	Q048	+	1	−	0	−	0	−	Transient infection
	X009	−	0	−	0	−	0	−	Protected
DNA controls	Q045	+	9	+	0.5	+	1	+	Infected
	Q054	+	27	+	27	+	3	+	Infected
Subunit protein controls	EP4	+	3	+	9	+	27	+	Infected
	WK2	+	81	+	9	+	1	+	Infected

^aPostchallenge virology was based on PCR of DNA from PBMC, with the results being scored as positive (+) or negative (−), and quantitative virus isolation (QVI), with values being the number of virus-producing cells per 10⁶ PBMC. All PCR results were negative and all QVI values were 0 at week 0. 

^bShown are PCR results for antibodies other than gp120. 

^cProtected, negative for virus by all parameters, including multiple nested PCR analysis of DNA from lymph node biopsy specimens; transient infection, positive at only one time point (week 2), by only one assay, and negative for virus by all parameters at all time points thereafter. 

^dPositive plasma RNA signal detected only at week 2 (see Fig. 3). 

DISCUSSION

In this study, a direct comparison of the protective immune responses elicited by three different vaccines: gp120/DNA, rgp120/MF59, and rgp120/ISCOM was performed. The types of immune responses generated by these three different vaccine strategies were distinctly different. While epidermal gp120/DNA immunization induced a type 1-like T-helper response (albeit weak) and a low-to-undetectable antibody response, the rgp120/MF59 vaccine induced a strong humoral response. In contrast, the rgp120/ISCOM vaccine induced both types of T-helper responses, in addition to a strong humoral response. The degrees of protection induced by each of these vaccine strategies were also clearly different. All rgp120/DNA group animals became infected, while all MF59 vaccinees exhibited evidence of transient infection. In contrast, two animals from the ISCOM group were fully protected while the remaining two animals had evidence of a transient infection that was successfully cleared. This finding is of importance based on current discussions that a vaccine-induced sterilizing immunity for HIV-1 may not be feasible.

While protection from infection was not observed and proviral DNA persisted in the gp120/DNA-immunized group, there were indications of a vaccine-induced effect in two of these vaccinees, in which plasma RNA loads were suppressed below the detection limit after the peak of primary viremia (Fig. 3). In this study, epidermal DNA immunization resulted in a low-level induction of both the humoral and cellular immune responses. Previous i.m. DNA immunizations of chimpanzees induced low levels of NA and provided evidence of protection from challenge with HIV-1_SF2 (5). In addition, induction of antienvelope antibodies, NA, and envelope-specific CTL in rhesus macaques have been observed with a gene gun DNA immunization protocol (30). CTL responses were not measured in this study, but numerous reports have also documented the ability of DNA immunization to elicit effective major histocompatibility complex class I-restricted CTL responses in nonhuman primates (6, 12, 31, 44, 56, 57). The reduction in virus load in two of four of the gp120/DNA-immunized animals might be attributable to CTL responses which cleared infected cells after infection. Indeed, although sterilizing immunity has not been observed in rhesus macaques immunized with SIV or HIV-1 DNA alone, a reduction in virus load and a delay in progression to disease have been found (17, 30), as has partial protection in cynomologous macaques (6). Antigen-specific T-helper responses were not measured in those studies, but the induction of CTL responses is likely. Induction in primates of type 1-like T-helper responses by i.m. delivery of gp120/DNA has been shown by Lekutis et al. (27), and additional support for our findings comes from previous mouse studies using a gene gun immunization protocol (16). More immunizations with DNA may be required to maximize immune responses in some settings (15). In this study, only after the fourth immunization did T-helper and humoral immune responses emerge in monkeys, suggesting that multiple DNA inoculations are required for maturation of HIV-specific immune responses. Modulation of the type of immune response induced by gp120/DNA immunization can also be accomplished by boosting with protein subunits (1, 15). This markedly increased NA titers and facilitated protection of two animals from SHIV_IIIB infection (28). Although SHIV or SIV vaccine protection with DNA immunization alone has been difficult to achieve, priming of type 1-like immune responses by DNA immunization followed by boosts with subunit proteins, virus-like particles, or recombinant viral vaccines to elicit mixed-type immune responses may be a more promising strategy for inducing protective immunity to HIV-1.

The rgp120/MF59 vaccination strategy was successful in preventing plasma viremia in three of four animals, while at only one time point, 4 weeks after challenge, were low viral RNA levels detected in animal Z64 (Fig. 3C). This was the only time point at which evidence of viral infection was found in this animal, which was negative by all other parameters (Table 2). The plasma specimens of the other three animals were negative for viral RNA, but proviral DNA could be detected in PBMC on a single occasion, 2 weeks after challenge. By all other criteria, including serological data, these animals remained negative and are thus classified as transiently infected (Table 2). The best NA responses, which were sustained at high levels immediately prior to challenge, were found in these rgp120/MF59-immunized animals (Fig. 1; Table 2). This is in agreement with other studies in which this adjuvant formulation elicited antibody responses far superior to those resulting from other vaccine formulations (51). Interestingly, T-helper responses induced by this formulation were weak, with gp120-specific IL-4 responses rising relatively early after the second immunization (Fig. 2D) but remaining low compared to those induced by the rgp120/ISCOM vaccine. Based on the strong humoral responses and the weak or absent IFN-γ and IL-2 gp120-specific responses early in the immunization period, we characterized these responses as being more type 2 like in nature. The Th-2 nature of immune responses induced by MF59 has also been described in previous reports of studies in which this adjuvant was used in other systems (47, 53).

Vaccine protection was most effective in rgp120/ISCOM-immunized animals. In none of the animals was viral RNA detected after challenge. Two animals remained free of provirus in mononuclear cells and viral RNA in plasma, while two had a transient proviral infection which was cleared by week 4 (Table 3). The ISCOM-based rgp120 vaccine induced potent and diverse T-helper immune responses. The profile of the cytokine-secreting T cells observed early in the immunization period revealed large numbers of both gp120-specific type 1 (IFN-γ and IL-2) and type 2 (IL-4) cytokine-secreting cells (Fig. 2). In this study, significant numbers of IL-2-secreting antigen-specific T cells were detectable only in ISCOM-immunized animals. Antigens incorporated in ISCOMs can elicit a variety of potent immune responses, both antibody and CTL (37), and have been shown to protect macaques against infection with HIV-2 (42) or SIVmac (21). ISCOMs in general induce stronger type 1-like T-cell responses than currently registered adjuvants. The immunomodulatory activities conferred by classical ISCOM formulations are primarily induced by the Quillaja saponin fractions QH-A and QH-C. ISCOMs made from QH-A have a potent immunomodulatory activity, enhancing antigen-specific proliferation and production of IL-2 and, above all, IFN-γ (2, 55). In mice, QH-C enhanced antibody production and was able to modulate the immunoglobulin G subclass 2a response, in spite of the fact that QH-C ISCOMs induced low levels of IFN-γ. The present ISCOM formulation was prepared from a mixture of QH-A and QH-C (QH703) which is currently being used safely in human trials. This formulation combines the activity of an active type 1 CD4⁺-T-cell response with type 2-like responses, as evidenced by enhanced IL-4 production (48).

In this study, the best immune response to the gp120_SF2 antigen was obtained by incorporating this antigen into ISCOMs. This vaccine induced potent gp120-specific IFN-γ, IL-2, and IL-4 responses, indicating that both type 1- and type 2-like responses were elicited. These findings are in agreement with those of previous studies with ISCOMs in which both CTL as well as NA responses were induced in rhesus macaques and vaccine protection was observed (21, 22). In contrast, potent NA responses were also induced by rgp120/MF59 immunization, but the T-helper responses after the first two immunizations, in particular IFN-γ and IL-2, were weak. Despite the fact that the rgp120/MF59 group had significantly higher levels of NA than the rgp120/ISCOM vaccinees (0.1 > P > 0.05) (Table 2), all animals became transiently infected, suggesting that a strong NA response alone was not sufficient for protection. It is unlikely that vaccine strategies which induce strong humoral responses without inducing potent helper as well as effector CTL responses will be effective in preventing de novo infection by cell-free virus. Our findings support previous observations that potent type 1-like as well as type 2-like T-helper responses are needed to drive multiple effector mechanisms of both arms of the immune system (20, 22, 34). The most-effective prophylactic HIV-1 vaccines may be a combination of approaches with different vectors and/or subunits capable of inducing multiple effector mechanisms against a number of conserved viral antigens. Our studies were initiated with gp120 as a common test antigen, to evaluate the nature of the T-helper immune responses elicited by three different vaccine strategies. Subsequent studies, in similar model systems, are needed to evaluate multivalent and multicomponent HIV-1 vaccine candidates with improved envelope antigens to determine if more-potent type 1 and type 2 CD4⁺-T-cell responses can be induced and if more-rigorous protection from highly pathogenic and diverse challenges can be achieved.