Hypervariable Epitope Constructs Representing Variability in ...

AIDS RESEARCH AND HUMAN RETROVIRUSES Volume 14, Number 9, 1998, pp. 751-760 Mary Ann Liebert, Inc.

Hypervariable Epitope Constructs Representing Variability in Envelope Glycoprotein of SIV Induce a Broad Humoral Immune Response in Rabbits and Rhesus Macaques DEBRA

MEYER,1

DAVID E.

ANDERSON,1

MURRAY B.

GARDNER,2

and

JOSÉ

V.

TORRES1

ABSTRACT

Using synthetic peptides, we developed an approach to account for protein epitope variability. We have prepared, in a single synthesis, a cocktail of peptides we have designated a hypervariable epitope construct (HEC), which collectively represents much of the in vivo variability seen in an epitope. Eight HECs representing the in vivo variability seen throughout the envelope glycoprotein of the simian immunodeficiency virus (SIV) were designed and synthesized. The constructs were collectively conjugated to KLH (HEC-KLH) or recombinant gpl30 (HEC-rgpl30) and used to immunize rabbits and rhesus macaques, respectively. Using sera collected from rabbits immunized with HEC-KLH, we demonstrated that individual components of the immunogen were recognized as antigen in ELISAs, and that the induced antibodies cross-reacted with several strains of SIV as well as with a strain of HIV-2. Following immunization of macaques with HEC-rgpl30 antiviral antibodies were induced. These antibodies were still present 9.5 months after the last boost and were also capable of recognizing several different strains of SIV, including SIVmac239, SIVmac251, and SIVsmH3, as well as a strain of HIV-2 (HIV-2ROD). In addition, the antibodies were also capable of neutralizing SIV viral infectivity in vitro. Peripheral blood lymphocytes (PBLs) from immunized macaques proliferated in response to whole proteins and virus. Finally, sera from monkeys immunized with SIV, rgpl30, and HIV-2 as well as sera from HIV-2-positive humans recognized HECs in ELISAs, demonstrating the relevance of these epitopes in vivo. This approach can be used as an effective method for generating a strong, broadly cross-reactive humoral response against HIV and can serve as an important component of combination vaccines against HIV and AIDS. CD4+ T cells3 and macrophages.4'5 These viruses are also genetically related; SIV shares approximately 50% homology with

INTRODUCTION

The

immune system of higher vertebrates has evolved to control a diversity of invading pathogens. The relative role that humoral and cellular immune responses play in protection and/or elimination of a pathogen depends on the biology of the pathogen and must be considered when designing vaccines.1 HIV exists in both a cell-free state and as an intracellular virus.2 Accordingly, any candidate HIV vaccine should elicit both strong cellular and humoral responses. SIV and HIV-2 infection of macaques serves as a valuable model of HIV infection in humans. Both viruses infect through the CD4 T cell receptor and are infectious and cytopathic to

'Department of Medical Microbiology and nia Davis, Davis, California 95616.

Immunology

and

HIV-1 and 80% homology with HIV-2. In SIV-infected rhesus macaques, as in HIV-infected humans, a predominantly asymptomatic infection is followed by a disease onset, which is characterized by decreasing CD4+ T cells, increasing susceptibility to opportunistic infections, and increasing viral load.67 The potential for using synthetic peptides as components of

protective vaccine or as immunotherapeutic agents is being explored extensively for a variety of viral and bacterial pathogens.8 Peptide immunogens that elicit some level of proa

tection in animal models or in clinical trials have been described for foot-and-mouth disease virus,9 mouse mammary tumor

department of Medical Pathology, School of Medicine, University of Califor751

752

MEYER ET AL.

virus,10 Venezuelan equine encephalitis virus,11 respiratory syncytial virus,12 Semliki Forest virus,13 Sendai virus,14 lymphocytic choriomeningitis virus,15 group A streptococci,16 and Plasmodium falciparum.17'18 Synthetic peptide vaccines have several

attenuated or inactivated viral the inability to revert back to an infectious

advantages relative

vaccines, including

to

form.

Hypervariability of the HIV envelope is a major problem that be overcome in the design of an effective vaccine.19 In this article, we investigate the use of synthetic peptide mixtures to prime the immune system for dealing with hypervariable pathogens. Although neutralization epitopes on the HIV envelope undergo extensive sequence variation, the sequence diversity of these epitopes appears to be limited, following a pattern of conservation and change. Exploiting this limited diversity, we used SIV as a model of HIV and developed a procedure in which peptide synthesis is manipulated to assemble a mixture of all the known sequences of a neutralization epitope.20 The resulting peptide mixture is referred to as a hypervariable epitope construct (HEC). We previously demonstrated the ability of one HEC to overcome variability by inducing the production of antibodies capable of recognizing native antigen.20 Following the success of this one construct representing the variability of an epitope we described (region 414-434 of the envelope of SIVmac), we explored the possibility of using HECs of all the variable regions in the envelope glycoprotein must

Table 1.

Peptide

cocktails

Amino Acid

Sequences

of the

to

induce

a

cross-reactive humoral immune response. We have

synthesized HECs representing the variability found in the envelope glycoprotein of several SIV isolates and have measured

the immune responses induced in immunized animals. Both rabbits and monkeys were immunized with HECs based on five hypervariable regions of the SIV envelope glycoprotein as well as peptides representing highly immunogenic but relatively conserved epitopes. Immunized animals were examined extensively to characterize the strength and diversity of humoral immune response. Sera from immunized rabbits contained antibodies directed at all eight epitopes, indicating that all the epitopes were immunogenic. Sera were then tested for reactivity with divergent strains of whole killed virus. High titers of antibody were present, which reacted with several strains of SIV as well as with a strain of HIV-2. Similar results were found in two monkeys immunized with the HECs and peripheral blood lymphocytes (PBLs) of these macaques proliferated when exposed to whole virus and recombinant gpl30. To confirm the significance of the immune response elicited by immunization of monkeys with the peptide cocktails, serum from monkeys immunized with SIV, HIV-2, and gpl30 was tested for reactivity with the HECs and found to contain high-titer antibodies that bound to the HECs. Finally, sera obtained from the HECimmunized monkeys neutralized SIV infectivity in vitro. This synthetic construct approach is an effective method for generating a strong, broadly reactive humoral response against HIV, Eight Hypervariable Epitope Constructs*1

Sequences

aa

t i t t aapt sapvseki dmvnet ssc STTTASTTTTTASPTDDMVNETSSCI a L TP AP AEAAIVMVNETSECITH I T S TKKKIDMVN SSC st

HEC 1

HEC 2

PVS VSEVNE kft mt gl kt dkt keynet wyst d KFNMTGLKRDKKKEYNETWYSTD

HEC 3

qrpkerhrrnyvp VEDRNTTNQKPKEQHKRNYVP

T

171-193

HEC 5

HEC 6

HEC 7

EVHKRNY dvkr yt t ggtsrnk TTI GLAPTGVKRYTTGGTSRNK D N cht t vpwpnasl t pdwnndt wqewe CHTTVP WP NETLTP KWDNMTWQE V A S V N N E D N I D S i a n i d w t d g n q t si t m I A N I D W I D G NQTNI T M T N E rpgrnkt vi pvt i msgl vfhsqpi ndrpkqa wc RPGRNKTVLPVTI MSGLVFHSQPI NDRPQAAWC a

HEC 8

413-433

P

_DLNRTTQ HEC 4

A

t

1 e 1 g d y k 1 v ei L E L G D Y K L V El

• Represented

t

pi gl apt

k d q a q 1 n a w g c a f K D Q A Q L N A WG C A F S

i

A I

here

are

in

I

_

vedrdvt

positions

SIVmac251 127-153

the amino acids added at each

q

v

Q

V

cycle of synthesis

494-526

619-643

469-484

314-346

597-618

of the

eight peptide mixtures.

Lower-case letters

represent the sequence of the SIVmac 251 envelope glycoprotein; upper-case letters represent HEC sequences.

753

HECs INDUCE BROAD HUMORAL RESPONSE and can be used in combination with recombinant vectors aimed at eliciting a strong cellular immune response against HIV.

moved, the ether was evaporated using nitrogen gas, and the peptides were resuspended in water and dialyzed. After dialysis, the peptides were lyophilized and stored at room temperature in a desiccator under vacuum.

MATERIALS AND METHODS

Conjugation

Hypervariable epitope

construct

concept and

design

The design and synthesis of one of these constructs was described previously.20 Briefly, protein sequences of 70 different isolates of SIV were aligned. Sequence information was obtained from several databases and publications including GenBank and SwissProt, as well as the Los Alamos Human Retroviruses and AIDS Database. All the variable regions were compared and five HECs were designed to account for hypervariability. Three more HECs based on highly immunogenic epitopes were also prepared. Subsequently, several amino acid coupling steps in the synthesis of an epitope were performed with a mixture of the amino acids occurring at that position as determined from sequence comparisons. This method of synthesis produces additional random permutations in growing peptide chains. Therefore, in a single synthesis, one HEC consisting of a mixture of peptides representing the observed variants of an epitope was produced. The sequences represented by each construct are shown in Table 1. The average size of a peptide in an HEC mixture is 25 amino acids in length. Using a recursive induction formula it was determined that the eight HECs collectively represent 2 X 1020 envelope sequence possibilities.

Peptide synthesis Solid-phase peptide synthesis using 9-fluorenylmethylcarbonyl (fmoc) chemistry and the RaMPs peptide system (Du Pont, Boston, MA) was employed. High-capacity (0.7-1.2 mEq/g) p-methylbenzhydrylamine resin-HCl with 1% crosslinking (Fisher, Pittsburgh, PA) was used for coupling amino acids. We have previously described the synthesis procedure.20 Briefly, the resin was initially neutralized with two additions of 50% (v/v) piperidine-dimethylformamide (DMF) solution (Sigma, St. Louis, MO). It was then washed with DMF and methanol. The appropriate amino acids (Advanced Chemtech, Louisville, KY) were allowed to couple for 2 hr at room temperature. The resin

was

washed four times with DMF and de-

protected with 50% piperidine-DMF for 9 min. Following coupling of the last amino acid, the resin was deprotected as before and washed again with DMF and methanol. The peptides were cleaved and deprotected by the addition

of 90% trifluoroacetic acid (TFA), 5% 1,2-ethanediol, and 5% water solution to the resin. The resin was incubated at room temperature for 14 hr and then washed several times with TFA. All of the peptides were extracted with cold ether. The peptide-TFA solutions were reduced to a volume of 1.0 ml with nitrogen gas. After adding 25 ml of ether, the peptide solutions were mixed and incubated on dry ice for 5 min. Samples were then centrifuged at 1000 X g for 5 min, the ether was removed, and the extraction with ethyl acetate-ether (1.5:1, v/v) on dry ice was repeated three times. Finally, 1.0 ml of water and 25 ml of ether were added to the peptides, which were then incubated on dry ice and centrifuged again. The top layer was re-

to

carrier proteins

All eight HECs were conjugated as a mixture to purified rgp 130 (produced using recombinant vaccinia virus) and to keyhole limpet hemocyanin (KLH) (Sigma Chemical Co.), using a modified carbodiimide method.21 Briefly, both peptides and proteins were dissolved separately at 1 mg/ml in 0.5 M Nmethylimidazole, pH 6.0, and then combined. A 100:1 mol ratio of peptide to carrier was used. To the carrier-peptide solution a 50-mol/peptide concentration of l-ethyl-3-(dimethylaminopropyl)carbodiimide (EDC) was added and the mixture incubated at room temperature for 30 min while being stirred. The immunogen was then purified by extensive dialysis in phosphate-buffered saline and subsequently in double-distilled water. Individual constructs were also conjugated to KLH.

Immunization

of experimental and control animals

Preimmune serum samples were obtained from each animal. New Zealand White rabbits (Grimauds, Stockton, CA) were immunized with the HEC-KLH immunogen and rhesus macaques were immunized with HEC-rgpl30. Monkeys were immunized intramuscularly and rabbits intradermally, using constructs emulsified in monophosphoryl lipid A (MPL) (Ribi Immunochem Research, Hamilton MT) and Freund's complete ad-

juvant (FCA), respectively. Following immunization, sera and lymphocytes were collected for testing. The first immunization of the rabbits was followed by the administration of a booster immunization 21 days later and postimmunization sera were

collected at 2- and 4-week intervals. The first immunization of the rhesus macaques was followed by second and third immunizations 8 and 13 weeks later. Sera were collected 12 days after a fourth immunization and then at 2-week intervals. A fifth immunization was administered 9 months later. Control macaques were inoculated intravenously with 10 MID5o (50% median infective dose) of the respective viral isolates (HIV-2ROD [LAV-2], SIVmac239 molecular clone, SIVmac251 biological isolate). Two other macaques received 100 pg of rgpl30 emulsified in adjuvant. The first three immunizations were administered at 4-week intervals. The monkeys then received two more immunizations 15 and 21 weeks later. Serum samples were collected at 2-week intervals.

ELISA

Testing for anti-peptide antibody titers was performed by the solid-phase enzyme-linked immunosorbent assay (ELISA), using a method previously described.20 Briefly, all antigens (peptide conjugates, proteins, and virus) were dissolved in 0.05 M sodium bicarbonate buffer, pH 9.5, and applied to flat-bottomed microtiter plates (Corning, Corning, NY). Heat-inactivated or sodium dodecyl sulfate-treated viral antigens were plated starting at 500 ng/well and peptide conjugates and/or proteins were plated at 1 /ug/well. After incubation with the test serum, antigen-bound primary antibodies were detected with alkaline phosphatase-labeled secondary antibodies (anti-rabbit or anti-

754

MEYER ET AL.

rhesus) (Fisher). Optical density was measured at 405 nm, using an automatic plate reader (model 3550; Bio-Rad, Hercules, CA). SIVsmH3, HIV-2ROD(LAV-2), SIVmac239, and SIVmac251 antigens were prepared as follows: virus isolates were grown in CEMX174 cells and the supernatant was collected by centrifugation. The supernatant was sterile filtered through 0.4pm pore size cell filters to remove cellular debris and viral particles were then pelleted by ultracentrifugation (113,200 X g for 1 hr). The crude pellet was purified on a 10-50% continuous sucrose

gradient by centrifugation (152,000

X

g).

Serum samples were collected from experimental and control rabbits and macaques. An HIV-2 human serum panel was obtained through the AIDS Research and Reference Reagent Program (Division of AIDS, NIAID, NIH, Bethesda, MD): it is the HIV-2 serum reference panel from S. Osmanov (World Health Organization, Geneva, Switzerland).

Neutralization of viral

infectivity

Viral stocks

were grown in CEMX 174 cells obtained from the American Type Culture Collection (ATCC, Rockville, MD). Cells from this line were used to determine virus titers and also to indicate viral infectivity in the neutralization assay. Serum samples were serially diluted ( 1:16 to 1:2043) and added in triplicate to a 96-well plate. As positive controls, we used the heatinactivated sera of SIV-positive monkeys; as negative controls we used the sera of naive monkeys. Preimmunization sera of all the test animals, including sera from HEC-rgpl30-immunized macaques, were used as additional negative controls in all the assays. SIVmac239 was added at 50 TCID50 (50% tissue culture infective doses) and the plates were incubated for 1 hr. After the incubation, CEMX 174 cells were added to control wells (cells alone) as well as to the virus-antibody wells at a concentration of 1 X 105 cells/well. Plates were incubated at 37°C in a C02 incubator and checked daily for syncytium formation. The neutralizing capabilities of the sera were assessed by testing the reverse transcriptase (RT) activity22 of a portion of the supernatant on day 8 and the cells were exposed to 3(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT; Sigma Chemical Co.) 2 days later to measure their viability. Percentages were calculated as follows: the counts per minute (cpm) indicating RT activity obtained with preimmune sera (i.e., cpm of 300-1000) was considered to be no reduction in RT activity and the reduction in activity caused by postimmune sera (cpm of 50-100) was subsequently expressed as the percentage inhibition or reduction in RT activity. All the data shown here were obtained from experiments that were repeated at least three times, with each data point representing the average of triplicate samples.

T cell

proliferation

Monkey whole blood was collected in heparinized vials. Following centrifugation, the plasma was removed and stored for later use. The remaining cellular pellet was diluted 1:4 with incomplete RPMI (no supplements added). This diluted cell mixture was carefully overlaid onto Ficoll (lymphocyte separation medium from Organon Teknika Corporation, Durham, NC) and was then centrifuged at 1000 rpm for 30 min. The lymphocyte layer was slowly removed from the Ficoll gradient, diluted with

incomplete medium, and washed twice. The pellet was suspended in complete RPMI (supplemented with 10% fetal calf serum [FCS], L-glutamine, and antibiotics) and cell number and viability were determined. The cells were then added at concentrations of 1 X 105 cells/well to 96-well plates containing several dilutions of mitogen (0.1-1 pg/well) or antigen (1-10 /xg/well). Each dilution was tested in triplicate. The mitogen used was phytohemagglutinin (PHA-P; Sigma Chemical Co.), a lectin isolated from the red kidney bean Phaseolus vulgaris. The antigens were SIVmac239, SIVmac251, and rgpl30. The cells were incubated at 37°C with mitogen for 72 hr or with antigen for 120 hr. Eighteen to 24 hr before ending the incubation, 1 pCi of [3H] thymidine was added to each well. Cellular DNA was then harvested onto glass fiber papers, using distilled H20 to lyse the cells. The radioactivity on the glass fiber paper was eluted into scintillation fluid and counted on a scintillation counter. Counts per minute from experimental wells were compared with those of control wells (cells and media only, no antigen or mitogen added) and a stimulation index (SI) was calculated by averaging the replicates and dividing the experimental mean by the control mean. Mitogen Sis were between 20 and 50 and were determined to confirm that assay conditions were suitable.

RESULTS Rabbits were immunized with the eight HECs collectively conjugated to KLH (HEC-KLH) to evaluate the immunogenicity of the epitopes when injected together. Table 1 shows the amino acid composition of each construct as well as its position in SIVmac251 (Los Alamos Database). We prepared cocktails of peptides by adding different proportions of one to four amino acids to be incorporated at specific positions in the growing peptide chain. The frequencies at which particular amino acids occurred at a particular position were determined by sequence comparisons of about 70 viral isolates. For example, at position 1 in HEC 1 we added 52% S, 28% L, and 20% I, representing the frequency at which these amino acids occurred at this position as determined by sequence comparison. These constructs represent a spectrum of the variability seen in vivo with each HEC representing a variable or antigenic region of gpl30 (e.g., HEC 3 corresponds to amino acids 414-434 in the fourth variable region of SIVmac251). HECs 1-5 were based on variable regions of SIV envelope while HECs 6-8 were based on previously reported antigenic regions. Rabbits were immunized with a mixture of the HECs and when tested again individual constructs as antigen in ELISAs, each individual construct elicited a response (Fig. 1). The response to HEC 6 was lower than responses to other HECs. However, HEC 6 is immunogenic as it induced a stronger response in two immunized macaques. The reason for the lower response to HEC 6 might be related to the different epitope recognition patterns of rabbits and macaques. This construct is not based on one of the hypervariable epitopes but instead on an epitope described by others to be highly immunogenic in macaques.23-26 Even though the response to HEC 6 in both rabbits was significantly lower than the response to other HECs, it was still threefold higher than that of the preimmune samples tested at the dilution shown in Fig. 1.

755

HECs INDUCE BROAD HUMORAL RESPONSE 2.52-

ft-

Q

d 0.50-

B

25 -A

¿S^

ff

Q

d 0.5

£^l 1

2

3

4

5

6

7

8

M^ NRP

Hypervariable Epitope Constructs FIG. 1. Antibody response of rabbits immunized with HEC-KLH, as measured by ELISA. (A) Serum from rabbit 1 was obtained 2 weeks after a second immunization with HEC-KLH (two immunizations for this study). (B) Serum from rabbit 2 was collected 4 weeks after a boost (four immunizations in this case). Antibodies present were capable of recognizing the individual HEC components of the immunogen and data shown here are the means of triplicate experiments. As a control we show the response to a nonrelevant protein (NRP). The binding of preimmune sera has been subtracted out as background. The individual HEC constructs used as antigen in this assay were unconjugated. The experiment was repeated three times and the standard deviation was less than 10% of the mean. In a series of dilutions (1:100-1:5000), the titer of the samples indicated here was 1:1000.

B

1000

5000

Serum Dilutions

FIG. 2.

Ability of the antibodies induced by HEC-KLH in immunized rabbits to recognize and bind to rgpl30, SIVsm, SIVmac239, and HIV-2ROD as well as to a mixture of carrier-free HEC peptides (HEC: 1-8). Sera used were obtained approximately 2 weeks (A) and 4 weeks (B) after a second immunization. Serum dilutions of 1:500, 1:1000, 1:5000, and 1:10,000 are shown and the responses of preimmune sera have been subtracted. (A) data for rabbit 1; (B) data for rabbit 2.

MEYER ET AL.

756

Sera

were

tested for the

ability

to

recognize

recombinant

to

humans. Because

an

anticarrier response is

generated when

gpl30 (rgpl30) as well as several strains of SIV and a strain immunizing with any antigen conjugated to a carrier protein, of HIV-2. SIVmac239, SIVmac251, and HIV-2ROD were heat inactivated and/or treated with detergent, while SIVsm never received either treatment. Figure 2 demonstrates the induction of antibodies that reacted with rgpl30 as well as with all strains of SIV tested and with HIV-2. SIV isolates were used without inactivation or following inactivation by sodium dodecyl sulfate (SDS) or heat. The antibodies recognized native protein as well as whole virus and detergent-disrupted virus, indicating that HECs induced antibodies capable of binding to native epitopes. In Fig. 1, the overall response to the HEC antigens appears to be similar for both rabbits; Fig. 2 shows that both rabbits respond to all antigens but at different titers. Overall, the data obtained with the rabbits demonstrated the ability to induce cross-reactive antibodies using HEC antigens. Encouraged by the significant antibody cross-reactivity against whole virus induced by the HEC-KLH construct, we immunized two rhesus macaques with a similar construct to evaluate the humoral response in a species more closely related

decided to use rgpl30 as carrier for the eight HECs. In this way, the anticarrier response to this HEC-rgpl30 construct could prove beneficial to the overall response. Animals immunized with rpgl30 became important controls. Figure 3 indicates that immunization of both monkeys resulted in an antibody response directed against all the individual HECs. This result is consistent with that obtained after immunization of rabbits, except that the response of macaques to HEC 6 was fivefold higher. As was true for the rabbits in Fig. 1, monkeys immunized with the complete HEC mixture recognized individual HEC components of the immunogen (Fig. 3). Binding decreased with higher dilutions of serum, suggesting that antibody binding was specific. Antibodies recognized unconjugated as well as KLH-conjugated individual constructs; the data shown here depict the response to unconjugated constructs. These results also represent the type of response observed when conjugated constructs are used as antigen in ELISAs. A complete immune profile measured over 4 years we

Prebleed

Prebleed

1:100

1:500

1:1000

1:5000

1:1000

1:5000

Serum dilutions FIG. 3. Immunization of macaques produced antibodies that recognized all of the individual HECs as free peptides. Antibodies also recognized KLH-conjugated constructs to the same extent (data not shown). The data shown is an indication of the response we obtained for every date tested and the response shown here was obtained for sera collected 12 days after a fourth immunization. The preimmune sera response obtained at 1:100 dilution is shown. (A) and (B) show the responses of rhesus macaques 1 and 2, respectively. The standard deviation of replicates from one experiment was less than 10% of the mean.

757

HECs INDUCE BROAD HUMORAL RESPONSE showed persistence of high antibody levels lasting 9.5 months, followed by a decline, making a booster immunization necessary (data not shown). In an attempt to demonstrate the in vivo significance of the eight epitopes on which the HECs were based, we compared antibody responses of HEC-gpl30- and rgpl30-immunized animals as well as those of SIV- and HIV-2-infected animals to various antigens (Table 2). The rgpl30 used as antigen in ELISA and proliferation assays, and as immunogen for animals, was based on the SIVmac239 sequence. Other control animals were infected with HIV-2ROD(LAV-2), SIVmac251, and SIVmac239. All infected animals were slow progressors and were still healthy at the time points used in this study. There was a strong humoral response of HEC-gpl30-immunized animals to all antigens tested; it was significantly higher than the response of animals immunized only with rgpl30 but was similar to the response of SIV- and HIV-2-infected animals. At this point, HEC-gpl30 animals had received a total of two immunizations since the start of the project and rgpl30 animals a total of five immunizations, further emphasizing the immunogenicity of the HEC-gpl30 construct. PBLs from HEC-gpl30immunized monkeys (collected 4.4 weeks after a booster immunization) proliferated at positive SI values (2-7) in response to SIVmac251, SIVmac239, and rgpl30, confirming the specificity observed in the ELISA and clearly demonstrating the induction of a helper T cell response. These results are consistent with our earlier results, which indicated that the HECs evoked a strong antiviral humoral immune response. The strong response to rgpl30 by rabbits immunized with HECs conjugated to KLH proved that recognition of gpl30 was due to antibodies elicited by HECs (Fig. 2). In previous projects sera from KLH-immunized rabbits were tested for binding to rgpl30 and found to be negative (data not shown). We conclude that

while an anticarrier response generated to our HEC-rgpl30 immunogen probably contributed to the antiviral humoral immune response of the immunized macaques, it was the presence of the HEC that significantly enhanced the magnitude and breadth of the response. When antibody responses against individual HECs were tested, the SIV-infected and gpl30-immunized monkeys had significant reactivity to four of the eight HECs, while sera from the HIV-2-positive monkey reacted mainly with three of the HECs (Fig. 4). These results demonstrate that at least some of the epitopes chosen for HEC design were recognized in vivo by hosts infected with virus. Because these monkeys were exposed to a single strain of virus, it is possible that infection with different strains might have resulted in recognition of additional epitopes. Furthermore, the antibody response to the various epitopes could change with time such that only some of the epitopes are recognized by the humoral response at a given stage of infection. The humoral response generated by immunization with the HEC-rgpl30 construct was long-lasting and cross-reactive with several strains of whole SIV virus, as indicated in Table 2. Table 2 shows the greatest antibody dilution at which the test sera had a binding value twofold higher than that of the preimmunization sera (as measured by ELISA). A plus symbol (+) in Table 2 indicates that antibody binding was significantly greater than 2.0 at a given dilution. From a complete immune profile of the HEC-gpl30 animals it is clear that every sample collected was able to recognize all antigens shown in Table 2. Furthermore, rgpl30-immunized animals did not respond as well as HECimmunized animals in ELISA or neutralization assays, suggesting that the cross-reactive immune response of HEC-gpl30 animals was not due to rgpl30 alone. Because immunizations of rabbits with an HEC-KLH con-

Table 2. Antibody Responses of HEC-gpl30- and rgpl30-lMMUNizED Rhesus Macaques, as Well as Those of SIV- and HIV-2-Infected Macaques, to Various Antigens8

ELISA

Animals immunized with:

HEC-gpl30 1 2

gpl30 1 2

SIVmac239 SIVmac251 HIV-2ROD

Neutralization

Sample time

(after a boost)

HECmix

rgpl30

SIVmac239

SIVmac251

10,000+ 10,000+

5,000+ 10,000+

10,000 10,000+

10,000 5,000+

5 weeks 3 weeks