Vaccine 20 (2002) 2263–2277

Safety, tolerability and immunogenicity of new formulations of thePlasmodium falciparum malaria peptide vaccine SPf66

combined with the immunological adjuvant QS-21

Oscar Kashalaa,∗, Roberto Amadorb, Maria V. Valerob, Alberto Morenob,Arnoldo Barbosaa,b,1, Beatrice Nickelc, Claudia A. Daubenbergerc,

Fanny Guzmanb, Gerd Pluschkec, Manuel. E. Patarroyoba Aquila Biopharmaceuticals Inc., 175 Crossing Boulevard, Framingham, MA 01702, USA

b Fundacion de Instituto de Inmunologia, Bogota, Colombiac Swiss Tropical Institute, Socinstrasse 57-59, CH 4002 Basel, Switzerland

Received 1 March 2001; received in revised form 20 December 2001; accepted 31 January 2002

Abstract

SPf66 is a synthetic malaria peptide vaccine, which has been widely tested in combination with aluminium hydroxide (alum) as theadjuvant. Since this formulation is weakly immunogenic, we sought to improve its immunogenicity by using the saponin adjuvant QS-21.SPf66/QS-21 vaccines were evaluated for safety, tolerability and immunogenicity in healthy adults. The vaccines were found to be safe in87/89 (97.8%) volunteers studied. However, two individuals developed severe vaccine allergy following the third dose of 1/3 SPf66/QS-21formulations tested. Vaccine formulations containing QS-21 induced a 45- to over 200-fold increase in anti-SPf66 IgG titres over the alumformulation after the second and third doses, respectively. Anti-SPf66 antibody from some subjects reacted against asexual blood stageparasites, as demonstrated by immunofluorescence and immunoblotting. Antibody responses generated by the QS-21 formulations were oflonger duration compared to those evoked by the alum formulation. While SPf66/alum has been found to induce only CD4+ T cell response,the QS-21 formulations exhibited the potential to also elicit SPf66-specific CD8+ responses. These observations demonstrate that the use ofQS-21 can substantially enhance the immunogenicity of peptide vaccines, such as SPf66. © 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Malaria peptide vaccine; QS-21; Clinical trial

1. Introduction

The search for a malaria vaccine, especially one againstPlasmodium falciparum, has been fuelled in recent years,owing to increased resistance of the parasite to anti-malarialdrugs, and that of the vectors to most insecticides [1]. Thesefactors have reinforced the view that a vaccine againstP. falciparum is an urgently required tool for the preven-tion and control of malaria world-wide and especially insub-Saharan Africa [2].

The malaria vaccine SPf66 was developed by Patarroyoet al. [3,4]. SPf66 consists of a polymeric synthetic pep-

∗ Corresponding author. Present address: EMD Pharmaceuticals Inc.,P.O. Box 21339, Research Triangle Park, NC 27709, USA.Tel.: +1-919-401-7154; fax:+1-919-401-7166.

E-mail address: oscar.kashala@emdpharmaceuticals.com (O. Kashala).1 Present address: Department of Immunology, Walter-Reed Institute of

Research, Washington, DC, USA.

tide incorporating amino acid sequences derived from threeproteins isolated fromP. falciparum infected erythrocytes.The peptide epitopes are linked by Pro-Asn-Ala-Asn-Pro(PNANP) sequences from theP. falciparum circumsporo-zoite protein (CSP) repeat unit, and all four peptide se-quences are assembled into a 45-amino acid long monomerunit. The monomer has cysteine residues added at the C- andN-terminal ends to allow for polymerisation. Current SPf66vaccine formulations use aluminium hydroxide (alum) asthe adjuvant.

Several malaria vaccines are under clinical development,and a few have reached advanced evaluation in field trialsin humans [2]. The SPf66 vaccine is the only candidate tohave been evaluated extensively for safety, immunogenicityand efficacy in several countries, including Colombia [5–7],Ecuador [8], Venezuela [9], Tanzania [10–12], The Gam-bia [13], Thailand [14] and Brazil [15]. Results from allthese trials have shown the alum-formulated vaccine to besafe in malaria-naıve adults [10,16], and semi-immune and

0264-410X/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved.PII: S0264-410X(02)00115-9

2264 O. Kashala et al. / Vaccine 20 (2002) 2263–2277

immune children and adults [5–14]. Efficacy evaluation ofthe vaccine has nevertheless led to mixed results, demon-strating some effects againstP. falciparum malaria in someof the major trials [17], while producing no significant ef-fect in Gambian and Tanzanian children under 1 year of age[12,13] and in older children from northern Thailand [14].Interestingly, a reduction in multiple infections was foundin asymptomatic vaccine recipients compared with those inasymptomatic placebo recipients [18]. This observation wascorroborated in Gambian children [19].

One commonality to all SPf66 vaccine trials has been theobservation that the alum-adsorbed SPf66 vaccine is poorlyimmunogenic in humans. For instance, induction of mea-surable levels of antibody against SPf66 often requires athree-dose regimen with a large antigen dose of 2 mg for eachadult immunisation. Moreover, the antibody levels evoked byvaccination are generally of short duration [16], returning tobaseline levels within 15–24 months following the immuni-sation [11,14,16]. Also, SPf66-specific T cell responses tendto be very low [20]. There are a number of reasons that couldexplain the mixed efficacy results observed with SPf66/alumvaccines. However, given the observation that modest im-mune responses to SPf66 resulted in promising, althoughpartial efficacy in some studies, we hypothesised that im-proving the immune responses to SPf66 could result in im-proved and more consistent immune protection. This hasprovided us with an incentive for replacing alum with a morepotent adjuvant, QS-21 and analysing the immunogenicity ofvarious formulations of QS-21-adjuvanted SPf66 vaccines.

QS-21 has been evaluated for adjuvant effects whencombined with a variety of antigens, and has been shown toenhance immune responses to subunit vaccines in laboratoryanimals [21] and humans [22–26]. In human challenge stud-ies of theP. falciparum CSP vaccine RTS,S [24], the addi-tion of QS-21 to a vaccine formulation markedly improvedthe vaccine immunogenicity, and enhanced the vaccine ef-ficacy, resulting in 50–80% protection against experimentalchallenge withP. falciparum sporozoites [24,27]. QS-21has also been shown to enhance the antibody responsesagainst aP. falciparum CSP-derived multiple antigen pep-tide (MAP) vaccine in healthy adults [28]. The additionof QS-21 to alum-precipitated MAP vaccines inducedhigher levels of IgG anti-MAP antibody compared to thealum-bound preparation. Our recent experiments inAotusnancymai monkeys immunised with SPf66 vaccines formu-lated with either alum or QS-21 showed a higher degreeof protection in animals administered QS-21-containingvaccines than those immunised with alum-bound peptide(unpublished results), suggesting that an SPf66 vaccine for-mulated with QS-21 may be highly desirable for effectiveimmunisation of individuals living in areas where malaria isendemic. These observations provided an incentive for theassessment of the safety, tolerability and immunogenicityof QS-21-adjuvanted SPf66 vaccines. We report here theresults of the first phase I clinical trial of the SPf66/QS-21formulations in healthy adults from Colombia.

2. Materials and methods

2.1. Volunteers

A total of 120 male volunteers aged 18–23 years at en-try were screened for eligibility, and 99 evaluable subjectswere enrolled based on inclusion and exclusion criteriaspecified in Section 2.2. All the subjects were recruitedby non-coercive means under a protocol approved bythe Colombian armed services institutional review board(IRB).

2.2. Study design

The study was designed as a randomised, double-blinded,placebo-controlled phase I trial of the safety, tolerability andimmunogenicity of a 2 mg dose of SPf66 formulated witheither alum, (standard vaccine or group A), 50�g of QS-21(group B), 100�g of QS-21 (group C) or a combination ofalum and 100�g QS-21 (group D). Subjects in the placebogroup were given 50�g of QS-21 alone (group E).

The trial’s main objectives were to determine whetherthree subcutaneous doses of SPf66 vaccines were well toler-ated, and whether the addition of QS-21 to the formulationswould result in induction of anti-SPf66 IgG titres at leasttwice as high as those seen among recipients of the alumformulation. Safety and tolerability were judged by the ab-sence of significant haematological and biochemical abnor-malities, and lack of serious adverse effects.

Participants were recruited from the Colombian navy per-sonnel in Tumaco by non-coercive means under interna-tional scientific and ethical standards embodied in the Dec-laration of Helsinki [29], and International Conference onHarmonisation guidelines on technical requirements for theregistration of pharmaceuticals for human use [30]. Studysubjects were provided verbal explanations on the goals, de-sign, risks and conduct of the trial. Participation in the studywas conditional upon voluntary acceptance and signing ofan informed consent form.

Subjects were excluded if they had a history of allergicreactions to vaccination, vaccination with live virus within4 weeks of the beginning of the trial; had undergonesplenectomy; or had a history of malaria, malaria para-sitemia, or malaria vaccination, atopy, immunosuppressivedrug therapy such as steroids; abnormal haematology and/orblood chemistry values. Volunteers were also excluded ifthey had positive markers of active auto-immunity; hadcardiovascular, renal, and hepatic dysfunction or any othermedical or psychiatric condition that in the opinion of theprotocol chairman would compromise the patient’s abilityto tolerate the treatment. Furthermore, they were excludedif they were unwilling to give informed consent, or werewilling to withdraw from the study. The study protocol wasapproved by the appropriate IRB. The trial was conductedby Instituto de Inmunologia, Bogota and monitored by localand World Health Organisation-appointed clinical monitors.

O. Kashala et al. / Vaccine 20 (2002) 2263–2277 2265

Good laboratory procedures and clinical-grade materialswere used throughout the study.

2.3. Vaccine formulations and immunisation

All the vaccines used in this study were based onclinical-grade materials produced under good manufactur-ing practices (GMPs) in Colombia (SPf66) and US (QS-21),and contained a fixed dose of SPf66 mixed with either alu-minium hydroxide, 50 or 100�g of QS-21 or a combinationof alum and 100�g of QS-21.

SPf66 peptide (lot 15.5) was synthesised at Instituto deInmunologia, Bogota, Colombia using methods reportedpreviously [4]. The finished product was characterisedchemically by size exclusion chromatography, mass spec-trometry, sodium dodecyl sulphate (SDS) polyacrylamidegel electrophoresis, and immunologically by the analysis ofthe antibody response against SPf66 in mice. Aluminiumhydroxide (alum) was purchased from Superfos Biosector(Vedvaek, Denmark). QS-21 vials (lot B6831, 0.5 mg/ml,0.6 ml, phosphate-buffered saline (PBS), pH 6.8) were sup-plied by Aquila Biopharmaceuticals, Framingham, MA,USA.

The standard alum-bound SPf66 vaccine was suppliedin ready-to-use two-dose sterile vials (4 mg/ml on alum;1 ml total volume). SPf66 vaccines containing QS-21 wereprepared at the clinical site on the day of the immunisa-tions. The vaccine for the SPf66/QS-21 groups (groups Band C) and the SPf66/alum/QS-21 group (group D) wereformulated from bottled SPf66 stocks (4 mg/ml in saline;1 ml total volume), and ready-to-use SPf66/alum vials(4 mg/ml; 1 ml total volume), respectively. Appropriate vol-umes of SPf66, SPf66/alum and QS-21 were removed fromthe respective vials, and were added to empty sterile 2 mlborosilicate vials to prepare a two-dose formulation (1.4 mltotal volume) for each study subject in group B–D. A 0.5 mland 0.7 ml of vaccine formulations were withdrawn intoindividual tuberculin syringes equipped with hypodermicneedles, and were used for subjects in group A and groupsB and C, respectively. Each vaccine dose consisted of 2 mgof SPf66 mixed with either 1.25 mg of alum (group A),50�g of QS-21 (group B), 100�g of QS-21 (group C),and 1.25 mg of alum mixed with 100�g of QS-21 (groupD). The formulations were injected subcutaneously intoalternate deltoid areas during dosings one–three.

Subjects in group E were given 50�g of QS-21 alone.These formulations were prepared from ready-to-use sterileQS-21 vials (0.5 mg/ml, 0.6 ml, pH 6.8) and sterile physio-logic saline diluent. The appropriate volumes of QS-21 andsaline were added to empty borosilicate vials to make twodoses of QS-21 in 1.4 ml solution. A volume of 0.7 ml wasdrawn into a syringe, and was injected in the same way asthe vaccines.

Three vaccinations were given on Days 0, 30 and 180as has been done routinely for SPf66 vaccine studies [7].Following each dose, all subjects remained at the clinic for

a 60 min period for the assessment of local and systemicreactions, and were again examined at 24, 48 and 72 h, andon Day 20 post vaccination.

2.4. Safety assessment

Assessment of safety included evaluation of the reacto-genicity, tolerability and toxicity. Reactogenicity was as-sessed by observing the incidence and severity of local andsystemic reactions within 1, 24–48, 72 h and 20 days follow-ing the injections. Severity was rated as none, mild, moderateor severe for measurable outcome. A safety pattern for eachformulation was determined by comparing the frequency andseverity of the reactions between the various formulations,focusing on the comparative assessment of local reactogenic-ity between the standard vaccine, and the QS-21-containingvaccines. Assessment of local reactions included erythema,tenderness/induration, pain, local pruritus, arm motion lim-itation related to pain and presence of local adenopathy.Erythema, tenderness/induration were scored as none, mild(<30 mm in the largest diameter), moderate (30–120 mm inthe largest diameter) and severe (>120 mm in the largest di-ameter). The limitation of the range of motion was consid-ered a surrogate measurement of local pain, and was usedto determine the severity of the local pain. Pain was scoredas none or mild, if the range of motion was unchanged, andthe subject could raise his arm to greater than a 90◦ an-gle, moderate if the subject could raise his arm to no morethan 45–90◦ angle, and severe if the subject could not raisehis/her arm to a 45◦ angle. Systemic reactions were alsorated, including allergic reactions such as rash/hives, gen-eralised pruritus, bronchospasm, with or without hypoten-sion, fever (≶38◦C), anorexia, vomiting, abdominal pain,malaise/dizziness, chills, diarrhoea, headache, hypotension,cyst/nodules, abscess, necrosis, axillary adenopathy, myal-gia, arthralgia and arthritis.

Tolerability was defined as the effect of the injectionson the subject’s basic daily life activities such as eating,walking, lifting objects, etc. Tolerability data were en-tered at four category levels, and was rated as excellent ifthere was no reaction and the subject had normal phys-ical activity; good if there were some reactions, but thesubject still had normal physical activity; acceptable ifthere were some reactions, and the subject had reducedphysical activity; and finally minimal if there were somereactions, and the subject had severely impaired physicalactivity.

To monitor toxicity, the blood specimens and serawere obtained from each subject at baseline, 48 h and20 days post vaccination for haematology and bloodchemistry. The following laboratory parameters wereanalysed: white blood counts (WBCs) with differentials,platelet counts, haemoglobin, blood urea nitrogen, crea-tinine, total protein, albumin, bilirubin (total, direct andindirect), alanine aminotransferase, aspartate aminotrans-ferase.

2266 O. Kashala et al. / Vaccine 20 (2002) 2263–2277

2.5. Immunological analyses

Serum specimens for immunological analyses were col-lected at baseline, on Day 20 after each dose, at approxi-mately 4 months after the second dose (Day 160 of the study)and 4 months after the third dose (Day 300 of the study) formeasurement of vaccine-induced immune responses. Serumsamples obtained at baseline were also analysed in the con-text of a routine screening for non-malarial infections. Sero-logical assays were performed for the detection of anti-SPf66antibody as well as antibody against human immunodefi-ciency virus (HIV), hepatitis B surface antigen, hepatitis Cvirus, Chagas and syphilis (special consent forms and po-tential counselling were included). HIV, HBsAg, and HCVantibodies were detected by enzyme-linked immunosorbentassays (ELISA). Chagas antibodies were detected by indi-rect immunofluorescence assays (IFA). Appropriate positiveand negative control sera were included in each assay.

Anti-SPf66 IgG, and IgG directed against the 83.1, 55.1and 35.1 peptide building blocks of SPf66 were detected byFalcon assays screening tests (FAST-ELISA, Becton Dick-inson, NJ, USA) using synthetic polymeric peptides as re-ported elsewhere [31]. Titres of SPf66-specific IgG weremeasured against the SPf66 lot 15.5 used for vaccinationin this study. Titres were also measured against the refer-ence SPf66 peptide lot 10.4. The antibody reactivity againstthese two lots was comparable. Therefore, only the reactiv-ity against the peptide lot 15.5 is reported here. IFA wasused to detect the antibodies reactive againstP. falciparumasexual blood stage parasites of the FCB-2 strain, and theassay was run as described previously [7]. A subset of 40sera was randomly selected and analysed by Western blotfor reactivity against parasite proteins using FCB-2 lysatesas described [31]. All the immunological analyses were con-ducted on samples obtained at baseline, on Day 20 aftereach dose and at approximately 4 months after the secondand third doses, respectively.

For a detailed analysis of cellular immune responses,blood specimens were obtained from four subjects (subjects21, 29, 34 and 89) 4 months after the third immunisation.These subjects were from group B (SPf66/50�g of QS-21),and were selected based on high IgG titres in the pres-ence (subjects 21 and 34) or absence (subjects 29 and 89)of allergic reactions. Peripheral blood mononuclear cells(PBMCs) were isolated from heparinized peripheral bloodby Ficoll–Hypaque gradient centrifugation and cryopre-served. Limiting dilution microcultures were set up fromcryopreserved PBMC in 96-well round-bottomed plates indoubling dilutions starting from 3×104 cells per well in 36replicate wells with or without SPf66 (20�g/ml). Culturemedium consisted of RPMI 1640 supplemented with 10%heat-inactivated pooled human AB serum, 2 mM glutamine,10 mM HEPES, and 100 U penicillin/streptomycin. Lim-iting dilution cultures were incubated for 7 days at 37◦Cin a humidified atmosphere containing 5% CO2. On Day7, proliferation of individual cultures was determined by

addition of 1�Ci per well [3H]thymidine (Amersham) foran additional 16 h. The amount of radioisotope incorporatedinto cultures was determined by scintillation counting. Thefrequencies of proliferating cells were calculated accordingto the Poisson distribution relationship between respondingcells seeded per well and the fraction of non-respondingwells. Individual wells were considered positive only iftheir counts per minute exceeded the mean of 36 con-trol wells containing PBMC without antigen by at leastthree S.D. Frequencies withP < 0.05 were accepted asaccurate.

T cell lines reactive to SPf66 were established from sub-jects 21, 29, 34 and 89 as follows: PBMC were adjustedto 1 × 106/ml in RPMI 1640 supplemented with 10%heat-inactivated pooled human AB serum, 2 mM glutamine,10 mM HEPES, 100 U/ml penicillin, 100�g/ml strepto-mycin (Gibco/Life Technologies, Rockville, MD 20850).Cells were cultured in 24-well plates in the presence of20�g/ml of SPf66 for 8 days at 37◦C, 5% CO2. Blastoidcells were recovered and re-stimulated in complete mediumat 1× 106 cells per well in 24-well plates in the presenceof 20�g/ml SPf66 and 1× 106 irradiated (30 Gy) autolo-gous PBMC for additional 8 days. On Day 2, interleukin-2(IL-2) was added at a final concentration of 20 IU IL-2/ml.Antigen specificity of the resulting cell lines was assessedin a proliferation assay. A total of 2× 104 blastoid cellswere co-cultured with 2× 104 irradiated autologous PBMCwith 20�g/ml SPf66 for 3 days in 96-well round-bottomplates. For the last 16 h 1�Ci per well of [3H]thymidinewas added and incorporated radioactivity was measuredwith scintillation counting. Mean cpm values of triplicatecultures were used for the data analysis.

Surface phenotype of cell lines was tested by FAC-Scan analysis using phycoerythrin (PE)-labelled anti-CD3and fluorescein-isothiocyanate (FITC)-labelled anti-CD4 oranti-CD8 monoclonal antibodies (BD Pharmingen, FranklinLakes, NJ 07417).

2.6. Statistical methods

All data were entered using Epi Info 6 software. Therecords were entered twice on an IBM compatible computer,and were checked for accuracy and consistency. The com-mand merge of Epi Info was used to compare databasesand correct mistakes. Statistical analyses were performed bymeans of the Stata version 6.0 and Epi Info version 6.02bCDC. Data were summarised by tabulation of the number ofsubjects (n), mean, median, S.D. and range (for continuousvariables), and by tabulation of the number of subjects, andproportion (categorical data) by study groups, by dose, andby time point of collection. Geometric mean titres (GMTs)of antibody raised against the vaccines were rounded to twosignificant figures. For those samples with undetectable an-tibody at 1:100 dilution, a titre of 1 was assigned. Becauseof the reduction of the sample size at various time points, theBarrett’s test of homogeneity was performed to ensure that

O. Kashala et al. / Vaccine 20 (2002) 2263–2277 2267

Table 1Background and demographic characteristics of study subjects

Parameter Study groups

A SPf66/alum (N = 17)

B SPf66/50�gQS-21 (N = 20)

C SPf66/100�gQS-21 (N = 19)

D SPf66/alum 100�gQS-21 (N = 20)

E 50�g QS-21(N = 13)

Age 19 (1) 20 (1) 19 (1) 19 (1) 19 (1)Weight (kg) 62.0 (4.9) 60.4 (6.9) 60.9 (5.6) 61.5 (8.5) 63.2 (8.9)Height (m) 1.73 (0.06) 1.70 (0.06) 1.70 (0.05) 1.72 (0.07) 1.74 (0.07)Hemoglobin (g/dl) 13.85 (1.01) 13.65 (0.96) 12.84 (1.01) 13.59 (1.17) 13.93 (0.67)WBC (× 1000/�l) 8.40 (2.22) 8.29 (2.18) 5.53 (1.66) 8.43 (2.47) 7.95 (1.36)Lymphocytes (%) 36.90 41.47 36.56 34.55 37.49

N: number of subjects. The values inside and outside the parenthesis represent the S.D. and mean, respectively.

the sample sizes did not deviate unacceptably. AP-value<0.05 was considered significant.

3. Results

3.1. Study groups and enrolment

One hundred-twenty subjects were screened for eligibil-ity, and 99 were enrolled. Table 1 shows the study groups,enrolment and baseline characteristics of the study sub-jects. All subjects were males, aged 18–23 years. The meanage for the groups was approximately 19 years. The meanhaemoglobin, WBC and lymphocyte counts ranged from12.84 to 13.93 g/dl; 7.95 to 8.43 × 103/�l, and 34.5 to41.47%, respectively (Table 1). There were no statistical dif-ferences between the mean values of the various groups.Subjects’ accountability and disposition appear in Table 2.Of the 99 enrolled subjects, 89 received the first dose, 85 sub-jects received the second dose, and 42 subjects were admin-istered the third dose. Because of very high antibody titresobserved in most subjects following the second dose of vac-cine, and the potential for greater reactogenicity at the thirddose of vaccine, it was agreed that only half of the subjectswould initially be given the third dose, and that all immunisa-

Table 2Subjects accountability and disposition

Parameter Study groups asn (%)

A SPf66/alum

B SPf66/50�gQS-21

C SPf66/100�gQS-21

D SPf66/alum 50�gQS-21

E 50�gQS-21

All

Total no. of enrolled volunteers 20 20 20 20 19 99No. of subjects who received dose 1 17 20 19 20 13 89No. of subjects who received dose 2 17 18 19 18 13 85No. of subjects who received dose 3 8 11 6 8 9 42No. of subjects discontinued 12 9 14 12 10 57

Reasons for discontinuationNon-compliance with protocol 2 0 1 0 6 9Inter-current illness 1 0 0 0 0 1Did not return to follow-up 0 2 0 2 0 4Treatment discontinueda 9 7 13 10 4 43

a Treatment withdrawn due to occurrence of serious adverse events at the third administration.

tions would be discontinued should two or more serious ad-verse events occur. Two (both from group B (SPf66+50�gof QS-21)) of the 42 subjects administered the third dose de-veloped severe allergic reactions. Therefore, immunisationswere discontinued for the remaining subjects. An additional14 subjects enrolled in the trial were discontinued at varioustime points as shown in Table 2.

3.2. Safety assessment

Assessment of the vaccine safety included local and sys-temic reactogenicity, tolerability and toxicity. No serious ad-verse events were noted after the first two doses as shownin Tables 3 and 4. At the first dose, 30/89 subjects fromall groups (33.7%) experienced mild to moderate local ery-thema. Of these 30 subjects, 25 had pain, and 11 had indura-tion at 24–48 h (Table 4). Induration was slightly more preva-lent among recipients of alum-formulated vaccines than vac-cines formulated with QS-21 alone. Indeed, induration wasobserved in 4/17 (23.5%) subjects immunised with SPf66on alum (group A), and 5/20 (25%) recipients of SPf66mixed with alum and 100�g of QS-21 (group D) versus1/20 (5%) subjects administered SPf66 plus 50�g of QS-21(group B), and 1/19 (5.3%) subjects administered the peptidemixed with 100�g of QS-21 (group C) as shown in Table 4.

2268 O. Kashala et al. / Vaccine 20 (2002) 2263–2277

O. Kashala et al. / Vaccine 20 (2002) 2263–2277 2269

2270 O. Kashala et al. / Vaccine 20 (2002) 2263–2277

There were no marked differences in the rate and severityof adverse events between the various formulations.

At the second dose, there were more reports of mild tomoderate erythema and induration in all groups, except forthe adjuvant alone group (group E) (Table 4). A total of 48/87(55.2%) subjects developed erythema. Of these 48 subjects,26 (54.2%) had induration, and 32 (66.7%) had pain. Again,no marked differences in reactogenicity were noted betweenthe various groups. One subject administered SPf66/100�gof QS-21 (group C) developed severe erythema and indura-tion at 24–48 h. The reaction resolved within 2 days withoutsequelae.

The third dose of QS-21-containing vaccines producedmore local reactions and more severe systemic adverseevents within the first hour following the vaccination thanseen with the previous doses. Sixteen subjects from thevarious groups experienced mild to moderate induration. Ofthese 16 subjects, three (18.7%) had developed erythemawithin the first hour of the immunisation (Table 3). A totalof 5/11 (45.5%) subjects in group B, 1/6 (16.7%) subjectsin group C, 2/8 (25%) subjects in group D, and 1/9 (11.1%)subjects in group E developed contralateral induration. Fourof the five subjects who had a contralateral reaction alsodeveloped mild localised pruritus. More than half of thesubjects administered SPf66 preparations formulated onlywith QS-21 (group B: 6/11 (54.5%) subjects; group C:5/6 (83.3%) subjects) developed some induration at 60 minafter vaccination. In contrast, only 25% of subjects admin-istered the alum-precipitated vaccine with or without QS-21developed such reactions (Table 3).

Other local reactions such as nodules, adenopathy andlimitation of the range of motion were infrequent. Subcu-taneous nodules were detected at the injection site on Day20 following immunisation with alum-precipitated vaccinesonly. Indeed, a total of 13 subjects (four from group A, andnine from group D), had small nodules at the site of injec-tion on Day 20 following the first vaccination. At the sec-ond dose, three subjects each from group A and group Ddeveloped nodules at the injection site. One subject fromgroup A had palpable nodules after the third dose. The nod-ules resolved completely within 72 h in most subjects. Smalladenopathies (<1 cm) were also infrequent, and were de-tected only in subjects administered QS-21-containing vac-cines (groups B–D). These were noted at 24–48 h in onesubject after dose 1, three subjects after dose 2, and one sub-ject after dose 3. The adenopathy resolved within 24–48 hfor most subjects, and waned completely by Day 20-postvaccination.

Systemic symptoms were infrequent after the first twodoses (Tables 3 and 4), and consisted mainly of headacheand myalgia. No systemic adverse event was noted fol-lowing the first vaccination. At the second dose, two re-cipients of vaccines formulated with QS-210, one eachfrom group B (1/18 subjects), and group C (1/19 subjects)experienced low-grade fever with chills with or withoutmyalgia within 24–48 h of the vaccinations (Table 4). The

reactions were short-lived and resolved within 24 h. Therewere more severe systemic reactions noted within the first60 min of the third dose administration than seen withthe first two doses. Five subjects from group B (4/11;36.4%) and group C (1/6; 16.7%) developed generaliseditching, mainly without rash (4/5 subjects). Of the foursubjects from group B who had generalised pruritus, onedeveloped mild bronchospasm, and another one had hy-potension and minor facial angioedema (Table 3). Thesetwo volunteers were treated with intravenous fluids, corti-costeroids and antihistamines. They recovered within 1 h,and returned to their normal activities. Because of the twoallergic events, a decision was made to stop further immu-nisations after the first 42 volunteers had received the thirddose.

Analysis of the vaccine tolerability showed all study for-mulations, with the exception of the SPf66/50�g QS-21preparation (group B), to be well tolerated by all subjects asshown in Table 5. Analysis of the toxicity results obtainedafter each dose showed no significant alteration in haema-tological and serum chemistry parameters.

3.3. Immunogenicity

Sera were obtained at baseline, on Day 20 followingeach immunisation, and at 4 months after the second andthird doses, respectively. The antibody response againstSPf66 has been characterised in previous reports basedon end-point titres, using two-fold increments, as low/noresponse (ELISA titres<100), intermediate (ELISA titre:200–800) or high (ELISA titre=1600) [6]. According tothese criteria, all subjects, but one volunteer each fromgroups A (SPf66/alum) and C (SPf66/100�g QS-21),had no pre-vaccination antibody titre. After dose 1, 4/14(28.6%) subjects from group A seroconverted (intermedi-ate responses), while 8/17 (47.1% ) subjects of group B(SPf66/50�g QS-21) and 10/18 (55.6%) subjects of groupC seroconverted with antibody responses characterised asintermediate or high. Interestingly, the SPf66 vaccine for-mulated with both alum and QS-21 (group D) inducedthe highest seroconversion rate after a single dose (15/17subjects). Nearly all the recipients of QS-21 formulationsseroconverted, and had a high antibody response after dose2 (94.6–100%) and dose 3 (100%). In the alum group, theproportion of high responders after the second (53.4%)and third doses (37.5%) was lower than that observed withQS-21-containing vaccines.

Table 6 shows GMT of IgG against SPf66 in the vari-ous groups. Vaccines formulated with QS-21 (groups B–D)evoked IgG titres that were significantly higher than thoseachieved by the standard alum-precipitated vaccine (groupA) after the second and third doses. GMT of anti-SPf66IgG were 20,400 and 31,300 for formulations containing50 and 100�g of QS-21, respectively, and only 1600 forthe alum-based vaccine after the second dose (P < 0.001).Levels of anti-SPf66 IgG peaked after the second dose for

O. Kashala et al. / Vaccine 20 (2002) 2263–2277 2271

Table 5Tolerability following each dosing in study subjects

Time post dosing Tolerability Study group A Study group B Study group C Study group D Study group E

1 2 3 1 2 3 1 2 3 1 2 3 1 2 3

1 h Excellent 17 16 6 20 17 4 19 18 1 20 17 5 13 13 8Good 1 2 1 5 1 5 1 3 1Acceptable 1Minimal 1

24–48 h Excellent 9 3 1 17 2 4 7 2 1 8 1 0 8 5 3Good 7 12 7 3 15 6 12 15 5 9 14 8 5 7 6Acceptable 1 2 1 1 2 3 3 1Minimal

72 h Excellent 16 17 8 20 18 11 19 19 6 15 16 8 13 13 9Good 1 5 0Acceptable 2Minimal

20 Days Excellent 17 17 8 20 18 11 19 19 6 20 18 8 13 13 9GoodAcceptableMinimal

A: SPf66/alum; B: SPf66/50�g of QS-21; C: SPf66/100�g of QS-21; D: SPf66/alum+ 100�g of QS-21; E: 50�g of QS-21. Excellent: no reaction/normalphysical activities; good: some reaction/normal physical activities; acceptable: some reaction/reduced physical activity; minimal: some reaction/impairedphysical activity. The numbers 1–3 represent the number of doses in each study group.

3/4 vaccine formulations (SPf66/alum, SPf66/50�g QS-21,and SPf66/alum+100�g QS-21), and could not be boostedfurther (Table 6). Moreover, formulations containing QS-21induced long-lasting antibody responses that could be de-tected at high titres in 83.3–94.6% of subjects 4 months af-ter the second dose. Overall, QS-21-containing vaccines in-duced IgG titres that were at least 2 logs higher than thoseevoked by the standard alum vaccine after doses 2 and 3(Figs. 1 and 2).

To determine the specificity of the anti-SPf66 response for3/4 building blocks included in SPf66, ELISA designed todetect the reactivity against the 35.1, 55.1, and the 83.1 pep-tide sequences was performed. Compared to alum, QS-21enhanced significantly the titres of antibody against all threepeptides, and induced a 1- to 2-log increase in anti-peptidetitres after two or three vaccinations. Administration of thethird dose boosted effectively the anti-peptide antibody re-sponse for both the SPf66/alum and the SPf66/100�g QS-21formulations.

Table 6Geometric mean titres of IgG against SPf66 after two (or three) vaccinations with SPf66 in various adjuvants

Studygroup

Pre-immune 20 Days after dose 1(Day 20 of study)

20 Days after dose 2(Day 50 of study)

4 Months after dose 2(Day 160 of study)

Day 200 of study Day 300 of study

A 2 (0) 20 (20) 1600 (530) 240 (20) 150 (530) 20 (190)B 0 (0) 6 (100) 20400 (37760) 8130 (4050) 6460 (25700) 3210 (4390)C 0 (3) 20 (400) 31330 (51290) 3920 (11220) 3220 (44670) 2090 (12890)D 0 (0) 2400 (830) 22390 (38100) 2510 (5100) 1580 (31330) 1100 (5850)E 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)

A: SPf66/alum; B: SPf66/50�g of QS-21; C: SPf66/100�g of QS-21; D: SPf66/alum+ 100�g of QS-21; E: 50�g of QS-21. Numbers in parenthesesindicate GMT after three doses.

In contrast to the antibody response against the immuno-gen, titres of merozoite-reactive antibody as determined byIFA were low in all groups (maximum IFA titre 320). Com-parative analyses of the IFA-reactive antibodies betweengroups showed no differences after doses 1 and 2. However,when considered as a group after the third dose, sera fromsubjects administered SPf66/QS-21 formulations demon-strated more reactivity againstP. falciparum FCB-2 bloodstage parasites than sera from subjects given SPf66/alumformulations. Indeed, 9/24 (37.5%) subjects immunisedwith QS-21-containing formulations (groups B–D) wereIFA-positive, as opposed to only 1/8 (12.5%) subjects giventhe alum-precipitated standard vaccine (group A) (Table 7).

Immunoblot analysis was performed on a subset of 40subjects chosen at random from all the groups, and fo-cused mainly on the demonstration of reactivity againstthe merozoite surface protein-1 (MSP-1). Previous stud-ies using a monoclonal antibody against theP. falciparumMSP-1 protein as a control, have shown reactivity against

2272 O. Kashala et al. / Vaccine 20 (2002) 2263–2277

Fig. 1. Kinetics of anti-SPf66 IgG responses among subjects administered two doses of vaccines or placebo. Subjects were immunised subcutaneouslyon Days 0 and 30 with 2 mg of peptide bound onto alum (�), or mixed with 50�g of QS-21 (�), 100�g of QS-21 (�), or a combination of alum and100�g of QS-21 (�). Subjects in the placebo group (�) were given two doses of formulations containing 50�g of QS-21, and no peptide. Geometricmean titres (GMT) of IgG against SPf66 were determined at baseline, on Day 20 following each vaccination, and 4 months after the second dose.

the 195 kDaP. falciparum MSP-1 protein in some recip-ients of the standard SPf66/alum vaccine [31]. Similarreactivity with the 195 kDaP. falciparum MSP-1 proteinand its cleavage products were observed with sera of someof the recipients of the different vaccine formulations eval-uated here. Specificity for MSP-1 was demonstrated bypre-incubation of sera with a 45-amino acid long syntheticpeptide derived from the 83 kDa fragment of MSP-1, which

Fig. 2. Kinetics of anti-SPf66 IgG responses among subjects administered three doses of vaccines or placebo. Subjects were immunised subcutaneouslyon Days 0, 30 and 180 with 2 mg of peptide bound onto alum (�), or mixed with 50�g of QS-21 (�), 100�g of QS-21 (�), or a combination ofalum and 100�g of QS-21 (�). Subjects in the placebo group (�) were given two doses of formulations containing 50�g of QS-21, and no peptide.GMT of IgG against SPf66 were determined at baseline, on Day 20 following each vaccination, and 4 months after the second and third doses.

completely abrogated the reactivity against the 195 and83 kDa bands (data not shown). The immunoblot reactivityagainst theseP. falciparum protein bands was lower in serafrom subjects given the standard alum vaccine as comparedto recipients of QS-21-containing formulations (data notshown). In addition to the reactivity against MSP-1, stain-ing of additional protein bands was observed with somesera.

O. Kashala et al. / Vaccine 20 (2002) 2263–2277 2273

Table 7P. falciparum FCB-2 IFA-reactive antibody following vaccination with SPf66 in various adjuvants

Time points (20 days after) Study groups

A SPf66/alum B SPf66/50�g QS-21 C SPf66/100�g QS-21 D SPf66/alum 50�g QS-21 E 50�g QS-21

Dose 1 1/16 0/17 1/18 1/17 0/13Dose 2 1/16 0/17 1/18 1/17 0/13Dose 3 1/8 3/11 3/5 3/8 0/8

Table 8Frequencies of SPf66-specific T cells in PBMC of volunteers from group B 4 months after receiving the third dose of SPf66/50�g QS-21

Volunteer ID

21 34 29 89

Reactogenicity Systemic allergic reaction Systemic allergic reaction None Local contralateralFrequency 1/3753 1/6115 <1/20000 1/3038Confidence interval (P > 0.05) 1/2993–1/5029 1/4732–1/8638 1/2507–1/3855

Four subjects from group B (SPf66/50�g QS-21) wereselected for an analysis of the magnitude and nature of thevaccine-induced T cell responses against SPf66. These in-clude the two individuals (subjects 21 and 34) with allergicevents after the third immunisation and two other individualsfrom group B (subjects 29 and 89) who had developed com-paratively high anti-SPf66 IgG responses, but no systemicsymptoms after the third vaccination. PBMC were collected4 months after the third immunisation and the frequencyof SPf66-reactive T cells was determined by limiting dilu-tion analysis. High frequencies (1/3753 and 1/6115) werefound in both the subjects with allergic events (Table 8). Acomparatively high frequency (1/3038) was also observed in1/2 other subjects, while the frequency of SPf66-reactive Tcells was low (<1/20000) in the fourth subject analysed (Ta-ble 8). These precursor frequencies were higher than thoseobserved with the standard SPf66/alum vaccine in a previ-ous trial in Switzerland (unpublished results).

For a preliminary analysis of the nature of the elicitedcellular immune responses, T cell lines were generatedfrom the four study subjects by repeated stimulation withSPf66 in the presence of autologous irradiated PBMC asantigen-presenting cells. A characterisation by flow cytom-etry revealed mixed populations of CD4+ and CD8+ Tcells in all four subjects (Fig. 3). The relative proportionof these T cell subtypes differed from line to line rangingfrom a clear dominance of CD8+ T cells in a line gener-ated from PBMC of one of the subjects (subject 21) withallergic reactions to a dominance of CD4+ T cells in a linefrom the other subject (subject 34) with allergic reactions.Cell lines from the two individuals who had not devel-oped allergic reactions after vaccination exhibited eitheran intermediate CD4+/CD8+ composition (subject 89) ora dominance of CD4+ T cells (subject 29). The findingthat the SPf66/QS-21 formulation has the potential to elicitSPf66-specific CD8+ T cell responses, is in contrast toour finding that vaccination with the standard SPf66/alum

vaccine induced a predominantly CD4+ T cell responseof the Th2 subtype in malaria-naıve Caucasians andsemi-immune individuals from Africa (unpublished results).

4. Discussion

Previously studied SPf66 vaccines were formulated withalum, a generally weak adjuvant that acts primarily byfavouring antigen deposition at the injection site, and bystimulating mainly a CD4-dependent Th-2 response [32].To induce measurable levels of antibody against SPf66,a large amount of antigen (2 mg) and a three-dose regi-men have generally been used with these alum-formulatedvaccines. Clinical trials of SPf66/alum vaccines have es-tablished the safety of the vaccines in children and adults[7,10,12,16]. Most subjects given the alum formulationsdevelop only mild local reactions (pain, erythema, and in-duration) that are self-limited. Unusual local side effectssuch as contralateral inflammation with or without pruritushave been observed in less than 1% of subjects enrolled inthe macroTumaco field trial in Colombia [6], 3/15 subjectsvaccinated with a US-manufactured peptide in the US [16],and approximately 4% of immune children from Tanzania[11] and Thailand [14]. These reactions have generallybeen observed after the administration of the third dose ofvaccine, and very rarely after that of alum alone [10].

Our results confirm these previous observations onSPf66/alum vaccines. The vaccines formulated with QS-21were generally well tolerated by most of the subjects(Tables 3–5) and had a local safety profile comparableto that of the alum formulation, except for a slight in-crease in the incidence of contralateral inflammation afterthe third dose. These reactions occurred mainly amongrecipients of the SPf66/50�gQS-21 vaccine (4/6 vol-unteers). Similar local reactions were reported in trialsof other malaria peptide vaccines (Edelman, personal

2274 O. Kashala et al. / Vaccine 20 (2002) 2263–2277

Fig. 3. Two-colour flow cytometric analysis of T cell lines generated by repeated in vitro stimulation of PBMC with SPf66. Panel A shows cells stainedwith PE- and FITC-labelled IgG isotype control antibodies. Panels B and C demonstrate the results obtained with PE-conjugated anti-CD3 (x-axis) versusFITC-labelled anti-CD4 or anti-CD8 mAb (y-axis), respectively. In addition to the cell lines obtained from the four volunteers 21, 34, 29 and 89 of studygroup B (SPf66/50�g QS-21), typical results with a cell line from a volunteer immunised with SPf66/alum (TS) in a previous trial are shown.

communication), and do appear to be associated with cer-tain malaria parasite epitopes. However, evidence pointsalso to parasite-unrelated mechanisms since recipients ofadjuvant alone formulations in previous SPf66/alum trialshave experienced contralateral inflammation [6,10]. In thisstudy, 1/11 subjects given QS-21 alone also developed con-tralateral inflammation. The mechanism of these reactions

is not yet fully understood. The rapid onset of the reaction(within 5–15 min of the injection) and the frequent associa-tion with pruritus suggest an IgE-mediated immediate typeI hypersensitivity. The finding of CSP-specific IgE in twosubjects with contralateral inflammation and generalisedpruritus following vaccination withP. falciparum CSP MAPvaccine lends support to a possible role for specific IgE

O. Kashala et al. / Vaccine 20 (2002) 2263–2277 2275

(Edelman, personal communication). However, other mech-anisms may be involved, especially when a contralateralreaction is not accompanied by pruritus, or when it occursat the first dose, at the time when specific IgE has not beeninduced. Occasional reports of contralateral reaction occur-ring after administration of adjuvant alone also suggest anantibody-independent mechanism.

In the study reported here, systemic reactions were in-frequent, and the incidence and severity of the reactionswere similar to those reported in previous trials of theSPf66/alum vaccine [6–14], and trials of other synthetic orgenetically-engineered vaccines [33]. However, two recipi-ents of the SPf66/50�g QS-21 vaccine formulation devel-oped severe allergic reactions after the third dose. Therewere no such allergic reactions in subjects administeredSPf66/100�g of QS-21. The reactions were characterisedby mild bronchospasm, generalised pruritus (one subject),facial angioedema with rash and hypotension (one subject).The symptoms resolved very quickly following systemictherapy with epinephrine and anti-allergy medication. Gen-eralised pruritus, bronchospasm and hypotension werealso reported in 0.28% of subjects in another SPf66/alumtrial in Colombia [6]. The reactions associated with thealum-formulations were nevertheless milder than thoseseen in this study. Systemic allergies have been observedin association with other vaccines [34]. Grotto et al. [35]have reviewed the literature on adverse events associatedwith yeast-derived HBV vaccines, and found evidence forlocal and systemic reactions such as anaphylaxis in a fewsubjects. Systemic allergic reactions have also been ob-served following administration of Japanese encephalitisvirus (JEV) vaccines [36]. The causal agents for the re-actions are not yet known for SPf66, but have been welldocumented for both HBV and JEV vaccines [35]. Gelatinewas shown to induce anti-gelatine IgE antibody in six chil-dren who developed systemic allergic reactions, includingone mild anaphylaxis following JEV vaccination in Japan[36]. Preservatives such as thimerosal have been implicatedin vaccine-induced allergies following administration ofHBV vaccines [35]. Both gelatine and thimerosal have notbeen used in association with SPf66 vaccines. Preliminaryanalyses of the cellular immune responses in the two sub-jects, who had developed severe allergy after three doses ofSPf66/50�g QS-21 (group B) revealed no consistent andunique feature, as compared to two other study group Bvolunteers without allergic reactions. However, 4 monthsafter the third dose, frequencies of SPf66-specific T cellswere still remarkably high in 3/4 individuals analysed.Phenotypic characterisation of T cell lines from these fouranalysed subjects suggests that QS-21 can support the stim-ulation both of CD4+ and CD8+ T cells in humans. Incontrast to alum, this adjuvant has previously been shownto have a strong potential to support CD8+ cytotoxic T cellresponses in mice [22].

Overall, the SPf66/alum, SPf66/100�g QS-21, andSPf66/alum/100�g QS-21 vaccine formulations were all

well tolerated whereas the SPf66/50�g QS-21 vaccine wasnot. The reasons for the increased reactogenicity of the50�g QS-21 formulation is still unclear. It is not unrea-sonable to suspect that physicochemical differences (i.e.micellar formation) between SPf66 peptide formulationscontaining 50�g QS-21or 100�g QS-21 may have ac-counted for different toxicity profiles. Assuming greatermicellar formation occurs at the higher QS-21 dose than thelower one, then one could speculate that greater entrapmentof the SPf66 peptide into micelles would be induced, result-ing in lower amount of free SPf66, and less SPf66-inducedtoxicity. Numerous studies advocate the use of micelles toreduce side effects associated with drugs.

A main objective of this trial was to determine whetheradding QS-21 to SPf66 formulations would improve the an-tibody response to SPf66. We measured the anti-SPf66 IgGtitres, and compared the titres between alum and QS-21formulations. Our results show that QS-21 significantly en-hances anti-SPf66 IgG titres, and the enhancement is seenwith all QS-21 formulations, including the vaccine formu-lated with both alum and QS-21. Anti-SPf66 IgG titres wereat least 2 logs higher among recipients of QS-21 formula-tions compared to subjects vaccinated with the alum formu-lation, after the second (GMT of SPf66/alum: 1600; GMT ofSPf66/QS-21: 20,400–31,300), and the third doses (GMT ofSPf66/alum: 530; GMT of SPf66/QS-21: 25,700–31,300).The antibody titres after three doses SPf66/QS-21 were thehighest titres ever seen with SPf66 formulations, and were atleast 2 logs higher than the GMT reported in previous trialsin Colombia (GMT 538) (unpublished results), and Tanza-nia (GMT 2783) [11].

To determine the specificity of the anti-SPf66 responsefor 3/4 SPf66 peptide building blocks, IgG titres againstthe 35.1, 55.1, and the 83.1 peptide sequences were anal-ysed. Again, SPf66/QS-21 induced significantly higheranti-peptide IgG titres than the alum formulation after twoand three doses. IgG titres against the 55.1 and the 83.1peptides were higher than those against the 35.1 peptide.All vaccine formulations elicited little IFA reactivity againstthe P. falciparumFCB-2 blood stages after the first twodoses. After the third dose, 9/24 (37.5%) subjects givenSPf66/QS-21 formulations developed detectable IFA titres.Only 1/8 subjects administered the SPf66/alum formulationhad such antibody reactivity. The low IFA reactivity elicitedby SPf66/alum vaccines has also been observed by Gordonet al. [16], who reported that only 1/15 subject adminis-tered SPf66/alum vaccine manufactured in the US devel-oped anti-P. falciparum asexual blood stage antibodies. Theincidence of IFA-reactive antibody has not been studied ex-tensively in other SPf66 vaccine trials. Thus, evaluation ofSPf66 vaccine performance based on IFA reactivity is diffi-cult. We have performed immunoblot analyses to investigatefurther the vaccine-induced antibody reactivity with parasiteproteins. We found that both alum and QS-21 formulationsinduced antibody reactive against several parasite proteins,including MSP-1. Specificity for MSP-1 was demonstrated

2276 O. Kashala et al. / Vaccine 20 (2002) 2263–2277

by immunoadsorption with a soluble MSP-1 derived pep-tide. There was increased recognition of MSP-1 in subjectsvaccinated with SPf66/QS-21 compared to SPf66/alum re-cipients. These results confirm earlier findings that SPf66can evoke genuine anti-P. falciparum asexual blood stageantibodies in some vaccinated subjects [16], and establishthe superior adjuvanticity of QS-21 over alum.

Another important goal for this trial was to examine thesafety and immunogenicity of the new formulations in orderto select a single formulation for further studies. Compari-son of the immunogenicity of the two QS-21 doses (50 and100�g) show similar effects on antibody response after twodoses (GMT 24,550 versus 25,000, respectively). However,after three doses, the 100�g dose induced a significantlyhigher anti-SPf66 IgG titre (GMT 16,980 versus 38,900,respectively), and a more frequent IFA reactivity than the50�g dose. A further under performance of the 50�gQS-21 was its lower level of tolerability when compared tothe 100�g QS-21 dose. Thus, the SPf66 formulation con-taining 100�g of QS-21 appears to be more suitable for in-duction of humoral responses to SPf66. Concerning the doseregimen, it is important to determine whether two ratherthan three doses of vaccine would be optimal for genera-tion of an effective immune response. Our data suggest thata two-dose regimen would be ideal since high anti-SPf66titres can be evoked with minimum risk of reactogenicity.Studies on antigen dose were not undertaken in this work.Such studies will be critical since they will help determinewhether the immune response induced by two doses ofSPf66 vaccines containing 100�g of QS-21 could be op-timised further by reducing the amount of antigen in thepreparation. Results of recent studies of HIV-1MN gp120vaccines combined with QS-21 in healthy uninfected adultshave shown that reducing the antigen dose from 30–600 to3�g resulted in the generation of the highest neutralisingantibody titre ever seen with a range of adjuvants tested[37]. Indeed, HIV-1-specific proliferative T cell responses,V3 loop-specific and fusion inhibiting antibodies at a 3�gantigen dose were equal or superior to the responses seenwith formulations containing 30–600�g of antigen.

A central question for this trial is whether or not im-provement in anti-SPf66 titres has a predictive value on theefficacy of the new formulation. Although critical, such aquestion can only be answered in a phase III efficacy trialsetting. Knowledge of protective immunity in malaria isstill incomplete. It is generally believed that induction ofmalaria-specific antibodies is critical, but perhaps not suf-ficient for an effective control of human malarial infection[38,39]. An effective malaria vaccine is likely to requirethe induction of antibody and cellular responses againstkey parasite antigen targets. Three peptide sequences de-rived from proteins isolated from infected erythrocytes andone CSP-derived epitope are built into the SPf66 vaccine.Anti-SPf66 antibodies induced by QS-21-containing for-mulations have specificity for all three blood stage antigenepitopes, and may have the potential to interfere with par-

asite entry into erythrocytes [40,41]. Studies inAotus mon-keys vaccinated with SPf66/QS-21 formulations showedthat these vaccines were more effective against experi-mental challenge withP. falciparum blood stage parasitesthan both SPf66/alum and SPf66/Freund’s adjuvant vaccineformulations (unpublished results). This demonstrates thatQS-21 has a great potential as an adjuvant for the deliveryof peptide-based vaccines in humans.

Acknowledgements

We are very grateful to Drs. Gerald Beltz, Cheryl Mur-phy and Charlotte Kensil for helpful discussion. We wouldlike to thank Ms. Annette Wertz for excellent secretarialhelp. Analyses of cellular immune responses were supportedin part by the Swiss National Science Foundation project31-52068.97. This work received financial support from theGovernment of Colombia, and UNDP/World Bank/WHOSpecial Programme for Research and Training in TropicalDiseases.

References

[1] WHO. World Malaria Situation in 1994. Part II. Wkly. Epidemiol.Rec. 1997;72:277–83.

[2] Engers HD, Godal T. Malaria vaccine development: current status.Parasitol Today 1998;14:56–64.

[3] Patarroyo ME, Romero P, Torres ML, Clavijo P, Moreno A, MartinezA, et al. Induction of protective immunity against experimentalinfection with malaria using synthetic peptides. Nature 1987;328:629.

[4] Patarroyo ME, Amador R, Clavijo P, Moreno A, Guzman F,Romero P, et al. A synthetic vaccine protects humans againstchallenge with asexual blood stages ofPlasmodium falciparum.Nature 1988;332:158.

[5] Amador R, Moreno R, Murillo LA, Sierra O, Saavedra D, Rojas M,et al. Safety and immunogenicity of the synthetic peptide malariavaccine SPf66 in a large field trial. J Infect Dis 1992;166:139.

[6] Amador R, Moreno A, Valero MV, Murillo M, Mora AL, Rojas M, etal. The first field trials of the chemically-synthesised malaria vaccineSPf66: safety, immunogenicity and protectivity. Vaccine 1992;10:179.

[7] Valero MV, Amador R, Galindo C, Figueroa J, Bello MS, MurilloLA, et al. Vaccination with SPf66, a chemically synthesisedvaccine, againstPlasmodium falciparum malaria in Colombia. Lancet1993;341:705–10.

[8] Sempértegui F, Estrella B, Moscoso J, Piedrahita CL, HernándezD, Gaybor J, et al. Safety, immunogenicity and protective effect ofthe SPf66 malaria synthetic vaccine againstPlasmodium falciparuminfection in randomised double-blind placebo-controlled field trial inan endemic area of Ecuador. Vaccine 1994;12:337–42.

[9] Noya GO, Gabaldón Berti Y, Alarcón de Noya B, Borges RE, ZerpaN, Urbáez JD, et al. A population-based clinical trial with the SPf66synthetic Plasmodium falciparum Malaria vaccine in Venezuela. JInfect Dis 1994;170:396–402.

[10] Teuscher T, Armstrong Schellenberg JRM, Bastos de Azevedo I, HurtN, Smith T, Alonso P, et al. SPf66, a chemically synthesised subunitmalaria vaccine is safe and immunogenic in Tanzanians exposed tomalaria transmission. Vaccine 1994;12:328.

[11] Alonso P, Smith T, Armstrong Schellenberg JRM, Masanja H,Mwankusye S, et al. Randomised trial of SPf66 vaccine againstPlasmodium falciparum malaria in children in southern Tanzania.Lancet 1994;344:1175.

O. Kashala et al. / Vaccine 20 (2002) 2263–2277 2277

[12] Acosta CJ, Galindo CM, Schellenberg D, Aponte JJ, Kahigwa E,Urassa H, et al. Evaluation of the SPf66 vaccine for malaria controlwhen delivered through the EPI scheme in Tanzania. Trop Med IntHealth 1999;4:368.

[13] D’Allessandro U, Leach A, Drakeley CJ, Bennett S, Olaleye BO,Fegan GW, et al. Efficacy trial of malaria vaccine SPf66 in Gambianinfants. Lancet 1995;346:492.

[14] Nosten F, Luxemburger C, Kyle DE, Ballou WR, Wittes J,Wah E, et al. Randomised double-blind placebo-controlled trial ofSPf66 malaria vaccine in children in northwestern Thailand. Lancet1996;348:701.

[15] Urdaneta M, Prata A, Struchiner CJ, Tosta CE, Tauil P, Boulos M.Evaluation of SPf66 malaria vaccine efficacy in Brazil. Am J TropMed Hyg 1998;58:378–85.

[16] Gordon DM, Duffy PE, Heppner GD, Lyon JA, Williams JS,Scheumann D, et al. Phase I safety and immunogenicity testingof chemical lots of the syntheticPlasmodium falciparum vaccineSPf66 produced under good manufacturing procedure conditions inthe United States. Am J Trop Med Hyg 1996;55:63.

[17] Graves P, Gelband H, Garner P. The SPf66 malaria vaccine: what isthe evidence for efficacy? Parasitol Today 1998;14:218.

[18] Beck HP, Felger I, Huber W, Steiger S, Smith T, Weiss N, et al.Analysis of multiple Plasmodium falciparum infections in Tanzanianchildren during the phase III trial of the malaria vaccine SPf66. JInfect Dis 1997;175:921.

[19] Haywood M, Conway DJ, Weiss H, Metzger W, D’Alessandro U,Snounou G, et al. Reduction in the mean number ofPlasmodiumfalciparum genotypes in Gambian children immunized with themalaria vaccine SPf66. Trans R Soc Trop Med Hyg 1999;1:65.

[20] Alonso PL, Lopez MC, Bordmann G, Smith TA, Aponte JJ, WeissNA, et al. Immune responses toPlasmodium falciparum antigensduring a malaria vaccine trial in Tanzanian children. Parasite Immunol1998;20:63–71.

[21] Kensil CR, Patel U, Lennick M, Marciani D. Separation andcharacterisation of saponins with adjuvant activity fromQuillajasaponaria Molina cortex. Immunology 1991;146:431–7.

[22] Kensil CR. Saponins as vaccine adjuvants. Crit Rev Ther DrugCarrier Syst 1996;13:1–55.

[23] Livingston PO, Adluri S, Helling F, Yao TJ, Kensil CR, NewmanMJ, et al. Phase 1 trial of immunological adjuvant QS-21 with aGM2 ganglioside-keyhole limpet haemocyanin conjugate vaccine inpatients with malignant melanoma. Vaccine 1994;12:1275–80.

[24] Stoute JA, Slaoui M, Heppner DG, Momin P, Kester KE, DesmonsP, et al. A preliminary evaluation of a recombinant circumsporozoiteprotein vaccine against Plasmodium falciparum malaria. RTS,SMalaria Vaccine Evaluation Group. N Engl J Med 1997;336:86–91.

[25] Foon KA, Sen G, Hutchins L, Kashala OL, Baral R, Banerjee M,et al. Antibody responses in melanoma patients immunised withan anti-idiotype antibody mimicking disialoganglioside GD2. ClinCancer Res 1998;4:1117–24.

[26] Kensil CR, Kammer R. QS-21: a water-soluble triterpene glycosideadjuvant. Exp Opin Invest Drugs 1998;7:1475–82.

[27] Kester KE, McKinney DA, Tornieporth N, Ockenhouse CF, HeppnerDG, Hall T, et al. Efficacy of recombinant circumsporozoite proteinvaccine regimens against experimentalPlasmodium falciparummalaria. J Infect Dis 2001;183:640.

[28] Nardin EH, Oliveira GA, Calvo-Calle JM, Castro ZR, NussenzweigRS, Schmeckpeper B, et al. Synthetic malaria peptide vaccine elicitshigh levels of antibodies in vaccinees of defined HLA genotypes. JInfect Dis 2000;182:1486.

[29] Council for International Organizations of Medical Sciences(CIOMS) and World Health Organisation (WHO). Internationalethical guidelines for biomedical research involving human subjects.1 ed. Geneva: CIOMS, 1994.

[30] International Conference on Harmonization. International Conferenceon Harmonization of Technical Requirements for Registration ofPharmaceuticals for Human Use. ICHGCPD9.WPs 27 April 1995.CPMP, 1995.

[31] Salcedo M, Barreto L, Rojas M, Moya R, Cote J, PatarroyoME. Studies on the humoral immune response to a syntheticvaccine againstPlasmodium falciparum malaria. Clin Exp Immunol1991;84:122–8.

[32] Gupta RK, Griffin Jr P, Chang AC, Rivera R, Anderson R, Rost B,et al. The role of adjuvants and delivery systems in modulation ofimmune response to vaccines. Adv Exp Med Biol 1996;397:105–13.

[33] Braun MM, Ellenberg SS. Descriptive epidemiology of adverseevents after immunisation: reports to the vaccine adverse eventreporting system (VAERS), 1991–1994. J Pediatr 1997;131:529–35.

[34] Center for Disease Control and Prevention. Update: vaccineside effects, adverse reactions, contraindications, and precautions.Recommendations of the Advisory Committee on ImmunisationPractices. Morb Mortal Wkly Rep 1996;45:1–35.

[35] Grotto I, Mandel Y, Ephros M, Ashkenazi I, Shemer J. Major adversereactions to yeast-derived hepatitis B vaccines: a review. Vaccine1998;19:329–34.

[36] Sakaguchi M, Sakae I. Two patterns of systemic reactions to Japaneseencephalitis vaccines. Vaccine 1998;19:68–9.

[37] Kallas EG, Evans TG, Gorse G, et al. 1998. QS-21 is superiorto alum as adjuvant with low dose recombinant MN gp120 forimmunisation of uninfected adults: humoral and cellular responses.In: Proceedings of the 5th International Conference on Retrovirusesand Opportunistic Infections, Chicago, IL, 1998 Feb. 1–5.

[38] Miller LH, Good MF, Kaslow DC. Vaccines against the blood stagesof falciparum malaria. Adv Exp Med Biol 1998;452:193–205.

[39] Troye-Blomberg M, Berzins K, Perlmann P. T-cell control ofimmunity to the asexual blood stages of the malaria parasite. CritRev Immunol 1994;14:131–55.

[40] Calvo M, Guzman F, Perez E, Segura CH, Molano A, Patarroyo ME.Specific interactions of synthetic peptides derived fromP. falciparummerozoite proteins with human red blood cells. Pept Res 1991;4:324–33.

[41] Moreno R, Pöltl-Frank F, Stüber F, Matile H, Mutz M, Weiss N,et al. A rhoptry-associated protein 1-binding monoclonal antibodyraised against a heterologous peptide sequence inhibitsPlasmodiumfalciparum growth in vitro. Infect Immun 2001;69:2558–68.