Inhibition of Erythrocyte Invasion and Plasmodium falciparum Merozoite Surface Protein 1 Processing by Human Immunoglobulin G1 (IgG1) and IgG3 Antibodies

Abstract

Antigen-specific antibodies (Abs) to the 19-kDa carboxy-terminal region of Plasmodium falciparum merozoite surface protein 1 (MSP1₁₉) play an important role in protective immunity to malaria. Mouse monoclonal Abs (MAbs) 12.10 and 12.8 recognizing MSP1₁₉ can inhibit red cell invasion by interfering with MSP1 processing on the merozoite surface. We show here that this ability is dependent on the intact Ab since Fab and F(ab′)₂ fragments derived from MAb 12.10, although capable of binding MSP1 with high affinity and competing with the intact antibody for binding to MSP1, were unable to inhibit erythrocyte invasion or MSP1 processing. The DNA sequences of the variable (V) regions of both MAbs 12.8 and 12.10 were obtained, and partial amino acid sequences of the same regions were confirmed by mass spectrometry. Human chimeric Abs constructed by using these sequences, which combine the original mouse V regions with human γ1 and γ3 constant regions, retain the ability to bind to both parasites and recombinant MSP1₁₉, and both chimeric human immunoglobulin G1s (IgG1s) were at least as good at inhibiting erythrocyte invasion as the parental murine MAbs 12.8 and 12.10. Furthermore, the human chimeric Abs of the IgG1 class (but not the corresponding human IgG3), induced significant NADPH-mediated oxidative bursts and degranulation from human neutrophils. These chimeric human Abs will enable investigators to examine the role of human Fcγ receptors in immunity to malaria using a transgenic parasite and mouse model and may prove useful in humans for neutralizing parasites as an adjunct to antimalarial drug therapy.

Plasmodium falciparum merozoite surface protein 1 (MSP1) is a major surface protein and a leading candidate antigen for a malaria vaccine. MSP1 is present on the surface of the merozoite as a polypeptide complex formed, in part, by the proteolytic cleavage, or primary processing, of the precursor protein. The complex is anchored to the merozoite by the C-terminal 42-kDa fragment attached to the membrane by a glycophosphoinositol anchor. At the time of erythrocyte invasion, the 42-kDa fragment undergoes a proteolytic cleavage, known as secondary processing, leaving a 19-kDa carboxy-terminal fragment (MSP1₁₉) attached to the merozoite and carried into the erythrocyte as the parasite invades (3).

Humans develop antibodies (Abs) against MSP1₁₉ through natural P. falciparum infection, and such Abs have been associated with protection from clinical malaria in a number of studies (5, 10, 11). There is also good evidence that Abs against MSP1₁₉ provide the major component of the invasion-inhibitory activity of human immune sera (26).

Secondary processing is essential for successful erythrocyte invasion and is vulnerable to interruption by some MSP1₁₉ specific Abs. For example, two murine monoclonal Abs (MAbs) specific for epitopes within MSP1₁₉, MAb 12.8 and MAb 12.10 (20) have been shown to prevent secondary processing of MSP1, and these MAbs also inhibit erythrocyte invasion (3, 4), suggesting that their biological activity, the inhibition of invasion, is due to the inhibition of secondary processing. It is therefore very important to understand why the antibodies are effective and which of their features are important. For example, their fine specificity of binding or their size may be key properties.

Determining the fine specificity of Abs against MSP1₁₉ is of crucial importance in understanding their biological activity. Significant associations have been demonstrated between the fine specificity of anti-MSP1₁₉ Abs and protection from subsequent malaria (27). Interestingly, MAbs 12.8 and 12.10 bind to epitopes that are conserved in both dimorphic forms of MSP1₁₉ and are found on parasites derived from all geographic locations examined (20). The structure of P. falciparum MSP1₁₉ has been determined (25, 28), and this information facilitates understanding the interaction between the antigen and the various Abs, which has been explored using binding studies with native or mutant forms of recombinant MSP1₁₉ (9, 21, 23, 24, 35) and by computational approaches (2). These studies have shown that MAbs 12.8 and 12.10 both bind to the same surface of MSP1₁₉ and their epitopes overlap, findings consistent with them having the same biological properties (reviewed in reference 16).

The size of the Ab binding to the antigen may also be an important factor in its activity. For example, Fab fragments of MAbs binding to apical membrane antigen 1 (AMA1) are more effective than the corresponding immunoglobulin Gs (IgGs) at inhibiting erythrocyte invasion (34). Furthermore, Fc-mediated effects have been shown to be an important component of the function of MSP1₁₉-specific antibodies in an in vivo transgenic model (21). We have prepared here F(ab′)₂ and Fab fragments from MAb 12.10 and examined their biological activity compared to the intact antibody. In addition, we have determined the DNA and amino acid sequences of the variable regions of the heavy and light chains of MAbs 12.8 and 12.10 by cloning and mass spectrometry and used this information to create recombinant chimeric human IgG1 and IgG3 versions of these MAbs. When these sequences are expressed by mammalian cell lines, they retain both their ability to bind to MSP1₁₉ and to inhibit erythrocyte invasion by P. falciparum in vitro. The recombinant chimeric human IgG1 and IgG3 Abs may prove suitable for in vivo studies of protection from P. falciparum in both primate and rodent transgenic models (14, 21).

MATERIALS AND METHODS

Production of MAbs and their fragments.

The mouse MAb 12.8 (IgG2b) and MAb 12.10 (IgG1) hybridomas were cultured in Dulbecco modified Eagle medium containing 10% fetal calf serum (depleted of bovine IgG) and Ab purified from culture supernatants by protein G-Sepharose affinity chromatography. F(ab′)₂ fragments were generated from MAb 12.10 by pepsin digestion and gel filtration chromatography. Fab fragments were then derived from F(ab′)₂ by mild reduction and alkylation, followed by further purification by gel filtration, essentially as described previously (19, 24). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of the final intact MAb, F(ab′)₂, and Fab showed that the preparations contained polypeptides of the expected molecular masses and were highly pure (see Fig. 1A).

FIG. 1.

Analysis of murine MAb 12.10 and its F(ab′)₂ and Fab fragments: effects on erythrocyte invasion and MSP1 processing. (A) Portions (10 μg) of MAb 12.10 and its F(ab′)₂ and Fab fragments were solubilized in SDS in the presence (tracks 4, 5, and 6) or absence (tracks 1, 2, and 3) of 50 μM dithiothreitol and analyzed by SDS-PAGE. The migration of standard size markers (in kilodaltons) is indicated. (B) Effect of Ab binding on MSP1 processing using purified T9/96 merozoites: MAb 12.10 but not its fragments is able to inhibit MSP1 processing. Merozoites were either solubilized into SDS (track 1) or incubated for 1 h at 37°C in the presence of no Ab (track 2), phenylmethylsulfonyl fluoride (track 3), 400 μg of MAb 12.10 ml⁻¹ (track 4), 250 μg of F(ab′)₂ ml⁻¹ (track 5), or 250 μg of Fab ml⁻¹ (track 6). In a competitive assay, merozoites were preincubated with 250 μg of F(ab′)₂ or Fab ml⁻¹, followed by 400 μg of MAb 12.10 ml⁻¹ (tracks 7 and 8, respectively). All samples were analyzed by Western blotting with MAb X509, specific for an epitope within MSP1₃₃, followed by the addition of horseradish peroxidase-conjugated anti-human polyclonal Ab. (C) Effect of Ab binding on erythrocyte invasion: MAb 12.10 but not its fragments is able to inhibit invasion. Invasion in individual microcultures was assessed by counting the number of ring-stage parasites in 5,000 red cells. Invasion is expressed as a percentage of the ring-stage parasitemia (10%) obtained in a control culture in the presence of nonimmune human sera. The culture was divided into equal aliquots, followed by incubation for 6 h at 37°C in the presence of either 5 mM EGTA, nonimmune human serum (NI Hs), 400 μg of MAb 12.10 ml⁻¹, 250 μg of F(ab′)₂ ml⁻¹, or 250 μg of Fab ml⁻¹. In competition experiments, 250 μg of either F(ab′)₂ or Fab ml⁻¹ was added to individual cultures, followed by the addition of 400 μg of MAb 12.10 ml⁻¹ to each culture. Maximum inhibition was obtained in a control culture with 5 mM EGTA. MAb 12.10, F(ab′)₂, and Fab inhibited erythrocyte invasion by 96.5, 1, and 2%, respectively. In a competitive assay, inhibition of erythrocyte invasion mediated by MAb 12.10 was abrogated by the presence of F(ab′)₂ or Fab. All samples were assayed in triplicate, and standard error bars are indicated. The results shown are representative of at least two individual experiments. (D) Effect of antibody binding on the shedding of the MSP1 complex into parasite cultures: MAb 12.10 but not its fragments is able to inhibit MSP1 processing and shedding. Metabolically radiolabeled T9/96 schizonts were supplemented with fresh erythrocytes and medium to obtain a parasitemia of ca. 2% and a hematocrit of 1%. The culture was then divided into equal aliquots, followed by incubation for 6 h at 37°C in the presence of either 5 mM EGTA, no MAb, 400 μg of MAb 12.10 ml⁻¹, 250 μg of F(ab′)₂ ml⁻¹, or 250 μg of Fab ml⁻¹. In competition experiments, 400, 250, 125, 62.2, and 31 μg of either F(ab′)₂ or Fab ml⁻¹ was added to individual cultures, followed by the addition of 400 μg of MAb 12.10 ml⁻¹ to each culture. Culture supernatant was analyzed by immunoprecipitation with MAb X509, and the products were resolved by SDS-PAGE. The presence of the shed MSP1₃₃ is indicated. In the competitive assay, inhibition of MSP1 processing and shedding mediated by MAb 12.10 was abrogated by the presence of high concentrations of F(ab′)₂ or Fab.

Ab-mediated inhibition of MSP1 processing and erythrocyte invasion.

Two protocols, both described previously, were used to assess Ab-mediated inhibition of MSP1 secondary processing. In the first method, purified, naturally released merozoites were washed, incubated on ice in the presence of test Abs, transferred to 37°C for 1 h to allow processing to proceed, and then analyzed by Western blotting for de novo production of the MSP1₃₃ product of secondary processing (4, 15). Alternatively, mature schizonts were biosynthetically radiolabeled with [³⁵S]methionine-cysteine and then placed back into culture with the addition of fresh erythrocytes, plus the Abs or Ab fragments being tested, and allowed to undergo merozoite release for 6 h. Cultures were then harvested; processing was quantified by immunoprecipitation of labeled MSP1₃₃ from the culture medium by using MAb X509 coupled to Sepharose, while ring-stage parasitemia in the cultures was assessed by microscopic examination of Giemsa-stained thin blood films (15).

Invasion assays were carried out essentially as previously described (3), using schizonts from the 3D7, FCB-1, or T9/96 lines of P. falciparum. Schizonts, collected and purified at 40 h postinvasion, were added to the culture with fresh erythrocytes. The plates were placed in a gassed box containing an atmosphere of 7% CO₂, 5% O₂, and 88% N₂ and incubated at 37°C for 20 h.

Antigen binding studies.

For quantitative comparisons of binding by surface plasmon resonance (SPR) analysis, recombinant MSP1₁₉-glutathione S-transferase (GST) (7) or GST alone (negative control) were amine coupled to each flow cell of a CM5 sensor chip. Various concentrations (as indicated) of purified MAbs, their F(ab′)₂ and Fab fragments, or human chimeric anti-MSP1₁₉ IgG1 and IgG3 were then injected over each antigen, and association and dissociation observed. The data from a BIAcore machine were analyzed by using BIAevaluation 3.0 software. For immunofluorescence (IF) studies, washed erythrocytes infected with the P. falciparum 3D7, T9/96, and FCB-1 parasite lines were fixed on slides in methanol-acetone (1:1 [vol/vol]) for 10 min. After being blocked in phosphate-buffered saline (PBS)-5% (vol/vol) goat serum, the slides were incubated with Abs at 5 μg ml⁻¹ in blocking buffer for 1 h, washed, and then incubated for 1 h with fluorescein isothiocyanate (FITC)-conjugated goat F(ab′)₂ anti-human IgG (Sigma or Caltag) diluted 1:500 in blocking buffer. After a washing step, the slides were mounted with DAPI (4′,6′-diamidino-2-phenylindole) antifade, and Ab binding was examined by IF microscopy on a Zeiss Axioscope 40 microscope.

Amino acid sequencing by mass spectrometry.

IgG heavy and light chains were fractionated by SDS-PAGE, stained, excised from the gel, and then subjected to trypsin digestion. Peptides were analyzed on LCQ and Q-TOF tandem mass spectrometers at the National Institute for Medical Research and the Biopolymer Synthesis and Analysis Unit (University of Nottingham), and the sequences were determined by using an online database (Matrix Science). These sequences were used in a search of related murine immunoglobulin sequences using the basic local alignment search tool (BLAST) at the National Center for Biotechnology Information (1), and this information was used to design specific primers to amplify the V genes from the cDNA.

Amplification of DNA encoding the variable regions of MAb 12.10 and 12.8, construction of expression vectors, and production of human chimeric antibodies.

Variable heavy (V_H) and light (V_L) complementary DNAs (cDNAs) were obtained from the mouse hybridomas 12.10 and 12.8 by reverse transcriptase-PCR by using degenerate primers (5′-CAGGCGCGCACTCCSARGTNMARCTGCAGSAGTCWG-3′ and 5′-TGARGAGACGGTGACCGDGGTBCCTKGGCCCCA-3′ for V_H and 5′-GAYATTGAGCTCACMCARWCTMCA-3′ and 5′-CCGTTTBAKCTCGAGCTTKGTSCC-3′ for the 12.8 V_L or 5′-GACATTGAGCTCACMCAGTCTC-3′ and 5′-CCGTTTCAGCTCGAGCTTGGTCCC-3′) for the 12.10 V_L. The V_H and V_L genes were cloned into pCR2.1-Topo or pGEM-T easy plasmids, respectively. The V_H gene was subcloned as a BssHII-BstEII fragment into pVH-C1-γ1 or pVH-C1-γ3 vectors upstream of the IgG1 or IgG3 constant regions, respectively (21). The human IgG3 constant gene was amplified from human genomic DNA by overlap PCR using two sets of primers (G3FOR1, 5′-CCTGGATCCTCGTGGATAGACAAG-3′; G3REV1, 5′-GAAGATATCGAGTTACTCAGATCTGGG-3′; G3FOR2, 5′-CTCAGCTCAGACACCTTCTC-3′; G3REV2, 5′-TCCCTTCTAGAACTCAGGCCTCAGAC-3′), cloned into the pCR2.1 TOPO vector (17), sequenced, and subcloned as a BamHI-XbaI fragment into pVH-C1-γ1 to create pVH-C1-γ3. The V_L gene was subcloned as an ApaLI-XhoI fragment into pVK-C1-Express upstream of the human κ gene as previously described (21, 29). Chinese hamster ovary (CHO-K1) cells were transfected with corresponding heavy- and light-chain constructs, and positive clones were selected (21, 29). Clones secreting MSP1₁₉-specific IgG1 or IgG3 were detected by enzyme-linked immunosorbent assay (ELISA) with recombinant MSP1₁₉ coated plates and by immunoblotting with goat anti-human IgG Abs conjugated to horseradish peroxidase. Recombinant MSP1₁₉ used in all ELISAs was generated as previously described (7). From large-scale cultures, human IgG1 or IgG3 were purified on HiTrap protein G-Sepharose (GE Healthcare) by fast performance liquid chromatography. The integrity and purity of the Abs was verified on SDS-PAGE gels.

Luminol chemiluminescence assay of respiratory burst and myeloperoxidase release.

Neutrophils were isolated from heparinized blood taken from healthy volunteers by sedimentation of erythrocytes in 6% (wt/vol) dextran T70 (GE Healthcare, United Kingdom) in 0.9% (wt/vol) saline at 37°C for 30 min, followed by leukocyte separation on a discontinuous density gradient of Lymphoprep (ρ = 1.077 g/cm³; Nycomed, Birmingham, United Kingdom) over Ficoll-Hypaque (ρ = 1.119 g/cm³) and centrifugation at 700 × g for 20 min at room temperature. Approval for the use of human cells was obtained from the local QMC ethical committee. Wells of chemiluminescence microtiter plates (Dynatech Laboratories, Billinghurst, Sussex, United Kingdom) were coated with 150 μl of MSP1₁₉ at 5 μg ml⁻¹ in coating buffer (0.1 M carbonate buffer [pH 9.6]) and incubated overnight at 4°C. After three washes with PBS, 150-μl portions of anti-MSP1₁₉ Abs (MAbs 12.8, 12.10, and their chimeric human IgG1 and IgG3 derivatives at 50 μg ml⁻¹) were added to the wells. In each case, triplicate wells were prepared and left for 2 h at room temperature. After a wash as described above, 100 μl of luminol (67 μg ml⁻¹ in Hanks buffered saline solution [HBSS] containing 20 mM HEPES and 0.1 g of globulin-free bovine serum albumin [BSA]/100 ml [HBSS/BSA]) was added to each well. After the addition of 50 μl of purified neutrophils (10⁶/ml in HBSS/BSA) to each well, the plates were transferred to a Microlumat LB96P luminometer, and the chemiluminescence was measured at 2-min intervals for 120 min at 37°C. The data were analyzed by using Excel software. Similarly coated wells not used for the chemiluminescence assay were tested by ELISA for reactivity with anti-mouse IgG (Pierce) or anti-human IgG (Sigma) reagents conjugated to peroxidase as appropriate, and coating levels for each Ab were found to be equivalent prior to entry into the respiratory burst assay.

RESULTS

Inhibition of erythrocyte invasion by MAb12.10 but not by its F(ab′)₂ or Fab fragments.

To examine the contribution of the size of the antibody to inhibition of processing and red cell invasion, we removed the Fc region and generated purified bivalent F(ab′)₂ and monovalent Fab fragments of MAb 12.10 (Fig. 1A). Both F(ab′)₂ and Fab fragments retained the specificity and binding properties of the parental MAb 12.10 as determined by SPR analysis, albeit with expected changes in affinity that are associated with the valency of mono- and bispecific reagents (Table 1). In addition, the F(ab′)₂ and Fab fragments effectively competed with intact MAb 12.10 for antigen binding to live merozoites or MSP1₁₉ coated onto plates (data not shown). However, whereas MAb 12.10 inhibited MSP1 processing in vitro (Fig. 1B, lane 4), when used at equimolar concentrations (2.6 μM each), neither the F(ab′)₂ nor the Fab fragments inhibited MSP1 processing (Fig. 1B, lanes 5 and 6). However, when merozoites were preincubated with the F(ab′)₂ or Fab, the processing-inhibitory activity of MAb 12.10 was significantly reduced, presumably because they competed effectively for binding to the antigen (Fig. 1B, lanes 7 and 8).

TABLE 1.

Analysis of the binding properties of MAb 12.10 and its F(ab′)₂ and Fab fragments to immobilized GST-MSP1₁₉ by SPR

MAb 12.10 or its MAb 12.10 Fab fragment	Kassociation (M−1 s−1)	Kdissociation (s−1)	Kd (M)a	t1/2 (s)b
MAb 12.10	5.37 × 105	5.5 × 10−6	1 × 10−11	126,026.7
F(ab′)2	7.1 × 105	4.9 × 10−6	6.9 × 10−12	141,458.6
Fab	6.7 × 105	8.9 × 10−5	1.3 × 10−10	7,788.1

^aK_d = K_dissociation/K_association.

^bt_1/2 = ln2/K_dissociation.

The next set of experiments was designed to investigate the ability of MAb 12.10 and its F(ab′)₂ and Fab fragments to inhibit MSP1 processing and to interfere with erythrocyte invasion under conditions of active merozoite release and erythrocyte invasion in parasite cultures. For this purpose, an in vitro parasite invasion assay was used in which MSP1 secondary processing was measured in parallel with erythrocyte invasion. Mature T9/96 schizonts were biosynthetically radiolabeled with [³⁵S]methionine-cysteine, and then merozoite release and red cell invasion was allowed to proceed in the presence of intact MAb 12.10, or the F(ab′)₂ or Fab preparations. Erythrocyte invasion over the course of the experiment was then assessed by counting the number of new ring-stage parasites formed (Fig. 1C), and MSP1 processing was assessed by direct immunoprecipitation of MSP1₃₃ from the culture supernatants (Fig. 1D). Significant inhibition of both erythrocyte invasion and MSP1 processing was observed when either MAb 12.10 or EGTA (a potent inhibitor of MSP1 processing) was present in cultures. The F(ab′)₂ and Fab preparations had no discernible effect on either erythrocyte invasion or MSP1 processing.

In competition experiments, in which increasing concentrations of F(ab′)₂ or Fab were added to microcultures containing intact MAb 12.10 at 400 μg ml⁻¹, it was observed that both fragments can interfere with the processing-inhibitory activity of MAb 12.10 in a dose-dependent manner and at high concentrations can effectively reverse both the invasion-inhibitory and the processing-inhibitory activity of MAb 12.10 (Fig. 1C and D).

Amplification of DNA encoding the variable regions of MAb 12.8 and 12.10.

Having established that intact IgG was necessary to inhibit invasion and MSP1 processing, we wanted to produce chimeric human IgG1 and IgG3 versions of 12.8 and 12.10 and examine their properties. As a first step, and in anticipation of nonproductive pseudogene expression in both hybridomas, to be certain of cloning the correct variable (V) gene sequences, we obtained amino acid sequence data for the heavy and light chains (separated by SDS-PAGE) from MAbs 12.8 and 12.10 by tryptic digestion and mass spectrometric analysis of the generated fragments (Fig. 2). For each of the heavy and light chains, at least two unique peptide sequences were derived from the variable regions, allowing assignment of the variable regions to specific families. Primers were then designed to amplify the correct sequences from cDNA.

FIG. 2.

Nucleotide and corresponding amino acid sequences from V_H and V_L genes of MAb 12.8 (A) and MAb 12.10 (B). The primer sequences used in the amplification of variable (V) gene sequences are underlined. DNA sequence coding for framework regions is highlighted in yellow; pink denotes DNA regions of hypervariability or coding for complementarity determining regions. Amino acid sequences confirmed by mass spectrometry of tryptic peptides from heavy and light chains are indicated in italic blue letters.

The MAb heavy- and light-chain V regions were amplified by PCR from cDNA derived from each hybridoma and cloned into plasmid vectors by using restriction enzyme sites included in the amplification primers and sequenced (Fig. 2). Searches against the ImMunoGeneTics (IMGT) (http://imgt.cines.fr/) databases indicated that the heavy chain of MAb 12.8 is a member of the mouse IgHV9 family (IGHV9-1*02), whereas that of MAb 12.10 belongs to the IgHV1 family (IGHV1S81*02). Comparisons with existing mouse immunoglobulin light-chain databases at IMGT also found close matches for each of the light chains. Thus, the MAb 12.8 light chain is a member of the IgKV4 family (IGKV4-70*01), and MAb 12.10 is a member of the IgKV6 family (IGKV6-17*01). The V regions of both the heavy and light chains differ significantly in DNA sequence between the two MAbs, which is consistent with them belonging to different subclasses and V gene families, and with their binding to nonidentical epitopes on MSP1₁₉ (23).

Development of chimeric human IgG1 and IgG3 recognizing MSP1₁₉.

We subcloned the murine variable region sequences derived from the two MAbs and linked them to human IgG1 and IgG3 constant domains in expression vectors used previously to generate fully human anti-P. falciparum or chimeric anti-P. yoelii MSP1₁₉ IgGs (21, 29). Both Abs, when purified from transfected CHO-K1 cell culture supernatants by protein G-Sepharose chromatography, contained polypeptides of the expected size on SDS-PAGE (Fig. 3, left panel) and the anticipated reactivity with anti-human IgG Fc-specific Abs (Fig. 3, right panel). As a consequence of the extended hinge region in human IgG3, the heavy chain has the expected higher apparent molecular mass compared to human IgG1.

FIG. 3.

SDS-PAGE, immunoblot, and IF analysis of purified Abs. Portions (5 μg) of purified MSP1₁₉-specific humanized IgG1, IgG3, and the mouse MAbs were subjected to SDS-PAGE under nonreduced or reduced conditions on 4 to 12% polyacrylamide gradient gels and immunoblotted with Fc-specific goat anti-human IgG conjugated to horseradish peroxidase. The migration of standard size markers (in kilodaltons) is indicated at the side of each gel. The left panel shows the Coomassie blue-stained gel and the right panel shows the Western blot. The secondary antibody is specific for the Fc region of human IgG and detects the intact IgG1 and IgG3 antibodies or heavy chains but not the mouse MAb antibodies. In the lower panel, recombinant IgG1 and IgG3 produce a characteristic pattern of MSP1₁₉ reactivity with schizonts, merozoites, and ring-stage parasites in IF assays of P. falciparum 3D7, identical to that obtained with the control human IgG1 C1 (JS1).

Chimeric human IgG1 and IgG3 bind MSP1₁₉ with high affinity.

Both recombinant IgG1 and IgG3 Abs bound to MSP1₁₉ in ELISAs and immunoblot analyses (data not shown) and produced a characteristic pattern of MSP1₁₉ reactivity with schizonts, merozoites, and ring-stage parasites in IF assays of P. falciparum 3D7, identical to that of both the parental murine MAbs (not shown) and the MSP1₁₉-specific human IgG1 C1 (JS1) (Fig. 3, lower panel). Importantly, the chimeric Abs interacted with MSP1₁₉ with very high affinity compared to that of the original parental mouse MAbs or the human IgG1 or IgG3 C1 (JS1) generated from a previous study (21) and did not bind to GST or irrelevant malaria antigens (e.g., AMA1 [data not shown]) when investigated by Biacore SPR analysis of antigen-coated sensor chips (Fig. 4).

FIG. 4.

SPR-derived association and dissociation curves of Ab binding to MSP1₁₉. Abs in buffer were injected at various concentrations into flow at time zero and replaced with buffer alone at the points indicated by vertical black arrows. (A) Murine MAb 12.8; (B) chimeric hIgG1 12.8; (C) chimeric hIgG3 12.8; (D) fully human IgG1 C1 positive control (JS1); (E) chimeric hIgG1 12.10; (F) chimeric hIgG3 C1. No binding was seen by any of the Abs to a control antigen, AMA1, coated onto flow cell 2.

The anti-MSP1₁₉ chimeric human IgG1 is fully functional.

The chimeric IgG1 antibodies were examined for their ability to inhibit erythrocyte invasion, using the parental murine MAbs 12.8 and 12.10 and EGTA as positive controls and the non-Plasmodium specific MAb BC1 as a negative control. The chimeric human IgG1s, 12.10 and 12.8, at concentrations of 560 and 1,200 μg ml⁻¹, respectively, reduced the percentage of red cells invaded by P. falciparum by 57 and 54% (Fig. 5). Interestingly, greater inhibition of invasion was observed with the chimeric IgG1 12.10 than with the parental MAb 12.10 used at 4,130 μg ml⁻¹, which may reflect variation in the activity of MAb 12.10 we observed in different batches prepared from the hybridoma on different occasions. We were unable to do these assays with the chimeric human IgG3 versions because of difficulties in generating sufficient quantities of Ab.

FIG. 5.

Inhibition of merozoite invasion of erythrocytes by chimeric human antibodies. Schizonts of P. falciparum clone 3D7 at a starting parasitemia of 1% and a final hematocrit of 3 to 5% were incubated in the presence of test (MAbs 12.8, 12.10, and chimeric human IgG1 derived from them) or control (MAb BC1, not directed against any P. falciparum) antibodies at a final concentration of Ab of 0.5 to 1 mg ml⁻¹, and the inhibition of merozoite invasion was measured. EGTA was used as a positive control. The percentage of erythrocytes invaded was determined separately from two Giemsa-stained thin-film slides from each well. A total of 1,000 erythrocytes were counted on each slide, and the slides were examined blindly. The mean of these four independent counts was taken to be the final parasitemia. Using medium alone, the parasitemia had increased to ∼3.1% after 24 h. Inhibition of invasion was calculated as a percentage of this control parasitemia. Error bars represent maximum and minimum parasitemias obtained. Differences between the control groups and samples incubated with chimeric human Abs were analyzed by using Dunnett's multiple comparison test. Statistically significant differences between the groups are indicated.

We next assessed the ability of these novel IgG reagents to interact with human Fcγ receptors (FcγRs) and induce NADPH oxidase activation in human blood neutrophils. For neutrophils, luminol chemiluminescence provides a readout of NADPH oxidase activation (respiratory burst) and myeloperoxidase release (22). In order to more directly compare the relative abilities of MSP1₁₉ specific Abs to induce functional responses, recombinant chimeric IgG1 and IgG3 with identical specificities for MSP1₁₉ and at equivalent concentrations, as determined by ELISA, were aggregated on MSP1₁₉-coated wells. Surprisingly, although both chimeric human IgG1s very effectively induced luminol chemiluminescence responses in resting neutrophils, the chimeric human IgG3 versions were without effect (Fig. 6). As anticipated from our findings with a mouse IgG2b specific to MSP1₁₉ from P. yoelii (29), murine MAb 12.8, which is also of the IgG2b subclass, was able to induce a weak burst, presumably through lower-affinity interactions with human FcγRs. Intriguingly, MAb 12.10, a mouse IgG1, was as effective as the human IgG1 at inducing bursts from human neutrophils.

FIG. 6.

Fc-mediated function of anti-MSP1₁₉ human chimeric Abs. Human neutrophil NADPH oxidative bursts were stimulated using the chimeric human Ab 12.10 (panel A, open symbols) or 12.8 (panel B, filled symbols) attached to GST-MSP1₁₉-coated microtiter plates. Chemiluminescence (light index; arbitary units) was induced by parental murine MAbs (triangles), human chimera IgG1 (squares), human chimera IgG3 (diamonds), or no Ab (circles). Each point shown is the mean of triplicate determinations. Each experiment was performed twice with cells from different healthy donors. The results shown are those from a typical experiment.

DISCUSSION

MAbs 12.8 and 12.10 inhibit MSP1 processing and erythrocyte invasion by a mechanism that is independent of FcγRs (3). To better understand the contributing factors within the structure of an Ab involved in this process, we generated bivalent (Fab′)₂ and monovalent Fab fragments from MAb 12.10. Removal of the Fc region abolished the processing-inhibitory activity of 12.10 without affecting the ability of both the (Fab′)₂ and the Fab fragments to bind to MSP1. This result indicates that the Fc part of the molecule and perhaps the overall size of the antibody are essential for inhibition. This result is in contrast to what has been reported for antibodies that bind to AMA1 and inhibit erythrocyte invasion: (Fab′)₂ and Fab fragments of MAbs specific for this protein are more effective than the intact IgG in mediating inhibition of invasion (8, 34). In this case, the smaller size may be important, and clearly in this case the Fc is not crucial to Ab function. On the other hand, recent work has shown that an anti-MSP1₁₉ human IgG1 (JS1) that cannot inhibit processing or invasion in vitro can protect animals from malaria in vivo by recruiting human FcγRs (21).

In order to understand the role of human Abs and their cognate FcγRs in immunity to malaria, we generated chimeric human IgG1 and IgG3 recognizing the candidate vaccine antigen MSP1₁₉. These reagents were generated through cloning of immunoglobulin variable (V) domains from murine hybridomas 12.8 and 12.10 and grafting them onto human constant region genes of the two most common IgG1 and IgG3 allotypes found in African populations (17) for expression in mammalian cell lines. The engineered chimeric human Abs recognized parasites in infected erythrocytes and bound specifically to MSP1₁₉ compared to a panel of control Abs. Reassuringly, SPR analysis revealed no reduction in affinity for MSP1₁₉ for the chimeric IgG1 or IgG3 compared to the parental murine MAbs. The chimeric human IgG1 Abs also very effectively inhibited red cell invasion and triggered potent FcγR-mediated respiratory bursts from human neutrophils. Human neutrophils express the activating FcγRIIA and the glycophosphoinositol-linked FcγRIIIB, and human IgG1 is known to bind to the former with very high affinity (6). Furthermore, FcγRIIIB cross-linking does not induce nuclear migration or Erk or MEK, suggesting that these two receptors mediate human IgG1 signaling through different pathways (13). Unexpectedly, these assays also indicated that human IgG3 may be a poor inducer of respiratory bursts despite possessing the highest affinity for FcγRs of all of the human IgG subclasses (6, 31). Since antigen-specific antimalarial Abs from clinically immune individuals can belong to the human IgG3 subclass (12, 30), these data suggest that the effectiveness of human IgG3 (and indeed human neutrophils) in controlling malaria parasites may not be mediated via respiratory bursts but through alternative modes of action. Neutrophils are abundant in blood, and products of their respiratory bursts have been shown to kill merozoites, although the exact mechanisms are unclear (18). Recent results with recombinant human Abs generated against P. falciparum AMA1 support these observations and show that human IgG3 may make up for this deficiency by being more efficient at promoting phagocytosis of malaria parasites than human IgG1 (unpublished data). It is known that human IgG3 immune complexes are unable to stimulate human neutrophil respiratory bursts, although they provide a strong stimulus for inducing neutrophil degranulation, presumably via FcγRIIIB triggering (36). Further work is therefore required with the human IgG3 to ascertain its mode of action in malaria.

Believing that antiparasite Abs that both inhibit processing and recruit FcγR mediated clearance may be more effective than Abs limited to a single mode of action, we postulate that the human chimeric Abs generated here would be more effective at clearing malaria parasites in vivo than the human MAb JS1 generated earlier (21). JS1, a human IgG1, does not inhibit MSP1 processing but nevertheless effectively cleared parasites when passively transferred into human FcγR1 transgenic mice infected with P. berghei parasites transgenic for P. falciparum MSP1₁₉. However, neither MAb 12.8 nor 12.10 could bind to the MSP1₁₉ transgenic P. berghei parasite (data not shown), probably because the first 7 amino acids (Gly-Ile-Asp-Pro-Lys-His-Val) of the transgenic MSP1₁₉ sequence following the secondary cleavage site are from P. berghei, which differs substantially from the corresponding P. falciparum sequence (Asn-Ile-Ser-Gln-His-Gln). In addition, passive immunization with MAb 12.8 and 12.10 had no effect on parasite development (data not shown), as expected if the antibodies do not bind to the parasite. In order to test the protective capacity of these human chimeric 12.8 and 12.10 antibodies in vivo prior to any clinical trial, it will be necessary to generate a new transgenic P. berghei parasite containing the complete MSP1₁₉ sequence, including the epitopes seen by both MAbs 12.8 and 12.10. For such a transgenic parasite to be viable the MSP1₁₉ cleavage site would need to be a suitable substrate for the P. berghei MSP1 processing enzyme, likely to be an orthologue of P. falciparum subtilisin 2. In the absence of an appropriate in vivo mouse model, Aotus monkeys have been used as a primate model for P. falciparum infection in passive immunization studies (14) and when testing MSP1₁₉-based vaccines (33). The same model could be used to examine the ability of the invasion-inhibitory Abs to ameliorate P. falciparum infection in vivo; the partly human chimeric Abs should have reduced immunogenicity in primates compared to the parental murine MAbs. Although nonhuman primates are widely used in malaria research, care must be taken in evaluating human IgG antibodies in these species. For example, although FcγRIII (CD16) from sooty mangabey's is capable of binding human IgG1 and IgG2, it cannot bind human IgG3 or IgG4 like the human CD16 orthologue, predicting difficulties in assessment of these reagents in the Aotus model (32). Thus, despite strong sequence conservation in the FcγRs, nonhuman primate and human FcγRs differ in their ability to bind human IgG subclasses. Clearly, differences in the interaction of human IgG with nonhuman primate FcγRs should be considered when these models are used for the in vivo evaluation of therapeutic antibodies. Nonetheless, chimeric human versions of MAbs 12.8 and 12.10 clearly directly inhibit invasion of human erythrocytes by parasites and therefore may be useful therapeutically when a malaria infection is nonresponsive to available drugs (31).

In conclusion, we have shown that intact antibody is essential for inhibition of MSP1 processing and red cell invasion, and have developed recombinant partly human chimeric IgG1/κ and IgG3/κ Abs targeting the malaria vaccine candidate antigen MSP1₁₉. These novel reagents will be useful in standardized immunological assays with which to rigorously examine their antiparasitic mechanisms of action, be they FcγR dependent or independent.