A multivalent Plasmodium falciparum circumsporozoite protein‐based nanoparticle malaria vaccine elicits a robust and durable antibody response against the junctional epitope and the major repeats

Abstract Plasmodium falciparum (Pf) malaria continues to cause considerable morbidity and mortality worldwide. The circumsporozoite protein (CSP) is a particularly attractive candidate for designing vaccines that target sporozoites—the first vertebrate stage in a malaria infection. Current PfCSP‐based vaccines, however, do not include epitopes that have recently been shown to be the target of potent neutralizing antibodies. We report the design of a SpyCatcher‐mi3‐nanoparticle‐based vaccine presenting multiple copies of a chimeric PfCSP (cPfCSP) antigen that incorporates these important “T1/junctional” epitopes as well as a reduced number of (NANP) n repeats. cPfCSP‐SpyCatcher‐mi3 was immunogenic in mice eliciting high and durable IgG antibody levels as well as a balanced antibody response against the T1/junctional region and the (NANP) n repeats. Notably, the antibody concentration elicited by immunization was significantly greater than the reported protective threshold defined in a murine challenge model. Refocusing the immune response toward functionally relevant subdominant epitopes to induce a more balanced and durable immune response may enable the design of a more effective second generation PfCSP‐based vaccine.

species and is the cause of more than 90% of the mortality due to malaria in humans. 2 Therefore, there is an imminent need for an effective vaccine that protects from Pf infections.
The malaria parasite life cycle involves two hosts-an Anopheles mosquito vector and a susceptible vertebrate. When an infected female Anopheles mosquito draws blood from a mammalian host, it injects Plasmodium sporozoites that travel to the liver to infect hepatocytes to start the pre-erythrocytic cycle. Sporozoites in the liver mature into hepatic schizonts releasing thousands of merozoites into the bloodstream that infect red blood cells to initiate the blood-stage cycle. In red blood cells, merozoites transform into blood-stage schizonts, some of which differentiate into gametocytes. Malaria symptoms and clinical complications in humans occur during the bloodstage cycle. The sporogonic cycle is initiated when a mosquito draws blood from a malaria-infected host that contains gametocytes. The gametocytes fertilize to form ookinetes that can traverse the mosquito midgut and transform into oocysts, each containing thousands of sporozoites. Finally, sporozoites migrate to the mosquito's salivary glands to initiate a new infectious cycle in a susceptible host. 3 The pre-erythrocytic stage, involving the migration of sporozoites to the liver leading to hepatic infection, is the first stage in a malaria infection in the vertebrate and represents a population bottleneck in the life cycle. 4 Therefore, a vaccine that targets antigens in this stage has the potential to prevent severe clinical outcomes. The circumsporozoite protein (CSP) is a particularly attractive candidate for a sporozoite-targeting vaccine. CSP is a surface protein densely covering sporozoites and is required for hepatocyte infection. 5,6 Due to these reasons, several attempts have been made to design a CSP-based malaria vaccine. 7,8 CSP consists of three domains: an N-terminal domain (NTD), the repeat region that comprises repetitive units unique to each malaria parasite, and a C-terminal thrombospondin-like type-I repeat (TSR) domain. 6 In PfCSP, the repeat region consists of [35][36][37][38][39][40][41] NANP tetramers. In the 3D7/NF54 Pf isolate, the prototypical Pf malaria strain, the CSP repeat consists of 37 NANP tetramers (major repeats) and three alternating NVDP and NANP tetramers (minor repeats).
PfCSP-based vaccines have come a long way due to extensive research over 30 years with multiple clinical trials. 9 One vaccine-RTS,S in the AS01 adjuvant system-has been approved for use by the World Health Organization (WHO). 1

This vaccine is expressed in
Saccharomyces cerevisiae and consists of 19 NANP repeats followed by the carboxy terminus of PfCSP fused to the Hepatitis B surface antigen (HBsAg). The fused protein is expressed along with free HBsAg 10 and the proteins spontaneously assemble upon cell lysis into RTS,S particles composed of a 1:4 ratio of PfCSP to HBsAg. 11,12 This vaccine induces high but short-lived antibody responses that wane after a year of administration with an efficacy of 30%-50% within the first 12 months of administration. 13,14 Collins et al. 11 developed a RTS,S-like vaccine (R21) that uses the PfCSP-HBsAg fusion protein as the antigen without any free HBsAg so that the PfCSP to HBsAg ratio increases to 1:1. R21 administered with a saponin-based Matrix-M adjuvant has been tested in a phase 2 clinical trial and was found to have over 75% efficacy after a primary series of three immunizations. 15 Importantly, R21 has shown similar levels of protection when a booster immunization is used 1 year after the initial immunization regimen 16 suggesting that changes in the vaccine delivery system and increasing the density or exposure of relevant antigens have a significant impact on the outcome. However, current PfCSP-based vaccines do not include highly relevant epitopes upstream from the coding region of the RTS,S/R21 vaccines that are the target of protective neutralizing antibodies, supporting the concept that these vaccines can be further improved. 9 The protective antibodies elicited by RTS,S vaccines primarily target the (NANP) n repeat region of PfCSP. Recent studies, however, have identified a junctional epitope consisting of alternating NVDP and NANP repeats that is present between the N-terminal region and the (NANP) n repeats in PfCSP and is a target of neutralizing antibodies. [17][18][19][20][21] Recent studies by Tan et al. 22 and Kisalu et al. 23 suggested that dual-specific antibodies targeting both this junctional epitope and the (NANP) n repeat region are more potent than the antibodies targeting only (NANP) n . Results from immunization studies conducted with junctional epitope-based vaccines suggest the potential value of including this epitope in a PfCSP-based vaccine. 21,24 Calvo-Calle et al. 21 immunized mice with a T1 (DPNANPNVDPNANPNV)-based vaccine, that contains two copies of the junctional epitopes DPNANP, NPN, and NPNV, and found higher neutralizing activity as compared to a (NANP) n repeat-based vaccine. Jelínková et al. 24 used the Qβ virus-like particle (VLP) platform to develop a junctional epitope-based vaccine and reported high antibody titers and long-lasting anti-PfCSP antibody responses. We have also previously evaluated the responses of subunit peptide vaccines and VLPs containing the T1 epitope. 25,26 These studies have demonstrated that these immunogens induce high antibody titers in mice, non-human primates, and human volunteers.
In the present work, we developed a chimeric PfCSP (cPfCSP) displayed multivalently on SpyCatcher-mi3 VLPs as a strategy to generate a robust and durable immune response targeting both the T1/junctional epitopes and the (NANP) n repeats. Based on previous studies with Plasmodium vivax CSP, 27  PfCSP. 35 Further, we hypothesized that we could enhance the Bcell response by displaying this version of PfCSP multivalently from VLPs. We chose SpyCatcher-mi3 VLPs for the multivalent display of antigens. [36][37][38][39][40] These VLPs allow for the conjugation of proteins incorporating a SpyTag peptide sequence through the formation of an isopeptide bond, making them a versatile and efficient platform. 36,38,41 Adding the SpyTag peptide to the C-terminus of cPfCSP allowed us to conjugate multiple copies of cPfCSP to SpyCatcher-mi3 ( Figure 1b). We immunized mice with cPfCSP-Spy-Catcher-mi3 using a simplified two dose regimen and characterized the elicited antibody response. cPfCSP-SpyCatcher-mi3 elicited a balanced response with high antibody titers against both the junctional epitope and the (NANP) n repeats. Importantly, the antibody response was durable, with high antibody levels maintained for more than 9 months after the second immunization.

| RESULTS AND DISCUSSION
We expressed cPfCSP, having a SpyTag sequence followed by a 6x-His tag at the C-terminus in Escherichia coli.  Next, we generated SpyCatcher-mi3 nanoparticles-selfassembled 60-mer virus-like particles based on a mutated aldolase protein from a thermophilic bacterium 36  nanoparticles was 48 nm, as compared to 34 nm for SpyCatcher-mi3, confirming that the SpyTag-SpyCatcher reaction was successful.
Next, to verify that the cPfCSP and cPfCSP-SpyCatcher-mi3 retained their antigenicity, we characterized the binding of cPfCSP and the conjugated product to dual-specific antibodies that bind to both the T1/junctional region and the (NANP) n repeats by an enzymelinked immunosorbent assay (ELISA). Figure 2c shows the successful recognition of cPfCSP and cPfCSP-SpyCatcher-mi3 by the CIS43 23 (PDB: 6B5M) and 2A10 42 (PDB: 5T0Y) antibodies.
We next characterized the immunogenicity of cPfCSP-Spy-Catcher-mi3 in mice. We subcutaneously inoculated mice (n = 5) with cPfCSP-SpyCatcher-mi3 or SpyCatcher-mi3 alone (control), each mixed with an equal volume of AddaVax, a vaccine adjuvant consisting of a squalene-based oil-in-water nano-emulsion. Addavax is similar to MF59, which has been licensed for use in humans, 44,45 and we have used it successfully in our previous work. 37,46-48 A boost was carried out 21 days later. Mice were bled 20 days after the initial immunization to characterize the immune response. The schedule for mice vaccination and serum collection is shown in Figure 3a. The collected sera were tested against two synthetic peptides-the T1 peptide and a peptide composed of three NANP repeats (B3)-as well as a full-length recombinant PfCSP expressed in E. coli. After a single immunization with cPfCSP-SpyCatcher-mi3, IgG antibody titers against PfCSP had a geometric mean titer (GMT) of $132,000 that increased by more than 24-fold after the boosting immunization to 3,200,000 (Figure 3b). Consistent with high antibody titers against the recombinant protein, high anti-peptide IgG antibody titers were also elicited. GMT against T1 rose more than 48-fold, from 50,000 to 2,425,147, and GMT against B3 increased more than 18-fold, from $43,500 to 800,000 (Figure 3b).
It has been reported that a 200 μg/mL concentration of the monoclonal antibody 2A10, which has been reported to recognize both (NANP) n and T1, 21 protected a majority of mice against experimental infection with a transgenic rodent malaria parasite expressing the full-length PfCSP. 49 Although differences in this protective serum concentration threshold of 2A10 have been described, 23 suggesting that it is dependent on the parasite-host combination used, we decided to calibrate the antibody response elicited by immunization with cPfCSP-SpyCatcher-mi3 by estimating the levels of anti-repeat region antibodies using a standard curve that was generated using serial dilution with the 2A10 antibody. Immunization with cPfCSP-SpyCatcher-mi3 induced high anti-repeat antibody concentrations that ranged between 305 and 1014 μg/mL, suggesting the induction of high titers of sporozoite-neutralizing antibodies ( Figure 3c). 49 We also used an ELISA competition assay to characterize the specificity of antibodies elicited by immunization with cPfCSP-Spy-Catcher-mi3. The (NANP) 3 peptide inhibited close to 50% of the serum binding to PfCSP at 50 μg/mL concentration (Figure 3d). The T1 peptide inhibited greater than 60% of the binding of the serum to the recombinant PfCSP at 50 μg/mL (Figure 3d). As expected, both synthetic peptides also inhibited the binding of 2A10 antibody to PfCSP (Figure 3d). These results suggest that the immunization has elicited the desired balanced response targeting both the T1/junctional and (NANP) n epitopes within PfCSP.
Finally, to characterize the longevity of the antibody response elicited by immunization with SpyCatcher-mi3-cPfCSP, we estimated the levels of anti-repeat region antibodies in sera 118, 188, and 300 days after the first immunization (i.e., 97, 167, and 279 days after the booster immunization). As seen in Figure 3e, anti-repeat antibody levels were maintained above the protective threshold 49 even on day 300, more than 9 months after the booster immunization. We note that there is variation in anti-repeat antibody levels F I G U R E 3 Characterization of immunogenicity elicited by cPfCSP-SpyCatcher-mi3. (a) Schedule for mice vaccination and serum collection. (b) Antibody endpoint titers of sera from mice immunized with cPfCSP-SpyCatcher-mi3 (gray), or SpyCatcher-mi3 (white) (geometric mean with geometric SD, n = 5) (i) after the first immunization (day 20) and (ii) after the second immunization (day 36) against peptides T1, B3 ((NANP) 3 ), and PfCSP: one assay with sera from 5 mice; n = 5. ****P < 0.0001, determined by unpaired t test for (i) and, unpaired t test (against T1 and B3), and Welch's t test (against PfCSP) for (ii). (c) Anti-T1(NANP) 3 IgG antibody concentration in mice immunized with cPfCSP-SpyCatcher-mi3 was determined by comparison to a standard curve generated with the mAb 2A10; sera from mice immunized with SpyCatcher-mi3 alone were used as control. The broken line indicates the protective anti-repeat antibody level reported to confer protection in murine models: one assay with sera from 5 mice; n = 5; **P < 0.01, determined by Mann-Whitney test. (d) Characterization of antibody specificity. 5 ng/mL mAb 2A10 (▲) or a pool of sera collected from mice immunized with SpyCatcher-mi3-cPfCSP diluted at 1:60,000 (•) were pre-incubated with three different concentrations (0.5, 5, and 50 μg/mL) of (i) (NANP) 3 (B3) and (ii) T1 competitor peptides and then allowed to bind to PfCSP. The intensity was normalized relative to that in the absence of competing peptides. (e) Anti-T1(NANP) 3 IgG antibody concentration in mice 118, 188 and 300 days after first immunization with cPfCSP-SpyCatcher-mi3 or SpyCatcher-mi3 (control) was determined by comparison to a standard curve generated with the mAb 2A10. The broken line indicates the protective anti-repeat antibody level reported to confer protection in murine models; one assay with sera from 5 mice; n = 5; *P < 0.05, determined by Welch's t test; **P < 0.01, determined by Mann-Whitney test at 118 days, and Welch's t test for 188 and 300 days. from mouse to mouse. We have consistently observed variations in the antibody responses elicited by malaria vaccines in animal models and humans. 25,28,50 Collectively, the data show that immunization with SpyCatcher-mi3-cPfCSP elicited balanced, potent, and longlasting antibody responses against the relevant target epitopes in PfCSP.

| CONCLUSIONS
In this study, we have successfully designed and produced multivalent cPfCSP-SpyCatcher-mi3 nanoparticles. The chimeric multivalent vaccine construct incorporates the T1/junctional region, an epitope not included in RTS,S or R21 vaccines, that is the target of potent neutralizing antibodies. cPfCSP-SpyCatcher-mi3 also includes a reduced number of (NANP) n repeats to reduce its described immunodominance reported as detrimental for the response to subdominant epitopes within PfCSP. 35 cPfCSP-SpyCatcher-mi3 was immunogenic in mice eliciting high and durable IgG antibody titers with a twoimmunization regimen. Moreover, immunization elicited antibody responses against the T1/junctional region as well as the (NANP) n repeats. Importantly, the antibody concentration elicited by immunization was greater than the threshold for protection defined in a murine challenge model. 49   following a similar approach that we used to design and express a recombinant chimeric P. vivax CSP (cPvCSP). 27 The protein incorporates ( Figure 1a): (1) The segment D 82 -P 104 recognized by the protective monoclonal antibody 5D5 52 that includes the proteolytic cleavage site 17,18 and Region I (RI) and the stretch of positively charged residues upstream of RI, which contain the binding domain to heparan sulfate involved in the attachment of sporozoites to hepatocytes. 53,54 (2) The junctional domain T1, recognized by the protective monoclonal antibody CIS43. 23 (3) (NANP) 6 representing the major repeat region with a reduced copy number to avoid the reported induction of short-lived plasmablasts since the native protein with a high copy number of repeats acts as a T cell-independent antigen. 55 (4) The immunodominant Th2R/T* (P 311 -N 340 ), Th3R, and a C-terminal T cell epitope mapped in murine and human models (G 341 -N 397 ) 56,57 separated by GPGPG spacers as we have described with other Plasmodium chimeric proteins. 27,[58][59][60] The sequence representing Th3R is an extended version that includes the H-2K k restricted CTL epitope DYENDIEKKI (D 359 -I 368 ). 61,62 4.2 | Cloning of cPfCSP constructs, anti-CSP antibodies, and SpyCatcher-mi3

| cPfCSP
The gene encoding cPfCSP followed by a GPGPG spacer, a SpyTag, and a 6x His-Tag was optimized for bacterial expression, synthesized, and cloned into the pET-28b vector by Gene Universal Inc.

| SpyCatcher-mi3
The gene encoding the SpyCatcher-mi3 fusion protein 36   The following steps were performed at 4 C. On the next day, the cells were pelleted by centrifuging at 7000g for 10 min. They were then resuspended in 20 mL of IMAC binding buffer (100 mM Tris, 150 mM NaCl, 20 mM imidazole, pH 8.0) containing 0.5 mg/mL lysozyme, 125 units of benzonase, and half a tablet of EDTA-free protease inhibitor cocktail (Fisher Scientific). A 5% sodium deoxycholate solution (Alfa Aesar) was added to the cell suspension before sonicating it twice for 3 min at 30% amplitude with 3 s on and 3 s off pulses.

| Expression and purification of cPfCSP
The cell suspension was then centrifuged at 12,000g for 30 min to separate the supernatant containing cPfCSP from the cell debris.

| Expression and Purification of Anti-CSP antibodies
The antibodies (CIS43 and 2A10) were expressed in HEK293F cells using the manufacturer's protocol. In brief, the constructs were transfected using the ExpiFectamine™ 293 transfection kit (Gibco), and the antibodies were harvested 5 days later.
The antibody cultures were centrifuged at 7000g for 7 min, and the supernatant was dialyzed overnight against PBS (137 mM NaCl, 2.7 mM KCl, 8 mM Na 2 HPO 4 , and 2 mM KH 2 PO 4 ). The CIS43 and 2A10 antibodies were purified using a MabSelect SuRe column (GE) based on the manufacturer's protocol. Purified antibodies were then concentrated using Amicon spin filters and stored in 30% glycerol at À20 C for future use.

| Conjugation of cPfCSP to SpyCatcher-mi3 nanoparticles
An optimum stoichiometric ratio for the reaction between cPfCSP and SpyCatcher-mi3 was determined by setting up a series of small-scale reactions. cPfCSP and SpyCatcher-mi3 were mixed and allowed to react overnight at 4 C. The optimal ratio was determined by characterizing the reaction mixtures using SDS-PAGE. The gels were then destained and imaged using the ChemiDoc MP imaging system (Bio-Rad).

| Dynamic light scattering
One hundred microliters samples of cPfCSP, SpyCatcher-mi3, and cPfCSP-SpyCatcher-mi3 at a concentration of $0.5 mg/mL were added to a UVette (Eppendorf). Dynamics software and a DynaPro NanoStar Dynamic Light Scattering detector were used to collect three acquisitions for each measurement. Acquisitions were averaged, and results were displayed as % Mass using the Isotropic Sphere model.

| ELISA
To determine the immunoreactivity of the proteins by ELISA, the wells of a 96-well plate were coated with 0.1 μg of protein suspended in 100 μL of PBS and incubated at room temperature for 1 h. The wells were then blocked with 5% (w/v) bovine serum albumin (BSA) solution in 0.1% tween-added PBS (PBST). After incubating for 1 h and washing three times with PBST, 100 μL of the corresponding primary antibodies-CIS43 or 2A10-in PBST with 1% BSA were added to the wells. After three PBST washes following a 1 h incubation, 100 μL of the corresponding diluted secondary antibody was added to the wells.
The specificity of antibodies elicited by immunization was also determined by ELISA using Immulon 2HB ELISA plates (Thermo Scientific, Waltham, MA) coated with 1 μg/mL of the synthetic peptides T1 or B3 diluted in PBS or full-length PfCSP kindly provided by Gennova (Pune, India) 63 diluted in carbonate buffer and incubated overnight at 4 C. The synthetic peptides used for coating have been previously described and synthesized incorporating cysteine residues at the amino and carboxyl-terminal ends with the topology Cys-T1-Cys or Cys-B3-Cys. 58 The wells were then blocked with PBS-1% BSA blocking solution (SeraCare) for 2 h at 37 C. Sera samples from immunized mice were serially diluted in PBS-0.5% BSA containing 0.05% Tween-20 and incubated for 1 h at 37 C. Reactivity was determined using HRP-labeled goat anti-mouse IgG (SouthernBiotech) and SureBlue Reserve TMB substrate (SeraCare). Optical densities were determined using a 450 nm filter on a VERSAmax ELISA reader (Molecular Device Corporation, Sunnyvale, CA). The cutoff value was set at the highest dilution of sera resulting in an O.D. greater than three standard deviations (SD) above the mean obtained using sera from unimmunized mice. ELISA results are presented as the reciprocal of the endpoint dilution. To determine anti-PfCSP IgG concentration, standard curves for the monoclonal antibody 2A10 were generated using 10 twofold dilutions of the monoclonal Ab starting at 100 ng/mL. Sera samples were tested at 1:50,000, 1:100,000, and 1:200,000 dilutions. The concentrations were determined using GraphPad's four-parameter logistic (4PL) curves. For the ELISA competition assays, sera samples were tested at 1:30,000, 1:60,000, and 1:120,000 dilutions and preincubated with T1 or B3 synthetic peptides without cysteine residues at 0.5, 5, or 50 μg/mL. Samples were then tested for binding to plates coated with PfCSP.

| Mice and immunization regimens
Six-to eight-week-old female C57BL/6 mice (The Jackson Laboratory) were obtained and housed in micro-isolation cages. All animal experiments and procedures were performed following guidelines and approved by the Emory University Institutional Animal Care and Use Committee. Mice were randomized into two groups of 5 mice each.
Mice in group 1 received subcutaneous immunizations with cPfCSP-SpyCatcher-mi3 (2.57 μg of the cPfCSP per mouse), emulsified at a 1:1 ratio with Addavax adjuvant (InvivoGen, San Diego, CA) on days 0 and 21. Mice in group 2 were immunized with the 33.9 μg of SpyCatcher-mi3 per mouse (the same amount of SpyCatcher-mi3 as in group 1), also emulsified at a 1:1 ratio with Addavax. Sera samples were collected 24 h before each immunization, and 36, 118, 188, and 300 days after the first immunization to determine antibody responses. Antibody responses were assessed as total IgG antibody titers and determined by ELISA as described. 59