Whole-Proteome Differential Screening Identifies Novel Vaccine Candidates for Schistosomiasis japonica

Abstract

Schistosomiasis remains a leading cause of chronic morbidity in endemic regions despite decades of widespread mass chemotherapy with praziquantel. Using our whole proteome differential screening approach, and plasma and epidemiologic data from a longitudinal cohort of individuals living in a Schistosoma japonicum–endemic region of the Philippines, we interrogated the parasite proteome to identify novel vaccine candidates for Schistosoma japonicum. We identified 16 parasite genes which encoded proteins that were recognized by immunoglobulin G or immunoglobulin E antibodies in the plasma of individuals who had developed resistance to reinfection, but were not recognized by antibodies in the plasma of individuals who remained susceptible to reinfection. Antibody levels to Sj6-8 and Sj4-1 measured in the entire cohort (N = 505) 1 month after praziquantel treatment were associated with significantly decreased risk of reinfection and lower intensity of reinfection over 18 months of follow-up.

We identified 16 vaccine candidates related to resistance to Schistosoma japonicum infection using a whole-proteome differential screening approach. Our immunoepidemiologic data suggest that Sj6-8 and Sj4-1 merit further evaluation for their vaccine potential for S. japonicum.

Schistosomiasis, caused by 3 principal species of dioecious trematodes (flatworms), currently infects >200 million individuals, results in an estimated 2%–15% chronic disability, and contributes to poor health and economic stagnation in endemic areas [1]. Although schistosomiasis is effectively treated with praziquantel (PZQ), rapid reinfection coupled with insufficient compliance in mass drug administration programs precludes effective population-level control and justifies current efforts to develop vaccines for these parasites [1, 2].

While almost a quarter of the >100 vaccine candidates currently under investigation generate some level of protection in murine models, despite decades of development, only 1 has advanced to published phase 3 trials [3–6]. New antigen candidates are urgently needed, but, given concerns regarding the relevance of the murine model for vaccine development [7], strategies to identify novel antigens are limited. Although protective antibody-mediated resistance in humans has been described for all 3 species of schistosomes [8, 9], surprisingly few studies have capitalized on this immunity for vaccine antigen identification [5, 10].

In previous vaccine antigen discovery studies, we pioneered a differential, whole-proteome screening method using plasma and epidemiologic data from a cohort of Kenyan adolescents and Tanzanian children, to identify antigens associated with resistance to human falciparum malaria [11–13]. In the present study, we have applied this differential screening method using serum samples and parasitologic data collected from an extensively characterized longitudinal treatment-reinfection cohort of 7- to 30-year-old volunteers living in a Schistosoma japonicum–endemic region of the Philippines (n = 641) [14–18]. We identified 16 antigens that were uniquely recognized by antibodies expressed by highly resistant, but not by susceptible, individuals. Antibody levels to Sj6-8 and Sj4-1 measured in the entire cohort 1 month after PZQ treatment were associated with resistance to reinfection over 18 months of follow-up.

MATERIALS AND METHODS

Study Population

Cohort subjects were recruited from 3 rice-farming villages in Jaro, Leyte, the Philippines, as described previously [16]. A total of 641 eligible individuals between 7 and 30 years old were enrolled. At the time of enrollment, the community prevalence for this age range was 60% [17]. Active schistosome infection was assessed by 3 consecutive Kato–Katz stool examinations performed in duplicate and expressed as the average eggs per gram of feces (EPG). All study participants were treated with a split dose of 60 mg PZQ per kilogram of body weight at the start of the study period. Phlebotomy was performed 1 month after treatment using Vacutainer serum separation tubes (Becton Dickinson, Franklin Lakes, New Jersey). Serum was aliquoted and shipped to the Center for International Health Research on dry ice and stored at –80°C until use. Participants were included in the analysis (N = 505) if they provided 3 stool samples at baseline, tested negative 1 month after PZQ treatment, and had at least 1 stool examination during the 18-month follow-up period.

Subjects were assessed for reinfection every 3 months for 18 months with 3 Kato–Katz stool examinations performed in duplicate. Schistosoma japonicum and geohelminth (Ascaris lumbricoides, Trichuris trichiura, and hookworm) eggs were enumerated on each Kato–Katz slide. Each participant was scheduled for water contact observation on 12 separate occasions during the 18 months of follow-up as previously described [16]. A cumulative exposure index was calculated by averaging the individual’s available total water contact scores. Socioeconomic status was assessed by questionnaire as described previously [15].

The study was approved by the institutional review boards of Brown University and the Philippine Research Institute for Tropical Medicine.

Selection of Resistant and Susceptible Individuals for Differential Screening Assays

Using our whole-proteome differential screening method [11–13] and serum and epidemiologic data from our longitudinal cohort [19], we performed 3 independent library screens using 3 independent sets of serum pooled from individuals selected based on their resistance or susceptibility to reinfection with S. japonicum. Relative resistance was determined based on the S. japonicum EPG assessed by Kato–Katz examination of stool specimens collected at 3, 6, 9, 12, 15, and 18 months after PZQ treatment.

We made 3 independent selections of resistant and susceptible individuals to comprise our differential serum sets. For each differential set, we rank-ordered individuals based on their mean S. japonicum EPG on all stool samples collected during the 18-month follow-up period. Individuals from the low and high extremes of this distribution were chosen to comprise the resistant (n = 10) and susceptible (n = 10) groups. For all 3 serum sets, selections were made with matching for age, baseline S. japonicum intensity, sex, and cumulative exposure index. For the third set, we restricted our selections to individuals who had levels of antisoluble worm antigen preparation (SWAP) immunoglobulin E (IgE) reactivity above the mean [19]. Selected individuals’ serum was pooled to make resistant serum pool (RP) and susceptible serum pool (SP), respectively. No study subject was a member of >1 differential serum set.

Whole Proteome Differential Screening

We performed 3 differential screening experiments using both a lambda phage–based S. japonicum complementary DNA (cDNA) library as well as a T7 phage display–based library as described in the Supplementary Materials and Methods.

In Silico Analysis

We performed nucleotide-based BLAST (https://parasite.wormbase.org/) searches for the clones identified and sequenced in our differential screens. Using the complete cDNA sequence, we performed blastx (UniProtKB/Swiss-Prot) to identify and align with homologues. We evaluated each clone for signal peptides (http://www.cbs.dtu.dk/services/SignalP/index.php), transmembrane motifs (http://topcons.cbr.su.se), and predicted function using both literature searches as well as the Gene Ontology database (http://geneontology.org).

Characterization of Candidate Antigens

We codon optimized and cloned the open reading frames encoded by the 16 uniquely reactive clones into the prokaryotic expression plasmid pJ411 (http://www.atum.com) and, using a combination of metal chelate, hydrophobic interaction, anion exchange, and size exclusion chromatography, successfully purified 10 of these recombinant proteins according to our published protocols [12, 13]. We generated antigen-specific antisera in mice and performed Western blots to confirm the presence of candidate proteins in crude worm lysates of S. japonicum adult worm antigen (SWAP), soluble egg antigen (SEA), and/or excretory/secretory proteins (ESP). Immunolocalization of Sj6-8 was performed by immunofluorescence microscopy on mechanically transformed schistosomula and adult worms. We used dicer-substrate RNA (DsiRNA) and RNA interference technology (RNAi) to knock down Sj6-8 in vivo (Supplementary Materials and Methods).

Isotype-Specific Antibody Assays

We detected antigen specific immunoglobulin G (IgG), IgG₁, IgG₂, IgG₃, IgG₄, and immunoglobulin E (IgE) using a bead-based platform (Bio-plex; Bio-Rad, Hercules, California). We covalently bound 100 μg of rSj97, rSj4-1, rSjALD, rSjERBP1a, rSjERBP1b, rSjG10, rSjF1, rSjF2, rSj6-8, rSj191, SEA, SWAP, and Bovine Serum Albumin to 1.25 × 10⁷ microspheres per the manufacturer’s protocol (Luminex, Austin, Texas). We followed our previously described protocol for antibody assay optimization and performance [19]. Optimized serum dilutions were 1:20 for IgE and 1:100 for IgG, IgG₁, IgG₂, IgG₃, and IgG₄. Optimized biotinylated detection antibodies (Pharmingen, San Diego, California) were 1:1000 for IgE, IgG₁, IgG₂, and IgG₃ and 1:5000 for IgG₄ and IgG. Streptavidin-phycoerythrin was used at 1:500 dilution (Pharmingen). Fluorescence was quantified using the BioPlex analyzer (Bio-Rad). All liquid handling was performed by a high-speed pipetting robot (Genesis; Tecan, Research Triangle Park, North Carolina).

Statistical Analysis

RNAi-mediated knockdown of Sj6-8 gene expression and its impact on worm fecundity were evaluated with analysis of variance followed by Tukey post-hoc comparison for multigroup comparisons, or Student t test for 2 group comparisons (GraphPad Prism version 8, GraphPad Software, La Jolla, California). To assess the relationship between antigen-specific antibody responses and the risk and intensity of reinfection, we developed generalized estimating equation (GEE)–based repeated-measures models for each antigen and isotype-specific antibody measurement, which was coded as a continuous variable (SAS version 9.4 software, SAS Institute, Cary, North Carolina). In each of these models, potential confounders and effect modifiers, including age, sex, water contact, socioeconomic status, and natural transformed intensity of geohelminths, were evaluated in a backward-elimination process until only variables with P values < .1 remained. To facilitate comparisons across the final models, potential confounders were retained if they were significant in more than one-third of the GEE models evaluated. We also evaluated these models with antibody levels coded as highest vs lowest quartile. A P value < .05 was considered statistically significant.

RESULTS

Identification and In Silico Evaluation of Vaccine Candidates

In the first screen, to minimize bias based on water exposure, we restricted our selection of resistant and susceptible individuals to those who were in the upper tertile for water exposure (n = 140) as assessed by direct observation (Supplementary Table 1). We differentially screened 500 000 clones from a S. japonicum adult worm cDNA library [20] to identify S. japonicum proteins uniquely recognized by IgG antibodies contained in the pool of resistant serum but not in the pool of susceptible serum. We identified 477 clones recognized by both RP and SP, 5 clones recognized only by SP, and 2 clones uniquely recognized by RP.

In the second screen (Supplementary Table 2), we differentially screened 400 000 clones from the same cDNA library used in the first screen. We identified 156 clones recognized by both RP and SP, 73 clones recognized only by SP, and 9 clones uniquely recognized by RP.

In the third screen (Supplementary Table 3), we restricted our selection to individuals with higher-than-average IgE anti-SWAP antibody levels. We differentially biopanned 7 × 10⁷ phages, sequenced 42 selected clones, and performed plaque lift assays. Among the 42 clones evaluated, 5 clones were uniquely reactive with IgE in RP but not IgE in SP.

The 16 clones that were identified in the 3 screening experiments as uniquely reactive with RP were sequenced (Table 1). Of these 16 genes, 11 have known function or high homology to genes of known function while the remaining 5 are of unknown function. Two genes contained signal sequences.

Table 1.

Bioinformatic Characteristics of the 16 Differentially Recognized Clones

Protein ID	Screening Method	Accession No.	Aligned Protein Name and Function	Cloned Region (AA)	SignalPa
Sj6-8	cDNA IgG	FN316753	Hypothetical protein	1–157	AA21-22
Sj4-1	cDNA IgG	FN317105	Saposin-like, IPR008139 SaposinB domain-containing protein	72–204	AA19–20
SjALDfl	PhD IgE	FN316243	Aldolas: Sm vaccine candidate	344–363	None
SjTPx-E7	PhD IgE	FN322861	TPx1: schistosomiasis vaccine candidate	137–207	None
SjSERBP1a	PhD IgE	FN321623	PAI-RBP1; product = Plasminogen activator inhibitor 1 RNA-binding protein	4–196	None
SjSERBP1b	PhD IgE	FN321620	PAI-RBP1; product = Plasminogen activator inhibitor 1 RNA-binding protein	4–278	None
SjF1a	cDNA IgG	FN321658	Ferritin-1 heavy chain	1–173	None
SjF2	cDNA IgG	FN314908	Ferritin light chain	1–172	None
Sj97	cDNA IgG	SJU11825	Paramyosin: Sj vaccine candidate	1–866	None
SjG51	cDNA IgG	FN314383	Zinc finger CCCH domain-containing protein 3	342–393	None
SjG52	cDNA IgG	FN329486	C-terminal of the “noncoding mRNA clone,” unknown function	128–242	None
SjG10	cDNA IgG	AY811267	SJCHGC04682 protein, unknown function	1–111	None
SjGDPDfl	PhD IgE	AY814571	Glycerophosphoryl diester phosphodiesterase	434–458	None
SjG191	cDNA IgG	FN320700	DEAD (Asp-Glu-Ala-Asp) box polypeptide 50	134–420	None
SjTSP-G194	cDNA IgG	FN314805	N-terminal of 25 KDA integral membrane protein; Tetraspanin family	1–224	None
SjG2332	cDNA IgG	AY813208	SJCHGC05456 protein, hypothetical protein Smp_194650	38–188	None

Abbreviations: AA, Amino Acid; cDNA, complementary DNA; IgE, immunoglobulin E; IgG, immunoglobulin G; mRNA, messenger RNA; PhD, phage display.

^aWebsite access: http://www.cbs.dtu.dk/services/SignalP/.

Expression, Purification, and Antibody Response to Vaccine Candidates

We codon optimized and cloned the open reading frames encoded by the 16 uniquely reactive clones into the prokaryotic expression plasmid pJ411 (http://www.atum.com) and, using a combination of metal chelate, hydrophobic interaction, anion exchange, and size exclusion chromatography, successfully purified 10 of these recombinant proteins (Supplementary Figure 1 and Supplementary Table 4).

Antibodies to Sj6-8 and Sj4-1 Are Associated With Resistance to Reinfection

To evaluate the impact of naturally acquired antibodies to the vaccine candidates identified in our differential screening on risk and intensity of reinfection with S. japonicum, we measured IgG subclasses and IgE antibody levels to the 10 purified recombinant candidates using a fluorescent, bead-based assay in our longitudinal treatment reinfection cohort. Antibody levels were measured 4 weeks after PZQ treatment and we related these levels to subsequent reinfection data collected at 3, 6, 9, 12, 15, and 18 months posttreatment (N = 505).

When analyzed as a continuous variable, increasing levels of IgE anti-Sj6-8 were associated with lower intensity of reinfection. Similarly, levels of IgE and IgG₃ anti-Sj4-1 were also associated with lower intensity of reinfection. Increasing levels of IgE and IgG₃ anti-Sj4-1 were associated with decreased risk of reinfection (Supplementary Table 5). These results remained significant even after adjusting for sex, socioeconomic status, geohelminths intensity, and directly observed water contact (Table 2).

Table 2.

Incidence and Intensity of Reinfection With Schistosoma japonicum Is Predicted by Antibody Responses to Sj6-8, SjERBP1b, and Sj4-1 Using Generalized Estimating Equation Models^a

Outcome and Antibody	Sj6-8		SjERBP1b		Sj4-1
	B or RR (95% CI)b	P Value	B or RR (95% CI)b	P Value	B or RR (95% CI)b	P Value
Intensity of infection, EPG
IgE	–0.097 (–.177 to –.018)	.016	–0.174 (–.490 to .142)	.28	–0.350 (–.554 to –.146)	.001
IgG	0.243 (.067–.419)	.007	0.305 (.117–.494)	.001	0.014 (–.211 to .239)	.90
IgG1	0.063 (–.086 to .212)	.40	0.225 (.057–.393)	.009	–0.364 (–1.016 to .288)	.27
IgG2	0.117 (.005–.228)	.040	0.160 (.037–.284)	.011	0.048 (–.067 to .164)	.41
IgG3	0.018 (–.074 to .109)	.71	0.083 (–.032 to .198)	.16	–0.161 (–.275 to –.047)	.006
IgG4	0.227 (.130–.324)	<.001	0.268 (.169–.368)	<.001	0.075 (.007–.144)	.031
Reinfection (yes vs no)
IgE	0.93 (.86–1.01)	.09	0.89 (.63–1.25)	.50	0.71 (.58–.86)	<.001
IgG	1.20 (.99–1.44)	.06	1.33 (1.10–1.61)	.003	0.92 (.75–1.13)	.42
IgG1	1.11 (.98–1.25)	.11	1.23 (1.04–1.45)	.015	0.68 (.34–1.39)	.29
IgG2	1.14 (1.02–1.27)	.020	1.21 (1.07–1.36)	.002	1.04 (.94–1.15)	.44
IgG3	1.02 (.94–1.11)	.62	1.11 (.99–1.24)	.06	.82 (.74–.92)	<.001
IgG4	1.23 (1.10–1.38)	<.001	1.25 (1.14–1.38)	<.001	1.07 (1.00–1.14)	.043

Abbreviations: CI, confidence interval; EPG, eggs per gram of stool; IgE, immunoglobulin E; IgG, immunoglobulin G; RR, relative risk.

^aAntibody levels and intensity of reinfection were natural log-transformed and generalized estimating equation models were adjusted for sex, socioeconomic status, and Trichuris trichiura and hookworm intensity.

^b B is the unstandardized regression coefficients of reinfection intensity.

When analyzed dichotomously (highest quartile of antibody responders vs lower levels), individuals with high isotype-specific antibody levels to these antigens had 26.5% to 38.7% lower intensity of reinfection over 18 months of follow-up (Figure 1A), even after adjusting for sex, socioeconomic status, geohelminths intensity, and directly observed water contact (Figure 1B).

Figure 1.

Isotype-specific antibody responses to Sj6-8 and Sj4-1 predict decreased intensity and incidence of reinfection with Schistosoma japonicum. In generalized estimating equation models, individuals with high levels (upper quartile) of anti-Sj6-8 immunoglobulin E, anti-Sj4-1 IgE, and anti-Sj4-1 immunoglobulin G₃ predict lower refection intensity (A) and incidence (B) compared to individuals with lower levels, even after accounting for sex, socioeconomic status, intensity of Trichuris trichiura and hookworm infection, and directly observed water contact. Abbreviations: EPG, eggs per gram of stool; IgE, immunoglobulin E; IgG₃, immunoglobulin G₃.

Stage-Specific Expression and Localization of Sj6-8

Based on the relationship between anti-Sj6-8 antibodies and resistance to reinfection with S. japonicum (see below), we characterized the stage-specific expression of Sj6-8 and determined its localization in adult worms and schistosomula. In Western blot analysis, murine polyclonal anti-rSj6-8 recognized a 25 kDa protein in extracts of adult worms, but not in extracts prepared from eggs (Supplementary Figure 2A). To determine whether Sj6-8 is released from the surface of adult worms, we performed Western blots on excretory/secretory supernatant collected from adult worms cultured in vitro. Murine polyclonal anti-rSj6-8 recognized a 25 kDa protein in excretory/secretory supernatant, consistent with the release of this tegumental protein (Supplementary Figure 2B). The apparent molecular weight was substantially greater than the 16 kDa predicted size based on its amino acid composition, consistent with its acidic composition [21].

To determine whether Sj6-8 is located on the exofacial surface of schistosomula, we performed immunofluorescence assays on live, mechanically transformed schistosomula probed with anti-rSj6-8. Living, motile schistosomula did not label with anti-rSj6-8, while nonmotile schistosomula showed abundant staining within the parenchyma of the parasite (Figure 2). In contrast, Sj6-8 did localize to the tegument and gut epithelium in freshly perfused, live, motile adult worms probed with anti-rSj6-8 (Figure 3).

Figure 2.

Immunolocalization of Sj6-8 in mechanically transformed Schistosoma japonicum schistosomula. Schistosomula were incubated with Sj infected rabbit antisera (A and B) or anti-Sj6-8 rabbit antisera (C and D). Schistosomula were imaged by differential interference contrast (A and C) and fluorescence microscopy (B and D). Positive control specifically stained the surface of living schistosomula (green), while anti-Sj6-8 sera was nonreactive (black).

Figure 3.

Immunolocalization of Sj6-8 in Schistosoma japonicum adult worms. Adult S. japonicum worms were incubated with preimmune rabbit sera (A) or polyclonal anti-rSj6-8 antisera (B and C). Nuclei were stained with DAPI. Images are presented as maximum Z projections (2.0-µm-distance intervals) (B) or stacked (C).

RNAi-Based Knockdown of Sj6-8 Reduces Worm Fecundity

To investigate the biologic role of Sj6-8, we performed RNAi-directed knockdown in adult worms and evaluated the impact on expression levels of Sj6-8 transcripts, Sj6-8 protein, worm viability, and fecundity. Following Sj6-8–specific RNAi transfection, Sj6-8–specific messenger RNA transcripts levels, measured 4 days and 48 days after transfection, were reduced by 44.5%–58.9% and 52.8%–70.5%, respectively, compared to controls (Figure 4). The reduction in expression level of Sj6-8 transcript was paralleled by a reduction of 36.8% in protein levels (Supplementary Figure 3).

Figure 4.

RNA interference (RNAi)–mediated knockdown of Sj6-8 gene expression. Three unique dicer-substrate RNA (DsiRNA) fragments, namely Sj68 DsiRNA-1, -2, and -5 were used in the experiments. Control worms were electroporated with scrambled negative control DsiRNA. A, Relative gene expression (mean ± standard error of the mean [SEM]) in Schistosoma japonicum adult worms treated with DsiRNAs and cultured 4 days in vitro. Target gene suppression is significantly different (P < .001) from control with DsiRNA2 (41%, n = 4 worms) and DsiRNA5 (55%, n = 4 worms), but not DsiRNA1 (96%, n = 3 worms). B, Relative gene expression (mean ± SEM) in S. japonicum adult worms treated with DsiRNAs and cultured for 48 days in vitro. Target gene suppression is significantly different (P < .001) from control with DsiRNA2 (53%, n = 3 worms) and DsiRNA5 (70%, n = 3 worms). Data were evaluated by analysis of variance followed by Tukey post-hoc comparison.

We assessed the impact of Sj6-8 knockdown on worm viability as assessed by vital dye exclusion and motility. Knockdown of Sj6-8 did not impact worm viability assessed on days 0, 2, or 4 posttransfection (data not shown). Electroporation of worm pairs with Sj6-8 siRNA5 decreased Sj6-8 expression (Figure 5A) but did not reduce worm pair fecundity as assessed by the proportion of worms that remained paired (Figure 5B) compared to controls. However, knockdown of Sj6-8 expression was associated with a trend toward reduced egg production per pair of worms (Figure 5C).

Figure 5.

RNA interference–mediated knockdown of Sj6-8 gene expression decreases egg production, but not worm pairing. Schistosoma japonicum worm pairs were perfused on day 42 postinfection and electroporated with small interfering RNA (siRNA) or control (8 worm pairs per group) on days 0 and 3. Eggs per worm pair and number of paired worms were assessed on days 5–9. A, Bars represent the average eggs/worm pair on days 5–9. B, Bars represent the average number of paired worms on days 5–9. Error bars represent the standard error of the mean. Data were evaluated by Student t test.

Discussion

Schistosomiasis, caused by parasitic helminths of the genus Schistosoma, remains a major public health concern and currently infects >200 million individuals, with >800 million individuals at risk of infection in 78 developing countries. National control strategies focusing on mass chemotherapy with PZQ have significantly reduced severe liver and urinary tract pathology in many endemic areas. Unfortunately, despite decades of mass drug administration, rapid reinfection, poor compliance, and, in the case of S. japonicum, zoonotic vectors result in prevalence plateauing at unacceptably high levels. This residual prevalence and intensity of infection results in subtle morbidities such as anemia [17, 22, 23], malnutrition [14, 15, 18], and cognitive impairment [22, 24] which persist despite decades of annual mass treatment [25]. Continued exposure to contaminated water sources mandates alternative control strategies such as vaccine-linked chemotherapy [26].

With few exceptions, vaccine development for schistosomiasis has utilized rodent models of Schistosoma mansoni during the process of antigen identification. Following antigen discovery, cross-sectional and longitudinal analyses of human immune responses to these candidates have been utilized to evaluate their role as targets of protective human immunity. Despite decades of development, only 1 of the >100 candidates identified in animal models has advanced to published phase 3 trials [6]. This disappointing performance, coupled with the continuing burden of morbidity, mandates alternative approaches to identify vaccine antigens for schistosomiasis.

While naturally acquired protective immune responses in humans have been exploited as tools for vaccine antigen discovery [27], the overwhelming focus has been on rodent models for vaccine identification. This is particularly concerning because the types of acquired immune responses that generally appear to protect rodents (Th1) are different than the immune responses that appear to protect humans (Th2) [8, 9, 28–34], thus impugning the utility of rodent models for antigen discovery [35].

In previous antigen discovery studies, we pioneered a differential, whole proteome screening method using serum and epidemiologic data from a cohort of Kenyan adolescents and Tanzanian children [13], to identify antigens associated with resistance to human falciparum malaria [11–13]. In the current study, we utilized a similar approach to identify schistosome antigens recognized by IgE or IgG antibodies expressed by resistant but not susceptible individuals. Using serum and epidemiologic data collected from our longitudinal treatment-reinfection cohort living in an endemic area of the Philippines, we conducted 3 independent, differential cDNA library screens. For each screen, we selected and pooled serum collected 1 month after PZQ treatment from highly resistant or highly susceptible individuals with matching for potential nonimmunologic determinants of resistance, such as water contact. Resistance was quantified based on reinfection intensity after treatment.

Our screening experiments identified 16 genes that encoded proteins uniquely recognized by resistant individuals. IgE responses to Sj6-8 and IgE and IgG₃ responses to Sj4-1 measured 1 month after PZQ treatment were associated with significantly reduced intensity of reinfection measured over 18 months of follow-up. Similarly, IgE and IgG₃ responses to Sj4-1 measured 1 month after PZQ treatment were associated with significantly reduced incidence of reinfection measured over 18 months of follow-up.

Because anti-Sj6-8 antibody responses predicted decreased incidence and intensity of reinfection with S. japonicum, we explored its localization and evaluated its essentiality by conducting siRNA-directed knockdown in adult worms. Sj6-8 was not found on the surface of newly transformed schistosomula. In contrast, Sj6-8 localized to the tegumental surface of living adult schistosomes, including the gastrodermis. These data suggest that the relationship between anti-Sj6-8 and protection from reinfection seen in our longitudinal study is predicated on targeting adult worms, not migrating schistosomula.

Importantly, Sj6-8 was released into the supernatant of cultured adult worms. The potential for excretory/secretory proteins to serve as vaccine candidates for helminthiasis has been recognized for decades [36]. In schistosomes, excretory/secretory products contain several known vaccine candidates and have been used in successful schistosome vaccine trials [37–39]. The localization of Sj6-8 to the gastrodermis of both male and female worms suggests a role in nutrient acquisition, and we note that siRNA-directed knockdown of Sj6-8 expression produced a mild defect in fecundity.

Sj4-1 encodes a saposin B domain–containing protein with unknown function. Saposins are a family of small lipid-binding proteins that facilitate lysosomal lipid degradation. Some saposin-like proteins (SAPLIPs), such as surfactant protein B, are secreted and function in extracellular lipid biology [40]. A SAPLIP from Clonorchis sinensis is expressed in the gut and functions in host cell lysis and nutrient acquisition [41]. Two SAPLIPs have been identified in S. mansoni (Sm-SLP-1 and Sm-SLP-2) and localized to the gastrodermis [42]. Vaccination with Sm-SLP-1 did not confer resistance to infection with S. mansoni in mice. Sj4-1 has 26% identity with Sm-SLP-1 and retains the 6 conserved cysteines characteristic of the SAPLIP family. We are currently evaluating the localization of Sj4-1 and its potential role in erythrocyte lysis.

Several of our vaccine candidates are the target of protective IgE responses in our cohort studies. Vaccinating previously exposed individuals with these antigens raises concerns regarding vaccine-induced, IgE-mediated hypersensitivity in sensitized individuals. These concerns were highlighted following the induction of short-duration generalized urticaria without laryngotracheal involvement in 3 previously exposed individuals receiving an experimental hookworm vaccine [43]. Evidence from experimental models [44, 45] and epidemiologic surveys [46–48], however, indicates that a state of allergic hyporeactivity is observed in the context of schistosome infection. Specifically, a 5-fold reduction in the prevalence of skin test positivity was observed in S. mansoni–infected (n = 42) vs uninfected individuals (n = 133) [47]. Similarly, a 68% reduced prevalence of skin test positivity was observed in Schistosoma hematobium–infected (n = 207) vs uninfected individuals (n = 306) [48]. Despite these reassuring epidemiologic data, comprehensive evaluation for potential immunotoxicity would be necessary to advance an IgE-inducing vaccine candidate to phase 1 trials.

A potential limitation of our study was the use of Kato–Katz (3 stool specimens, each examined in duplicate) for the diagnosis of S. japonicum. In S. japonicum, the sensitivity of a point-of-care Circulating Cathodic Antigen (CCA) urine assay has been reported as “in the same order as that of a single Kato–Katz thick smear examination” [49]. The sensitivity “approached that of triplicate Kato–Katz when a combination of both CAA [Circulating Anodic Antigen] and CCA assays was used.” Similarly, for S. mansoni under field conditions, polymerase chain reaction (PCR) from a single stool has similar or slightly better sensitivity than Kato–Katz on 3 stool specimens [50]. We note that neither point-of-care CCA nor PCR was broadly available in 2002–2003 when our field samples were collected. Importantly, the same diagnostic approach was utilized throughout our study (Kato–Katz on 3 stools); thus, the assessments of reinfection remain internally valid. In addition, when selecting individuals to comprise our resistant and susceptible groups for differential library screening, the potentially low diagnostic sensitivity of Kato–Katz would have biased us away from identifying differentially recognized genes and is therefore not a threat to validity.

Together, our data validate our whole proteome differential screening strategy for the rational identification of vaccine candidates for schistosomiasis. Our data provide compelling evidence that Sj6-8 and Sj4-1 represent critical additions to the limited pool of candidate antigens currently being explored for their vaccine potential for S. japonicum.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. The authors thank all of the study participants in Leyte, the Philippines; they also thank the staff from the Immunology Department at the Research Institute of Tropical Medicine for their continuous support and assistance during the population survey and participant recruitment and treatment, as well as sample collection and treatment. We thank Ms. Ginny Hovanesian at the Lifespan Molecular Pathology Facility Core for technical assistance with confocal microscopy. We are grateful to Ms. Sarah Li for helping with Schistosomiasis japonicum lifecycle maintenance as well as O. hupensis snails provided by the NIAID Schistosomiasis Resource Center of the Biomedical Research Institute (Rockville, MD) through NIH-NIAID Contract HHSN272201700014I for distribution through BEI Resources.

Financial support. This work was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health (grant numbers 1R01AI1048123 and R01AI101274 to J. D. K.).

Potential conflicts of interest. All authors: No reported conflicts of interest.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Presented in part: 62nd Annual Meeting of the American Society of Tropical Medicine and Hygiene, Washington, D.C., November 2013. Abstract 1346.