Schistosomiasis treatment relies on the use of a single drug, praziquantel, which is insufficient to control transmission in highly endemic areas1. Novel medicines and vaccines are urgently needed2,3. An experimental human model for schistosomiasis could accelerate the development of these products. We performed a dose-escalating clinical safety trial in 17 volunteers with male Schistosoma mansoni cercariae, which do not produce eggs (clinicaltrials.gov NCT02755324), at the Leiden University Medical Center, the Netherlands. The primary endpoints were adverse events and infectivity. We found a dose-related increase in adverse events related to acute schistosomiasis syndrome, which occurred in 9 of 17 volunteers. Overall, 5 volunteers (all 3 of the high dose group and 2 of 11 of the medium dose group) reported severe adverse events. Worm-derived circulating anodic antigen, the biomarker of the primary infection endpoint, peaked in 82% of volunteers at 3–10 weeks following exposure. All volunteers showed IgM and IgG1 seroconversion and worm-specific cytokine production by CD4+ T cells. All volunteers were cured with praziquantel provided at 12 weeks after exposure. Infection with 20 Schistosoma mansoni cercariae led to severe adverse events in 18% of volunteers and high infection rates. This infection model paves the way for fast-track product development for treatment and prevention of schistosomiasis.
Worldwide, 290 million people are infected with schistosomes, mainly Schistosoma haematobium and Schistosoma mansoni4. The endemicity is determined by the presence of the fresh water snail intermediate host. Snail-derived cercariae penetrate the human skin and migrate into the vascular system, where mature male and female worms mate and produce ~300 eggs per day. S. mansoni eggs provoke inflammatory responses, which can lead to liver cirrhosis and portal hypertension5.
Current treatment and control of schistosomiasis relies on the use of a single drug, praziquantel. Mass drug administration with praziquantel does not protect from reinfection1 and thus provides insufficient control in highly endemic areas, creating the need for a vaccine2.
Several schistosome antigens have been put forward as possible S. mansoni vaccine candidates, of which three are in clinical development: Sm-TSP-2, rSm14/GLA-SE and Sm-p806,7,8. These candidates aim for >40% reduction in worm load for World Health Organization endorsement9, but higher levels are preferred10,11. To obtain efficacy data, large phase II and III field trials in Schistosoma-endemic areas are needed12. In addition, increasing concerns of praziquantel resistance create a need for anti-schistosomal drug development3. Controlled human infection (CHI) trials can select drug and vaccine candidates early in clinical development and help prevent late clinical failure12. We thus aimed to develop a schistosome CHI model to aid vaccine and drug development and better characterize human anti-schistosome immune responses.
Between September 2016 and January 2018, 35 healthy adult volunteers were screened, of which 17 were included in the trial and completed follow up. (Fig. 1). Baseline demographics between dose groups were comparable (Supplementary Tables 1 and 2).
Volunteers were exposed to 10, 20 or 30 cercariae in water on the forearm skin for 30 min, after which they were followed bi-weekly for adverse events and infectivity. After exposure, water was pipetted off the skin and inspected for remaining cercariae. We found tails of roughly half the number of the cercariae that they were exposed to, with clear differences between exposure groups (r = 0.70, P = 0.002; Supplementary Table 1).
The majority of volunteers (15 of 17, 88%) experienced pruritus during or after exposure, but no topical treatment was required (Supplementary Table 3). All but one volunteer (16 of 17, 94%) developed a mild local cercarial dermatitis within 2 d after exposure (Fig. 2a–c), which lasted longer in higher dose groups (10 cercariae: median 1 week (range 1–2); 20 cercariae: 3 weeks (1–9); 30 cercariae: 6 weeks (2–7), r = 0.45, P = 0.07).
There were no serious adverse events (AEs) (an event that is life-threatening or requires hospitalization), but nine severe related AEs (resulting in the inability to perform daily activity) were reported in volunteers from the 20 (n = 2) or 30 cercariae (n = 3) group. Seven of these severe AEs were symptoms of an acute schistosomiasis syndrome (n = 5). In the 30 cercariae group all volunteers (n = 3) experienced severe AEs, starting 2.5 to 5.0 weeks after exposure (Fig. 2d–f) as follows: headache (n = 2), fever (n = 2), syncope (n = 1), nausea (n = 1) and elevated liver enzymes (n = 1) (Supplementary Table 3). In one volunteer this episode was followed by mild to moderate headaches, malaise, fatigue and nocturnal sweats for up to 6 weeks as previously described13, but steroid treatment was declined. Given the burden of these prolonged symptoms of an acute schistosomiasis syndrome (Katayama symptoms), defined as moderate to severe symptoms of malaise, fatigue, fever, night sweats, flu-like symptoms or headache, between 2 to 7 weeks after exposure, the dose was de-escalated to 20 cercariae. Subsequently, 2 of 11 (18%) volunteers reported one severe symptom of an acute schistosomiasis syndrome (headache, nocturnal fever and sweats).
In addition to these five volunteers with severe AEs there were four volunteers, all exposed to 20 cercariae, with moderate symptoms of an acute schistosomiasis syndrome (flu-like symptoms, n = 3; and malaise, n = 2). Eight volunteers did not experience any symptoms of an acute schistosomiasis syndrome.
Eosinophils increased in 11 of 17 (65%) volunteers peaking between week 2 and 8 after infection (0.1–8.4 × 109 l−1; Extended Data Fig. 1). Eosinophils were not related to the dose or symptoms (Extended Data Fig. 2a).
Infection rates by antigen detection assays
Both worm-excreted circulating anodic antigen (CAA) and circulating cathodic antigen (CCA) were measured to determine the presence and degree of infection. In two of three volunteers exposed to 10 cercariae the serum CAA was higher than 1 pg ml−1. In the higher dose groups, 9 out of 11 (82%) volunteers exposed to 20 cercariae and all volunteers exposed to 30 cercariae crossed this threshold. The time to patency was comparable between these groups (range 3–8 weeks, Fig. 3a).
Cercarial dose and serum CAA levels were related (10 cercariae median at week 7–12: 0.4 pg ml−1 (range 0.3–0.8); 20 cercariae: 1.2 pg ml−1(0.3–1.9); and 30 cercariae: 3.6 pg ml−1(2.0–6.5), r = 0.70, P = 0.002) (Fig. 3b). The volunteer experiencing the most severe AEs had at least sevenfold higher serum CAA levels (maximum 49.9 pg ml−1) than other volunteers. There was no significant correlation between serum CAA levels and symptoms of an acute schistosomiasis syndrome (Extended Data Fig. 2b) or eosinophils (data not shown).
Urine CAA levels were variable (Extended Data Fig. 3a,b), but correlated with serum CAA (r = 0.58, P < 0.0001). The point-of-care rapid test for CCA (POC-CCA) was positive in 12% (2 of 17) of volunteers within 12 weeks after exposure and did not correlate with serum or urine CAA levels.
After a single 40 mg kg–1 dose of praziquantel treatment 12 weeks after exposure, serum CAA levels dropped below the detection limit in 8 out of 14 (57%) positive volunteers. The remaining 43% of volunteers were treated again with praziquantel, after which all remained undetectable until 1 year after exposure (Extended Data Fig. 4a–c).
All volunteers showed seroconversion of IgM against adult worms by immunofluorescence assay (IFA) (Fig. 4a,b). Seroconversion became apparent at week 4 in two volunteers and by week 6 in all. In addition, seven volunteers showed anti-soluble egg antigen (SEA) IgG seroconversion respectively at week 4 (1 of 16), week 12 (1 of 16) or week 20 (5 of 16). The absence of eggs was confirmed by a negative feces Schistosoma PCR at week 12 in all volunteers.
All volunteers showed an increase in adult worm antigen (AWA)-specific total IgG over time, with seroconversion above 2× s.d. of all baseline values in 12 of 16 volunteers at week 20. These responses were primarily IgG1, with seroconversion in all volunteers by week 16 (Fig. 4c,d). There was a clear dose response in AWA-specific total IgG and IgG1 levels (IgG week 20: r = 0.70, P = 0.003; IgG1 week 16: r = 0.56, P = 0.02), and a trend toward higher IgG1 levels and symptoms of an acute schistosomiasis syndrome (P = 0.08) (Extended Data Fig. 2c). No changes over time were found in total IgE or AWA-specific IgE and IgG4 compared to baseline (data not shown).
Cytokine and chemokine measurements in serum revealed increases in innate chemokines interferon (IFN)-γ-inducible protein (IP)-10 (Fig. 5a,b) and macrophage inflammatory protein (MIP)-1β (Fig. 5c), the latter of which was significantly higher in volunteers with symptoms of an acute schistosomiasis syndrome (P = 0.01 at week 8, Fig. 5d). There were no detectable changes in the other circulating chemokines or cytokines measured in serum (data not shown).
Overall, the frequency of antigen-specific IFN-γ (P = 0.01) and Th2 cytokine (interleukin (IL)-4, IL-5 and IL-13, P = 0.004)-producing CD4+ T cells increased over time (Fig. 5e and Extended Data Fig. 5a,b), but FOXP3+CD25+CD4+ regulatory T cells (P = 0.49) or the production of the regulatory cytokine IL-10 (P = 0.91, Fig. 5e) did not. However, in volunteers with symptoms of an acute schistosomiasis syndrome, both IFN-γ and Th2 cytokine-producing CD4+ T cells and CD25+FOXP3+ regulatory T cells were higher at week 4 and week 8 respectively, although not significantly for all (P = 0.28, P = 0.01 and P = 0.02, respectively; Fig. 5f). There were no differences in IL-10-cytokine-producing CD4+ T cells at week 4 (Fig. 5f). At all timepoints 17% of IFN-γ+CD4+ T cells also produced IL-2.
To understand which immunological and microbiological datasets were associated with the occurrence of symptoms of an acute schistosomiasis syndrome, we performed data integration using parallel generalized canonical correlation and partial least squares discriminant analysis. Eosinophil data decreased model accuracy and were thus removed (Extended Data Fig. 6a,b). At baseline, 9 of 16 (56.3%) participants were accurately predicted by the model with leave-one-out cross-validation, which increased to 13 of 16 (81.3%) correct classifications at week 12 and further to 15 of 16 (93.8%) when data from all timepoints were included (Fig. 6a,b). Thus, the model was able to accurately classify participants by the presence of an acute schistosomiasis syndrome. Permutation analysis confirmed that symptoms were strongly associated with the measured immunological and microbiological parameters over the infection course (n = 1,000; 99.6th percentile, Extended Data Fig. 6c). We identified Th2 cytokines at week 4 and FOXP3 regulatory T cells at weeks 12 and 16, MIP-1β at weeks 8 and 12 and levels of serum CAA at week 9 as important features elevated in symptomatic participants by leave-one-out cross-validation (Fig. 6c). Levels of Th2 cytokines upon stimulation at week 4 correlated with serum CAA at week 9, whereas concentrations of MIP-1β in plasma correlated with FOXP3 regulatory T cell numbers (Extended Data Fig. 6d).
This pilot study shows that experimental exposure to 20 male cercariae results in a detectable and well-tolerated S. mansoni infection in 82% of volunteers. This infection rate resembles that of other human infection models14. All volunteers were cured after 1–2 doses of praziquantel.
The dose-escalating design of the study revealed a concomitant increase in AEs. The occurrence of a severe acute schistosomiasis (Katayama) syndrome in one volunteer prompted us to lower the dose. At a dose of 20 cercariae, 2 of 11 volunteers reported severe AEs, which is comparable with other human infection models. For example, in experimental malaria infections, volunteers generally experience three to four AEs, of which one would be severe for several days15. In typhoid infections ~50% of volunteers report severe symptoms and 55% experiences fever16 and in cholera infection studies 40% of volunteers experience 1.6–8.0 liters of diarrhea17. We thus conclude that 20 cercariae may be the optimal dose that was both well tolerated and detectable. A relationship between dose and symptoms, as suggested by epidemiological data18, cannot be statistically confirmed with the current number of volunteers. Regardless of dose, all symptoms had resolved without sequelae at 12 weeks.
The follow up of volunteers in our study unequivocally showed that anti-adult worm IgM (100% seroconversion by week 6) or serum CAA (10 of 17 positive by week 6) are the earliest diagnostic markers currently available. This is in accordance with previous studies in travelers19,20,21. We found levels of serum CAA comparable to those in low endemic settings22. In nonhuman primate models these levels reflect 5–24 cercariae reaching adulthood8,23. Although serum CAA levels were more stable compared to urine CAA, the higher volume input of urine allows for more sensitive detection24. As suggested in previous studies, the urine POC-CCA rapid test was not suitable to detect very low intensity infection25. The main advantage of measuring serum CAA levels, as opposed to antibody detection, is the potential to follow up after treatment and confirm cure. In doing so, we found low cure rates with a single dose of 40 mg kg−1 praziquantel. Although a 60 mg kg−1 dose seemed more effective, the subsequent changes in pharmacokinetic and pharmacodynamic parameters need to be evaluated to conclude this with certainty. Because of the synchronous development of adult worms and the absence of reinfection in our model, the CHI design allows for screening of existing and new (stage-specific) anthelmintics.
The controlled schistosome infection model presented here clearly differs from infections in endemic settings, where doses are not controlled and infection occurs repeatedly. In addition, the single-sex infection lacks production of eggs that drive chronic regulatory and modified Th2 responses26. The presence of IgG to soluble egg antigen, indicates cross-reactive epitopes between eggs and worms. The induction of chemokines IP-10 and MIP-1β aligned with the increase in both antigen-specific CD4+ T cell IFN-γ and Th2 cytokine production, respectively. The IFN-γ production contrasts with predominantly Th2 profiles in epidemiological studies, which may be driven by egg-related responses26. Although we detected increases in FOXP3+CD25+CD4+ T cells, these mainly occurred in symptomatic volunteers and were found at a later timepoint than IFN-γ- or Th2 cytokine-producing T cells. This would be in line with the induction of regulatory T cells to prevent deleterious immune responses27, different from the chronic regulation found in endemic infections. These important differences in antigen exposure and subsequent immunological responses between single-sex infections and endemic chronic egg production may limit the use of the model to study anti-fecundity effects. However, the immunological observations from our model are comparable to acute infection models in travelers and baboons, where a mixed Th1 and Th2 response also dominates28,29. Integrated data analysis revealed that acute schistosomiasis symptoms were particularly associated with antigen-specific Th2 cytokine production and circulating MIP-1β and IgM production, but could not be predicted at baseline. In contrast to volunteers with symptoms, those without symptoms predominantly showed IgG1 antibody responses. Similarly to what has been performed for other infectious diseases30, a trial with repeated controlled infections could address whether these IgM or IgG1 antibodies have a protective effect. Currently, the clinical and immunological data from baseline alone cannot predict who will become symptomatic. However, the analysis of longitudinal responses reveals a clear profile predicting presence or absence of symptoms in 15 out of 16 participants. Unraveling the interplay between symptoms, immune responses and resistance to schistosome reinfection provides an opportunity for identification of new antigens for vaccine development31.
Immunological models for protection against S. mansoni were developed in rodents and nonhuman primates by repeated exposure to radiation attenuated cercariae23,32. In humans, three candidate S. mansoni vaccines are currently in clinical development6. A phase I study with rSm14/GLA-SE showed a good safety profile and immunogenicity7. Phase I safety results for Sm-TSP-2 are expected soon (ClinicalTrials.gov identifier: NCT02337855), while Sm-p80 is currently entering phase I testing on the basis of promising efficacy data in baboons8. The limited funding opportunities and large sample size required for phase III testing preclude testing of all three candidates in phase III trials. Despite the differences between chronic S. mansoni infection in the field and the controlled human S. mansoni infection, the model provides an opportunity to obtain preliminary efficacy data on these vaccines and reduce the costs by allowing selection of the most promising candidates, which may be co-formulated to maximize benefit12.
Future vaccine studies aim at a 75% infection reduction in worm burden and egg output10. As there is a clear relationship between worm burden and serum CAA levels24,33, we propose that the reduction in worm burden can be determined by measuring the median serum CAA level between week 7–12. Assuming an 80% power and α = 0.05, this would require a group size of 11 individuals per arm. Currently the main limitation of our model is the use of male schistosomes only. Consequently vaccine targets that are more commonly expressed on females, such as the Sm-p80 vaccine candidate8, cannot be fully evaluated. A female worm infection model would be of value to dissect mechanisms of action and sex-specificity of vaccines.
We conclude that this controlled human S. mansoni infection model results in an 82% infection rate with few severe side effects at a dose of 20 cercariae. In addition, this model provides insight into the onset of symptoms of a schistosome infection, the ensuing immune response and the performance of diagnostic tests over time. Notably, this model paves the way for cost-effective and rapid proof-of-concept testing of new vaccines and drugs.
This phase I trial (ClinicalTrials.gov identifier: NCT02755324) was an open-label dose-escalation study at the Leiden University Medical Center (LUMC).
Male cercariae were produced as previously reported34. The initial dose was 10 cercariae (n = 3), which was escalated to 30 (n = 3) and subsequently de-escalated to 20 cercariae (n = 3) on the basis of AEs. The 20 cercariae infection was then validated in another group (n = 8). The viability of cercariae was confirmed by imaging of cercariae penetrating skin explants35. The cercariae were applied to the volunteer’s forearm in 0.5–1 ml of water for 30 min, the number of remaining cercariae were counted by microscopy. Volunteers were observed for 30 min after exposure.
Volunteers were followed bi-weekly between week 0–24 and on week 52. During each visit, AEs were recorded. Symptoms of an acute schistosomiasis syndrome were defined as moderate to severe symptoms of malaise, fatigue, fever, night sweats, flu-like symptoms or headache, between 2–7 weeks after exposure. Safety reports were reviewed regularly by an external safety monitor, who advised on dose escalation. Blood and urine samples were collected at all visits.
The pre-patent period was defined as the time until serum CAA levels were above 1.0 pg ml−1 (ref. 24). At week 12 all volunteers were treated with 40 mg kg−1 praziquantel in two doses. A second regimen of 60 mg kg−1 praziquantel in two doses was provided if serum CAA levels persisted 3–6 weeks after treatment. Cure was defined as serum CAA levels ≤0.5 pg ml−1.
The study was approved by the LUMC Institutional Medical Ethical Research Committee (Institutional Review Board P16.111). It was performed according to the European Clinical Trial Directive 2001/20/EC, in accordance with ICH-GCP guidelines and the Declaration of Helsinki36,37.
Informed consent procedure
Healthy 18–45-year-old Schistosoma-naive volunteers were screened by medical history, general physical examination and safety laboratory tests. Informed consent was obtained from all volunteers.
Through advertisements, volunteers provided their email address and received written information. When they so wished, volunteers could schedule a screening visit at least 3 d after having received the information. They were then requested to complete an application form, which included a questionnaire regarding their health.
During the 1.5–2-h screening visit, the study purpose and procedures were explained and questions answered. The possible AEs and right of withdrawal was explained to the volunteers. The informed consent form was signed and a full physical exam was performed. All volunteers were required to consent to an HIV, hepatitis B (HBV) and hepatitis C (HCV) serological screening, urine toxicology and (for females), a pregnancy test at screening.
At the infection day (14 d to 23 weeks after screening), informed consent was reconfirmed, and a final check of inclusion and exclusion criteria was performed, including a focused physical exam. Volunteers were exposed to male Schistosoma mansoni cercariae after baseline assessment and safety laboratory tests.
The first volunteer was included on 27 October 2015 and the last volunteer was included on 1 February 2018.
All three volunteers gave permission to use the photographs (Fig. 2a–c) taken of their skin after cercarial exposure for publication.
Inclusion and exclusion criteria
Volunteer is aged ≥18 and ≤45 years and is in good health.
Volunteer has adequate understanding of the procedures of the study and agrees to abide strictly thereby.
Volunteer is able to communicate well with the investigator and is available to attend all study visits.
Volunteer will remain within Europe (excluding Corsica) during the study period and is reachable by mobile telephone from week 3 to week 12 of the study period.
Volunteer agrees to refrain from blood donation to Sanquin or for other purposes throughout the study period.
For females: volunteer agrees to use adequate contraception and not to breastfeed for the duration of study.
Volunteer has signed informed consent form.
Any history, or evidence at screening, of clinically significant symptoms, physical signs or abnormal laboratory values suggestive of systemic conditions, such as cardiovascular, pulmonary, renal, hepatic, neurological, dermatological, endocrine, malignant, hematological, infectious, immune-deficient, psychiatric and other disorders, which could compromise the health of the volunteer during the study or interfere with the interpretation of the study results. These include, but are not limited to, any of the following:
body weight <50 kg or body mass index <18 or >30 kg m−2 at screening;
positive HIV, HBV or HCV screening tests;
the use of immune-modifying drugs within 3 months before study onset (inhaled and topical corticosteroids and oral anti-histamines exempted) or expected use of such during the study period;
history of malignancy of any organ system (other than localized basal cell carcinoma of the skin), treated or untreated, within the past 5 years;
any history of treatment for severe psychiatric disease by a psychiatrist in the past year;
history of drug or alcohol abuse interfering with normal social function in the period of 1 year before study onset;
any clinically significant abnormalities (including extended QT interval) on electrocardiogram.
The chronic use of any drug known to interact with praziquantel, artesunate or lumefantrine (artesunate combined with lumefantrine served as alternative treatment of schistosomiasis in an earlier phase of infection) metabolism (for example, phenytoin, carbamazepine, phenobarbital, primidone, dexamethasone, rifampicin, cimetidine, flecainide, metoprolol, imipramine, amitriptyline, clomipramine, class IA and III anti-arrythmics, antipsychotics, antidepressants, macrolides, fluorquinolones, imidazole- and triazole antimycotics and anti-histamines). Because lumefantrine may cause extension of QT-time, chronic use of drugs with effect on QT interval are excluded from the study.
For female volunteers: positive urine pregnancy test at screening
Any history of schistosomiasis or treatment for schistosomiasis
Positive serology for schistosomiasis or elevated serum or urine CAA at baseline
Known hypersensitivity to or contraindications (including co-medication) for use of praziquantel, artesunate or lumefantrine
Being an employee or student of the Department of Parasitology or Infectious Diseases of the LUMC
Mild (grade 1): awareness of symptoms that are easily tolerated and do not interfere with usual daily activity
Moderate (grade 2): discomfort that interferes with or limits usual daily activity
Severe (grade 3): disabling, with subsequent inability to perform usual daily activity, resulting in absence or required bed rest
Serious AE: any untoward medical occurrence in a patient or trial participant, which does not have a causal relationship with the treatment, and:
is fatal, and/or
is life-threatening for the volunteer, and/or
makes hospital admission or an extension of the admission necessary, and/or
causes persistent or significant invalidity or work disability, and/or
manifests itself in a congenital abnormality or malformation, and/or
could, according to the person that carries out the research, have developed to a serious undesired medical event, but was, however, prevented due to premature interference.
Not related: a relationship to the administration of the S. mansoni male cercariae cannot be reasonably established; another etiology is known to have caused the AE or is highly likely to have caused it.
Unlikely related: a relationship to the administration of S. mansoni male cercariae is unlikely; however, it cannot be ruled out.
Possibly related: there is a potential association between the event and administration of the S. mansoni male cercariae; however, there is an alternative etiology that is more likely.
Probably related: administration of the S. mansoni male cercariae is the most likely cause; however, there are alternative reasonable explanations, even though less likely.
Definitely related: administration of the S. mansoni male cercariae is the cause; another etiology causing the adverse event is not known.
CAA was measured in serum and urine by the previously described upconverting phosphor lateral flow (UCP-LF CAA) assay24,38. The presence of urine CCA was determined by the POC-CCA (Rapid Medical Diagnostics). The Schistosoma PCR on feces was performed according to previous descriptions (ISO 15189:2012-certified)39.
Antigen detection assays
Upconverting phosphor lateral flow assay for circulating anodic antigen
Both serum and urine were analyzed for schistosome CAA using a UCP-LF CAA assay. The assay was performed as described previously24,38. In brief, 500 µl of serum (or 4 ml of urine) samples or standards were diluted 1:1 in 4% trichloro-acetic acid (TCA) (or diluted 5:1 in 12% TCA for urine samples) and incubated for 5 min at room temperature. Serum samples were centrifuged for 10 min at 13,000 r.p.m. (and urine samples were centrifuged for 45 min at 4,000 r.p.m.). Then 500 µl of supernatants of serum (or 4 ml of urine) was applied to 0.5-ml (or 4.0-ml) Amicon filtration devices (Amicon Ultra-0.5 (or Amicon Ultra4), Millipore) and concentrated to approximately 20 μl by centrifugation for 25 min at 13,000 r.p.m. for serum (or 60 min at 4,000 r.p.m. for urine). The concentrates were diluted 1:5 in LF assay buffer and incubated in microtiter plate wells at 37 °C for 1 h while shaking. LF strips were inserted into the wells and incubated for 3 h or overnight, before being read on a modified Packard FluoroCount microtiter plate reader24. A TCA-soluble fraction of S. mansoni adult worm antigen with known CAA concentration was used as a reference standard for the quantification of the antigen. Predefined cutoff values were used, where a serum CAA concentration above 1 pg ml−1 was defined as positive, below 0.5 pg ml−1 as negative and between 0.5 and 1.0 as undecisive24.
Point-of-care circulating cathodic antigen
Urine samples were tested for the presence of schistosome CCA using a commercially available rapid diagnostic test (POC-CCA, batch no. 170622073, Rapid Medical Diagnostics) according to the manufacturer’s procedure. Readings above a trace line were considered as positive.
Schistosome-specific antibodies, IgM against adult worms and anti-SEA IgG were determined according to ISO 15189:2012-certified routine diagnostic in-house IFA and ELISA respectively, which have been in use for the clinical diagnosis of schistosomiasis at the LUMC for decades40,41.
IgM adult worm antibodies by IFA
IgM antibodies against adult worms were detected by an in-house IFA assay as previously described42. This IFA, as well as the ELISA described below, are currently in use at LUMC as the routine antibody detecting assays for the diagnosis of imported schistosomiasis and feature in all laboratory quality assessment requirements (ISO 15189:2012-certified), including successful participation in an external quality assessment scheme (UK-NEQAS). In brief, sections of Rossman’s fixed male adult worms were incubated with a twofold dilution series of serum samples starting at 1:8 dilution. Following incubation with goat anti-human IgM (u-chain specific)-FITC antibody (Sigma-Aldrich; F5384), slides were examined using a fluorescence microscope. A negative control and a positive reference serum were run in parallel at each slide. The titer was determined as the dilution of the sample at which the fluorescence of adult worm gut epithelium was still visible. Samples were considered positive if titers were above 1:8. Previous studies showed that the IgM detected in this IFA is mainly directed against CCA42.
IgG against soluble egg antigen
IgG antibodies directed against S. mansoni SEA were detected by a previously described in-house ELISA with some minor modifications40. Crude SEA was prepared from S. mansoni eggs collected from the livers of infected hamsters43,44. A concentration of 5 µg protein ml−1 was diluted in 100 µl of 0.1 M sodium carbonate buffer (pH 9.6) and coated overnight at 4 °C in 96-well plates (Polysorb NUNC). Plates were stored at −80 °C until use. After thawing, plates were washed and blocked in 1% BSA in PBS for 1 h at 37 °C. A dilution series of serum samples from 1:16 to 1:2,048 in a solution with 5% FCS and 0.05% Tween in PBS were incubated for 1 h at 37 °C. Plates were washed and incubated for 1 h at 37 °C with mouse anti-human IgG alkaline phosphatase 1:10,000 (Sigma-Aldrich; A2064) in 4% BSA and 0.05% Tween/PBS. Para-nitrophenyl phosphate (pNPP) substrate (Sigma-Aldrich; P5994) in 0.1 M diethanolamine buffer (pH 9.6) (Merck) was added for 1 h at room temperature after washing. Plates were read with the Multiskan EX reader at 405 nm. A negative control and a positive reference serum were run in parallel at each plate. The titer was determined as the dilution of the sample at which the extinction is higher or equal to the reference standard. Samples were considered positive if titers were above 1:16.
Experimental immunological assays
Serum was analyzed for total IgE, S. mansoni AWA-specific IgE and IgG by ELISA45,46,47. Data were expressed as arbitrary units (AU ml−1). Peripheral blood mononuclear cells (PBMCs) were evaluated for their phenotype and function after 24-h stimulation with crude S. mansoni AWA by flow cytometry.
Adult worm antigen-specific IgE, IgG, IgG1 and IgG4
S. mansoni AWA-specific IgE and IgG was measured by ELISA modified from previous protocols45,46. In brief, adult S. mansoni worms were collected from hamsters and crude AWA was prepared as described previously48. MaxiSorp plates (Nunc) were coated overnight with 5 μg ml−1 AWA diluted in carbonate buffer pH 9.6. After blocking with 5% BSA/PBS, the sera were diluted with a solution with 5% FCS and 0.05% Tween in PBS and the presence of IgG1 or IgG4 was shown by using horseradish peroxidase (HRP)-labeled anti-human IgG1 or HRP-labeled anti-human IgG4 (1:3,000 dilution; Sanquin). For measuring total anti-AWA IgG antibody, alkaline phosphatase-conjugated anti-human IgG (1:5,000 dilution; Sigma) was added, whereas for the anti-AWA IgE assay, the plate was incubated with biotinylated goat anti-human IgE (1:1,000 dilution; Vector Laboratories) followed by streptavidin HRP conjugate (1:10,000 dilution; Sanquin). IgG1 and IgG4 assays were developed using tetramethylbenzine, stopped with 10% H2SO4 and the absorbencies were measured at 450 nm. For total IgG and IgE, the color was developed by addition of pNPP (Sigma) diluted in diethanolamine buffer and optical density was measured at 405 nm. For all four assays, the levels of antibody present in a given sample were expressed in AU ml−1 or International Units (IU ml−1) according to the standard curve of pooled sera from inhabitants of an S. mansoni endemic area in Ghana. The levels of antibody present in a given samples were expressed in AU ml−1 for total IgG, IgG1 and IgG4 or in IU ml−1 for IgE. Seroconversion was defined as antibody levels above 2× s.d. of the baseline.
Total IgE levels were measured as previously described47. Briefly, MaxiSorp plates were coated overnight with rabbit anti-human IgE (Dako). Plates were blocked with PBS 5% BSA followed by incubation of diluted samples in PBS 0.05% Tween-20. As a reference, the World Health Organization standard of human serum IgE (NIBSC) was used, starting at a concentration of 90 IU ml−1. After a washing step, the plates were incubated with IgE biotinylated goat anti-human IgE antibody (Vector Laboratories) followed by an incubation with streptavidin alkaline phosphatase conjugate (Boehringer Mannheim). The color was developed by addition of pNPP (Boehringer Mannheim) diluted in diethanolamine buffer and optical density was measured at 405 nm. The results were expressed in IU ml−1. Seroconversion was defined as antibody levels above 2× s.d. of the baseline.
Ex vivo cytokines
Serum samples were tested for the presence of different cytokines using a commercially ProcartaPlex Multiplex Immunoassay (17-plex, lot 178863000, Invitrogen) according to the manufacturer instructions. The following cytokines were measured: IFN-α, IFN-β, IL-1β, IL-10, IL-12p70, IL-13, IL-15, IL-2, IL-22, IL-23, IL-4, IL-6, IP-10, MCP-1, MIP-1α, MIP-1β and TNF-α using the Bio-plex 200 Luminex (Bio-Rad).
Between week 0 and 24 after exposure every 4 weeks, human PBMCs were isolated from whole blood collected in heparin. Cells of two heparin tubes were diluted at least 1:2 with HBSS (ThermoFisher) at room temperature. Ten milliliters of ficoll at room temperature was added, followed by 25 min centrifugation at 400g with low brake. Cells were collected and washed with HBSS, counted and frozen in 10% DMSO, in RPMI Hepes (Invitrogen), with 100 U ml−1 penicillin, 100 μg ml−1 streptomycin, 1 mM pyruvate/2 mM glutamate and 10% FCS (Bodinco). Subsequently, PBMCs were thawed in RPMI Hepes, with 100 U ml−1 penicillin, 100 μg ml−1 streptomycin, 1 mM pyruvate/2 mM glutamate and 10% FCS and rested overnight at 37 °C with 5% CO2. Cells were counted and transferred to a 96-well round bottom plate (Corning) with 500,000 cells per well. Cells were stimulated with AWA (50 µg ml−1) for 24 h. Staphylococcal enterotoxin B (SEB) (Sigma-Aldrich) 200 ng ml−1 was used as a positive control and RPMI Hepes (Invitrogen), with 100 U ml−1 penicillin, 100 μg ml−1 streptomycin, 1 mM pyruvate/2 mM glutamate and 10% FCS as a negative control. After 4 h of incubation 5 mg ml−1 brefeldin A (Sigma) was added to SEB-stimulated wells and after 20 h to AWA- and medium-stimulated wells. After a total stimulation of 24 h, the cells were stained with Aqua (Invitrogen) and fixed with 3.9% formaldehyde (Sigma). After fixating, the cells were stained with the following antibodies: CD3, CD4, IFN-γ, IL-2, Th2-cytokines (IL-4, IL-5, IL-13), TNF and IL-10 (Supplementary Table 4). Human FC block was used to avoid nonspecific interactions. The cells were measured with the FACSCanto II (BD Biosciences; Supplementary Fig. 1). The data were analyzed with FlowJo 10.5 software for MAC OS. The gating was placed with the help of fluorescence minus one controls, the medium as a negative control and SEB as a positive control. The leftover, aqua-stained and fixed cells were frozen in 10% DMSO, in RPMI Hepes (Invitrogen), with 100 U ml−1 penicillin, 100 μg ml−1 streptomycin, 1 mM pyruvate/2 mM glutamate and 10% FCS (Bodinco) and stored at −80 °C. The cells were thawed at 37 °C and stained with the following antibodies: CD3, CD4, CD25, CD127 and FOXP3 (Supplementary Table 4). As before, human FC block was used to avoid nonspecific interactions. The cells were measured with the FACSCanto II (BD Biosciences; Supplementary Fig. 1). The data were analyzed with FlowJo 10.5 software for MAC OS. The gating was placed with help of fluorescence minus one controls, the medium as a negative control and SEB as a positive control.
All 17 volunteers were included in the intention-to-treat analysis (safety analysis and parasitological assays). One volunteer was excluded from the per-protocol analysis (all immunological readouts) on the basis of high baseline AWA-specific IgG and IgG1 levels. Samples from individual volunteers were measured once and plotted as single values.
Demographics and the presence of symptoms between groups were analyzed with a Mann–Whitney U-test, time to patency with a log-rank test and correlations with a Spearman’s rank test. Changes in the frequency of cytokine-producing cells over time were analyzed using a linear mixed model. Time was considered as the fixed effect and the volunteer ID as a random effect for the intercept. P values based on Student’s t-tests were obtained using the Satterthwaite’s degrees of freedom method. In the cytokine boxplots, the negative values (after subtracting the medium condition) were set to zero to prevent a negative cytokine response. However, the statistical analysis was performed on the unaltered data. All statistical tests were two-tailed with α set at the 0.05 level.
P values were considered significant when P < 0.05.
Mixomics and data integration
Data integration was performed using the mixOmics package in R (v.6.8.0)52,53. This method allowed us to correlate across datasets, while associating features with outcome. Feature selection was performed using Lasso-like penalization for each of the datasets. The number of components was set at two and tuning was performed to find the minimum number of features needed per dataset (in a range from 1–3 per component) and the correlation between datasets was entered into the design matrix. The number of features included in the final model was 3, 3, 2 and 5 for cytokines, antigens, antibodies and cellular responses, respectively. A correlation of 0.75 between datasets was used in the design matrix.
Extended information can be found in the Nature Research Reporting Summary.
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
The data that support this publication are available at ImmPort (https://www.immport.org) under study accession SDY1609. All data will be made available for further research, provided that reference is made to the LUMC source.
King, C. H. et al. Utility of repeated praziquantel dosing in the treatment of schistosomiasis in high-risk communities in Africa: a systematic review. PLoS Negl. Trop. Dis. 5, e1321 (2011).
Tukahebwa, E. M., Vennervald, B. J., Nuwaha, F., Kabatereine, N. B. & Magnussen, P. Comparative efficacy of one versus two doses of praziquantel on cure rate of Schistosoma mansoni infection and re-infection in Mayuge District, Uganda. Trans. R. Soc. Trop. Med. Hyg. 107, 397–404 (2013).
Bergquist, R., Utzinger, J. & Keiser, J. Controlling schistosomiasis with praziquantel: how much longer without a viable alternative? Infect. Dis. Poverty 6, 74 (2017).
Global Burden of Disease Study 2013 Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 386, 743–800 (2015).
Colley, D. G., Bustinduy, A. L., Secor, W. E. & King, C. H. Human schistosomiasis. Lancet 383, 2253–2264 (2014).
Merrifield, M. et al. Advancing a vaccine to prevent human schistosomiasis. Vaccine 34, 2988–2991 (2016).
Santini-Oliveira, M. et al. Schistosomiasis vaccine candidate Sm14/GLA-SE: phase 1 safety and immunogenicity clinical trial in healthy male adults. Vaccine 34, 586–594 (2016).
Zhang, W. et al. Sm-p80-based schistosomiasis vaccine: double-blind preclinical trial in baboons demonstrates comprehensive prophylactic and parasite transmission-blocking efficacy. Ann. NY Acad. Sci. 1425, 38–51 (2018).
Status of Vaccine Research and Development of Vaccines for Schistosomiasis (World Health Organization, 2014); http://who.int/immunization/research/meetings_workshops/Schistosomiasis_VaccineRD_Sept2014.pdf
Mo, A. X. & Colley, D. G. Workshop report: schistosomiasis vaccine clinical development and product characteristics. Vaccine 34, 995–1001 (2016).
Siddiqui, A. J. et al. Sm-p80-based vaccine trial in baboons: efficacy when mimicking natural conditions of chronic disease, praziquantel therapy, immunization, and Schistosoma mansoni re-encounter. Ann. NY Acad. Sci. 1425, 19–37 (2018).
Roestenberg, M., Hoogerwerf, M. A., Ferreira, D. M., Mordmuller, B. & Yazdanbakhsh, M. Experimental infection of human volunteers. Lancet Infect. Dis. 18, e312–e322 (2018).
Langenberg, M. C. C. et al. Katayama syndrome without Schistosoma mansoni eggs. Ann. Intern. Med. 170, 732–733 (2019).
Balasingam, S. & Wilder-Smith, A. Randomized controlled trials for influenza drugs and vaccines: a review of controlled human infection studies. Int. J. Infect. Dis. 49, 18–29 (2016).
Walk, J. et al. Diagnosis and treatment based on quantitative PCR after controlled human malaria infection. Malar. J. 15, 398 (2016).
Jin, C. et al. Efficacy and immunogenicity of a Vi-tetanus toxoid conjugate vaccine in the prevention of typhoid fever using a controlled human infection model of Salmonella typhi: a randomised controlled, phase 2b trial. Lancet 390, 2472–2480 (2017).
Chen, W. H. et al. Single-dose live oral cholera vaccine CVD 103-HgR protects against human experimental infection with Vibrio cholerae O1 El Tor. Clin. Infect. Dis. 62, 1329–1335 (2016).
Rocha, M. O. et al. Pathogenetic factors of acute Schistosomia mansoni: correlation of worm burden, IgE, blood eosinophilia and intensity of clinical manifestations. Trop. Med. Int. Health 1, 213–220 (1996).
van Grootveld, R. et al. Improved diagnosis of active Schistosoma infection in travellers and migrants using the ultra-sensitive in-house lateral flow test for detection of circulating anodic antigen (CAA) in serum. Eur. J. Clin. Microbiol. Infect. Dis. 37, 1709–1716 (2018).
Meltzer, E. & Schwartz, E. Schistosomiasis: current epidemiology and management in travelers. Curr. Infect. Dis. Rep. 15, 211–215 (2013).
Visser, L. G., Polderman, A. M. & Stuiver, P. C. Outbreak of schistosomiasis among travelers returning from Mali, West Africa. Clin. Infect. Dis. 20, 280–285 (1995).
Sousa, M. S. et al. Performance of an ultra-sensitive assay targeting the circulating anodic antigen (CAA) for detection of Schistosoma mansoni infection in a low endemic area in Brazil. Front. Immunol. 10, 682 (2019).
Kariuki, T. M. et al. Parameters of the attenuated schistosome vaccine evaluated in the olive baboon. Infect. Immun. 72, 5526–5529 (2004).
Corstjens, P. L. et al. Tools for diagnosis, monitoring and screening of Schistosoma infections utilizing lateral-flow based assays and upconverting phosphor labels. Parasitology 141, 1841–1855 (2014).
Colley, D. G. et al. A five-country evaluation of a point-of-care circulating cathodic antigen urine assay for the prevalence of Schistosoma mansoni. Am. J. Trop. Med. Hyg. 88, 426–432 (2013).
Caldas, I. R. et al. Human Schistosoma mansoni: immune responses during acute and chronic phases of the infection. Acta Trop. 108, 109–117 (2008).
Thompson, L. J. et al. Conditioning of naive CD4(+) T cells for enhanced peripheral Foxp3 induction by nonspecific bystander inflammation. Nat. Immunol. 17, 297–303 (2016).
Kariuki, T. M. & Farah, I. O. Resistance to re-infection after exposure to normal and attenuated schistosome parasites in the baboon model. Parasite Immunol. 27, 281–288 (2005).
Soonawala, D., Geerts, J. W., de Mos, M., Yazdanbakhsh, M. & Visser, L. G. The immune response to schistosome antigens in formerly infected travelers. Am. J. Trop. Med. Hyg. 84, 43–47 (2011).
Parrino, T. A., Schreiber, D. S., Trier, J. S., Kapikian, A. Z. & Blacklow, N. R. Clinical immunity in acute gastroenteritis caused by Norwalk agent. N. Engl. J. Med. 297, 86–89 (1977).
Sombetzki, M. et al. Host defense versus immunosuppression: unisexual infection with male or female Schistosoma mansoni differentially impacts the immune response against invading cercariae. Front. Immunol. 9, 861 (2018).
Richter, D., Harn, D. A. & Matuschka, F. R. The irradiated cercariae vaccine model: looking on the bright side of radiation. Parasitol. Today 11, 288–293 (1995).
Agnew, A. et al. The relationship between worm burden and levels of a circulating antigen (CAA) of five species of Schistosoma in mice. Parasitology 111, 67–76 (1995).
Janse, J. J. et al. Establishing the production of male Schistosoma mansoni cercariae for a controlled human infection model. J. Infect. Dis. 218, 1142–1146 (2018).
Winkel, B. M. F. et al. Early induction of human regulatory dermal antigen presenting cells by skin-penetrating Schistosoma mansoni cercariae. Front. Immunol. 9, 2510 (2018).
World Medical Association. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. JAMA 310, 2191–2194 (2013).
European Commission. Clinical Trials - Directive 2001/20/EC. Official Journal of the European Communities https://ec.europa.eu/health/sites/health/files/files/eudralex/vol-1/dir_2001_20/dir_2001_20_en.pdf (2001).
Corstjens, P. L. et al. Improved sensitivity of the urine CAA lateral-flow assay for diagnosing active Schistosoma infections by using larger sample volumes. Parasit. Vectors 8, 241 (2015).
Obeng, B. B. et al. Application of a circulating-cathodic-antigen (CCA) strip test and real-time PCR, in comparison with microscopy, for the detection of Schistosoma haematobium in urine samples from Ghana. Ann. Trop. Med. Parasitol. 102, 625–633 (2008).
Deelder, A. M. et al. Applicability of different antigen preparations in the enzyme-linked immunosorbent assay for Schistosomiasis mansoni. Am. J. Trop. Med. Hyg. 29, 401–410 (1980).
Nash, T. E., Ottesen, E. A. & Cheever, A. W. Antibody response to a polysaccharide antigen present in the schistosome gut. II. Modulation of antibody response. Am. J. Trop. Med. Hyg. 27, 944–950 (1978).
Deelder, A. M., van Zeyl, R. J., Fillie, Y. E., Rotmans, J. P. & Duchenne, W. Recognition of gut-associated antigens by immunoglobulin M in the indirect fluorescent antibody test for Schistosoma mansoni. Trans. R. Soc. Trop. Med. Hyg. 83, 364–367 (1989).
Dalton, J. P., Day, S. R., Drew, A. C. & Brindley, P. J. A method for the isolation of schistosome eggs and miracidia free of contaminating host tissues. Parasitology 115, 29–32 (1997).
Deelder, A. M. Immunology of experimental infections with Schistosoma mansoni in the Swiss mouse and with Fasciola hepatica in the rabbit. Acta Leiden 39, 5–107 (1973).
van den Biggelaar, A. H., Borrmann, S., Kremsner, P. & Yazdanbakhsh, M. Immune responses induced by repeated treatment do not result in protective immunity to Schistosoma haematobium: interleukin (IL)-5 and IL-10 responses. J. Infect. Dis. 186, 1474–1482 (2002).
Faulkner, H. et al. Antibody responses in onchocerciasis as a function of age and infection intensity. Parasite Immunol. 23, 509–516 (2001).
Staal, S. L. et al. Prevalence of atopy following mass drug administration with albendazole: a study in school children on Flores Island, Indonesia. Int. Arch. Allergy Immunol. 177, 192–198 (2018).
van Dam, G. J. et al. Antibody response patterns against Schistosoma mansoni in a recently exposed community in Senegal. J. Infect. Dis. 173, 1232–1241 (1996).
Bates, D., Machler, M., Bolker, B. M. & Walker, S. C. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48 (2015).
Kuznetsova, A., Brockhoff, P. B. & Christensen, R. H. B. lmerTest package: tests in linear mixed effects models. J. Stat. Softw. 82, 1–26 (2017).
Wickham, H. ggplot2: Elegant Graphics for Data Analysis 2nd edn (Springer, 2016).
Rohart, F., Gautier, B., Singh, A. & Le Cao, K. A. mixOmics: an R package for ‘omics feature selection and multiple data integration. PLoS Comput. Biol. 13, e1005752 (2017).
Singh, A. et al. DIABLO: an integrative approach for identifying key molecular drivers from multi-omic assays. Bioinformatics 35, 3055–3062 (2019).
We thank P. van Genderen for reviewing the safety data and providing his advice as safety monitor of our study. We thank M. Casacuberta Partal, M.A.A. Erkens, J.L. Fehrmann-Naumann, M.S. Ganesh, H. Gerritsma, G.C. Hardeman, P.T. Hoekstra-Mevius, Y.C.M. Kruize, Y.D. Mouwenda, H.H. Smits, K. Suijk-Benschop, J.J.C. de Vries and C.J.G. van Zeijl-van der Ham for their laboratory, clinical and data-analyzing support during the study. Most of all we thank all volunteers participating in the study, without whom the study could not have been performed. Dr. Roestenberg was supported by a Veni grant (no. 016.156.076) from the Netherlands Organization for Health Research and Development and a Gisela Thier Fellowship (no. 14-0645) from LUMC. Dr. Langenberg and Dr. Hoogerwerf were supported by a grant from Dioraphte Foundation (no. 16020405). The funding sources had no role in collecting, analyzing, interpreting or reporting the data. Dr. Jochems has received funding from the European Union’s Horizon 2020 Research and Innovation Programme under Marie Skłodowska-Curie grant agreement no. 707404. The opinions expressed in this document reflect only the author’s view. The European Commission is not responsible for any use that may be made of the information that it contains.
The authors declare no competing interests.
Peer review information Alison Farrell was the primary editor on this article and managed its editorial process and peer review in collaboration with the rest of the editorial team.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
a-c. Eosinophil counts (x109/L) per volunteer and the median per group. The thin green, red, and blue lines represent data of individual volunteers infected with 10 (n = 3), 20 (n = 11) or 30 (n = 3) cercariae respectively, while thick lines represent the median of each group.
Extended Data Fig. 2 Relation between symptoms of an acute schistosomiasis syndrome and immunological readouts.
The relation between the presence of symptoms of an acute schistosomiasis infection and a. the highest eosinophil count (n = 17), b. the median serum CAA level from week 7 to 12 (n = 17), or c. the AWA specific IgG1 response at week 16 (n = 16). All using the two-sided Mann-Whitney U test. Individual data is presented as dots, the line represents the median, while the error bars represent the interquartile range of the groups.
a. Urine CAA levels after exposure and b. after first and second praziquantel treatment (week 0) and at week 52 after exposure. The thin green, red, and blue lines represent of individual volunteers infected with 10 (n = 3), 20 (n = 11) or 30 (n = 3) cercariae respectively, while thick lines represent the median of each group before treatment.
a-c. Serum CAA levels in pg/ml after the first treatment, second treatment, and at week 52. All values below the detection threshold of 0.5 pg/mL, are plotted at 0.25 pg/mL. The gray, yellow and blue lines represent data of individual volunteers infected with 10 (n = 3), 20 (n = 11) or 30 (n = 3) cercariae respectively.
The percentage of a. IFNγ and b. Th2-cytokine producing CD4+ T-cells over time in weeks after exposure in the 10 cercariae (gray, n = 3), 20 cercariae (yellow, n = 10) and 30 cercariae (blue, n = 3) groups. Dotted lines are linear regression lines, gray areas are confidence intervals, light dots are individual data, and horizontal lines with dots are the average values.
a. Individual predictions across folds for the full model per dataset including majority vote. Each symbol represents one prediction, with volunteers in columns and datasets in rows. Filled circles indicate a correct prediction and open circles a false prediction. Symptomatic and asymptomatic volunteers are indicated in red and the number of correct predictions per dataset is indicated. b. Mean Pearson correlation score between datasets using the first component of the projection onto the latent space across all folds from the model including all datasets. Size and color of circles reflect the mean rho value. c. Permutations analysis (n = 1000) with leave-one-out cross-validation on the full model using all subjects and including the four datasets without eosinophils. Blue and red dashed lines indicate the 99th percentile and the accuracy when comparing symptomatic and asymptomatic individuals (99.6%), respectively. d. Spearman correlation matrix of the seven consensus features selected in > 75% of folds in the leave-one-out cross-validation. Features were clustered using hierarchical clustering with complete linkage on Euclidean distance. All graphs are based on n = 16.
About this article
Cite this article
Langenberg, M.C.C., Hoogerwerf, MA., Koopman, J.P.R. et al. A controlled human Schistosoma mansoni infection model to advance novel drugs, vaccines and diagnostics. Nat Med 26, 326–332 (2020). https://doi.org/10.1038/s41591-020-0759-x
This article is cited by
Mucosal Immunology (2022)
Evaluation of microscopy, serology, circulating anodic antigen (CAA), and eosinophil counts for the follow-up of migrants with chronic schistosomiasis: a prospective cohort study
Parasites & Vectors (2021)