Alongside the wide distribution throughout sub Saharan Africa of schistosomiasis, a chronic parasitic disease caused mainly by Schistosoma mansoni and Schistosoma haeamatobium, the morbidity associated with chronic infection remains characterized by immunomodulation and a hyporesponsiveness dependent on parasite burden (1). One potential driver of this phenomenon is in utero exposure to circulating schistosome antigens during maternal infection, which could influence the developing immune system of children, and modify anti-schistosome immunity. Immunomodulatory effects of schistosome infection are well-described in the context of bystander immune responses (1, 2), as are modulated immune responses following exposure to maternal infection (3–5), increased levels of schistosome adult-worm specific IgE and IgG in newborns cord blood whose mothers were infected during pregnancy (and not treated) (6). Schistosome antigens can drive immunomodulatory changes via presence in breastmilk (7), and have been detected as persisting in human children and murine offspring following exposure to maternal infection (8, 9). Murine models demonstrate that exposure to maternal infection with schistosomes consistently modifies immune development, with recent findings included impact upon B cell priming and DC-T cell interactions (5), as well as NKT cells (3).

Further infant outcomes of maternal parasite infection include earlier immune maturation (B cells), spontaneous (polyclonal) IgE, decreased infantile eczema episodes, anti-parasite IgG, lower vaccine-induced Ig-titres, suppressed T cell responses (anti-parasite and against vaccine antigens), altered T helper cell responses, increased regulatory responses (IL-10), and modified APC status (as reviewed in (10) and (11)). Considering that 40 million women of reproductive age are at risk of infection worldwide, treatment with Praziquantel (PZQ) during pregnancy could have significant impact on disease outcomes in these populations. PZQ has been effectively used for decades against adult schistosomes, via a recently described mode of action targeting a transient receptor potential melastatin ion channel (12). However, there is hesitancy to follow the WHO recommendation for treatment during pregnancy, as PZQ, marketed as Biltricide^®, and donated for use within WHO-agreement as Cesol 600, is classed as a Category B drug for pregnancy, with animal studies not indicating a risk to the fetus, but with insufficient controlled studies in humans at this point (13). In existing studies, Praziquantel treatment has not been associated with adverse events regarding birth weight, fetal mortality, or congenital abnormalities (14). Anthelminthic strategies have been described to drive further changes associated with the re-awaking of the anti-parasite immune response, through disturbance of the regulatory states maintained by live parasites to ensure their survival within hosts (1). However, altered inflammatory outcomes have been associated with maternal treatment during pregnancy, including modified allergic parameters (15), and boosted immune responses to schistosome antigens after pregnancy (16). Enhanced immune priming during treatment is thought to occur due to death and degradation of worms, leading to increased load of worm antigens in the system, contributing to a natural partial-resistance hypothesized to develop of slowly as exposures to dead worms accumulate (17).

Treatment-induced shifts in schistosome-directed immune responses, particularly to antigens found in adult worms, can re-focus the tolerogenic modulations accumulated through chronic infection. Accordingly, treatment of S. haematobium infected children increased their IgG4 and IgE to adult worm antigen (AWA) as well as IgE to soluble egg antigen (SEA) (18), and a similar boost was seen following PZQ treatment of pregnant women infected with S. mansoni, where treatment caused a boost in IgG1, IgG2, IgG4 specific against worm antigens and enhanced IgE levels against SEA in the mothers (16), and enhanced cytokine levels (19). The well-established mouse model of S. mansoni [summarized in (20)], benefits from using a human pathogen in its preferred niche of the hepatoportal vasculature, and previously revealed materno-fetal effects including transcriptional changes in the placenta (21), later validated in human cohorts (6). Similarly, previous studies in murine systems of employing direct infection of previously schistosome-exposed offspring demonstrated transfer of schistosome antigens and anti-schistosome antibodies in utero and through breast milk, further associated with modified infections and reduced immune responsiveness (22), with detectable antibodies in neonatal mice arising from acute maternal infection further associated with reduced liver pathology and lower worm burden (9).

The shifting character of the anti-schistosome immune response across phases of infection is well described (23), with maternal infection phase, and likely associated cytokine profiles, further reflected in later offspring immune responses, able to differentially modify reactivity to allergens (21). Differences of priming effects also occur when comparing routes of maternal exposure, such as via breastmilk or in utero exposure (7). Parasite load presents another further variable, as-yet underexplored in maternal infection settings, yet determinants of worm burden are shown to impact host bystander immune responses (2). As such, the shared mechanisms behind transgenerational effects on cognate infection remain unclear, and with suggestions that these may involve processes linked to trained immunity (24) in innate cells, or specific priming in adaptive cells, such as in B cells, as has also been shown for other infections (25).

In line with these human cohort studies and murine models, here we focused on the interaction between immune imprinting effects of maternal helminth infection during pregnancy, and anthelminthic treatment with PZQ. Using this murine model, we explored the added implications of maternal treatment in transgenerational priming, and more specifically the hypothesis that the immune reactivation events triggered by PZQ-induced worm death during maternal treatment can imprint offspring immunity, with particular regard to cognate schistosome-direct immune responses. We link this to a profile of how prenatal treatment can disrupt the steady state immune priming imprinted via maternal parasite infection, changes that may not only underlie responses to schistosome infection, but also whether treatment of maternal infection can restore a more normalized immune profile to such offspring exposed in utero to parasite infection. These data provide supporting information regarding the WHO recommendation of anthelminthic treatment during pregnancy.

Environmental changes during early life are increasingly understood to impact proper development and education of the immune system. Of the many potential changes to the gestational environmental, maternal infection during pregnancy has been identified as a key factor driving heterologous immune responses (31). Maternal viral infection during pregnancy has been linked to numerous developmental changes, and has become explored experimentally through maternal immune activation models (26). Prenatal elimination of schistosomiasis through anthelminthic treatment has been shown to modify transgenerational effects of maternal schistosomiasis on atopy (15), and may poise the immune system for responsiveness to cognate infection. Indeed, recent discussions have highlighted the complexity of factors present during schistosome infection (11). Here, by means of a chronic schistosome infection mouse model, we uncover that the modified systemic immunological signature imparted by treating schistosomiasis is transferable from mothers to their offspring, distinguishable from active chronic infection by increased cytokine reactivity to schistosome antigens. Alongside improving pregnancy outcomes, prenatal treatment of schistosomiasis induced a form of maternal immune activation (otherwise described from maternal viral infection or models of this (26)) as signified by strongly increased antigen-specific T_H1 as well as T_H2 cytokines. We find that PZQ treatment boosted maternal cellular cytokine responses, in line with responses from treatment during pregnancy in human cohorts (19), and that this persists well beyond the point of weaning and therefore may impact the entire period of maternal contact for pups. The resultant improvements in offspring immunity to cognate schistosome infection indicate imprinting of this activation upon the developing immune system through transmaternal exposure. This altered offspring immune response by such treatment-exposed offspring to cognate schistosome infection, characterized by enhanced T_H2-type cytokine responses to schistosome antigens, with a notably boost in SEA-induced IL-10 levels, and modified SEA-specific antibody levels. It is interesting to note that one of the clearest differences induced by exposure to treated maternal schistosomiasis was SEA-induced IL-10 levels, which reflects previous human studies where responses to schistosome antigens involving this cytokine was notable as elevated in children of PZQ-treated women (32). That maternal treatment was associated with normalized anti-schistosome IgE levels during cognate infection, otherwise suppressed by exposure to live maternal infection, reflects previous findings of detrimental effects of maternal schistosomiasis on effective induction of specific antibody titers (3). Treatment of helminth infection in humans was previously associated with increased IgG4 and IgE antibody responses (18), although pregnancy itself has been suggested to have a dampening effect of such boosting of antibodies (16). Recent data from our group demonstrated altered antibody transfer to children of schistosome-infected (yet untreated) mothers), in association with placental inflammatory parameters (6). Other infection parameters appeared largely unaltered by exposure to maternal infection in the absence of treatment, indicating only relatively faint specific immune effects in this setting, but the activating effect of maternal anthelminthic treatment was further associated with a modified appearance of infection-associated pathology, namely with reduced weight loss, lower worm burden, and associated signs of lower egg burden. Alongside the more general immune activation that follows treatment, the differential effect of exposure to PZQ treated versus non-treated maternal schistosomiasis may derive from specific induction on anti-worm versus anti-schistosome egg immune responses, the former not developing in the absence of worm death (17) and therefore may be enriched in the PZQ-treated groups. As with treatment leading to development of some resistance for infected hosts (33), this may similarly contribute to the transmaternally-primed partial resistance indicated in this study. This positions PZQ-treated maternal infection as a particular priming stimulus, that may differ in effect from untreated parasite infection, generally thought to downmodulate offspring anti-parasite responses to optimize the cross talk with the next generation of potential hosts, as observed in maternal malarial infection, where active (but not treated) maternal infection enhanced IL-10+ regulatory CD4+ T cells in cord blood that inhibited parasite-specific IFN-gamma responses (34). Additional dynamics of treatment (e.g.: PZQ given later during pregnancy) therefore represents a further avenue for research into this effect, as the timing of antigen release from dying worms, and the accompanying cytokine activation following treatment, could modify the amplitude of the priming effect on the next generation.

Despite the observance of reduced weight-loss during cognate infection following exposure to treated maternal schistosomiasis, the absence of a clear impact on schistosomiasis-induced liver pathology (in contrast to reports of total absence of egg-induced granulomas in some settings (9) suggests limits to protective effect of maternal treatment, at least in the context of a experimentally induced, high-burden infection. The reduced induction of serum cytokine levels, however, further supports an overall profile of reduced inflammatory pathology during infection. That these treatment-exposed offspring showed modified infection responses may further be linked to altered cellular priming, particularly within B cell activation state and B cell subset proportion, which we have observed after exposure to maternal schistosomiasis. Our cluster-based analysis allowed us to identify differences across subpopulations not immediately identified by manual gating, and further to assess these differences in an unsupervised way, and revealed the influence of maternal PZQ treatment. This includes a modification to the transgenerational priming driven by the parasites, which in the steady state did leave a signature of increased activation markers and modified enrichment of B cell subsets, evident in the bone marrow, which may further indicate priming alterations to hematopoiesis and trained immunity that remain to be elucidated, and deepens our profiling of the activation-induced enhanced regulatory phenotype described using manual gating-based analysis of DCs in our previous work (5). One such important change may be the underrepresentation of CD1d^hi B cells in offspring primed by live infection, with SEA-loaded cells previously shown to activate NKT cells in a CD1d-dependent manner (35) and NKT functionality recently shown to be affected by maternal schistosomiasis and associated with dampened responsiveness to immunization in a cascade involving instead lower IL-4 and reduced B cell priming, when evaluated early at the onset egg-deposition in patent infection in mothers (3). This indicates a potential pathway for transgenerational immune interference by maternal infection, with possible links to the modified regulatory network postulated to alter offspring bystander immune responses. Late into chronic infection these suppressed offspring bystander responses are instead associated with an IL-4-skewed, modified type 2 phenotype (5), typical of the regulatory network present during chronic infection itself (36). Further, detailed analysis of differences in antigen-specific immunity and repertoire changes may further illuminate the transmaternal priming effect driven by PZQ-treatment, including potential differences that may exist in effector T cell and antibody-responses against worm-derived antigens compared to those direct against egg antigens. Findings from human studies of treatment during pregnancy have found that although treatment during pregnancy augmented cytokine responses to SEA, that overall schistosome-directed cytokine responses were dampened during pregnancy (19), indicating further cross talk between infection and the tolerogenic immune processes characteristic of pregnancy. The timepoint at which maternal immune responses were measured during our current study was well after when such potential dampening effects of pregnancy would be likely to be observed. As such, a further opportunity offered by this murine model would be to explore in detail the direct effects on fetal development that occur during the in utero period (that is, earlier end-analysis of maternal immune responses), and deeper profiling of how infection and treatment alter the dynamic of cytokine environment to induce transgenerational priming following the maternal immune activation.

In humans, the currently running, prospective freeBILy study (registration number NCT03779437) gathers information on PZQ usage in pregnant women, a vulnerable population not usually included in mass drug administration programs, assessing the impact of diagnostics and PZQ-based treatment upon health-related quality of life in children, as well as incidence of schistosome infection in these individuals. The planned assessment of infant infection in association with maternal PZQ treatment (37) may add information about the priming effects of maternal infection and immune activation in human children. Through improved application of diagnostics, the results of such studies will facilitate integration of pregnant women into schistosomiasis treatment programs, as well as providing additional important safety information that will further inform guidelines (38). Our current murine-based study shows that maternal PZQ treatment can lead to non-specific priming effects, as well as partial resistance to cognate infection in mice, and would thus support the current WHO recommendation of anthelminthic treatment during pregnancy which is still met with hesitancy in the field. Further studies in human cohorts, such as FreeBILy, may provide additional valuable information on whether the observed effects translate to clinical settings.

Animal experiments were performed under national and European Union guidelines 2010/63, and in accordance with specific approval by regional governmental authorities of Upper Bavaria (license AZ ROB-55.2-2532.Vet_02-17-145). C57BL/6J mice were purchased from Envigo (Germany). All mice were fed standard chow and housed under specific pathogen-free conditions at the Technical Univerity of Munich (TUM), Institute of Med. Microbiology, Immunology, and Hygiene, Munich, Germany.

C57BL/6 female mice were infected with 100 S. mansoni cercariae (using the in-house maintained Brazilian strain, with life cycle kept in mice), via subcutaneous injection into the flank (as demonstrated in Figure 2C , leading to on average 15-20 adult worms by week 8). These reached the 10^th week of infection (end of the peak T_H2 immune response) when half of the infected mice (8 out of 16) were then treated with a single oral dose of PZQ at 400mg/kg in 10% Cremaphor (in H₂O). This was given as a single treated at one-time point. Infected mice not receiving praziquantel received instead on the vehicle (10% Cremaphor), as did the control uninfected mice (all also via oral gavage). The following day, mice were mated with naïve male C57BL/6 mice. Offspring mice were weaned at 21-days of age. Upon end-analysis, dams were sacrificed, and successful infection plus treatment was confirmed by visual inspection of livers, in combination with absence of worms either retrieved via liver perfusion or inspection of intestines. For cognate infection experiments investigating offspring responses following maternal exposure, adult offspring cohorts (minimum of 6 weeks old) were infected via the same method, but were instead analyzed at week 8 after infection to assess peak responses. Infected mice were euthanized with sodium-pentobarbital (Narcoren, Boeringer Ingelheim) at a dose of 5ml/kg (800mg/kg) bodyweight.

For assessment of secreted cytokines in supernatant, cell suspensions were treated with SEA (produced in-house from hamster-derived schistosome eggs, at 20µg/ml), or anti-T cell receptor (TCR) stimulation via anti-CD3/CD28 (1µg/ml each). Unstimulated controls standardized to similar volumes using medium alone. Measurement of SEA-IgG1 and SEA-IgE in serum was performed with in-house ELISA as adapted from (39). Briefly, 1µg of SEA in 50µl PBS was used to coat each well of a 96-well microtiter plate overnight (NUNC). After washing 3x with PBST, these were blocked for 2h at RT with 3% (w/v) in PBS. After a further 3x washing, 100µl of serum diluted in BSA blocking buffer was probed for 2h at RT, washed again, then probed with 100µl detection antibody (biotinylated anti-mouse IgE or IgG1, Biolegend) diluted at 1:400 for 2h at RT. After washing, detection antibody was conjugated to 50µl Avidin-HRP peroxidase diluted 1:1000 for 30min at RT (in the dark), washed, then 50µl undiluted TMB substrate (eBioscience) was added per well and allowed to develop in the dark, before stopping with 50µl of 2 N H₂SO₄.

Granuloma area was calculated from formalin fixed (4%) paraffin-embedded cut sections (4µM) of liver and intestine samples after standard Masson’s trichrome staining and microscopic analysis as previously described (40) using up to 40 granuloma per mouse to calculate an average area µm² per mouse. Determination of S. mansoni egg numbers per liver performed described previously (21) using 5% (w/v) KOH digestion of liver samples. Multiplexed analysis of serum cytokine levels was performed using Bio-Plex Pro Mouse Cytokine 23-plex Assay #M60009RDPD (Bio-Rad), according to manufacturer’s instructions with an initial 1:4 dilution of sample, and run on a Bio-Plex 200 System (Bio-Rad). Culture supernatants were instead analyzed using Mouse Ready-Set-Go ELISA kits (IL-4, IL-5, IFNγ)(eBioscience) or Mouse Duo-Set ELISA kit (IL-10)(R&D), according to manufacturer’s instructions. Alanine aminotransferase (ALT) levels in serum, indicative of hepatic disease, and were measured using a 1:4 dilution in PBS, and Reflotron GPT/ALT tests (Roche Diagnostics, Mannheim, Germany).

Flow-cytometry was performed as previously described (5), with cells were acquired using the flow cytometer Cytoflex S (Beckman Coulter) and initially analyzed in FlowJo 10 (v10 TreeStar). For B-cell analysis of spleen and bone marrow, live cells were gated on by excluding cells that were positive for propidium iodide (PI) (Biolegend), and pre-gated on cells positive for B220-APCH7 and CD19-BV510 (Biolegend). Further staining for B-cell subsets included IgM-BV605, IgD-PE-Cy7, CD23-AF700, CD21-PE-Dazzle594, CD138-PE, CD1d-BB700, CD80-FITC, and CD86-APC (all Biolegend), and MHCII-ef450 (eBioscience). For DC analysis, live cells were gated on by excluding cells that were positive for PI (double positive for PE and ECD), monocytes and monocyte-derived macrophages were excluded by expression of CD64 PerCP5.5 (Biolegend), Tissue resident Macrophages were excluded by the expression of F4/80 PE (eBioscience). cDC were identified by the high expression of CD11c-PE-Cy7 (Biolegend) and MHCII-ef450 (eBioscience). Further staining for DC subsets included XCR1-BV650, CD11b-BV510, CD172a (Sirp-α)-APC/Fire 750, CD24-AF700, CD80-FITC, and CD86-APC (all Biolegend).

For the FlowSOM analysis, fcs files were pre-gated in FlowJo for CD11c⁺MHCII⁺ conventional DCs or B cells based on B220 and/or CD19 expression. The exported samples were quality controlled (removal of low-quality events) using the PeacoQC algorithm (41) then processed with the FlowSOM workflow (28) and concatenated from individual samples. The SOMs were visualized using minimal spanning trees (MST). For dendritic cells, 145 922 cells were used for producing a self-organizing map (SOM) with a 11x11 grid, producing 121 clusters and 10 metaclusters ( Supplement 4D ). For splenic B cells, 1 559 562 cells were clustered in a 9x9 grid, resulting in 81 clusters and 12 metaclusters ( Supplement 4E ). From the bone marrow, 335 603 B cells were used in a 10x10 grid, producing 100 clusters and 12 metaclusters ( Supplement 4F ). To find clusters that were differentially represented in the experimental groups, median percentages of clusters of total number of cells and fold changes per cluster were compared using a Wilcoxon test. Significant clusters where characterized by the mean fluorescence intensities (MFI) of the markers used for clustering, heatmaps show scaled MFI values (pheatmap package https://github.com/raivokolde/pheatmap).

Statistical tests were performed using GraphPad Prism 8.0.2 for Windows (GraphPad Software, San Diego California USA). Data was subjected to D’agostino and Pearson omnibus normality test. Where specified, normalization was calculated using standard or z-values, where Z = (x - mean of original experiment cohort)/standard deviation of original sample cohort, which allowed re-scaling of data between repeat experiments, with normalized value calculated as: Z*(repeat experiment standard deviation) + mean of repeat experiment cohort, to minimize artefacts when pooling repeat experiments. Differences between sample sets with a normal distribution were analyzed using an unpaired, two-tailed Student’s t-test, while non-parametric data was assessed using an unpaired, two-tailed Mann-Whitney U-test, where p-values ≤ 0.05 were considered statistically significant. Analysis between multiple groups was performed using ANOVA or Kruskall-Wallis test for non-parametric data, with Dunn’s post-hoc test for comparison between individual groups. Pearson and Spearman correlations calculated using Rstudio Version 1.4.1106 (^© 2009-2021 RStudio, PBC).

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

The animal study was reviewed and approved by Animal experiments were performed under national and European Union guidelines 2010/63, and in accordance with specific approval by regional governmental authorities of Upper Bavaria (Regierung von Oberbayern, license AZ ROB-55.2-2532.Vet_02-17-145).

ML conceptualized the project, carried out the experiments, analyzed the data, produced the figures and wrote the manuscript. RK performed the experiments and contributed to experimental design, and conceptualized data pipelines, performed analysis, interpreted data and contributed figures and writing of the manuscript. LS performed experiments. UP performed an experiment and contributed to data processing. YH contributed to figure development and visualization, as well as the editing of the manuscript. TS conceptualized the study, supervised its execution and contributed to the discussion and editing of the manuscript. CdC conceptualized the study, supervised its execution including development of the manuscript and figures, and edited the manuscript. All authors contributed to the article and approved the submitted version.

Financial support was provided by the Deutsche Forschungsgemeinschaft DFG CO 1469/16-1.

TS is an employee of Ares Trading SA, an affiliate of Merck KGaA, Darmstadt, Germany.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The authors declare that this study received funding from Merck KGaA, Darmstadt, Germany. The funder was involved in the design of the study, data interpretation and preparation of the manuscript.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.