Maternal schistosomiasis impairs offspring IL-4 production and B cell expansion

Maternal helminth infections are a global public health concern and correlate with altered infant immune responses to some childhood immunizations, but a mechanistic understanding of how maternal helminth infection alters the cellular immune responses of offspring is lacking. Here we establish a model of maternal Schistosoma mansoni infection in dual IL-4 reporter mice. We find that offspring born to mothers infected with S. mansoni have impaired production of IL-4 during homoeostasis, and following immunization with a Tetanus-Diphtheria vaccine. We identified that iNKT cells are the dominant source of IL-4 during early life homeostasis, and that diminished IL-4 production was associated with both reduced B cell and follicular dendritic cell responses. These defects were maintained long-term, affecting memory B and T cell responses. Single-cell RNASeq analysis of immunized offspring identified egg antigen-dependent reductions in B-cell cell cycle and proliferation-related genes. These data reveal that maternal infection leads to long-lasting defects in the cellular responses to heterologous antigens and provide vital insight into the influence of maternal infection on offspring immunity.

Schistosomiasis is an infectious disease caused by trematode parasites of the genus Schistosoma, with S. mansoni, S. haematobium and S. japonicum causing the most human morbidity [1]. This disease affects an estimated 779 million people who are at risk, and 207 million people infected annually [2], with the highest prevalence in adolescents and young adults (10 to 24 years of age) [3]. This leads to a unique impact on women of reproductive age, with 40 million women infected annually. Human studies on schistosomiasis and pregnancy have previously established a link between maternal infection and low birth weight, as well as premature birth and intrauterine growth restriction [4][5][6]. Moreover, examination of the effects of maternal schistosomiasis on the response of offspring to heterologous antigens has demonstrated impaired responses in childhood-vaccine induced immunity, raising the concern that a mother's parasitic status during pregnancy might render early childhood immunization ineffective for years and even decades post-immunization [7]. Decreased responses to bacillus Calmette-Guérin (BCG) have been shown in children sensitized in utero to Schistosoma haematobium [8].
Furthermore, children born to mothers infected with S. mansoni presented lower anti-measles antibody levels [7] at 2 years of age. Altogether, these data suggest that populations where schistosomiasis is endemic, are vulnerable to vaccine failure, and as a consequence, susceptible to deaths due to preventable infectious diseases. Endemic regions may require an adjusted immunization regimen in order to ensure optimal protection against diseases, but mechanistic studies are needed to examine this.
Similar to the human data, mouse studies have shown reduced protection from BCG vaccine against Mycobacterium tuberculosis challenge in mice infected with S. mansoni [9], while infection during gestation has shown increased susceptibility of offspring to subsequent S. mansoni infection [10]. Many authors have described a state of hyporesponsiveness to homologous antigens that has been attributed, in part, to immune factors acquired from the mother and/or previous antigen contact [11]. Previous sensitization of murine offspring to antigens from mothers infected with other parasitic infections such as Wuchereria bancrofti and/or Plasmodium falciparum have shown a Th-2 biased response (higher production of il-4 mRNA and IgE antibody) to diverse antigens when compared to mice from uninfected mothers [12]. However, the effects of schistosomiasis during pregnancy on the immune response in offspring during homeostasis, and when challenged with heterologous antigens remain poorly defined [13]. S. mansoni infection elicits host responses that are similar in humans and mice, so the murine model of schistosomiasis is useful to determine the mechanisms by which prenatal infections induce diminished postnatal immune responses.
In this study we aimed to define the effects of perinatal S. mansoni infection on offspring homeostatic immunity during early life, and determine whether maternal infection causes a deleterious or beneficial effect in response to immunization with heterologous antigens. We have established a novel experimental murine model of maternal schistosomiasis in IL-4 dual reporter mice. Our data from this model demonstrate that maternal schistosomiasis reduces steady-state iNKT and CD4-T cell production of IL-4. This reduction correlates with reduced follicular dendritic cells (FDCs) and circulating plasma and memory B cells at steady state Additionally, the cellular responses to immunization with commercially licensed tetanus/diphtheria vaccine are diminished following primary immunization, leading to a defect in memory TFH precursors. .
Single cell RNASeq analysis revealed reductions in multiple cell cycle/proliferation genes, as well as the critical B cell transcription factor EBF-1. Thus, in utero exposure to S. mansoni egg antigens induces long-lived modulation of the development of humoral and cellular responses via suppression of IL-4 production, and the B cell-stromal cell axis.

Maternal Schistosoma mansoni infection results in in utero egg antigen sensitization and a reduction in the B cell-stromal cell axis
We infected 4get homozygous (IL-4 reporter mice on a BALB/c background [14]) female mice with a low dose of S. mansoni ( Exposure in utero to S. mansoni egg antigens correlated to a significantly diminished bulk plasma cell population at steady-state in both popliteal ( Figure 1D) and hepatic lymph nodes ( Figure 1E), with a 42% reduction of plasma cells in popliteal lymph nodes and a 60% reduction of total plasma cells in hepatic lymph nodes. Follicular dendritic cells trap immunocomplexes via Fc and Complement receptors [17]. Previous studies have established that FDCs serve as sites of B cell antigen capture [18], and that persisting antigens trapped by FDC induce somatic hypermutation [19]. Since we found that pups born to infected mothers have reduced plasma cells in lymph nodes, we sought to determine whether this is accompanied by modifications in Our recent work demonstrated a role for IL-4 in the homeostatic maintenance of FDCs [20], so we analyzed IL-4 competent Th2 CD4-T cells (GFP + ) and found a reduction in Th2 cells in the blood (PBMCs Figure 1 H). The Baff-BaffR signaling pathway is known to be critical for long-term B cell survival, so we measured BaffR expression via FACS and found reduced BaffR in bulk CD19+, plasma, and memory B cells ( Figure 1I). These data suggest that egg antigen sensitization alters two critical pathways that provide survival signals to B cells, BaffR and immune complex presentation. infected mothers, with a 5-fold reduction in IL-4 secreting iNKT cells from pups that were antenatally exposed to SEA antigens (Figure 2 B). These data suggest that steady-state IL-4 production is attenuated by antenatal S. mansoni antigen exposure, and establishes that iNKT cells are the homeostatic source of lymph node IL-4 during early life.

Offspring born to infected mothers have reduced Th2 development following primary immunization with tetanus/diphtheria
Having confirmed that maternal infection induces changes in the immune response in offspring at steady state, we wondered whether the immune response to heterologous antigens could be affected by S. mansoni maternal infection. We immunized 28-35-day-old 4get/KN2 pups subcutaneously (rear footpad injection) from infected and uninfected mothers with ~1/10 th of the human dose of adjuvanted Tetanus/Diphtheria vaccine [20]. At eight days post-immunization we observed that pups from infected mothers exhibited reduced germinal centers in popliteal lymph nodes as well as reduced IL-4 production in the draining lymph node (Figure 3 A). Moreover, offspring born to infected mothers presented significantly reduced IL-4 production within germinal centers (arrowheads indicate HuCD2 + CD4 + cells, Figure 3 A). In addition to reduced frequency, immunofluorescence analysis revealed that germinal center area was reduced in the draining lymph node of mice born to infected mothers ( We then asked if the observed reduction in bulk plasma cells and reduced germinal center area led to reduced anti-tetanus or anti-diphtheria titers. We found no correlation between offspring anti-SEA IgG1 titers and anti-tetanus titers, but did find a correlation (R 2 =0.136, p=.044) between anti-SEA and anti-diphtheria IgG1 titers, suggesting that the diphtheria specific response is more adversely affected by antenatal schistosome exposure than the anti-tetanus response. This is likely tied to the reduction in IL-4 secretion in the germinal center reaction, as our recent work demonstrated that tetanus and diphtheria have a differential dependence on IL-4 for IgG1 class switching [20].

Schistosomiasis during pregnancy alters the memory response to heterologous antigens in the offspring
Immunological memory is key to a protective immune response. Our previous work has demonstrated that subcutaneous immunization induces a persistent germinal center that continues to generate antigen-specific B cells at a low level, and that these memory B cells are critical for an accelerated secondary immune response [16]. Examination of the response to tetanus/diphtheria during maternal schistosomiasis revealed that at 60-90 days post immunization, mice born to infected mothers exhibited a defect in germinal center persistence (

Maternal schistosomiasis transcriptomically modulates the B cell lineage in an egg antigendependent manner
In order to more fully understand the molecular effects of maternal schistosomiasis we performed single cell RNASeq (scRNASeq) using the 10x Genomics platform. Offspring from infected, uninfected, and single-sex infected mothers (6-8 offspring per group) were immunized with tetanus/diphtheria at 28 days of age. On day 8 post-immunization live CD45 + immune cells were sorted from the pooled draining lymph nodes of each group. We aggregated data from 13,437 individual cells (4,879 cells-uninfected mother; 4,388 cells-infected mother; 4,170 cellssingle-sex infected mother) and performed unsupervised clustering analysis based on the similarity of gene expression signatures by using the Seurat single-cell genomics R package [21].
This analysis revealed 23 distinct cell clusters representative of both lymphoid and myeloid lineages, and cluster identity was determined using CIPR [22]. For cell clusters in which the algorithm could not make a clear call such as TFH cells, we resorted to differential expression analysis between clusters to identify distinguishing markers (Supplementary Day 14 post-immunization, we find that lymph node plasma cells in offspring from uninfected mother expand, but the offspring from infected mother fail to expand the plasma cell compartment in response to tetanus and diphtheria antigens (Figure 6 E). These data strongly support the conclusion that in utero exposure to schistosome egg antigens modulates the cell cycle and proliferative capacity of B cells, resulting in deficient plasma cell production to both microbiota and immunization.

Discussion
In this study we report the generation of a maternal Schistosoma mansoni infection model in dual IL-4 reporter mice. Our results indicate that in utero egg antigen sensitization occurs, which is consistent to what has been published by other authors [24] and confers a state of immunehyporesponsiveness in the offspring at steady state, this study does not differentiate between trans-placental antigen exposure and sensitization through breastfeeding; however, it has been previously suggested that exposure to parasitic antigens, either in utero or via breast milk, diminishes the heterologous response [24], and in some other cases nursing by infected mothers protected offspring against infection with the same helminth [25]. We observed anti-S. mansoni egg antigen-specific IgG1 (the dominant isotype produced against egg antigens [16]) at 35, 90 and 180 days of age in mice from infected mothers and no detectable titers in mice born to uninfected controls, suggesting either maternal antibody, antibody-secreting cell transfer to the offspring, or in situ antibody production by the offspring. The placental transfer of different IgG subclasses has been well documented [26][27][28]. In our model, we observed maternal antibodies in KN2 homozygous (IL-4 deficient) pups at 35, 60, and 90 days of age, suggesting a sustained antibody response to SEA from either maternal antibodies or cells, since our previous work has demonstrated that IL-4 deficient mice fail to mount an IgG1 response to SEA [16]. The presence of an antigen specific humoral response can potentially translate to protection, or alternatively, inhibit vaccination induced protection, as is the case for measles and other vaccinations, in humans and other mammals [29]. In humans, the persistence of maternally-derived antibodies is  [32]. Recent work in our laboratory has demonstrated that IL-4 is required for FDC maintenance and positioning at steady via induction of lymphotoxin α/β [20]. In light of that, we sought to determine the cellular source of steady-state lymph node IL-4 production in this model. We found that iNKT cells are the major producers of IL-4 in control offspring during early life, and that steady-state secretion of IL-4 is almost eliminated in offspring born to infected mothers. This data suggests that the observed defects in FDC numbers and stimulatory capacity may be due to insufficient IL-4 during lymph node development and maturation. Bolstering this is the finding that there is a significant reduction in IL-4 competent (GFP + ) CD4 T cells in the peripheral blood of these mice. This evidence suggests that prenatal helminth sensitization has a lasting effect on offspring immunity, and that offspring from mothers that were infected during pregnancy could be at a higher risk of infection with various pathogens due to a defect in homeostatic immunity as has been previously postulated [33], which will be the focus of future studies in our lab.
We further evaluated the immune response following primary immunization with a commercially available alum adjuvanted Tetanus/Diphtheria vaccine used in clinical settings.
There is extensive evidence in humans of impaired response to immunization [7,34], but little is known of the mechanism(s) controlling this defect. We observed that following primary immunization, the pups from infected mothers exhibit a defect not only in expansion of FDCs, but also in the development of T follicular helper and germinal center B cells. We hypothesize that the combined FDC and Tfh deficiency leads to the diminished response to primary immunization, as well as diminished memory T and B cell responses. Indeed, at 8-days postimmunization, there is a reduction of bulk memory B cells and IgG1 + B cells, which could potentially have an impact in the long-term maintenance of humoral immunity necessary for protection following vaccination. This is bolstered by our data at day 14 post-immunization where germinal center formation in a reactive lymph node was impaired in 4get/KN2 pups from infected mothers in comparison to pups to uninfected mothers. Acute antigen exposure causes lymphocyte proliferation, which is followed by a pool of quiescent long-lived IL-7R + memory cells capable of potent response to challenge with the same antigen. [35]. We have previously determined that in a type 2 response, secondary TFH cells are generated in large part by recruitment of circulating Th2 memory T cells (IL-7R + GFP + ) back to the lymph node, and that their interaction with memory B cells is critical to the secondary plasma cell response [16]. In cycle/proliferation, such as Snpre and Snprg, which are part of the spliceosome machinery whose overexpression have been shown to drive cell proliferation in multiple models [36,37]; Rbx1, a known modulator of DNA replication licensing proteins [38]; Tmsb10, a regulator of actin cytoskeleton with roles in B cells [39,40]. Concomitantly, Jun and Junb, which are involved in both B cell differentiation and B cell receptor signaling [41,42] were increased in these clusters.
We validated that Ebf-1 expression is reduced in both bulk CD19 + and plasma cells at steadystate (28-35 days of age) and found that a significantly lower proportion of steady-state plasma cells express Ki-67, a state that persists following immunization, strongly indicating that there is indeed a defect in B cell expansion and differentiation in response to foreign antigens.
Concomitantly with defects in B cell proliferation, we find a consistent defect in follicular dendritic cell expansion and expression of C21/35. While our data suggests a role for iNKT cell IL-4 production in FDC maintenance, the data is not definitive. Future work will focus on identifying the specific homeostatic role of iNKT cell derived IL-4, and the molecular mechanism that underlie both iNKT cell production of IL-4, and diminished FDC cell function in the context of maternal infection. There is a body of literature supporting the many effects of the prenatal environment on inflammatory diseases during adolescence and adulthood. It has been hypothesized that inflammatory responses and infections during pregnancy might alter epigenetic profiles in the fetus [43]. Future studies linking the epigenetic effects of schistosomiasis infection during pregnancy and immune responses in offspring will help elucidate potential pathways involved in the immune hypo-responsiveness observed in this model.
One of the key goals of vaccination is to induce an immunological memory that is protective upon re-challenge with the same antigen [44][45][46]. Hence, we explored whether there was a difference between pups from infected and control mothers in the development of a sustained germinal center and maintenance of memory B and T cell pools. Pups were immunized and their immunological responses were assessed at >70 days post immunization. Surprisingly, offspring from infected mothers mounted a weaker cellular response, with significantly smaller germinal centers that correlated to a significant reduction in long lived Tfh, memory Th2 cell, and FDCs, as well as reduced complement receptor 2/1 (CD21/35) expression. Importantly, our previous work has established precursor bulk memory T cells (CXCR5 + PD-1 -) as critical to the secondary TFH cell response, and this cell population was significantly reduced at the memory time-points in pups from infected mothers in comparison to pups from uninfected mothers. These data strongly support the premise that in utero exposure to schistosome egg antigens transcriptionally reprograms the proliferative potential of the B cell lineage to foreign antigens, as well as the immune complex presenting capacity of FDCs, leading to long-lived suppression of the stromallymphocyte axis and the ability to mount a B cell response to immunization.

Conflict of Interest
The authors declare no competing interests.        Statistical analysis for C-E was calculated with Student's t-test with Welch's correction.

Isolation of cells and flow cytometry
Cells were isolated from popliteal lymph node (pLN), hepatic lymph node (hLN), whole blood, and spleen. For follicular dendritic cells (FDC) analysis, tissues were digested as previously described ( [20,47]. Briefly, lymph nodes were digested at 37°C for 20 minutes, with occasional inversion of the contents and filtered through 100 μ m filters. Single-cell suspensions were stained with surface markers and intracellular staining was performed as previously described [48].

Immunofluorescence microscopy
Whole tissue (popliteal lymph node or hepatic lymph node) was collected and placed in Tissue- Tek optimum cutting temperature compound (OCT) (Thermo Scientific) and frozen in liquid nitrogen. Serial cryostat sections (10μm) were collected using a Leica CM 1850. Sections were then air-dried and fixed in ice-cold 75% acetone/25% ethanol for 5 mins. Sections were rehydrated in PBS for 5 to 10 minutes and blocked using biotin blocking kit (Vector Laboratories) followed by incubation with 1%(v/v) in PBS of rat and rabbit serum. Staining with appropriate antibodies was done overnight at 4°C followed by secondary staining for 1 hour at room temperature. Coverslips were mounted using ProLong anti-fade reagents (Life Technologies).
Images were acquired with a Leica TCS Sp5 Laser Scanning Microscopy with an average grid size of 3x3. Images were taken with a 20x objective at a resolution of 1024x1024. Image postprocessing was done using Fiji is Just ImageJ software (1.47v).

ELISA
Schistosoma mansoni egg antigens (SEA) and tetanus and diphtheria-specific IgG1 endpoint titers were determined by enzyme-Linked immunosorbent assay (ELISA) using the mAb X56 (BD) and Immulon 4HB plates (Thermo Fisher Scientific) as previously described [20]. Briefly, plates were coated with 2μg/mL tetanus (List Labs), diphtheria (Sigma), or SEA. The following morning plates were blocked with 1% milk and incubated with serial dilutions of primary antibody, followed by incubation with anti-mouse IgG1 ads-HRP antibody (Southern Biotech) and ABTS substrate. Plates were read at 405 nm at room temperature on a BioTek plate reader.

Single Cell RNASeq
Sample preparation and sequencing. Single-cell suspension form 6-8 draining lymph nodes per group were pooled and stained with Dapi (Sigma) and anti-CD45. Live CD45+ cells were sorted via BD FACSAria cell sorter and washed once in PBS containing 0.04% BSA. Samples were then processed for SCseq via a 10× platform according to the manufacturer's instructions (10× Genomics). Paired-end RNASeq (125 cycles) was performed via an Agilent HiSeq nextgeneration sequencer. Sequencing reads were processed by using 10× Genomics CellRanger pipeline and further analyzed with the Seurat R package. The effect of mitochondrial gene representation and the variance of unique molecular identifier (UMI) counts were regressed out from the data set prior to analysis. Gene expression signatures defining cell clusters were analyzed after aggregating 3 samples (uninfected mother, infected mother, single-sex infected mother). The raw data from scRNASeq experiments in this manuscript can be found in the NCBI's Gene Expression Omnibus database (TBD).

Identification of cell clusters
Cells in our data set were clustered by using the FindClusters function of the Seurat analysis package, which identifies clusters via a shared nearest neighbor (SNN) modularity optimization- multiplies ImmGen log-ratio values with the log-ratio of matching genes that are differentially expressed in each cell cluster in the SCseq dataset; and (d) sums up scores from all the genes to yield an aggregate identity score for each ImmGen cell type for a given SCseq cluster. In this approach, genes that are differentially upregulated or downregulated in both ImmGen and SCseq data sets contribute to the immune identity score more heavily (a positive number is obtained when 2 log-ratio values with the same sign are multiplied). In contrast, if a gene is inversely regulated in ImmGen and SCseq clusters, the immune identity score is reduced. Through this method, the correlation between the gene expression signatures of SCseq cell clusters in our study and ImmGen data subsets assists in determining the cluster identities. In cases where this algorithm is unable to make a clear call (as T follicular helper cells and myeloid cells), we surveyed the expression of known genes in the data set and performed differential expression analyses between closely related cell clusters. This approach allowed us to further differentiate subsets. Upon naming the clusters, the Seurat R package was used to create plots for the expression of selected genes. GSEA analysis was performed by using fgsea R package [49], after ranking genes using a signal-to-noise metric [50].

Statistical analysis
Statistical analyses were performed using either a non-parametric Mann-Whitney test, unpaired