H. bakeri infects the small intestine of mice where the ingested infective larvae (L3) invade the submucosa of the duodenum and upper jejunum for development into the pre-adult stage, which returns to the small intestinal lumen at day 8 post-infection (19). The tissue damage provoked during the development of the histotropic larval stage results in a strong immune response reflected by the substantial increase in size and cellularity of the mesenteric lymph nodes (MLN) draining the duodenum and jejunum ( Figures 1A, B ). We further noticed that the size of the two lymph nodes draining the liver, namely celiac (CLN) and portal lymph nodes (PLN), were increased in mice infected with H. bakeri for 2 weeks ( Figures 1A–C ). Earlier studies demonstrated the multi-organ drainage by the CLN, comprising the liver and pancreas, but also the upper duodenum (5, 8, 13, 14), whereas the PLN solely drains the liver and pancreas (5). In accordance with the drainage of the upper SI shared between MLN and CLN, both sites comprised significantly elevated and similar frequencies of IL-4 and IL-13 competent Th2 cells during larval development (day 6 p.i.) and after transition to the patent stage of infection (day 14 p.i., Figure 1D ). However, Th2 cells were also significantly increased in the PLN of infected mice at both timepoints, albeit at lower levels compared to MLN and CLN ( Figure 1D ). In parallel to the induction of Th2 cells, elevated IFN-g production by CD4+ T cells was seen in all LN samples at day 6, but the percentage of T-bet+ Th1 cells remained unchanged, and the rate of IFN-g production leveled down in CD4+ T cells at day 14 post-infection ( Figure 1E ).

We further sought to compare the expression of Th2 effector functions in the different lymph nodes and therefore assessed the activity of CD8+ T cells, as earlier work reported the IL-4 dependent expansion of CD8+ memory-like T cells which was increased along the expansion of CD4+ Th2 cells in H. bakeri-infected mice (20, 21). Confirming these earlier studies, we determined a highly significant increase in the proliferation of CD8+ T cells in all active lymph nodes of infected mice ( Figure 1F ). Infection-derived CD8+ T cells also displayed elevated T-bet expression, which translated to significantly stronger IFN-g responses upon in vitro stimulation compared to naïve controls ( Figure 1G and data not shown). Interestingly, the comparable profiles observed across different lymph nodes suggest that Th2 cells are not only present but also actively induced in the PLN, where they contribute to the expansion of CD8+ VM cells as early as day 6 post-infection. This finding makes it unlikely that Th2 cells were initially induced in MLN or CLN and subsequently migrated to the PLN, as the PLN lacks a direct connection to the gut, a key site for Th2 induction.

Despite our claims that Th2 cell levels in the PLN are reduced by half compared to MLN and CLN ( Figure 1D ), the PLN exhibits a comparable magnitude of CD8+ T cell responses. This observation suggests that lower Th2 levels in the PLN, if present, do not negatively impact CD8+ T cell responses at day 6. Furthermore, IFN-g expression of CD8+ T cells was positively correlated with the extent of Th2 responses across the LNs.

Hence, while the majority of Th2 cells induced in response to a strictly enteric nematode infection derive from the larger MLN and CLN draining the site of infection, a more limited number of Th2 cells derive from the liver-draining PLN. Furthermore, the correlation between Th2 response and local expansion of memory-like CD8+ T cells indicates robust Th2 effector functions expressed in the active LNs. This included the liver-draining PLN which is activated during enteric nematode infection despite the lack of the lymphatics directly connecting the PLN to the site of infection.

As we observed that the LLNs are a site of Th2 induction during H. bakeri infection, we next sought to determine if these T effector cells (Teff) contribute to anti-helminth immunity in the gut. Therefore, we assessed the expression of the gut-homing marker CCR9 (22) in Th2 cells in the different lymph nodes. CCR9-mediated gut homing of GATA-3+ cells is shown to correlate with high expression of aldehyde dehydrogenase (ALDH) in DCs isolated from the MLN of mice early during infection (22, 23). In addition, the early stage of infection drives the early recruitment of T cells to the infected gut and determines worm clearance (24). We, therefore, looked at the gut-homing marker CCR9 at day 6 to get an idea of how these cells behave in extra-lymphoid compartments and their contributions to controlling gut infection.

We observed that Th2 cells from the PLN express very low levels of CCR9 at days 6 and 14 compared to these cells in the CLN and MLN ( Figures 2A, B ). Higher RALDH activity (retinoic acid metabolism) in CD103+ MLN-DCs and an enhanced vitamin A metabolism during the early stage of infection resulted in higher CCR9+ T cells homing to the gut (25–27). Therefore, we assessed ALDH activity in CD103+ DCs at 6dpi using the Aldefluor kit. We found that low CCR9 expression in Th2 cells in the PLN was accompanied by low ALDH expression by CD103+ CD11c+ DCs. In contrast, the higher expression of gut-homing markers by Th2 cells from the CLN and MLN was mirrored by the higher expression of ALDH by DCs ( Figure 2C ). Together, this data shows that MLN and CLN express higher proportions of the gut-homing marker CCR9 compared to PLN as a result of higher levels of ALDH activity in their DCs.

The quantification of IL-4 and GATA-3 expression in CD4+ T cells in the liver versus gut-draining lymph nodes indicated a trend for less extensive Th2 effector cell generation in PLN compared to the ‘classical’ gut-draining lymph nodes. However, the three sites displayed similar features associated with IL-4 signaling in worm infections, such as the strong outgrowth of B cells and the expansion of VM CD8+ T cells ( Figure 1H ). Furthermore, both B cell proliferation and IFN-g competence of CD8+ T cells strongly correlated with the individual levels of Th2 responses in a given organ rather than with the organ itself ( Figure 1F ). We, therefore, asked if qualitative differences in the B cell responses might be associated with the extent of Th2 differentiation and/or the balance of Th2 effector versus follicular T helper cell differentiation across the different lymphatic sites. We hence quantified IgG1-expressing cells by intracellular stains at day 6 and 14 post-infection and found a significant rise of IgG1+ cells in the three lymph nodes at day 14 post-infection ( Figure 3A ). As shown for CLN-derived cells in Figures 3B, C , IgG1+ cells were composed of B220+ cells which displayed a germinal center (GC) B cell phenotype by the expression of the canonical GC transcription factor Bcl6, the binding of peanut agglutinin (PNA) and the elevated expression of GL7. In contrast, a smaller subset of B220-IgG1+ cells lacked GC markers and expressed CD138 and Ly6C associated with plasma cells (28). Furthermore, the B220- subset had lost the CXCR5 expression required for positioning in the B cell follicles ( Figures 3B, C ) (29–31). IgG1+ PC accounted for a prominent part of the IgG1+ cells in both MLN and CLN of infected compared to control mice ( Figure 3D ). In contrast, only 2 out of 8 mice harbored a prominent IgG1+ PC population in the liver-draining PLN at day 14 p.i, and these individuals also displayed the highest IgG1+ PC responses in MLN and CLN ( Figure 3D ). Because earlier studies demonstrated the capability of Th2 cells for the instruction of extrafollicular B cell responses (32, 33), we tested for associations between the magnitude of Th2 and overall IgG1 responses as well as early IgG1+ PC differentiation across the LN at day 14 p.i. As shown in Figures 3E, F , the frequencies of Th2 cells were strongly correlated with the overall outgrowth of IgG1+ cells ( Figure 3E ) as well as with the subset of IgG1+ B220- PC cells ( Figure 3F ). In contrast, the more homogenous GC B cell responses were not linked with the extent of Th2 differentiation ( Figures 3G, H ). Furthermore, while the PLN of most of the infected mice comprised low proportions of IgG1+ PC, the B220+IgG1+ subset was the most enriched in GL7+ Bcl6^high GC B cells in the liver-draining LN ( Figure 3I , Supplementary Figure 3 ). These data suggested that the liver-draining PLN provided an environment suited for the generation of germinal center reactions while being less prone to the rapid instruction of plasma cell differentiation in the context of enteric nematode infection ( Supplementary Figure 3 ).

We complemented the investigation of B and Th2 cell responses by the quantification of follicular T helper (TFH) cell response at days 6 and 14 post-infection. TFH cells are required for efficient germinal center reactions, resulting in the generation of high-affinity IgG and the development of memory B cells (34), whereas Th2 cells providing IL-4 were shown to be sufficient for extrafollicular IgG1 class switching and the generation of short-lived PC (35–37). As shown in Figure 3J , particularly high frequencies of TFH cells identified by the expression of the chemokine receptor CXCR5 together with the inhibitory signaling molecule PD-1 were detectable in CLN at day 6 post-infection. However, CD4 cells isolated from the three LN at day 14 post-infection comprised similar frequencies of TFH cells ( Figure 3J , Supplementary Figure 2 ), matching the homogenous proportions of B220+IgG1+ GC B cells ( Figure 3G ). TFH responses were not mirrored in the overall rate of IgG1 class switching but negatively correlated with the extent of early plasma cell responses ( Figures 3K, L ).

IL‐21 has been reported to be integral to protective humoral immunity, where it acts on B cells from the outset of an adaptive immune response to promote their expansion and contribution to germinal center reaction (38–40). Therefore, to assess the contribution of IL-21 to the observed B cell responses, single-cell suspensions from the MLN, CLN, and PLN were stimulated with H. bakeri excretory-secretory product (HES) or antibodies against CD3 and CD28 (strong T cell receptor stimulants), cultured for 72 hours, after which the supernatants were harvested and tested for IgG1 and IL-21. We found that all three LNs harbor B cells producing IgG1 at day 14. However, we detected IgG1 in the supernatants of PLN cells on day 6 ( Figure 3A ), suggesting an early and sustained IgG1 production by B cells in this lymph node. Interestingly, of all the LNs, we could only detect IL-21 in the supernatant of cultured PLN cells in response to T-cell receptor stimulation ( Figures 3B, C ). In summary, these data suggest that the PLN environment is conducive in terms of IL-21 response by TFH and could support GC and quality B cell responses.

To confirm the assertion that the PLN environment supports higher quality B cell responses compared to the MLN and CLN, we run an SDS-PAGE of ‘untouched’ HES and de-glycosylated HES using either PNGase F or O-glycosidase. The HES was de-glycosylated to remove carbohydrate moieties (present in adult ES) on antigenic epitope, the so-called Glycan A and B, which have been shown to induce non-protective antibodies and act as a decoy that generates ineffective humoral responses during primary H. bakeri infection (41). From the SDS-PAGE gel, ‘untouched HES’ as well as de-glycosylated HES show the two band sizes that correspond to Venom allergen/Ancylostoma secreted protein-Like-1 (VAL-1) and acetylcholinesterase-1 (ACE-1) proteins, depicted by the red and black boxes, respectively ( Figure 4A ). These proteins are suggested to represent two major vaccine targets that induce protective immunity against H. bakeri (42). Western blots were thereafter run using pooled sera from d14 H. bakeri infected BALB/c mice as well as the supernatant of MLN, CLN, and PLN cell suspensions after 72 hours of culture with HES. We found that the serum IgG1 reacted strongly to these two bands (red and black boxes – Figure 4B ).

Next, we probed whether the supernatants from MLN, CLN, and PLN show similar binding patterns. We show that the supernatants from the MLN and CLN only reacted to the second band, ACE-1 (black box) while the supernatant from the PLN reacted to the two bands just like the pooled sera (red and black boxes and red arrows – Figure 4C ). We also show that both the ‘untouched’ and de-glycosylated HES showed similar binding patterns when probed with pooled sera or supernatant ( Figures 4B, C ). Furthermore, we show that with the supernatants, the binding to the de-glycosylated HES was weaker shown in fainter bands than the ‘untouched’ HES. The binding intensities to the ‘untouched’ HES decreased in MLN, compared to CLN and PLN, with the PLN showing the highest intensity. The data together show that the PLN has a similar binding pattern to the pooled sera, reacting to the H. bakeri vaccine target proteins VAL-1 and ACE-1.

In summary, our data show that the PLN, even though it has restricted Th2 responses, compared to the MLN and CLN, possibly due to its high IFN-γ environment, is skewed towards TFH response. Even though the PLN shows lower PCs compared to the MLN and CLN at day 14, the PLN is able to form more GC, which enables it to produce more quality B cells and high-affinity antibody response ( Figures 3D, I ), evident in the binding of its supernatants to both VAL-1 and ACE-1 compared to the MLN and CLN ( Figures 4A–C ). Together, our data shows that the PLN is important for quality B cell responses during H. bakeri infection.

The animal experiments were done in accordance with the National Animal Protection Guidelines with approval from the German Animal Ethics Committee for the Protection of Animals (G0176/17, G0113/15, H0438/17). Female BALB/c mice were ordered from Janvier (Saint-Berthevin, France) and were infected with 200 3rd-stage larvae of H. bakeri by oral gavage. Mice used in the experiments were anesthetized with xylazine and ketamine and sacrificed by cervical dislocation. All animal experiments were performed in accordance with the National Animal Protection Guidelines and approved by the German Animal Ethics Committee for the Protection of Animals (LAGeSo, G0176/20).

MLN, CLN, and PLN were processed into single-cell suspensions using 70 µl cell strainers (BD Biosciences, San Jose, CA, USA). After washing in RPMI (1% FCS, 100U/ml penicillin, 100 mg/ml streptomycin (PAA, Wien, Austria), any residue red blood cells were removed with an erythrocyte lysis buffer, followed by two further washing steps.

For the detection of gut homing markers, cells isolated from lymphatic organs were stained for CCR9 (clone CW-1.2; PE-Cy7) and α4β7 (clone DATK32, biotin) at 37°C for 30 minutes, followed by incubation with a fixable viability dye (eFluor506 or eFluor780) on ice for 5 minutes. After labeling CD4 (clone RM4-5; Alexa 700, Brilliant Violet 510, or PerCP), cells were fixed using fix/perm buffer (Thermo Fisher) for 30 minutes at room temperature and stained intracellularly using the following reagents: FoxP3 (clone FJK-16s; PE-eFluor610, Alexa 488, or PerCP-Cy5.5), GATA-3 (clone TWAJ; eFluor 660), T-bet (clone 4B10; PE), and Ki-67 (SolA15, eFluor 450 or PE-Cy7). Streptavidin coupled with Brilliant Violet 605 was used as a secondary conjugate. Dendritic cells were analyzed using antibodies against MHCII (clone M5/114.15.2; APC), CD11c (clone N418; Brilliant Violet 421), and CD103 (clone 2E7, APC). Aldehyde dehydrogenase (ALDH) activity was determined using the Aldefluor kit (Stemcell Technologies). Non-specific binding of antibodies was prevented by adding FcγRII/III blocking antibody (clone 93). All antibodies and other reagents were from BioLegend, Thermo Fisher, or BD Biosciences. A catalog of the respective antibodies used in the multi-color flow cytometry can be seen in Supplementary Table 1.1 in the Supplementary Figures . Cells were acquired on FACSCantoTM II (BD Biosciences) or FACSAriaTM III (BD Biosciences), and data was analyzed using FlowJo (Tree Star Inc., Ashland, USA). The antibodies used for the detection of surface and intracellular markers are described in Supplementary Table S1 . Dead cells were excluded using eFluor780 or eF560 fixable viability dye (Thermo Fisher, Waltham, USA). For intracellular staining of cytokines and transcription factors, cells were fixed and permeabilized using the Fixation/Permeabilization kit and Permeabilization buffer from ThermoFisher/eBioscience. Analyses of the samples were done using a Canto II flow cytometer and an Aria cell sorter (BD Biosciences, Heidelberg, Germany) and FlowJo software Version 10 (Tree Star Inc., Ashland, OR, USA) was used to analyze the data. The data for naïve and infected mice (days 6 and 14) presented in this study are pooled from at least two independent experiments with 3-4 mice per group in each experiment.

For the analysis of cytokine production capacities of cells, MLN, CLN, and PLN cells were plated out per well in a round-bottom 96-well cell culture plate in a final volume of 200μL RPMI medium, containing 10% FCS, 100U/mL penicillin, and 100μg/mL streptomycin (all from PAA, Pasching, Austria). The cells were then stimulated with 1μg/mL of PMA/ionomycin and incubated for 3-4 hours at 37°C and 5% CO₂. The cells were further stained with a dead cell marker and other surface stains, fixed and permeabilized for intracellular staining, followed by acquisition via flow cytometry.

For the analysis of parasite-specific IgG1 and IL-21, 1x10⁵ of MLN, CLN, and PLN cells were plated out per well in a round-bottom 96-well cell culture plate in a final volume of 200μL RPMI medium containing 10% FCS, 100U/mL penicillin and 100μg/mL streptomycin (all from PAA, Pasching, Austria). The cells were stimulated with either anti-HES (10μg/mL) or without. The cells were then incubated for 72 hours at 37°C and 5% CO₂. The cell suspensions were then spun down, and the supernatant was harvested and measured via sandwich. Absorbance was measured on a Biotek Synergy H1 Hybrid Reader at a 450nm wavelength.

Adult H. bakeri excretory and secretory product (HES) were either deglycosylated with Peptide N-glycosidase F (PNGase F), which cleaves N-linked glycans from glycoproteins, or Endo-α-N-Acetylgalactosaminidase (O-glycosidase), that catalyzes the removal of O-disaccharides from glycoproteins. These enzymes are from New England Biolabs^®. Glycosylated and deglycosylated HES were run on an SDS-PAGE, and Western blot was done on these using either serum from d14 H. bakeri infected BALB/c mice or supernatants from cell suspension from MLN, CLN, and PLN after culturing with HES for 72 hours according to the protocol published in (65). Briefly, the SDS-PAGE was blotted unto a nitrocellulose membrane, blots were blocked in 2% dry skim milk in TBS with 0.05% Tween 20 for 2 hours at room temperature and probed with either pooled sera (1:200 dilutions) or supernatants from cell suspensions of MLN, CLN, and PLN (1:2 dilutions) at 4 ⁰C overnight. This was followed by 3 times washing in TBST, after which AP-conjugated secondary antibodies (anti-mouse IgG1 1:3000, antibodies-online.com) for 2 hours at room temperature. The blots were then washed 3 times in TBST and developed with SIGMAFAST™ BCIP^®/NBT substrate (Sigma-Aldrich^®).

All statistical analyses were performed using GraphPad Prism Software (San Diego, CA, USA). Normality was tested with the Shapiro-Wilk test, followed by ordinary one-way-ANOVA or Kruskal-Wallis test and Tukey’s or Dunn’s multiple comparison test. Comparisons of the two groups were performed with an unpaired t-test or Mann-Whitney test.