Oral Challenge with Wild-Type Salmonella Typhi Induces Distinct Changes in B Cell Subsets in Individuals Who Develop Typhoid Disease

A novel human oral challenge model with wild-type Salmonella Typhi (S. Typhi) was recently established by the Oxford Vaccine Group. In this model, 104 CFU of Salmonella resulted in 65% of participants developing typhoid fever (referred here as typhoid diagnosis -TD-) 6–9 days post-challenge. TD was diagnosed in participants meeting clinical (oral temperature ≥38°C for ≥12h) and/or microbiological (S. Typhi bacteremia) endpoints. Changes in B cell subpopulations following S. Typhi challenge remain undefined. To address this issue, a subset of volunteers (6 TD and 4 who did not develop TD -NoTD-) was evaluated. Notable changes included reduction in the frequency of B cells (cells/ml) of TD volunteers during disease days and increase in plasmablasts (PB) during the recovery phase (>day 14). Additionally, a portion of PB of TD volunteers showed a significant increase in activation (CD40, CD21) and gut homing (integrin α4β7) molecules. Furthermore, all BM subsets of TD volunteers showed changes induced by S. Typhi infections such as a decrease in CD21 in switched memory (Sm) CD27+ and Sm CD27- cells as well as upregulation of CD40 in unswitched memory (Um) and Naïve cells. Furthermore, changes in the signaling profile of some BM subsets were identified after S. Typhi-LPS stimulation around time of disease. Notably, naïve cells of TD (compared to NoTD) volunteers showed a higher percentage of cells phosphorylating Akt suggesting enhanced survival of these cells. Interestingly, most these changes were temporally associated with disease onset. This is the first study to describe differences in B cell subsets directly related to clinical outcome following oral challenge with wild-type S. Typhi in humans.


Introduction
Salmonella enterica serovar Typhi (S. Typhi) is a human-restricted pathogen and the agent responsible for typhoid fever, a disease that continues to be a major global public health problem [1][2][3]. Due in part to the absence of a suitable animal model, several aspects of the human response to S. Typhi infection remain to be explored [4,5]. A successful human oral infection model of S. Typhi, which allowed studying various aspects of the host-pathogen interaction as well as test vaccines and alternative treatment options, was developed forty years at the University of Maryland [4,[6][7][8][9][10]. A new controlled human infection model of S. Typhi was recently developed at the Centre for Clinical Vaccinology and Tropical Medicine, University of Oxford (Oxford Vaccine Group). In this new model, participants were challenged with up to 10 4 CFU of S. Typhi (Quailes strain) in a sodium bicarbonate buffered solution. This dose resulted in 65% of participants being diagnosed with typhoid fever (referred here as typhoid diagnosis -TD-) [11].
Immunity to S. Typhi is not well understood, though it is believed to be complex involving local and systemic antibody and cell mediated immunity (CMI) components. Until very recently, the principal role of B cells was considered to be antibody production and antigen presentation; however, recent reports have demonstrated that the B cell compartment is quite complex and involves multiple subsets [12][13][14][15][16]. Interestingly, various B memory (B M ) subsets have been associated with certain diseases and novel functions [12,[17][18][19]. For example, using one of the most widely accepted classification schemes (IgD/CD27), four B M subsets are defined [12]: (i) naïve [IgD+CD27-], (ii) unswitched memory (Um) [CD27+IgD+], (iii) switched memory CD27+ (Sm CD27+) [CD27+IgD-] and (iv) Sm CD27-[CD27-IgD-]. Among these subsets, the frequency of Sm CD27-cells has been shown to increase in patients with respiratory syncytial virus (RSV) infections [12]. Additionally, Um cells seem to play a crucial role in response to encapsulated pathogens (e.g., S. pneumonia or N. gonorrhea) [18,20,21]. In the case of S. Typhi, there is evidence of the importance of the B cell compartment in protection from disease. For example, the purified Vi polysaccharide administered as a parenteral vaccine, which is a T-independent antigen that activates only B cells, is efficacious in the prevention of typhoid fever; therefore, demonstrating that serum Vi antibodies can mediate protection. Additionally, another typhoid vaccine, Ty21a, elicits serum IgG antibodies against lipopolysaccharide (LPS) O-antigen, which correlates with the level of protection conferred by some, but not other, formulations and immunization schedules [4,22]. Further evidence comes from studies showing that B cells are able to cross-present Salmonella antigens and activate CD8+ T cells, a process that depends on CD4 T cell help [23]. In the rodent model of typhoid (S. Typhimurium) more evidence of the role of B cells has been reported using B celldeficient mice (Igh-6 -/or Igμ -/-) [24][25][26]. For example, B cell-deficient mice that were vaccinated with a live-attenuated strain of Salmonella and subsequently challenged with a virulent strain (SL1344) were unable to resist infection [27]. Of note, adoptive transfer of immune serum to vaccinated B cell-deficient mice (Igμ -/-) the day previous to challenge (virulent Salmonella SL1344) successfully reconstituted their immunity [25]. Moreover, in B cells, Toll-like receptor (TLR) stimulation appears to drive appropriate development of humoral responses, as demonstrated in mice with B cells deficient in MyD88. In these animals, Salmonella infections resulted in impaired IgG2b, IgG2c, IgA and IgM responses compared to mice with functional MyD88 [28]. These animals also showed impairment in the development of IFN-γ effector cells mainly due to deficient cytokine production by B cells [29], suggesting a role for B cells in T cell differentiation, which depended on TLR stimulation. Importantly, in human B cells, TLR stimulation (e.g., TLR-2, TLR-5, TLR-7 and TLR-9, but not TLR-4 since human B cells do not express this receptor) has also been suggested as a requirement for effective activation [30]. Other studies are providing insights into the interactions between Salmonella and B cells [31]. For example, B cell infection by S. Typhimurium was reported and this process depended on antigen-specific BCRs (on the B cell side) and a functional Type-III secretion system (T3SS) (on the bacterium side) [32][33][34]. Additionally, S. Typhimurium is able to modulate ongoing immune responses by facilitating the development of regulatory B cells (immune-suppressive) [35,36]. Finally, S. Typhimurium can induce B cells survival, a process that dependents on inhibition of the inflammasone and that requires the bacteria T3SS SPI-1 [37]. Induction of B cell survival benefits Salmonella because the bacteria use the cells as a survival and dissemination niche [33]. Finally, while the existence of human B M cells to S. Typhi was suspected for many years, only recently has our group provided the first direct evidence for the presence of S. Typhi-specific B M cells (IgA and IgG anti-LPS and -Vi) in volunteers immunized with vaccines for S. Typhi [38,39]. Despite these advances, our knowledge regarding human B cell responses in typhoid fever is still limited. For example, it is unknown whether a specific B cell subset has a predominant function in typhoid disease as described for other pathogens and the changes induced in these cells following immunization and/or infection. Furthermore, whether similar Salmonella-B cell interaction as described above for S. Typhimurium are operational in humans infected with S. Typhi remain to be explored. Evaluation of these phenomena in humans has been impaired since specimens from individuals infected with wild-type (wt) S. Typhi are difficult to obtain in field settings. The development of a new human infection model of typhoid fever has provided a unique opportunity to explore important questions about the role of circulating B cells and their various memory subsets in this disease. In the current study we report changes in frequency, activation and migration of various B M subsets in participants with typhoid diagnosis (TD) and those who did not developed disease (NoTD) following wild-type challenge with S. Typhi. Furthermore, we explore changes in activation of S. Typhi-LPS-specific B M cells and contrast the differences between TD and NoTD volunteers.

Human volunteers, clinical trial description and ethics statement
The specimens (peripheral blood mononuclear cells -PBMC-) used in the current study were collected as part of a clinical trial performed at the University of Oxford (Centre for Clinical Vaccinology and Tropical Medicine) aimed at developing a new human model of S. Typhi infection. The clinical results of this study have already been published [11]. In short, healthy adult (18-60 years-old) individuals without previous history of typhoid vaccination or residence (>6 months) in endemic areas were included in the study. Previous to oral challenge, the volunteers fasted for 90 minutes before ingesting 120 mL/2.1 g NaHCO 3 (aq). The bacteria inocula (S. Typhi -Quailes strain-10 4 CFU) were prepared in 30 mL/0.53 g NaHCO 3 (aq) which was administered 2 minutes after the volunteers ingested the 120 mL/2.1 g NaHCO 3 (aq). Following oral challenge, the participants were evaluated daily for at least 14 days. During this time, solicited and unsolicited symptoms experienced by the participants as well as oral temperature readings (2 times per day) were recorded. Typhoid fever diagnosis included reaching clinical (temperature 38°C sustained for 12 hours) and/or microbiological (blood culture confirmed S. Typhi bacteremia) endpoints. Antibiotic treatment (ciprofloxacin, 500 mg twice daily, 14 days) was indicated when (i) typhoid was diagnosed, (ii) unmanageable symptoms were present or (iii) due to clinical necessity. Additionally, all volunteers who did not develop typhoid fever received antibiotic treatment at day 14. Additional follow-up visits were completed at days 21 and 28 days post-challenge. In the current study a subset of individuals (6 TD and 4 NoTD) were evaluated for changes in B cells. These volunteers were selected based on specimen availability at critical time points to evaluate B cell responses.
All volunteers enrolled in the study provided a written informed consent and the procedures were approved by the Oxfordshire Research Ethics Committee A (10/H0604/53). This trial was registered on the UK Clinical Research Network (identifier UKCRN ID 9297). Additionally, in order to optimize flow cytometry panels and other assays, PBMC from healthy adult volunteers recruited from the Baltimore-Washington area and the Center for Vaccine Development (CVD) of the University of Maryland (UMB) were used. These volunteers also provided written informed consent and the procedures approved by the UMB IRB (HCR-HP-00040025).

Isolation of PBMC
PBMC isolation (density gradient centrifugation) and cell cryopreservation from blood samples of volunteers challenged with S. Typhi (Quailes strain) were performed before (day 0) and after (various time points) challenge as previously described [40,41]. The time points evaluated differed slightly between TD and NoTD volunteers. Days 0 (pre-challenge), 1, 2, 4, 7, 9, 14, 21 and 28 were evaluated in all subjects. Additional samples were collected in TD volunteers, which included the time at which typhoid was diagnosed (6-9 days after challenge [11]) as well as 48 and 96 hours later.

Statistical methods
The frequency of the cells (per ml of blood) or percentages of the various B cell subsets in TD and NoTD volunteers before challenge were compared using Mann-Whitney tests. The frequency of cells (cell/ml) was calculated using the number of lymphocytes (per ml of blood) as determined in the white cell counts (WCC) for each volunteer. WCC were performed at various time points after challenge and these data have already been published in [11]. Among those who acquired disease, the disease onset occurred 6-9 days post-challenge (10 4 CFU); however, since not all volunteers developed typhoid at the same time, the data was grouped in narrow time frames, defined by the exact onset date, to facilitate analysis and interpretation of the data. These time frames included: Around Time of Disease (AroundTD) and After Time of Disease (AfterTD). AroundTD in TD volunteers encompassed the day in which typhoid was diagnosed (TD+0h) until 96 hours post-diagnosis (TD+96h). Importantly, in TD volunteers two extra blood samples were collected at 48 and 96 hours post-diagnosis (TD+48h and TD +96h, respectively). In NoTD volunteers, the AroundTD time frame corresponds to days 7-11 (D7-11). The AfterTD time frame in TD volunteers involves all time points >TD+96h. Meanwhile, AfterTD in NoTD volunteers encompassed all time points >day 11 (>D11) after challenge. We used mixed effects models in order to compare mean values by time period (AroundTD and AfterTD) and group (TD and NoTD), while still accounting for the lack of independence between multiple measures from the same volunteer at the same time period and across time periods. The mixed effects models evaluation included a random effect for subject, fitted by restricted maximum likelihood. We have confirmed that this approach provides valid statistical inference for data sets of this size through various simulation experiments. All hypotheses in the study were evaluated using two-sided tests and p values <0.05 (two-sided) without adjustment for multiple comparisons were considered statistically significant. Finally, in TD volunteers, correlations (Spearman) between populations that demonstrated changes in markers and the fold-increase in antibody titers were evaluated at various time frames (Prism V6, CA).

Changes in B cells and plasmablasts (PB) induced by S. Typhi infection
B cells were identified using CD19 (CD19 + CD3 -) ( Fig 1A). Within B cells, plasmablasts (PB) were identified as cells expressing high levels of CD27 and CD38 (CD27 high and CD38 high ). Non-PB cells were further categorized into various B memory (B M ) subsets using the IgD/ CD27 classification scheme as proposed by Sanz et al [12], which delineates four subpopulations: and (iv) naïve [CD27-IgD+] cells. No significant differences were observed in the frequency of all B cells (B cells/ml) (Fig 1B), PB (PB/ml) (Fig 1C) or the various B M subsets (cells/ml) ( Fig 1D) before S. Typhi challenge between participants who were diagnosed with typhoid (TD) and those who were not (NoTD). We next assessed changes in the number of whole B cells (cell/ ml) after challenge (Fig 1E-1G). A significant reduction in the number of B cells in TD (compared to NoTD) volunteers was observed AroundTD (Fig 1F). The reduction of these cells was transitory as no differences between TD and NoTD volunteers was observed AfterTD ( Fig  1G). Of note, PB did not show statistically significant differences in the frequencies between TD and NoTD AroundTD (Fig 1H and 1I). However, an increase in frequency of these cells was noted AfterTD (Fig 1J). We then evaluated various molecules on PB including gut homing (integrin α4β7), activation (CD40 and CD21) and class switching markers (IgA) as well as the ability of these cells to bind S. Typhi (Fig 2 and S1 Fig). These data are reported as net differences (percentages or MFI) AroundTD or AfterTD relative to day 0. TD volunteers showed a significant upregulation of integrin α4β7, during the AroundTD time frame, compared to NoTD volunteers (Figs 2A, S1A and S1B). This homing marker returned to pre-challenge levels AfterTD (Fig 2B). An up-regulation of CD21 and CD40 was also noted ( Fig 2C-2F and S1D-S1F Fig) in TD volunteers and the increase was significant in the AfterTD time frame (Fig 2D and 2F). Moreover, the upregulation of CD21 by PB (AfterTD) correlated with the increase of anti-flagella IgG antibody titers, but not anti-LPS or anti-Vi, AfterTD (S1I Fig) [11]. Examples of the differences between pre-challenge and post-challenge in TD volunteers as well as the time courses of the changes in these markers are included in S1 Fig Finally,

Changes in B M subsets following wt S. Typhi infection
Evaluation of the frequency (cells/ml) of the B M subsets (IgD/CD27 classification) between TD and NoTD volunteers before challenge showed no significant differences among these groups ( Fig 1D). Changes in the frequency (cells/ml) of the various B M subsets after challenge (Fig 3) demonstrated that SmCD27+ and Um cells of TD volunteers decreased significantly AroundTD, compared to NoTD volunteers (Fig 3A, 3B, 3G and 3H). However, the decrease was transitory, since the frequency of the cells returned to pre-challenge levels by day 21 (Fig  3C and 3I). Changes in frequency were not evident in Sm CD27-or naïve cells (Fig 3D, 3E, 3J and 3K). When the data was analyzed as net percentage changes, only Um cells of TD volunteers (compared to NoTD) showed a significant decrease in AroundTD time frame (S2A- S2D  Fig). Next, we evaluated changes in the expression of various molecules (e.g., CD21, CD40, integrin α4β7 and IgA, as well as S. Typhi binding) within each B M subset (reported as net changes relative to day 0) (Figs 4, 5 and S2). Interestingly, despite that only SmCD27+ and Um cells showed reduction in their frequency AroundTD, all B M subsets showed significant changes in some of the evaluated markers and most of them were AroundTD. However, in each B M subset the marker(s) altered were different. TD volunteers downregulated CD21 in  AroundTD and AfterTD (compared to NoTD) (Fig 4E-4F and S2I and S2J Fig). Naïve cells of TD volunteers showed a significant down-regulation of integrin α4β7 AroundTD (compared to NoTD) and this marker returned to pre-challenge levels around day 21 (Figs 5A, 5B, S2K and S2L). In naïve cells, up-regulation of CD40 was also present. This marker started to increase AroundTD (S3A and S3B Fig) in TD volunteers, but statistically significant differences were demonstrated only AfterTD (Fig 5D). The levels of IgA remained unchanged in Sm CD27+ and Sm CD27-cells compared to pre-challenge and no differences between TD and NoTD volunteers were identified in any time frame (S3C and S3D Fig). Similarly, no B M subset showed significant changes in the binding to S. Typhi (S3E-S3H Fig).   [38,39]. To identify and evaluate S. Typhi-LPS-specific B cells, S. Typhi-LPS micelles of nanoparticle size (LPS-nanoparticles) were used [40]. These LPS-nanoparticles contain fluorescent Quantum dot (Qdot655) cores to allow identification of antigen-specific cells. Furthermore, LPS-nanoparticles are used as stimulants to determine if the receptor(s) interacting with this reagent induce activation of signaling pathways. Since human B cells lack TLR4 and CD14, these receptors cannot induce signaling. However, LPS-specific memory B cells have been reported suggesting that clones specific for this molecule are developed upon infection or vaccination and therefore, the B cell receptor (BCR), which drives the initial clonal selection, is the most likely candidate to interact with LPS. Thus, the main objective of these experiments was to determine changes in the intracellular signaling pathways of B cells induced by S. Typhi infection and identify whether differences were present between volunteers who developed disease (TD), and those who did not (NoTD). PBMC were stimulated with LPS-nanoparticles (~5-10 ug; 10 min; 37°C) and then fixed for analysis by phospho-flow. S. Typhi-LPS-specific B cells were identified and the distribution of these cells within the different B M subsets was evaluated ( Fig 6A). No differences in the frequency (cells/ml) of S. Typhi-LPS-specific B cells or within the B M subsets between TD and NoTD volunteers before challenge (day 0) were identified (Fig 6B and 6C) (S5A Fig display gating examples of S. Typhi-LPS-specific B cells from TD volunteers on day 0. Additionally, S5B Fig and S5C Fig show percentage comparison between TD and NoTD groups at day 0 and the percentage distribution in the different B M subsets). Furthermore, after challenge, no differences in the frequency of S. Typhi-LPS-specific B cells between TD and NoTD volunteers was identified (Fig 6D-6F). Subsequently, we examined changes in the frequency of B M subsets within CD20+ LPS+ B cells. Only Um (CD20 + LPS+ IgD+ CD27+) cells of TD volunteers (compared to NoTD) showed a significant decrease in their frequency AroundTD (Fig 6G and 6H). The Um cell frequency was back to pre-challenge levels AfterTD (Fig 6G-6I). The other B M subsets did not show significant changes (TD vs. NoTD) in their frequencies (S5D and S5F Fig). We next evaluated changes (net percentage relative to day 0) in the B M subsets phosphorylating signaling proteins associated with the BCR after challenge. The results were compared between TD and NoTD volunteers (Fig 7). The BCR-associated signaling proteins evaluated included: Spleen tyrosine kinase (Syk); Bruton's tyrosine kinase (Btk), Protein kinase B (Akt), p38 mitogen-activated protein kinase (p38MAPK), extracellular-signal-regulated kinase (Erk) 1/2 and nuclear factor kappa-  (TD+48h or TD+96h). The gates were set up based on cells stimulated with media only (unstimulated control) and the data was reported as percentage differential (% Net changes). Significant differences in the net percentage of cells phosphorylating signaling proteins were identified in TD volunteers (compared to NoTD volunteers) and the changes were concentrated in the AroundTD time frame. These changes were evident in Sm CD27+ (CD20+ LPS+ IgD-CD27+), Sm CD27-(CD20+ LPS+ IgD-  CD27-) and naïve (CD20+ LPS+ IgD+ CD27-) cells (Fig 7 and S6B-S6E Fig). In Sm CD27 + cells of TD volunteers, the percentage of cell phosphorylating Btk and NFκB was higher than NoTD volunteers AroundTD (Fig 7A and 7C). Sm CD27-cells (from TD volunteers) showed a significantly higher percentage of cells phosphorylating Erk1/2 ( Fig 7E). Finally, a significant higher percentage of naïve cells phosphorylating Akt was also noted (Fig 7G). S6B- S6E

Discussion
Because S. Typhi is a human restricted pathogen and no animal model is capable of reproducing all the clinical aspects of this disease, studies in humans are necessary to understand pathogen-human host immune interactions. These studies ultimately hold the potential to aid in the design of novel vaccines. The specimens used in the current study were collected as part of a clinical study aimed at developing a new human challenge model to study typhoid disease [11]. B cells are part of the adaptive immune system which, in addition to being responsible for antibody production, perform other important functions including antigen presentation and cytokine production. Recent advances in B cell biology have shown that the B cell compartment is composed of multiple subsets whose functions are not clearly understood, particularly in the context of infectious diseases. In the current study we aimed at studying changes in the frequency, activation, and signaling induced by S. Typhi in peripheral blood B cells of volunteers challenged with 10 4 CFU of wt S. Typhi (Quailes strain). Importantly, by comparing the results between TD and NoTD volunteers we aimed at furthering our understanding of the role of this important component of the human adaptive immune response when first encountering S. Typhi, and their potential to impact the development of disease as well as anamnestic responses.
No differences in the frequency (cells/ml) whole B cells (CD19 + CD3 -) or plasmablasts (PB) (CD3 -CD19 + CD27 high CD38 high ) were identified before wild-type challenge, suggesting that it is not possible to predict which volunteers will develop disease by simply looking at the frequency of these cells in circulation (Fig 1B and 1C). Of note, after challenge the frequency of whole B cells (Fig 1E-1G) was reduced in volunteers that developed disease (TD), mimicking the results reported in the parent study (decrease of whole cell counts (WCC) and lymphocytes) [11]. We further evaluated the effects of typhoid fever in various B cell subsets. Interestingly, plasmablasts (PB), a B cell subset that produces antibodies early after infection, did not show a reduction in the frequency AroundTD in TD volunteers, but an increase and AfterTD (Fig 1J). This increase in frequency was accompanied by upregulation of CD21 and CD40 ( Fig  2D and 2F), which are co-activator and activator markers, respectively. This suggests that these cells are activated and could be the responsible for giving rise to plasma cells and memory cells [47]. This is further supported by the positive correlation between CD21 upregulation by PB and the increase in anti-flagella antibody titers (S1I Fig). Of interest, no correlation was identified with anti-LPS antibody titers.
Integrin α4β7-expressing B and T cells identified in peripheral blood are re-circulating cells that left Peyer's patches (PP) and/or mesenteric lymph nodes (MLN) upon being primed by intestinal DC [48][49][50][51]. Upregulation of integrin α4β7 AroundTD by PB (Fig 2A) suggest that a proportion of these cells are getting ready to migrate to the gut, which is not surprising since it is the primary site of infection. This upregulation of integrin α4β7 and antibody production by PB is in agreement with oral vaccination/infection to other enteric diseases such as shigellosis [52]. Similar to observations with whole B cells and PB, no differences were noted in the frequency (cells/ml) of the B M subsets before challenge between TD and NoTD volunteers ( Fig  1D). However, following wild-type challenge, Sm CD27+ and Um cells showed a significant reduction in their frequency AroundTD in TD volunteers (compared to NoTD) (Fig 3B, 3H and 3K), which suggest that some B cell subsets are more affected than other by typhoid fever. Further evidence of the changes induced in B cells by typhoid came from the analysis of activation and homing markers. CD40 was significantly upregulated in a significant proportion of the naïve subset (Figs 4E-4F and 5D) showing that these cells were activated. Considering that naïve B cells give rise to all the other subsets, including PB, plasma cells and memory B cells, it was not surprising to observe that these cells were activated. Of note, naïve B cells also showed downregulation of integrin α4β7 (AroundTD; Fig 5A), which is intriguing since it was expected that these cells would migrate to the gut. These data might indicate that naïve B cells first migrate to other secondary organ(s) (currently undefined, but different from the gut) where they differentiate further (e.g., PB) and acquire the expression of integrin α4β7. Alternatively, these data might suggest that different B subsets use different gut homing receptors, for example naïve cells might use CCR9 or CD103 to reach the gut and once these cells commit to another lineage (e.g., PB) integrin α4β7 is upregulated. Yet another possibility is that naïve cells have already migrated to the gut using the integrin α4β7 receptor, resulting in lower levels of circulating integrin α4β7+ B naïve cells. Future studies will be directed to further explore these possibilities.
Um cells have been described to produce high-affinity IgM in the early phase of infections and the equivalent of B1 murine cells [53,54]. However, the exact origin and function of these cells in humans remains controversial. Here, we demonstrate that these cells play a role in typhoid fever, since despite that their frequency is reduced AroundTD (Fig 3H), some of these cells show signs of activation (Fig 4E-4F). Their role in typhoid and other enteric infections remains to be further explored.
CD21 works in conjunction with CD19 (CD19/CD21 complex) to enhance BCR signaling, leading to B cell activation. Downregulation of CD21 by Sm CD27+ and Sm CD27-cells AroundTD in TD volunteers (Fig 4A and 4C) might represent a physiological process during the development of a normal B cell response, particularly in its early stages. We observed that CD21 was upregulated in PB (Fig 2D-2F), which likely represent antigen-specific cells already producing antibodies and which require activation to continue their developmental process. In contrast, Sm CD27+ and Sm CD27-cells might contain memory B cells with affinity for antigens other than S. Typhi LPS. In the case of any infectious disease, the immune response can lead to an environment (e.g., cytokine milieu) that facilitates the activation of innate and antigen specific (adaptive) cells. In the case of B cells, downregulation of co-activators in cells that have not encountered an antigen specific for the invading microorganism might represent a mechanism to limit their activation. To our knowledge, CD21 downregulation has been reported in some autoimmune diseases and these cells are characterized by being anergic [55,56]. Therefore, these B cells (not S. Typhi-specific) might be refractory to activation.
One of the most abundant molecules on the surface of S. Typhi and to which antibodies are directed is LPS. LPS is a well-recognized mouse, but not human, B-cell mitogen. This is due to the fact that unlike mouse B cells, human B cells express neither TLR4 nor CD14. However, in humans exposed to S. Typhi, anti-LPS-O antibodies are detected and used as an indicator of successful immunization or challenge. This suggests that B cell clones specific for this antigen are generated during the development of immune responses to S. Typhi. Considering that clonal selection for B cells occurs through interaction of the antigen with the BCR, the presence of anti-LPS antibodies indicate that LPS was recognized by B cells that developed into antibody producing clones. Therefore, B cell clones that recognize S. Typhi-LPS though the BCR allow the identification of changes in signaling profiles elicited by typhoid fever. To this end, we used LPS-nanoparticles (S. Typhi-LPS), which contain a quantum dot (QDot 655) core, as stimulants and evaluated the BCR associated signaling pathways induced within the different B M subsets that interacted with the LPS-nanoparticle [40]. Importantly, LPS is a very "sticky" molecule and its "sticky" nature is thought to be dependent on ionic interaction [57]. Therefore, besides the BCR (specific interaction), it is likely that LPS interacts non-specifically with other molecules, especially if they have a net positive charge (LPS has a net negative charge). As expected, due to the large repetitive nature of LPS and its "stickiness", a relatively high percentage of B cells interacted with the LPS-nanoparticles (high background) (Fig 6A-6C). This is evidenced in pre-challenge samples (day 0) where LPS interaction of B cells raged from~10-25% (S5A Fig), despite the fact that these volunteers did not exhibit anti-S. Typhi antibodies (anti-LPS, anti-Vi or anti-flagellin) before challenge as previously reported [11]. We expected that only a small percentage of cells interacting with LPS will actually respond to the stimuli (specific interaction), which in our assay was determined by phosphorylation of signaling proteins. To identify the differential in activation of BCR signaling pathways induced by typhoid, each sample was corrected for background using the phosphorylation on each protein induced in day 0. Following wild-type S. Typhi challenge, naïve B cells from TD volunteers showed a significant difference (compared to NoTD) in the net percentage of cells phosphorylating Akt AroundTD (Fig 7G). Akt is part of the survival pathway and this suggests that naïve B cells of TD volunteers that interact with the LPS-nanoparticles are more resistant to apoptosis. These data complemented the observations that naïve B cells were activated (CD40 upregulation- Fig  5D). Interestingly, we also noticed differences in the Sm CD27+ population, in which NFκB and Btk phosphorylation showed differences between TD and NoTD volunteers AroundTD (Fig 7A and 7C). Sm CD27-B cells from TD volunteers also showed differential in the phosphorylation of Erk1/2 (compared to NoTD). Whether either one of these pathways leads to the downregulation of CD21, as identified in the Sm CD27+ and Sm CD27-AroundTD (Fig 4A  and 4C) remains to be explored. Also, whether activation of these signaling proteins indicate that these cells are memory cells, will be evaluated in future studies. We acknowledge, however, that since we evaluated phosphorylation at a single time point (10 minutes stimulation), we might have captured only a fraction of the changes that these cells are undergoing. Future studies will be directed to evaluate in more detail the kinetics of the signaling pathways associated with the BCR activation in TD and NoTD volunteers.
In summary, the changes induced by S. Typhi infection in volunteers that developed disease (TD) were analyzed in various B cell subpopulations. The cells that showed the most important changes were PB which exhibited significant increases in various key activation and gut homing molecules. Naïve B cells showed activation and enhanced survival, which are likely related to their differentiation into PB as well as plasma cells and possibly long term memory B cells. Interestingly, changes in Sm CD27+ and Sm CD27-cells were also identified including downregulating CD21 as a possible mechanism to avoid non-specific activation. All these changes were present only in TD volunteers, suggesting that these were induced by exposure to wt S. Typhi and are in response to the development of disease. To our knowledge this is the first study to describe these phenomena on human B cells of volunteers that developed typhoid disease after wild-type challenge. The changes that other S. Typhi antigens (e.g., flagella, Vi) induce in the different B M subsets will be addressed in future studies. The statistical analyses displayed were performed using a Mixed Effects Model (as described in materials and methods). Changes (increase or decrease) in a specific marker were evaluated in TD volunteers with respect to NoTD volunteers during the specific time frames indicated. Information found in the boxes include the time frame in which the changes were identified (e.g., AroundTD and/or AfterTD); how the marker was evaluated (e.g., % net change compared to day 0) and the P value from the Mixed Effects Model analysis. Significant data are shown in highlighted boxes (light green). These boxes also include the Fig numbers in which the data is presented in the manuscript. (PDF) S5 Fig. S. Typhi-LPS-specific B cells. S. Typhi-specific B cells were identified and concomitantly stimulated using S. Typhi-LPS-nanoparticles. Panel A displays bi-exponential plots showing the gating of LPS+ cells within the CD20+ population. Volunteers evaluated in the TD group are shown and, as expected, a high non-specific binding of LPS-nanoparticles was identified due to the "sticky" nature of LPS. Panel B shows the percentage of CD20+ LPS+ B cells at day 0 (pre-challenge) in TD (blue symbols) and NoTD (brown symbols) groups and demonstrate that no differences between these groups existed. Panel C shows the percentage of B M cell subsets that are CD20+ LPS+ cells at day 0, demonstrating that no differences between TD (blue symbols) and NoTD (brown symbols) groups exist before challenge. Panels D-F display time courses of the frequency (cells/ml) of SmCD27+, Sm CD27-and Naïve cells. No differences between TD and NoTD groups were identified. Importantly, Um cells showed a significant difference (Fig 6) between TD and NoTD groups AroundTD. This is described in detail in the main body of the manuscript. In panels D-F AroundTD is indicated by the blue rectangles with dotted lines. In the same panels TD and NoTD volunteers are indicated by the blue and brown symbols, respectively. Panels D-F display Mean ± SD. (PDF) The data were reported as the differential in phosphorylation between days post-challenge and day 0. Panels B, C, D and E show time courses of the net changes (percentages) in phosphorylation of signaling proteins showing significant differences between TD and NoTD volunteers AroundTD. These changes were present in SmCD27+, Sm CD27-and Naïve cells. In panels B-E AroundTD is indicated by the blue rectangles with dotted lines. In the same panels TD and NoTD volunteers are indicated by the blue and brown symbols, respectively. Panels B-E display Mean ± SD. Panel F shows a table summarizing the results of the signaling proteins evaluated in the different B M subsets. Information found in the boxes include the time frame in which the changes were identified (e.g., TD+0h to TD+96h); how the marker was evaluated (e.g., % net change compared to day 0) and the P value from the Mixed Effects Model analysis. Significant data are shown in highlighted boxes (light green). These boxes also include the