Active mTORC2 Signaling in Naive T Cells Suppresses Bone Marrow Homing by Inhibiting CXCR4 Expression

The mammalian immune system is a highly specialized system that serves to defend the host against invading pathogens through dynamic interactions between innate and adaptive defense modules. Specific organized structures, termed lymphoid organs, facilitate the development, maturation, and subsequent encounters between components of the adaptive and innate immune system (1). Primary lymphoid organs, including the bone marrow and thymus, function primarily as sites that support the development of B and T lymphocyte populations, respectively. Following development, mature naive T cells exit the thymus into the periphery, where they enter into secondary lymphoid organs (SLOs), including the spleen, lymph nodes, Peyer patches, and MALT. These SLOs function primarily as sites that facilitate interactions between APCs and rare Ag-specific lymphocytes.

Naive T cells continuously recirculate through SLOs to scan for their cognate Ag, making multiple brief interactions with APCs (2–5), while they simultaneously receive signals in the form of tonic TCR signaling and cytokines such as IL-7, produced by fibroblastic reticular cells in the T cell zones of SLOs (6). These signals are essential for survival (7, 8), maintenance of basal metabolism (9–11), and control of Ag responsiveness (12–14). Thus, continuous recirculation of naive T cells through SLOs contributes to immune homeostasis by ensuring survival, efficient immune surveillance, and subsequent mounting of an effective immune response in the presence of an infection or cancer.

Although significant progress has been made in understanding the mechanisms that specifically direct naive T cell entry into, and migration within, SLOs during recirculation (reviewed in Ref. 15–17), the mechanisms that keep them away from fluxing into inappropriate or undesirable compartments, such as healthy tissues or primary lymphoid organs, are less well understood. For example, although the bone marrow, a primary lymphoid organ, is rich in T cell survival nutrients and cytokines such as IL-7, only a few naive T cells migrate into the bone marrow, such that their representation in this compartment is reduced by ∼100-fold when compared with SLOs (18–20). Trafficking of cells to the bone marrow is largely controlled by the chemokine CXCL12 and its receptor CXCR4, but the underlying molecular mechanisms remain to be fully understood. CXCL12 attracts and promotes the bone marrow retention of cells such as hematopoietic stem cells (HSCs) and plasma cells that express high levels of CXCR4 (reviewed in Ref. 21, 22). Developing B cells are also retained in the bone marrow through their high expression of CXCR4 (23–25).

In this study, we report an unexpected finding that under steady state, the expression of CXCR4 and consequent bone marrow homing of naive T cells is actively suppressed by mTORC2 signaling. The implication of this finding is that naive T cells divert and more readily recirculate through SLOs to receive survival signals, resulting in more efficient immune surveillance and prompt initiation of immune responses against invading pathogens. Together, our findings show that mTORC2 signaling in naive T cells promotes T cell homeostasis by preventing aberrant accumulation of naive T cells in the bone marrow, thereby promoting recirculation through SLOs.

Mice with loxP-flanked exon 4 of Sin1 alleles (Sin1^f/f) were recently established. Briefly, the Sin1 allele–targeted embryonic stem cell line (obtained from the European Conditional Mouse Mutagenesis Program) was injected into C57BL/6 blastocysts by South China Mouse Facility Center (Shanghai, China) to produce Sin1-FrtLacZNeoFrt-loxp chimeric mice. Seven founder lines were generated and bred to a CMV-Flp mouse line to obtain three germline-transmitted Sin1^f/+ founder lines (with a deletion of the FrtLacZNeoFrt cassette). Sin1^f/f lines were obtained from self-crossing of the Sin1^f/+ founder lines. Cre-mediated Sin1^f/f allele deletion has been confirmed by crossing Sin1^f/f lines to several Cre-expressing mouse lines, including the CD4-Cre line used in the current study (more details on the characterization of the Sin1^f/f line are described elsewhere [X. Ouyang and B. Su, unpublished observations]). In this study, the Sin1^f/f mice, after back crossing into C57BL/6 background for at least nine generations, were crossed with CD4-Cre and OTI TCR transgenic mice. All mice were bred and maintained in pathogen-free conditions. Age- and sex-matched mice between 8 and 12 wk old were used for experiments. All mouse work was performed in accordance with the Institutional Animal Care and Use Committee guidelines at Yale University and Shanghai Jiao Tong University School of Medicine.

Anti-mouse CD4 (clone RM4-5), CD8α (clone 53-6.7), CD3e (clone 145-2C11), CD44 (clone IM7), CD62L (clone MEL-14), CXCR4 (clone 2B11), CD127 (clone A7R34), and FOXP3 (clone FJK-16s) Abs were purchased from eBioscience. CD25 (clones 7D4 and PC61) Ab was purchased from BD Biosciences. All Abs used in Western blot analyses were obtained from Cell Signaling Technology. FOXO1 inhibitor AS1842856 (catalog no. 344355; Calbiochem) and CXCR4 antagonist AMD3100 (catalog no. A5602) were purchased from Sigma-Aldrich.

Wild-type (WT) or Cd4 promoter CD4-Cre⁺Sin1^fl/fl (T-Sin1^−/−) splenocytes were labeled with either Cell Trace Violet or Cell Trace Far Red (Thermo Fisher Scientific). The frequency of CD62Lhi CD44lo naive CD4⁺ T cells in each sample was determined by flow cytometry. Equal numbers of labeled naive WT or T-Sin1^−/− CD4⁺ T cells were mixed together and transferred into recipient C57BL/6 mice. Spleens, lymph nodes, and bone marrow of recipient mice were harvested 2 h after transfer, and the total number of CD62Lhi CD44lo naive WT and T-Sin1^−/− CD4⁺ T cells recovered from each tissue were determined by flow cytometry. Where indicated, cells were treated with AMD3100 prior to adoptive transfer.

Chemotaxis assay was performed as previously described (26). Briefly, single-cell suspensions of spleens were prepared, and RBCs were lysed using Sigma-Aldrich RBC Lysis Buffer (R7757; Sigma-Aldrich). Cells were incubated in serum-free migration medium (RPMI 1640, 0.5% fatty acid-free BSA [A9647; Sigma-Aldrich], 1% antibiotics, 10 mM HEPES) for 1 h at 37°C to promote resensitization of chemokine receptors. Cells were then plated onto Transwell inserts placed in 24-well tissue culture plates (3421; Corning) containing indicated concentrations of stromal cell–derived factor 1 (SDF) (CXCL12) in migration medium. Controls without SDF (null) or without the Transwell insert (input) were also made. Plates were then incubated at 37°C in 5% CO₂ for 3 h, followed by enumeration of migrated cells by flow cytometry.

For proliferation, freshly isolated OTI WT or T-Sin1^−/− splenocytes from naive mice were labeled with Cell Trace Violet or Cell Trace Far Red (Thermo Fisher Scientific) and stimulated with 200 ng/ml OVA 257–264 (SIINFEKL) peptide. Cells were harvested after 24, 48, and 72 h of stimulation for flow cytometric analyses of activation markers and dilution of Cell Trace dye in CD8⁺ T cells. For cytokine production, freshly isolated OTI WT or T-Sin1^−/− splenocytes, obtained from mice day 8 post Listeria monocytogenes–OVA infection, were stimulated with 200 ng/ml OVA 257–264 (SIINFEKL) peptide for 6 h in the presence of BD GolgiStop and GolgiPlug. For CD107a staining, anti-CD107a Ab was added during the 6-h peptide stimulation.

WT or T-Sin1^−/− mice were infected by i.v. injection of 15,000 CFU L. monocytogenes–OVA. Tissues were harvested and analyzed at day 7 or 8 postinfection to assess the endogenous polyclonal T cell response.

The graphs were prepared and statistical analyses were done using the GraphPad Prism software. Except where indicated, comparisons between the two groups were done using unpaired t test. Error bars represent mean ± SEM. A p value <0.05 was considered significant.

To investigate the role of mTORC2 in naive T cell homeostasis, we conditionally deleted an obligatory mTORC2 adaptor protein, Sin1, in T cells by breeding Sin1^fl/fl mice with transgenic mice expressing Cre recombinase enzyme under the Cd4 promoter T-Sin1^−/−. As shown in Fig. 1A, Sin1 protein was efficiently deleted from T cells, and consistent with its critical role for mTORC2 activity, phosphorylation of important mTORC2 targets Akt and protein kinase C (PKC) at the turn motif sites T450 and T638/641, respectively, as well as Akt S473, was severely impaired in T-Sin1^−/− cells compared with WT (Sin1^fl/fl) controls (Fig. 1A). In contrast, phosphorylation of ribosomal protein S6, a downstream target of the mTORC1 signaling pathway, was not impacted by Sin1 deletion (Fig. 1A). Because Sin1 is deleted in this system at the double-positive (DP) stage of thymic T cell development, we examined the impact of loss of Sin1/mTORC2 on late-stage T cell development. The frequencies of CD4⁺ and CD8⁺ single-positive (SP) T cells emerging from the T-Sin1^−/− DP population were comparable to WT controls (Fig. 1B, 1C). Moreover, loss of Sin1/mTORC2 did not result in increased death or defects in the cell size of emerging CD4⁺ and CD8⁺ SP T cells (data not shown). Together, these data demonstrate that deletion of Sin1 at the DP stage via CD4-Cre has little impact on the thymic T cell development.

In contrast to the thymus, analyses of mature T cells in the periphery revealed a reduction in total CD3⁺ T cell numbers in the spleen, inguinal lymph nodes, mesenteric lymph nodes, lamina propria, intraepithelial lymphocyte population, and blood of T-Sin1^−/− mice, when compared with that in WT control mice (Fig. 1D–F, 1I, 1J, data not shown), consistent with reports from groups studying the role of mTORC2 signaling in T cells using models of Rictor deletion (27, 28). Surprisingly, however, we observed a corresponding increase in total numbers of T cells in the bone marrow of T-Sin1^−/− mice (Fig. 1D–F), and both CD4⁺ and CD8⁺ T cells of the T-Sin1^−/− mice displayed this altered distribution pattern (Fig. 1G, 1H). These data indicated that the reduced percentage of periphery T cells was likely due to the abnormal accumulation of bone marrow T cells, although it was unclear if certain subsets or all subsets of T cells might be affected in this altered distribution.

Memory CD4⁺ and CD8⁺ T cells have been shown to preferentially home to the bone marrow, where they interact with IL-7–producing stromal cells and receive signals for survival and homeostatic proliferation (29–31). In addition, a large percentage of CD4⁺ T cells in the bone marrow of mice and humans are FOXP3⁺ regulatory T cells (Tregs) (32–34). Furthermore, ablation of mTOR signaling has been shown to enhance both memory (35) and Treg (36) formation, and specifically, ablation of mTORC2 in T cells by Rictor deletion has been shown to enhance memory formation (37). We, therefore, examined the composition of T cells homing to the bone marrow to determine if any of these two populations accounted for the increased accumulation of T cells in the bone marrow of T-Sin1^−/− mice. Surprisingly, the increased population of T cells in bone marrow of T-Sin1^−/− mice was largely composed of CD62L^hi CD44^lo naive T cells (Fig. 2A, 2B), and this phenotype was more striking in the CD4⁺ T cell population, with ∼5-fold increase in the numbers of bone marrow naive CD4⁺ T cells and no significant changes in the numbers of FOXP3⁺CD25⁺CD4⁺ Tregs or CD62L^lo CD44^hi memory CD4⁺ T cells in the T-Sin1^−/− mice when compared with that in WT control mice (Fig. 2B).

Consistently, we found that there was reciprocal reduction of naive T cells in the periphery, suggesting that there was a likely connection between the peripheral decrease and corresponding increase in bone marrow naive T cells. Although Sin1 deficiency results in increased CD62L and decreased CD44 expression in T cells, distinct separation between naive and memory populations was still apparent, and the two populations could easily be gated by flow cytometry (Fig. 2A). Taken together, our results show that loss of Sin1 results in altered lymphoid tissue distribution of naive T cells.

To further investigate the cellular and molecular mechanisms underlying the altered naive T cell distribution, we first examined if it was a result of abnormal T cell apoptosis or proliferation. However, we found that the overall level of proliferation and apoptosis among bone marrow naive T cells was low, <2 and <6%, respectively, in both T-Sin1^−/− T cells and WT T cells, although T-Sin1^−/− T cells consistently displayed lower levels of both parameters (data not shown). This suggests that the altered distribution of naive T cells observed in T-Sin1^−/− mice was not likely a result of significant changes in proliferation or survival of these cells. This led us to examine whether the accumulation of naive T-Sin1^−/− T cells in the bone marrow was an active homing process. To do this, we labeled naive WT and T-Sin1^−/− T cells and adoptively cotransferred them into the same recipient host. We reasoned that if the accumulation of naive T-Sin1^−/− T cells in the bone marrow was an active process, the cells would also show increased homing to the bone marrow shortly after adoptive transfer. Indeed, naive T-Sin1^−/− T cells displayed increased homing to the bone marrow compared with WT control T cells when the recipient hosts were analyzed 2 h after adoptive cotransfer (Fig. 3A). These data suggest that loss of Sin1 results in augmented trafficking of naive T cells to the bone marrow.

Entry of cells into the extravascular spaces of the medullary sinus of the bone marrow is largely regulated by the chemokine CXCL12, which mediates both the bone marrow recruitment and retention of cells expressing high levels of CXCR4, such as HSCs and developing B cells (23–25). We therefore examined the contribution of CXCR4 expression to bone marrow homing of naive T-Sin1^−/− T cells. T-Sin1^−/− T cells displayed higher surface expression of CXCR4 protein (Fig. 3B, 3C) as well as cxcr4 mRNA compared with WT controls (Fig. 4C). In addition, naive T-Sin1^−/− T cells displayed better chemotactic response to CXCL12 in in vitro Transwell assays (Fig. 3D). Furthermore, treatment of naive WT and T-Sin1^−/− T cells with a CXCR4 antagonist AMD3100 prior to adoptive cotransfer of the cells into recipient hosts resulted in a 50% reduction in the migration of transferred cells into the bone marrow (Fig. 3E). This observed reduction of T cell homing to bone marrow by AMD3100 treatment may be an underestimation of the requirement for CXCR4 for T cell bone marrow migration, given the short t_1/2 of AMD3100 and potential for the blockade of CXCR4 signaling to promote surface re-expression of CXCR4, which can have an opposite and counteracting effect on bone marrow homing if the drug is not simultaneously present in vivo after transfer of the naive T cells to continuously block CXCR4/CXCL12 signaling. We, however, favor the approach of treating the cells in vitro prior to transfer instead of directly treating WT or T-Sin1^−/− mice with AMD3100, as the drug has been shown to mobilize HSCs from the bone marrow into peripheral circulation, which might change the overall distribution of cells in the lymphoid compartments of treated mice, thereby complicating analysis and conclusions that can be drawn from such analysis. Genetic studies involving Sin1^fl/fl mice expressing Cre recombinase enzyme under the cxcr4 promoter will be instructive in this regard. Nevertheless, we can infer from our data that CXCR4 plays a critical role in bone marrow migration of naive T cells, and that loss of Sin1 results in increased CXCR4 expression on naive T cells, which promotes bone marrow homing in response to CXCL12.

In B cells, strong PI3K signaling in the germinal center light zone suppresses FOXO1 activity, thereby inhibiting CXCR4 expression. Conversely, increased nuclear FOXO1 activity in B cells in the dark zone promotes CXCR4 expression, highlighting the antagonistic relationship between PI3K signaling and FOXO1 activity in the control of CXCR4 expression (38, 39). In T cells, FOXO1 has been shown to bind directly to the CXCR4 promoter (40), suggesting that FOXO1 likely controls CXCR4 expression in T cells as well. Moreover, we previously established that PI3K signaling suppresses FOXO1 activity through an mTORC2-dependent activation of Akt and SGK1, which in turn phosphorylate FOXO1 to promote its nuclear exclusion and inhibit its transcriptional activity (41).

Consistent with this notion, T-Sin1^−/− T cells displayed reduced FOXO1 phosphorylation compared with WT controls following αCD3/αCD28 stimulation (Fig. 4A). We, however, only observed a very small but consistent reduction in basal FOXO1 phosphorylation in the T-Sin1^−/− T cells compared with WT controls. We believe this is likely because FOXO1 phosphorylation is better observed following an induction, and in the case of the naive unstimulated T cells, we do not yet know which signals induce mTORC2 activity in the naive cells or at what point to look. Nevertheless, we observed an increase in FOXO1 activity in naive T-Sin1^−/− T cells compared with WT control T cells, consistent with less suppression of FOXO1 activity in the naive T-Sin1^−/− T cells, which displayed increased expression of FOXO1 targets such as CD62L and IL-7R (Fig. 4B). In addition, treatment of naive T-Sin1^−/− and WT T cells with a FOXO1 inhibitor, AS1842856, which blocks transcriptional activity of FOXO1 (42), led to a reduction in mRNA expression of cxcr4 and other FOXO1 targets such as il7r and klf2 (Fig. 4C). Taken together, these data suggest that Sin1/mTORC2 suppresses CXCR4 expression in naive T cells by inhibiting FOXO1 activity.

To understand the physiological implications of the observed perturbations on T cell immune responses, we examined the ability of T-Sin1^−/− T cells to respond to antigenic stimulation in vitro as well as following Listeria infection in vivo. We reasoned that the reduction of peripheral naive T cells with increased accumulation in the bone marrow might result in reduced efficiency of immune surveillance and thus result in impaired immune responses. If this is the case, we would observe defects in the immune response in vivo, but not in vitro, when T cells and Ag-presenting cells are colocalized experimentally.

Indeed, when stimulated in vitro, the expression of markers of activation and the proliferation of T-Sin1^−/− T cells were comparable to those of WT control T cells, suggesting that loss of Sin1/mTORC2 signaling does not impair early activation and proliferative responses of T cells to their cognate Ag when they are colocalized (Supplemental Fig. 1A, 1B). In contrast, following Listeria infection in vivo, we observed reduced numbers of effector T-Sin1^−/− cells in various tissues at day 7–8 postinfection (Fig. 5), supporting the notion that the mislocalization of rare Ag-specific T cell populations from APCs in vivo, due to the accumulation of naive T cells in the bone marrow, likely contributed to the observed reduction in effector T cell numbers because no problems were detected in vitro when T cells were colocalized with Ag-bearing cells experimentally. We cannot, however, rule out the contributions of other defects that we observed in the T-Sin1^−/− cells, such as altered surface expression of several adhesion molecules, including CD44, LFA-1, integrin β 1 (CD29), and integrin β 2 (CD18) (Supplemental Fig. 2A–D).

Despite the defects observed in effector T cell numbers, levels of cytokine production and expression of the degranulation marker CD107a in the effector T cells that formed following Listeria infection were comparable between T-Sin1^−/− T cells and WT control T cells (Supplemental Fig. 1C and data not shown). Together, our data show that Sin1/mTORC2 deficiency results in reduced numbers of effector T cells against Listeria infection, likely owing in part to the defects in the proper localization of naive T cells at peripheral lymphoid organs, although Sin1/mTORC2 is not required for acquisition of CD8⁺ T cell effector function.

One of the most fundamental functions of naive T cells is to patrol the host by moving through various SLOs, where they interact with APCs to scan for infections and other pathologic conditions while simultaneously receiving survival cues. For this reason, naive T cells constitutively express appropriate adhesion molecules and receptors, such as L-selectin and CCR7, respectively, which permit their entry into and proper migration and positioning within various SLOs. In contrast, naive T cells generally migrate poorly into tissues such as the bone marrow, a primary lymphoid organ. The underlying mechanism(s) for this selective T cell tissue trafficking are poorly understood. In this study, we report a previously unknown mechanism underlying selective inhibition of migration of naive T cells into the bone marrow. We show that under physiological conditions, Sin1 via mTORC2 in naive T cells inhibits FOXO1 activity and, consequently, the expression of CXCR4 to prevent naive T cell bone marrow homing. We observed that loss of mTORC2 signaling via Sin1 deletion led to accumulation of naive T cells in the bone marrow and a corresponding reduction in numbers of peripheral T cells in the spleen, lymph node, and blood. Interestingly, unlike naive T cells, neither Tregs nor effector T cells show this differential bone marrow trafficking when Sin1 was deleted in T cells.

Our findings raise two fundamental questions. First, what is the physiological significance of preventing bone marrow homing of naive T cells in contrast to memory T cells and Tregs, which have been shown to reside in the bone marrow? We propose that one physiological advantage is to ensure efficient immune surveillance and induction of immune responses. By preventing sequestration of naive T cells in the bone marrow, naive T cells can more readily recirculate through SLOs, whose architecture better supports interactions between rare Ag-specific T cells and APCs, consequently leading to greater efficiency in the ability of naive T cells to encounter and respond to their cognate Ag in the presence of an infection. This notion is supported by our data showing that, in vivo across all tissues we examined, T-Sin1^−/− mice had reduced numbers of effector T cells as compared with WT control mice at the peak of immune responses to Listeria. In contrast, Sin1-deficient T cells displayed comparable activation and proliferation to WT control T cells when incubated with antigenic peptide and costimulation in vitro. This suggests that Sin1/mTORC2-deficient T cells are capable of responding normally to Ags, provided they are appropriately and efficiently colocalized with the APCs, and that the defect in effector T cell numbers that we observed following Listeria infection in vivo was likely due, at least in part, to the mislocalization of naive T cells to the bone marrow instead of SLOs, where efficient induction of immune responses take place. In addition to these defects of Sin1-deficient T cells, we also found reduction in surface expression of adhesion molecules, including CD44, LFA1, and integrin β 1, which may have roles in T cell activation and trafficking.

Second, our findings raise the question of what signals sustain or periodically activate mTORC2 signaling in naive T cells such that FOXO1 activity is sufficiently suppressed to prevent upregulation of CXCR4 and naive T cell bone marrow homing without enforcing exit of the cells from quiescence. The antagonistic relationship between PI3K signaling and FOXO1 activity is well established. Strong PI3K signaling in T cells, following stimulation by cognate Ag, costimulation, and cytokines such as IL-2, leads to FOXO1 phosphorylation by Akt and SGK1 in an mTORC2-dependent manner. This in turn, leads to the nuclear exclusion of FOXO1 and strong inhibition of its transcriptional activity. In contrast, nuclear localization as well as transcriptional activity of FOXO1 is observable in unstimulated T cells (43). Nevertheless, the fact that loss of mTORC2 signaling in naive T cells results in hyperactive FOXO1 signaling suggests that under normal conditions, FOXO1 activity is at least partially suppressed by mTORC2, and this partial suppression is sufficient to inhibit CXCR4 expression and bone marrow homing of naive T cells. Moreover, mTORC2-dependent phosphorylation of Akt and PKCα/βII at their turn motif sites T450 and T638/641, respectively (Fig. 1A), is observable in freshly isolated and unstimulated naive T cells, although phosphorylation of Akt at S473 is usually difficult to detect in freshly isolated naive T cells but is strongly induced upon T cell activation. This further supports the notion of a dichotomy in mTORC2 activity and, consequently, a dichotomy of FOXO1 suppression in naive versus activated T cells that could explain how mTORC2 can suppress CXCR4 expression and bone marrow homing of naive cells under steady state.

The molecular mechanisms controlling this dichotomy in mTORC2 signaling are currently unknown. One hypothesis is that molecular constraints on mTORC2 signaling exist in naive T cells to curtail the overall strength of mTORC2 signaling, leading to low-level mTORC2 signaling that is sufficient to partially suppress FOXO1 activity, whereas such constraints are overridden in activated T cells, allowing a more complete suppression of FOXO1. One such constraint could be the nature of upstream stimuli that activate mTORC2 signaling in naive versus activated T cells. Although stimuli that strongly induce PI3K and mTORC2 signaling in activated T cells, such as cognate TCR–Ag engagement, costimulation, and cytokines like IL-2 and IL-4, are well studied, less is understood about stimuli that regulate PI3K and mTORC2 signaling in naive T cells. Tonic TCR and IL-7R signaling are likely candidate stimuli in naive T cells. Naive T cells encounter and respond to IL-7 when they recirculate through SLOs. IL-7, in turn, promotes survival and is essential for maintenance of basal metabolic activity of these cells (9–11). Akt, a downstream target of mTORC2, has been shown to be an important mediator of the metabolic effects of IL-7 in naive T cells (44), suggesting that IL-7 signaling can activate mTORC2 in naive T cells. Thus, transient and periodic encounters with IL-7 in SLOs can potentially activate mTORC2 to a level sufficient for partial suppression of FOXO1 activity and CXCR4 expression in naive T cells to prevent its bone marrow homing. Future studies will be needed to address the molecular mechanisms regulating the dichotomy of mTORC2 signaling in naive versus activated T cells.

In summary, our study represents, to our knowledge, one of the first studies demonstrating an active homeostatic mechanism that prevents naive T cells from homing to the bone marrow, thereby ensuring their recirculation through SLOs for efficient immune surveillance and induction of immune responses. Future studies will be needed to fully address the physiological implications of these findings.

The authors have no financial conflicts of interest.