L-selectin-dependent and -independent homing of naïve lymphocytes through the lung draining lymph node support T cell response to pulmonary Mycobacterium tuberculosis infection

Recruiting large numbers of naïve lymphocytes to lymph nodes is critical for mounting an effective adaptive immune response. While most naïve lymphocytes utilize homing molecule L-selectin to enter lymph nodes, some circulating cells can traffic to the lung-draining mediastinal lymph node (mLN) through lymphatics via the intermediate organ, lung. However, whether this alternative trafficking mechanism operates in infection and contributes to T cell priming are unknown. We report that in pulmonary Mycobacterium tuberculosis-infected mice, homing of circulating lymphocytes to the mLN is significantly less efficient than to non-draining lymph node. CD62L blockade only partially reduced the homing of naïve T lymphocytes, consistent with L-selectin-independent routing of naïve lymphocytes to the site. We further demonstrated that lymphatic vessels in infected mLN expanded significantly and inhibiting lymphangiogenesis with a vascular endothelial growth factor receptor 3 kinase inhibitor reduced the recruitment of intravenously injected naïve lymphocytes to the mLN. Finally, mycobacterium-specific T cells entering via the L-selectin-independent route were readily activated in the mLN. Our study suggests that both L-selectin-dependent and -independent pathways contribute to naïve lymphocyte entry into mLN during M. tuberculosis infection and the latter pathway may represent an important mechanism for orchestrating host defence in the lungs.


Introduction
Lymph nodes (LN) are specialized secondary lymphoid organs, which act as the main sites for the generation of adaptive immunity. As naïve antigen-specific T cells are present at an extremely low frequency, the initiation of an efficient adaptive immune response relies on the recruitment of a large pool of circulating naïve lymphocytes into LN to increase the probability of encounter between antigen-specific lymphocytes and their cognate antigen presented by antigen presenting cells (APCs) [1][2][3]. Thus, in conjunction with priming adaptive immune response, effective recirculation of naïve lymphocyte through LN underlies the critical function for these secondary lymphoid organs in immunosurveillance.
LN have two main portals of entry, high endothelial venules (HEV) and afferent lymphatic vessels 4 . Naïve and memory lymphocytes are recruited to the LN from circulation through HEV. This recruitment process occurs via a multistep adhesion cascade that begins with the binding of L-selectin (CD62L) expressed by naïve and memory lymphocytes to peripheral node addressin (PNAd) expressed by HEV [4,5]. This mechanism is further strengthened by the interaction of CCR7 expressed on naïve lymphocytes with the arresting and homing chemokine CCL21 expressed by fibroblastic cells that line the venules [6][7][8]. It is generally believed that HEV act as portals that determine the entry of circulating L-selectin positive naïve lymphocytes, whereas lymphatic vessels (LV) and more specifically the afferent LV is the main pathway by which antigen, effector lymphocytes and APCs reach LN from the draining peripheral tissues [9,10]. One exception, however, has been reported for lung-draining LN where naïve T cells have been suggested to enter the LN directly via circulation and indirectly through lung tissue in naïve mice [11]. The significance of this dual-supply mechanism for the tissue-draining LN in the generation of adaptive immunity in a disease setting has yet to be investigated.
Under the homeostatic state, lymphocyte populations within the LN are maintained at a steady level. Following infection or immunization, the LN draining a site of inflammation undergoes several morphological changes that promote the increased recruitment of naïve lymphocytes [12,13]. These changes include the expansion of HEV and the arteriole that feeds into lymph node, and increased expression of CCL21 on HEV [12,[14][15][16]. This remodelling process is thought to enhance the efficiency of recruitment and screening for rare antigen (Ag)-specific lymphocytes [12]. Defects in these processes observed in tumour models and acute viral infection lead to transiently impaired delivery of naïve lymphocytes to the LN, which in turn hinders effective immunosurveillance and immune response activation [17,18]. The impact of chronic inflammation on the patterns of naïve lymphocyte recirculation remains incompletely understood.
Mycobacterium tuberculosis (M.tb) is a pathogenic bacteria that establishes a persistent pulmonary infection, with the lung-draining mediastinal LN (mLN) serving as the main site for the initiation of adaptive immunity [19][20][21]. In this study, we report that the recruitment of naïve lymphocytes via the L-selectin-dependent HEV pathway into the mLN of M.tb infected mice was significantly less efficient than that to LN draining a non-infection site. Consequently, a substantial fraction of naïve lymphocytes entered the mLN via a L-selectin independent pathway involving afferent lymphatics. Importantly, naïve Ag-specific T cells that entered the mLN via the L-selectin-independent mechanism were responsive to cognate Ag stimulation in vivo and were functionally comparable to those that entered via circulation. This study reveals an L-selectin-independent CD4 + T cell priming mechanism in mice with chronic pulmonary infection and suggests that the operation of this additional homing pathway is critical in maintaining immune surveillance and ongoing immune response to pulmonary infection.

The entry of circulating naïve lymphocytes to lung-draining mediastinal lymph node is reduced compared to non-draining LN in M.tb-infected mice
An important function of the lymph node is to increase the probability of encounter between rare antigen-specific T cells and their cognate antigen expressed on APCs by enabling large numbers of naïve lymphocytes circulating through the site [4]. To investigate the recirculation patterns of naïve lymphocytes in mLN during M.tb infection, we firstly examined the homing of lymphocytes into mLN versus non-draining inguinal LN (iLN). Labelled splenocytes were injected i.v. into M.tb-infected mice and their relative abundance in mLN or iLN were analyzed 2 h after cell transfer ( Fig 1A). This approach has been developed to analyse lymphocyte entry in LN as the time frame is sufficient for lymphocytes to migrate from circulation to the LN, but insufficient for them to egress [22,23]. We found that despite increased cellularity of mLN compared to iLN ( Fig 1B) in pulmonary M.tb-infected mice, there was a decrease in the frequency and absolute number of labelled lymphocytes that were recruited to the mLN ( Fig  1B). This contrasts with some previous studies where lymphocyte recruitment via HEV was shown to increase following inflammation, with the rate of lymphocyte entry being proportional to lymph node cellularity [12,13,24].
The impaired recruitment of lymphocytes to the mLN suggested that there might be local alteration in the LN environment. Therefore, we examined the main molecules involved in lymphocytes homing expressed on HEV including, L-selectin ligand PNAd identified by MECA-79 marker, as well as the homing chemokine CCL21 (Fig 1C-1E). First, we quantified the number of HEV per LN area across entire sections of naïve mLN and mLN and iLN from M.tb infected mice at weeks 3 and 8 p.i ( Fig 1C). Unexpectedly, while the density of HEV has been previously shown to increase as the LN enlarges following an infection or inflammation [14,25,26], the density of HEV in M.tb infected mLN were decreased compared to both naïve mLN and iLN. Examining CCL21 expression in iLN HEV, revealed a uniform distribution within the paracortex (Fig 1D). In contrast in mLN, CCL21 expression was variable with partial and/or complete loss of CCL21 surrounding 50% of HEV (Fig 1D and 1E). These data suggests that the entry of lymphocytes into the mLN is impaired in M.tb-infected mice, and this is associated with reduced HEV density and perturbed CCL21 expression pattern within the mLN.

Significant proportion of naïve lymphocytes enter the mLN of M.tbinfected mice via an L-selectin-independent mechanism
The reduced lymphocyte entry from the circulation accompanied by the increased cellularity of mLN raised the possibility of an alternative entry pathway for naïve lymphocytes. To experiments (n = 4-6 mice / group, 1-2 sections / mouse were analyzed). Data in (D-E) are from two independent experiments (n = 3 mice / group, 1-2 sections / mouse were analyzed) Statistical differences between groups in (B, E) were determined using Student's t test or in (C) using one-way ANOVA with Tukey's multiple comparison test. ** p < 0.01, ***p < 0.001, **** p < 0.0001.

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection determine whether lymphocytes rely on L-selectin for recruitment to the mLN, we assessed the cellularity of the LN after blocking L-selectin, an essential adhesion molecule for recirculation via HEV [27,28]. A single injection of anti-CD62L antibody (Mel-14) has been shown to prevent trafficking of naïve lymphocytes and in turn reduce LN cellularity by 90% within 3 days post injection [29,30]. We therefore examined the cellularity of the mLN and iLN in 3-or 8-wk M.tb-infected mice at 72 h after anti-CD62L injection (Fig 2A). Interestingly, we found a partial reduction in lymphocytes at week 3 p.i. in the mLN with a 47% reduction in CD4 + T cells, no significant decrease in CD8 + T cells, and a 65% reduction in the B cell population ( Fig  2B). A similar trend of partial reduction was also found in the mLN at week 8 p.i., with 50% reduction in T lymphocytes, and 69% reduction in the B cell population (Fig 2C).
Similar reductions were observed when naïve CD4 + and CD8 + T lymphocyte populations were identified as CD69 low CD44 low at 3-and 8-week p.i. (Fig 2F and 2G). Although reduction of naïve CD8 + T cell population following L-selectin treatment was not statistically significant at week 3 p.i. (Fig 2F). Nevertheless, these data indicate that L-selectin treatment only partially reduces lymphocyte accumulation in the mLN during early and late stage of M.tb infection. In contrast, we found a 85-96% reduction in the accumulation of total and naïve lymphocytes in iLN following inhibition of L-selectin at 3-and 8-week p.i. (Fig 2D and 2E and 2H and 2I). These findings suggest that a large proportion of naïve lymphocytes enter the mLN from an Lselectin independent pathway during M.tb infection.
To address whether lymphocytes from circulation can home into the lymph node independently of L-selectin through an alternative route, such as the afferent lymphatic vessels (LV) draining the infected lungs, we intravenously injected labelled lymphocytes together with anti-CD62L antibody and analyzed lymphocyte populations in the mLN and iLN 18 h post cell transfer ( Fig 3A). This 18 h time point was chosen as this is the period required for circulating naïve lymphocytes to traverse the lung [11]. We found that transferred lymphocytes were able to reach the mLN independently of L-selectin, while they failed to enter the iLN (Fig 3B and  3C). Examining further the phenotype of lymphocytes that reached the mLN, we observed that the majority of transferred lymphocytes that reached the mLN using this alternative L-selectin-independent pathway were naïve CD69 low CD44 low CD4 + T cells and CD8 + T cells (Figs 3D and 3E and S1A). These results indicate that while the iLN mainly relies on entry via HEV for naïve lymphocyte recruitment, the mLN in M.tb-infected mice can use an L-selectin independent pathway, which is likely via the afferent LV that drain from the lung.

Naïve lymphocytes migrate from M.tb-infected lungs to draining mLN independently of L-selectin
To determine directly whether naïve lymphocyte can exit M.tb-infected lungs and migrate to the mLN via the LV, we transferred CTV-labelled lymphocytes intratracheally into M.tbinfected mice. In additional, to exclude the possibility of intratracheally delivered cells entering the mLN from the blood, these cells were injected together with anti-CD62L mAb ( Fig 4A). Intratracheally administered lymphocytes were then analysed 24h post-transfer, a timeframe allowing accumulation of the injected lymphocyte in the mLN (Fig 4A). The greatest frequencies and numbers of labelled lymphocytes were identified in the lungs, followed by the mLN, and no cells were detected in the iLN in both isotype control and anti-CD62L antibody-treated mice (Fig 4B and 4D). Phenotypic analysis of T cells recovered from the lung and mLN showed that they were mainly naïve lymphocytes (CD69 low CD44 low ) ( Fig 4C). Consistent with previous studies [31], there was an increase in the number of cells recovered from the lungs and mLN in the group that received lymphocytes together with anti-CD62L antibody.

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection Given the recruitment of lymphocytes from the lung to mLN via an L-selectin independent pathway in M.tb-infected mice, we examined the LV network in iLN and mLN of M.tbinfected mice. As expected, LV structure, identified by Lyve-1 marker, were mainly restricted to the capsule and medullary region of iLN ( Fig 4E). In contrast, the LV network was markedly altered in mLN of the infected mice so that Lyve-1 + LV cells were no longer restricted to the capsule and medullary area but extended deeper into paracortical region. We also observed an increased expansion of the area occupied by LV within the mLN compared to the iLN ( Fig  4F). The expansion of LV network was associated with an increase in the expression of lymphangiogenic factors, VEGF-C and its receptor VEGFR3, in mLN compared to iLN in the infected mice ( Fig 4G). These findings collectively indicate that naïve lymphocytes can enter from the lung into the mLN of M.tb-infected mice.

Inhibiting lymphangiogenesis reduces the entry of naïve lymphocytes to the mLN
To examine the role of afferent lymphatics in lymphocyte transmigration, M.tb-infected mice were injected daily with a VEGFR3 kinase inhibitor, MAZ51, starting from 2 weeks post-M.tb infection. This inhibitor has been demonstrated to be effective in suppressing

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection lymphangiogenesis in mice [32]. Ten days after MAZ51 treatment, the mice received a transfer of tdTomato + lymphocytes i.v. from naïve donor mice, and the transferred cells were tracked 18 hours later ( Fig 5A). To determine the efficacy of this inhibitor in our model, we examined LV expansion in the mLNs of vehicle and MAZ51 treated mice (Fig 5B and 5C). We found that MAZ51 treatment resulted in a significant reduction in Lyve-1 + area compared to vehicletreated control mice ( Fig 5C). We also observed that, compared to vehicle control, MAZ51 treatment led to a significantly reduced recruitment of the total tdTomato + CD4 + and CD8 + T cells (65% and 70%, respectively) ( Fig 5D), as well as naïve CD69 low CD44 low tdTomato + CD4 + and CD8 + T cells (60% for both subsets) (Fig 5E), into the mLN. Interestingly, the B cell population was not significantly decreased (Fig 5D), which is consistent with our previous results (Figs 2B and 2C and 3C) that showed only a small number of B cells entering the mLN via the L-selectin-independent pathway.

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection in the numbers of endogenous tetramer-positive ESAT-6 4−17 -specific CD4 + T cells in mLN ( Fig 5G). The numbers of endogenous or exogenously injected (tdTomato + ) lymphocytes, as well as tetramer-positive ESAT-6 4−17 -specific CD4 + T cells, did not differ in the lungs of MAZ51-and vehicle-treated mice (S2A Fig). Interestingly, we found a decrease in the frequency of IFN-γ, TNF-α, but not IL-2, expressing CD4 + T cells in the lungs of MAZ51-treated mice compared to vehicle-treated controls mice (S2B Fig). However, this could be due to reduced output of effector cells from mLN to the lungs, as effector T cells utilize the efferent LV as an exit route [4,33]. We also found that the bacterial burden in mLN and the lungs were

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection also comparable regardless of the treatment (S2C Fig). We could not exclude the possibility that the minimal effect of MAZ51 treatment on the dynamics of pathogen-host interaction in the lungs and mLN is due to the short treatment duration used. Nonetheless, these data indicate that afferent lymphatic vessels contribute to the entry of naïve T lymphocytes to the mLN during M.tb infection.

Ag-specific CD4 + T cells entering in mLN independent of L-selectin are activated locally to become Th1 cells
To determine whether lymphocytes entering via an L-selectin independent pathway are antigen responsive, we adoptively transferred GFP-expressing, ESAT6 1-20 epitope-specific TCR transgenic CD4 + T cells (Clone 7) together with anti-CD62L mAb or isotype control into 3 or 8 week M.tb infected recipient mice, and donor cells (GFP + ) were analysed at 48 h and 72 h post-transfer (Figs 6A and S1B). Interestingly, flow cytometric analysis of mLN showed no difference in the frequency of E6 Tg cells in the mLN at 48 h between control and anti-CD62Ltreated LN, however, we saw an increase at 72 h in the anti-CD62L-treated group (Fig 6B). As anticipated no difference in the frequency of E6 Tg were detected in the spleen between anti-CD62L-treated and control groups at week 3 and 8 p.i., as lymphocyte recruitment in this secondary lymphoid organ is L-selectin independent [31]. In corroboration with our earlier

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection results, a significant decrease was found in the frequency of E6 Tg cells in iLN that received Lselectin compared to isotype control (Fig 6B). These data reaffirm our previous results, suggesting that lymphocytes can reach the mLN in L-selectin-independent pathway.
We next analysed the response of CTV-labelled E6-specific CD4 + T cells in the mLN by comparing the rate of cell division and expression of the early activation marker CD69 and Th1-specific transcription factor T-bet in anti-CD62L or isotype control mAb treated M.tbinfected mice at week 3 and 8 p.i. (Fig 6C-6E). We found that cell division, as measured by CTV dilution, began between 48 and 72 h post cell transfer, and the percentage of CTV Low E6-specific T cells was comparable between the two-treatment groups at both weeks 3 and 8 p. i. (Fig 6C and 6D). CD69 expression was upregulated at 48 h post cell transfer and began to decrease by 72 h at the beginning of cell division regardless of treatment (Fig 6C and 6D). Suggesting that L-selectin independent pathway provides antigen responsive cells in early and late stage of M.tb infection. Similarly, the kinetics of T-bet upregulation was equivalent between anti-CD62L and isotype control treated groups (Fig 6E). Finally, although we found an increase in total lymphocyte numbers and tetramer-positive ESAT-6 4−17 CD4 + T cells in the This could be due to the short-term treatment with anti-CD62L, as effector cells are likely to be already in the lungs after 3 weeks p.i. [21]. Taken together, these data indicate that the L-selectin-independent pathway is a physiologically functional route for supplying antigen responsive cells to the mLN during M.tb infection.

Discussion
A pivotal step in the initiation of the adaptive immune response is the recruitment of naïve Ag-specific lymphocytes to LN draining a site of infection or immunization. In this study, we showed that during pulmonary M.tb infection, the entry of naïve lymphocytes into the lungdraining LN via the L-selectin dependent HEV pathway was significantly impaired compared to entry into non-draining LN. Approximately half of the naïve T cells entered the mLN via a L-selectin-independent mechanism. We further showed that persistent M.tb infection promotes lymphangiogenesis and the afferent LVs play a previously unappreciated role in routing circulating naïve T cells to the mLN through the lung. These findings reveal that the L-selectin-independent mechanism controls the entry and priming of naïve T lymphocyte in the mLN during M.tb infection, and we propose that this alternative route of entry is important in sustaining adaptive immune responses to persistent pulmonary infection.
HEV have been well documented to be the main gateway for the homing of naïve lymphocyte from the circulation to LN [4,34,35]. Most studies examining changes in lymphocyte recirculation during infection or immunization have reported an increase in lymphocyte recruitment to the draining LN via HEV [12,36,37]. This enhanced recruitment is associated with an efficient initiation of the adaptive immune response, as it ensures the increase in antigen-specific T cell pool to support the ongoing immune response [12]. Unexpectedly, we found that homing of circulating naïve lymphocytes via HEV is reduced in mLN during M.tb infection. This reduced recruitment was associated with decreased expansion of HEV along with disrupted expression of the homing chemokine CCL21 in the mLN. Similar alterations in the lymphocyte homing machinery have been previously shown to lead to LN atrophy during infection with a pathogenic strain of chikungunya virus (AF15561) [17]. By contrast, we observed that the cellularity of mLN increased strikingly following M.tb infection, indicating that homing via HEV is unlikely to account solely for the accumulation of lymphocytes in mLN.
In addition to HEV, the afferent LV is a secondary portal for leukocyte recruitment into the LN. Unlike the HEV pathway, homing of lymphocytes via the afferent LV has long been considered to be a minor contributor to LN cellularity [9,33]. The current concept on naïve lymphocyte trafficking was largely based on cannulation studies in sheep or adoptive transfer experiments performed in mice [10,[38][39][40], in which inflammation promoted the migration of APCs and effector lymphocytes into the draining LN predominantly via the afferent LV. However, it is unclear whether these distinct trafficking patterns for naïve and effector lymphocytes are similar for LN draining other organs during infection. Previous studies have suggested that low numbers of naïve lymphocytes are known to circulate through highly vascularized nonlymphoid organs, such as the lung [11,41] and the liver [31]. It has also been demonstrated that following M.tb infection CD4 + T lymphocytes accumulate in lung parenchyma, peaking around week 3-4 p.i. and persisting until at least 6 months p.i. [41]. Naïve CD4 + T cells were further shown to be more efficient at entering the parenchyma from the blood vasculature than effector cells. The transit of naïve lymphocytes into the lung parenchyma suggests the possibility that these cells can exit the lung and traffic to the mLN through the LV, and this has

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection indeed been proposed to occur in naïve mice [11]. In this study, we found that a substantial proportion of naïve lymphocytes are capable of homing to the mLN even when L-selectin is blocked, which suggests that these naïve cells are likely entering into the mLN via the afferent LV. In addition to the contribution of the L-selectin-independent pathway to lymphocyte homing to the mLN at the peak of adaptive immunity (week 3) and the chronic phase (week 8), as demonstrated in this study, it would be interesting to investigate in the future the impact of L-selectin blockade and characterize early recirculation events at the onset of adaptive immunity in the mLN. This investigation could provide new insights into the delay in the generation of adaptive immunity following aerosol M.tb infection.
By using two distinct approaches, we were able to confirm that the naïve T lymphocytes that entered the mLN in the absence of L-selectin-HEV interaction were migrating from circulation via the lung-draining LV. First, we found that the LV network underwent extensive expansion in the lung-draining mLN, but not in distant LN draining a non-infection site. Through the inhibition of lymphangiogenic factor VEGFR3, we confirmed an essential role for the M.tb-induced lymphangiogenesis in the recruitment of both naïve and antigen-specific T lymphocytes to the mLN during M.tb infection. Second, by intratracheal transfer of lymphocytes we demonstrated that naïve lymphocytes could exit the lung and home into the mLN in an L-selectin-independent manner. Our finding is consistent with recent studies performed in a different respiratory disease model [42,43], establishing a critical role for lung-draining LV in supplying naïve lymphocytes to LN. Thus, under conditions where direct naïve T cell entry from blood circulation is impaired, as demonstrated in this study, T cell accumulation in the lung and the subsequent trafficking into the mLN can act as an alternative pathway to compensate for this defect. It would therefore be interesting to examine whether this alternative route is active during other respiratory pathogens, and assess its contribution to the supply of naïve lymphocytes to the mLN. In the future, it will also be important to investigate whether vaccination before M.tb challenge would modify the entry of naïve lymphocytes via the lung, and whether this L-selectin independent mechanism could be exploited in vaccine development strategies. It will also be interesting to investigate whether similar mechanism operates in other large, vascularized organs, such as the liver. Both the lung and liver are large solid organs with very limited numbers of draining lymph nodes in mice [44,45], and mechanisms that permit entry of naïve lymphocytes via alternative routes may facilitate the immune surveillance of these organs by ensuring adequate numbers of naïve lymphocytes circulating through LN.
We further demonstrate that the L-selectin-independent homing mechanism is functional in M.tb infection and contributes to the generation of Ag-specific effector T cells in the mLN. The kinetics of CD4 + T cell activation and differentiation into effector Th1 cells were comparable between isotype control and anti-CD62L antibody-treated mice. Interestingly, it has been previously reported in a mouse model of M.tb infection [46] that the frequency of IFN-γ-producing CD4 + T cells quantified following in vitro restimulation with plate bound CD3/CD28, were decreased in the mLN of selectin ligand deficient FtDKO mice. These mice lack the expression of fucosyltransferases IV and VII, which prevents the binding of both naïve lymphocytes via L-selectin ligand and activated T cells with E and P-selectin ligands [47]. Therefore, the observed reduction in IFN-γ-producing CD4 + T cells could be due to impaired recruitment of naïve T lymphocytes and/or activated T cells to the mLN. However, our study is consistent with recent findings in a mouse model of influenza virus infection, where no changes in the activation of CD8 + T cells in the mLN of L-selectin deficient and anti-L-selectin antibody treated mice was observed [42]. All together these data indicate that the L-selectin independent pathway can act as a unique complementary pathway for recruiting a pool of naïve lymphocytes to ensure the efficient supply of antigen-responsive CD4 + T cells to the mLN during M.tb infection.

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection While the generation of effector CD4 + T cells is known to restrict M.tb infection, the pathogen survives in the face of host adaptive immunity leading to persistent infection [21]. Mechanisms regulating this host-pathogen equilibrium are incompletely understood. Our study reveals an alternative homing route for supplying naïve lymphocytes to the lung draining LN in murine TB. This mechanism may be particularly important in the setting of chronic respiratory infection where L-selectin-dependent pathway is restricted due to local inflammation. Understanding the regulatory mechanisms governing naïve T cell trafficking in chronic infection and how these may differ between different organs may ultimately help the development of more specific, host-directed therapies for pathogen elimination.

Ethics approval
All murine experiments were conducted with the approval of the Sydney Local Health District Animal Welfare Committee (protocols 2015-037 and 2020-007).

Treatments with anti-CD62L and VEGFR3 kinase inhibitor
For blockade of CD62L, mice received a single dose (250 μg/mouse Mel-14; BioXcell) intraperitoneally at weeks 3 or 8 p.i. Mice were culled 72 h after anti-CD62L treatment for analysis. For VEGFR3 kinase blockade mice received daily doses of MAZ51 (10 mg/kg/day; Calbiochem) intraperitoneally for 10 days before harvest at week 3 p.i.

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection

Preparation of single suspension from tissues
Lymph nodes, spleens, and lungs were collected in RPMI supplemented with 2% foetal calf serum (FCS) for flow cytometric analysis. LN and spleens were mechanically dissociated through 70 μm strainers. Lungs were perfused with PBS by cardiac cannulation and collected in 2 mL RPMI media supplemented with 0.1 mg/mL DNase I (Sigma Aldrich) and 10 U/mL collagenase type I (Sigma Aldrich). Lungs were dissociated using GentleMacs Dissociater (Miltenyi Biotec) and incubated for 30 min at 37˚C. Lungs were then further dissociated with Gen-tleMacs Dissociater and filtered through a 40 μm strainer. Single cell suspensions were treated with ACK lysis buffer to remove erythrocytes. Cells were washed with RPMI media supplemented with 2% FCS before viable cells were counted using trypan blue exclusion on a hemocytometer.

Organ collection and processing for imaging
LN were collected and fixed in 1% paraformaldehyde (Alfa Aesar) at 4˚C for 24 hours. Fixed tissues were then placed in 30% (v/v) sucrose (Sigma Aldrich) in PBS. Twenty-four hours later tissues were frozen in Optimal cutting temperature compound (VWR chemicals). Tissues were sectioned at 10-14 μm on a Shandon Cryotome E (ThemoFisher). Sections were blocked and permeabilised with 3% normal goat serum and 0.1% Triton X-100 diluted in PBS (Cell Signaling Technologies) for 30 minutes at room temperature. Sections were then stained with the appropriate primary antibodies diluted in 3% normal goat serum (NGS) in PBS for 2 h at room temperature or overnight at 4˚C. Sections were washed in 1x tris-buffered saline with 0.5% tween-20 (TBST) before staining with secondary antibodies diluted 3% NGS in PBS for 1 h at room temperature. The sections were washed 3x in 1x TBST, and subsequently mounted with prolong gold antifade (Thermo Fisher). The following antibodies (clone number; source) were used in this study: CD3 (17A2; BioLegend), B220 (RA3-6B2; BD Biosciences), Lyve-1

Image analysis
The number of HEV (MECA79 + CD31 + ) were counted on whole LN sections and were normalised to the total area of the LN section. The number of HEV with continuous (CCL21 + ) or intermittent (partial and/or no detection) (CCL21 -) expression were counted on whole LN sections. All cell counting was performed using the cell counter plugin in Fiji software. The area covered by Lyve-1 was measured in mm 2 on whole LN sections and was normalised to the total area of the LN sections. The area was measured using Fiji software.

mRNA preparation and qRT-PCR
LN were collected and submerged in RNAlater (Ambion) for 24 h (4˚C) prior to long term storage at -80˚C. RNA was purified from lymph nodes using Trisure (Bioline) as per manufacturer's instructions (Bioline). RNA (2 μg) was reverse transcribed using the Tetro cDNA synthesis kit with random primers according to the manufacturer's instructions (Bioline). Quantitative reverse transcriptase PCR (qRT-PCR) was performed using SensiFast SYBR Green (Bioline) on a Roche LightCycler480. Data are expressed as fold change over uninfected control lymph nodes. Data were calculated by ΔΔCT method using 18S as the reference gene. Forward and reverse qRT-PCR primers used in this study: 18S Fw GTAACCCGTTGAACCC CATT, Rv CCATCCAATCGGTAGTAGCG, vegfr3 Fw GCGACAGGGTTCTCATAA, Rv CGTTGCCTCATTGTGATTAG, vegfc Fw CAGCCCACCCTCAATACCAG, Rv GCTGCTCCAAACTCCTTCC.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 9 software (GraphPad Software, La Jolla, CA). Differences between two groups were analyzed by Student's t-test. Analysis between multiple groups was performed using one-way or two-way analysis of variance (ANOVA) with Tukey's or Sidak multiple comparisons test respectively. Results with p < 0.05 were deemed statistically significant.
Supporting information S1 Fig. Flow cytometry gating strategy. (A) Lymph node samples were gated for total lymphocytes, singles cells, and live cells. Live cells were then gated for donor CD45.1 + . CD45.1 + were further gated on CD4 + and CD8 + T cells. The expression of CD44 and CD62L was then identified on either CD4 + or CD8 + T cells. (B) Lymph node samples were gated on total lymphocytes, single cells, and live cells as shown in (A). Live cells were then gated on CD4+ and GFP+ to identify donor E6 Tg GFP + CD4 + T cells. The expression of CD69 or Tbet against CTV was then analysed on E6 Tg CD4 + T cells.

PLOS PATHOGENS
Trafficking dynamics of lymphocytes in lymph nodes during Mycobacterium tuberculosis infection IFN-γ, TNF-α and IL-2 expression in CD4 + CD44 + T cells in the lung of control and MAZ51-treated mice at wk3 p.i. Lung cells were restimulated in vitro with ESAT6 1-20 peptide. (C) Bacterial loads in the lung and mLN of control and MAZ51-treated mice measured at wk 3 p.i. Data shown are representative of two independent experiments with similar results (n = 4 mice/group). Statistical differences between groups were determined by Student's t test.