Lung type 3 innate lymphoid cells respond early following Mycobacterium tuberculosis infection

ABSTRACT Tuberculosis is the leading cause of death due to an infectious disease worldwide. Innate lymphoid type 3 cells (ILC3s) mediate early protection during Mycobacterium tuberculosis (Mtb) infection. However, the early signaling mechanisms that govern ILC3 activation or recruitment within the lung during Mtb infection are unclear. scRNA-seq analysis of Mtb-infected mouse lung innate lymphoid cells (ILCs) has revealed the presence of different clusters of ILC populations, suggesting heterogeneity. Using mouse models, we show that during Mtb infection, interleukin-1 receptor (IL-1R) signaling on epithelial cells drives ILC3 expansion and regulates ILC3 accumulation in the lung. Furthermore, our data revealed that C-X-C motif chemokine receptor 5 (CXCR5) signaling plays a crucial role in ILC3 recruitment from periphery during Mtb infection. Our study thus establishes the early responses that drive ILC3 accumulation during Mtb infection and points to ILC3s as a potential vaccine target. IMPORTANCE Tuberculosis is a leading cause of death due to a single infectious agent accounting for 1.6 million deaths each year. In our study, we determined the role of type 3 innate lymphoid cells in early immune events necessary for achieving protection during Mtb infection. Our study reveals distinct clusters of ILC2, ILC3, and ILC3/ILC1-like cells in Mtb infection. Moreover, our study reveal that IL-1R signaling on lung type 2 epithelial cells plays a key role in lung ILC3 accumulation during Mtb infection. CXCR5 on ILC3s is involved in ILC3 homing from periphery during Mtb infection. Thus, our study provides novel insights into the early immune mechanisms governed by innate lymphoid cells that can be targeted for potential vaccine-induced protection.

T uberculosis (TB), caused by the bacterium Mycobacterium tuberculosis (Mtb), infects approximately one-fourth of the world's population, resulting in 1.6 million deaths in 2021 globally (1).In 2021 alone, an estimated 10.6 million people fell ill with TB.A major hurdle to eradicate the disease is our poor understanding of the early immune mechanisms mediating protective immunity against Mtb infection.
In recent studies, innate immunity has gained importance in controlling Mtb infection prior to or despite the engagement of adaptive immune responses (2)(3)(4)(5).Our own recent study has established a protective role for innate lymphoid cells (ILCs) in the lungs of mice challenged with Mtb (2).As another important arm of innate immunity, ILCs share features with both adaptive and innate immune cells and comprise three main subsets.Group 1 ILCs produce interferon γ (IFN-γ) and include natural killer (NK) cells and non-cytotoxic, non-NK type 1 ILCs, which are related to type 1 immune response (Th1) (6).Group 2 ILCs, which produce interleukin-4 (IL-4), IL-5, and IL-13, are involved in inflammatory-linked airway hyperactivity, tissue repair (7), and helminth clearance by promoting type 2 immune responses (8).Innate lymphoid type 3 cells (ILC3s) produce IL-17 and/or IL-22 (related to Th17 immune response) (9) and can participate in the strategic positioning of immune cells in ectopic lymphoid structures (10).ILCs are also critical for tissue repair of the lungs following infection-induced lung damage (11) and for generating hepatic granulomas (12).Indeed, our recent study reported that ILC3 mediated early protection against Mtb through enhanced secretion of IL-17 and IL-22 (2).However, not much is known about the early signaling mechanisms that govern ILC3 accumulation and control of Mtb infection.
In the current study, we show that ILC3s are early responders upon exposure to Mtb.During early infection, we demonstrate that interleukin-1 receptor (IL-1R) signaling plays a crucial role in regulating lung ILC3 accumulation, formation of granulomas, and Mtb control.Depletion of IL-23, together with IL-1R signaling, results in increased suscepti bility to Mtb infection, with reduced accumulation of lung ILC3 subsets and myeloid cells, and loss of Mtb control.Importantly, following Mtb infection, IL-1R signaling within the epithelial cells provides critical signals for CXCR5 + ILC3 accumulation, myeloid cell accumulation, and early control of Mtb replication.Taken together, our data provide novel evidence for an early signaling mechanism in mediating ILC3 localization and immunity against emerging Mtb infections.

Mtb infection induced early accumulation of ILC subsets
During Mtb infection, lung ILC populations, specifically ILC3s, increased early follow ing Mtb infection in the murine lung (2).Therefore, to gain better insight about the functional heterogeneity of lung ILCs, we characterized them at the single-cell level following Mtb infection.We sorted for highly pure innate lymphoid cells (CD45 + CD127 + Lin − NK1.1 − ) from Mtb-infected lung at 5 and 14 days post-infection (dpi) and performed scRNA-seq.Following single-cell alignment and quality control process, we found a total of 5,084 high-quality cells (1,886 cells in 5 dpi and 3,198 cells in 14 dpi).Cluster analysis found 12 distinct clusters (Fig. 1A) consisting of 78.9% of total cells, with six clusters of ILC population in addition to other types of cells.The identification of the cell types is based on the expression of known markers (Fig. 1B).Clusters 0, 1, 4, 5, and 10 are ILC2 (Gata + ); cluster 2 is mix of ILC1 (Cxcr3 + ), ILC2 (Gata + ), and ILC3 (Rorc + ); cluster 3 is T cells (Cd3d + , Cd4 + , and Cd8a + ); cluster 6 is dendritic cells; cluster 7 is an unknown type of cells; cluster 8 is NK (Ncr1 + ) cells; cluster 9 is dendritic macrophage (Cd14 + ); and cluster 11 is B (Cd79a + ) cells (Fig. 1B).
Therefore, our single-cell transcriptomic analysis not only confirmed the early and rapid accumulation of ILC3 and ILC3/ILC1-like clusters in the Mtb-infected lungs but also provided important information about the gene expression profiles within the clusters that changed over the course of infection.

IL-1/IL-23 cytokine axis regulates the accumulation of ILC3s during Mtb infection
Lung scRNA-seq data revealed expression of Il1r and Il23r in the Mtb-infected lung ILC3 cluster.Therefore, we next determined the importance of IL-1R and IL-23R signal ing in the accumulation of ILC populations following Mtb infection.As known, the absence of IL-1R signaling (Il1r −/− ) in mice resulted in increased susceptibility due to Mtb infection (13) (Fig. 2A).Moreover, this increased Mtb susceptibility was also associated with reduced accumulation of lung ILC1s, ILC3s, and NKp46-expressing ILC3s (NKP-ILC3s) at 14 dpi (Fig. 2B, D and E; Fig. S1C and D).However, the accumulation of ILC2s was not impacted in the Mtb-infected Il1r −/− mice (Fig. 2C).Importantly, the reduced ILC3 accumulation coincided with reduced accumulation of alveolar macrophages (AMs) and monocytes in the Mtb-infected Il1r −/− mice lung at 14 dpi (Fig. 2F and G; Fig. S1B).These changes in myeloid cell accumulation were also reflected in poorly formed granulomaassociated lymphoid tissue (GrALT) (Fig. 2H; Fig. S3A) and associated with decreased IL-6, CXCL-1 (KC) and macrophage inflammatory protein 1a (MIP1α) (MIP1a) production in absence of IL-1R signaling (Fig. S3B through D).Therefore, these results suggest that early IL-1R signaling is important for accumulation of ILC3s and downstream activation of myeloid cells as well as formation of protective GrALT.
To evaluate the combined importance of IL-23R and IL-1R signaling in mediating ILC3 accumulation, we utilized IL-23 neutralizing antibody either in C57BL/6 (B6) Mtb-infected mice or IL-23 neutralizing antibodies in Il1r −/− Mtb-infected mice.As before (2), early IL-23 depletion in the B6 Mtb-infected mice resulted in increased lung Mtb burden as com pared to isotype-treated B6 Mtb-infected mice at 14 dpi (Fig. 2I).However, depletion of IL-23 in Il1r −/− mice did not further increase early Mtb susceptibility as compared with isotype-treated Il1r −/− Mtb-infected mice.Additionally, we did not observe any significant reduction in either ILC1 or ILC2 compartments in Il1r −/− Mtb-infected mice that also received IL-23 neutralizing antibody as compared with the isotype-treated Il1r −/− Mtb-infected mice (Fig. 2J and K).Indeed, we saw significant decrease in the accumula tion of lung ILC3 and NKP-ILC3s in Il1r −/− mice following IL-23 neutralization at 14 dpi, suggesting that ILC3 accumulation is mediated synergistically by both IL-23/IL-1R cytokine axis (Fig. 2L and M).In addition, we observed significantly reduced AMs (Fig. 2N) as well as monocyte (Fig. 2O) accumulation following IL-23 depletion in Il1r −/− Mtbinfected mice at 14 dpi.Therefore, our data point toward synergistic role for IL-1R/IL-23R signaling in optimal ILC3 accumulation but not ILC1 or ILC2 accumulation, and that IL-23 or IL-1R signaling alone is sufficient to mediate early Mtb control.

IL1r/NFκβ signaling in lung epithelial cells regulates the accumulation of ILC3s and Mtb control
As shown above, the IL-1/IL-23 cytokine axis regulated the accumulation of early lung ILC3s to mediate Mtb control.Epithelial cells are early responders to infection and induce downstream effectors such as IL-6, IL-8, TNF-α, IFN-γ, and nitric oxide (14)(15)(16).To address whether IL-1R signaling in the epithelial cells mediated downstream responses for ILC3 accumulation following Mtb infection, we infected Il1r f/f Sftpc cre (lung type 2 epithelial cells will lack IL-1R signaling) and littermate control mice with Mtb.Similar to Il1r −/− mice, Il1r f/ f Sftpc cre mice showed early Mtb susceptibility with increased lung Mtb CFU at 14 dpi when compared to littermate control Mtb-infected mice (Fig. 3A).Moreover, this coincided with decreased accumulation of lung ILC1s, ILC3s, and NKP-ILC3s (Fig. 3B, D, and E), although no changes in the accumulation of lung ILC2s were noted at 14 dpi (Fig. 3C).The lung AM accumulation was significantly hampered in Il1r f/f Sftpc cre mice as compared to the Il1r f/f Mtb-infected mice at 14 dpi (Fig. 3F); however, we did not observe any defects in monocyte accumulation (Fig. 3G).No differences in the baseline lung immune cell populations were observed in Il1r f/f Sftpc cre mice (Fig. S3E through M).Taken together, these data suggest that IL-1R signaling in lung type 2 epithelial cells plays an early role in accumulation of ILC3s and mediated downstream protective responses and Mtb control.
To further address the role played by lung epithelial cells in the early recruitment of ILC3s and Mtb control, we sublethally irradiated Il1r f/f Sftpc cre and littermate control mice and reconstituted both groups with B6 bone marrow (BM).After complete reconstitution (Fig. S3N), we infected the BMC (bone marrow chimeric) mice with Mtb.Our data show increased lung Mtb burden in Il1r f/f Sftpc cre recipient mice receiving wild-type BM (nonhematopoietic BMC) as compared to Il1r f/f recipient mice receiving B6 BM (control BMC mice) at 14 dpi (Fig. 3H).In addition, we observed reduced accumulation of lung ILC1s, ILC3s, and NKP-ILC3s (Fig. 3I, K and L), along with reduced AMs and monocytes in nonhematopoietic BMC Mtb-infected mice (Fig. 3M and N).As before, the accumulation of lung ILC2 was unaffected in non-hematopoietic BMC mice compared to control BMC Mtb-infected mice (Fig. 3J).Together, these data project an important role for IL-1R on lung epithelial cells in driving downstream accumulation of lung ILC3s and early Mtb control.
NF-κβ is a master regulator of lung epithelial cell-mediated immune function against pulmonary infection (17).To test if this pathway independent of IL-1R signaling is involved in accumulation of ILC3s during Mtb infection, we generated conditional knockout mice that lacked both IL-1R and NFκβ signaling on lung type 2 lung epithelial cells.Following Mtb infection, our results revealed that the Mtb susceptibility in the Il1r f/ f Ikk2 f/f Sftpc cre mice phenocopied the effect we observed in the Il1r f/f Sftpc cre mice at 14 dpi (Fig. S4A).We did not observe any further defect in ILC1 and ILC2 accumulation (Fig. S4B  and C) in Il1r f/f Ikk2 f/f Sftpc cre mice as compared with littermate control (Il1r f/f Ikk2 f/f ) Mtbinfected mice.However, we observed significantly reduced accumulation of ILC3 and NKP-ILC3s in Mtb-infected Il1r f/f Ikk2 f/f Sftpc cre mice as compared with littermate control (Il1r f/f Ikk2 f/f ) Mtb-infected mice (Fig. S4D and E).We did not observe any change in AMs and monocyte population (Fig. S4F and G) at 14 dpi.Due to lack of IL-1R/NFκβ signaling, there was significantly increased lung inflammation (Fig. S4H) in Il1r f/f Ikk2 f/f Sftpc cre mice as compared to Il1r f/f Ikk2 f/f mice during Mtb infection.We did not detect any abnormalities in the immune cell population at baseline in both Il1r f/f Ikk2 f/f Sftpc cre and littermate control mice (Fig. S4I through Q).These results support a critical role for IL-1R signaling and that the protective response driven by this pathway was NFκβ-independent.

CXCR5 mediates ILC3 recruitment to the lung and early Mtb control
Cxcr5 −/− mice exhibited increased lung Mtb CFU and decreased accumulation of ILC3s within lymphoid follicles (2).Upon kinetic analysis of the timing of lung ILC3 accumula tion, deficiency of Cxcr5 led to a delay in lung ILC3 accumulation and also resulted in fewer lung ILC3s in Cxcr5 −/− mice as compared to the control Mtb-infected lungs (Fig. 4A).Similar findings were observed in case of both IL-17 and IL-22 cytokines lacking (Il17/ Il22 −/− ) Mtb-infected mice (Fig. 4A).As noted before, both Cxcr5 −/− and Il17/Il22 −/− mice show early Mtb susceptibility (2).These data therefore support a role for CXCR5 and IL-17/IL-22 in driving CXCR5 expression for early accumulation of ILC3s and Mtb control.To understand the role of CXCR5 in recruitment of ILC3 from circulation, we generated BMC by irradiating B6 mice and adoptively transferring BM from either B6 or Cxcr5 −/− mice, allowed mice to reconstitute fully, and then Mtb infected the BMC mice.While the control BMC mice retained accumulation of early lung ILC3s, we observed lower accumulation of lung ILC3s in hematopoietic Cxcr5 −/− BMC mice at 14 dpi (Fig. 4B).Thus, CXCR5-mediated accumulation of ILC3s from the bone marrow is a prominent axis in accumulation of lung ILC3s to mediate Mtb control.To mechanistically address the role of ILC3-intrinsic CXCR5 signaling in accumulation and function of ILC3, we generated Cxcr5 f/ f Rorγt cre mice and infected with Mtb.We found that Cxcr5 f/f Rorγt cre mice were more susceptible as compared to the littermate control mice (Fig. 4C and K) at both 24 and 30 dpi.Moreover, absence of functional CXCR5 signaling in early ILC3s resulted in decreased accumulation of ILC1, ILC3, and NKP-ILC3s (Fig. 4D, F, G, L, N, and O) in Mtb-infected mice lung at both time points.Similar to results in Il1r and IL-23 neutralized Mtb-infected mice, ILC2 compartment was unaltered in Cxcr5 f/f Rorγt cre mice at both time points (Fig. 4E and  M).Although, there was no significant defects in the AMs and monocyte accumulation at 24 dpi (Fig. 4H and I), we observed significant decrease in accumulation of AMs and monocytes (Fig. 4P and Q) in Cxcr5 f/f Rorγt cre mice as compared to the littermate control mice at 30 dpi.In addition, we did not observe any GrALT formation in Cxcr5 f/f Rorγt cre mice as compared with littermate control Mtb-infected mice at 24 dpi (Fig. 4J).No differences in the baseline lung immune cell populations were observed in Cxcr5 f/f Rorγt cre mice (Fig. S5A through I).Thus, CXCR5 signaling in ILC3s is important for both early ILC3 accumula tion, likely from the circulation and subsequent Mtb control.

DISCUSSION
ILC3s are important mediators in early innate immunity against Mtb infection (2,18).However, the signaling mechanisms that regulate the early recruitment and accumula tion of lung ILC3s at the site of Mtb infection are not well defined.In the current study, we show that subsets of ILC3 and ILC3/ILC1s are increased early following Mtb infection, and that IL-23 and IL-1 cytokine signaling synergize to mediate early ILC3 accumulation.Furthermore, expression of CXCR5 receptor on ILC3s likely sustain lung ILC3 accumula tion and mediate early Mtb control.Together, these findings provide detailed insights into the mechanisms that regulate ILC3 recruitment and accumulation within the lung, thereby identifying novel immune pathways that can be targeted for vaccines against TB.
Our scRNA-seq analysis revealed different clusters of ILC populations over time probably due to infection-induced changes in the microenvironment.ILC2s were the predominant lung population as reported by many other groups (11,18,19).Despite being the predominant ILC population, ILC2s do not play a significant role in the early protection following Mtb infection as IL-13 deficient mice do not show increased early susceptibility (2).The scRNA-seq analysis also identified ILC3 as well as ILC1/ILC3-like populations in the Mtb-infected lung.Our in depth analysis show that markers such as Jaml, Gtf3a, Zbtb16, Rorc, and Ccr6 are upregulated in the ILC3 cluster but not in ILC3/ILC1-like cluster, while there are unique markers such as Cxcr3 and Lck that are upregulated in the ILC3/ILC1-like cluster.The presence of IL-23 and IL-1R expression on ILC3s in our data set suggests that this population can directly respond to the cytokine stimuli and amplify responses following infection.However, this needs to be tested in future studies.IL23r expression within ILC3 subset has previously been reported by another group (20).Moreover, it has been shown that continuous IL-23 stimulation led to ILC3 depletion in the gastrointestinal tract (21).Therefore, the presence of this receptor could also act as a regulator of ILC3-mediated signaling during the course of infection.Interestingly, combined IL-23/TNFα stimulation led to conversion of ILC1-type population from ILC3 (21).Therefore, it would be interesting to know whether lung ILC3s are also similarly responsive to cytokine stimuli and convert to another ILC population during Mtb infection.Corral et al. also reported the presence of ILC1-like populations which were shown to drive Mtb-induced type1 immune responses (18).In the gut, ILC1-like populations have been found to be derived from ILC3 populations (22).Therefore, the diversity and plasticity within the ILC populations could be attributed to the local environment and presence of microbial stimuli.ILC3 generation from ILC1-like cells likely occurs in the presence of polarizing cytokines such as IL-1β and IL-23 (23).During Mtb infection, we and others have reported induction of both IL-1β and IL-23 by DCs or macrophages which may lead to the conversion of ILC1s to ILC3s or, alternatively, plastic populations that express markers of both ILC1s and ILC3s.Although Corral et al. reported the presence of ILC1-like cells following Mtb infection (18), our results do not support a protective role for ILC1 cells in the early immunity following Mtb infection as Ifnγ −/− mice did not show early innate susceptibility to Mtb infection (2).The discrepancy between our study, describing a protective role for ILC3s (2), and the Corral study, describing an ILC1-like cell type (18) as protective, could be due to the different infecting Mtb strains used.Our studies have used a clinically relevant hyper-virulent Mtb HN878 strain which induces more robust IL-1β responses (24,25), as compared to lab adapted H37Rv Mtb strain, which was used in the Corral et al. study (18).Future work using different lineages of Mtb and infecting strains to study accumulation of ILC subsets in the lung following infection should shed new light on the early differentiation and protective nature of ILCs during Mtb infection.
One of the consistent effects of the early defect in accumulation of lung ILC3s was the decreased accumulation of AMs and monocytes in the Mtb-infected lung (2) [our published work (2) and this study].The early activation of AMs has been shown to be important in their migration form the airways into the parenchyma to initiate granuloma formation (26).The localization of Mtb-infected AMs into the lung interstitium is also shown to be dependent on IL-1R signaling acting on lung epithelial cells (27).
Additionally, in an in vitro co-culture system of macrophages and small airway epithelial cells, macrophages produced IL-1β in response to Mtb infection, which eventually acted on the epithelial cells, inducing their production of the antimicrobial peptide DEFB4/ HBD2, which was effective in controlling Mtb replication inside the infected macrophage (14).These results are consistent with our results, suggesting that ILC3s are early innate cells that drive downstream responses to regulate AM accumulation and Mtb control.It is also possible that during Mtb infections, ILC3s can also secrete IL-1β along with other cytokines, which would then act on epithelial cells to induce downstream effector function.We also expect that Mtb infection-induced IL-1 will act on the ILC3s and induce ILC3 effector function or plasticity, depending on the microenvironment.Collectively, these findings point toward a novel role for ILC3s in mediating protective mechanisms by which macrophages can respond, undergo activation, and mediate early and innate Mtb control.
In our previous study, we demonstrated that CXCR5 signaling is important in accumulation of lung ILC3s following infection for optimal early Mtb control (2).Indeed, our current study confirms that CXCR5 signaling on ILC3s is important for localization of ILC3s from the periphery into the lung and supported by the increased early suscepti bility seen in Cxcr5 f/f Ror cre Mtb-infected mice.Lung ILC3s are likely the first responder cells following Mtb infection and CXCR5 + ILC3s likely accumulate in the lung from the periphery in response to Mtb infection.We have not observed any defects in terms of Mtb clearance in Cxcr5 f/f Ror cre mice at 15 dpi.However, we did find lower accumulation of ILC3s at 15 dpi in Cxcr5 f/f Ror cre mice, with increased susceptibility at 24 dpi.Therefore, we speculate that the resident ILC3s and local proliferation of these resident cells are enough to provide early immunity following Mtb infection, while CXCR5-expressing ILC3s may be recruited from the periphery and contribute to early but sequential protection at 24 dpi.The ligand for CXCR5, namely, CXCL13, is inducible in the murine lung following infection with Mtb (28)(29)(30).In our previous study, we demonstrated that mouse and human ILCs migrate in response to CXCL13 in a CXCR5-dependent manner (2).Epithelial cells are a potential source of CXCL13 as reported by us and many other groups (31,32).Therefore, it is possible that the early induction of CXCL13 within lungs of Mtb-infec ted mice is responsible for the recruitment of CXCR5-expressing ILC3s from circulation.Furthermore, in our previous study, we demonstrated that Il17/Il22 −/− double-knockout mice displayed a significant early increase in lung Mtb burden, decreased number of lung ILC3s as well as CXCR5-expressing ILC3s, and decreased expression of Cxcl13 mRNA within GrALT (2).Therefore, further detailed mechanistic analysis of the distinct ILC3 subsets in the future will provide additional insights into understanding the role of lung versus recruited populations of ILC3s in Mtb infection, and other pulmonary infections.
Additionally, ILC3s can present antigen to T cells, thereby helping in further T-cell activation (33), and ILC3s are known to interact with B cells (34).ILC3s can control B-cell function and homeostasis in a T cell-independent manner.Human, (RORγt)+ ILCs (mostly ILC3s) activate naïve, marginal zone, and plasma B cells by expressing cell surface molecules such as CD40L, and Delta-like 2 (a notch ligand) (35), thereby further inducing IgA-class switching in lymphoid structures (36).Our scRNA-seq showed the expression of Cd40lg in ILC3 clusters, likely induced due to ongoing infection.Therefore, it is also possible that the ILC3s play an early role in activation and accumulation of T and B cells and formation of protective GrALT.Recently, another report described intestinal "trained" ILC3s (Tr-ILC3) that emerged and persisted after initial encounter with Citrobacter rodentium (37).Further infection led to Tr-ILC3s proliferation and robust production of IL-22, thus promoting mucosal defense.Therefore, it is exciting to consider that ILC3s can be targeted for activation by adjuvants for vaccine design against TB.
In summary, our data revealed Mtb infection-induced ILC populations, especially ILC3 subsets, which are potentially being regulated by the lung epithelial signaling.Moreover, we demonstrated novel role played by IL-1, IL-23, and CXCR5 signaling in ILC3 accu mulation within Mtb-infected mice lung.Therefore, understanding the early signaling events responsible for ILC3-mediated immune response will help in the development of targeted therapeutic strategies aimed at enhancing ILC3-mediated protection against Mtb infection.

Mice
C57BL/6 (B6), B6.129S7-Il1r1 tm1Imx /J (Il1r −/− ), and B6.129(Cg)-Il1r1 tm1.1Rbl /J (Il1r f/f ) mice were obtained from Jackson Laboratory (Bar Harbor, ME) and bred at Washington University in St. Louis.Il17 −/− ( 38 ) and Il22 −/− ( 39 ) mice were crossed at Washington University in St Louis to obtain double knockout mice.Cxcr5 f/f mice were generously donated by Dr. Nell Mabbott and Dr. Barry Bradford from University of Edinburgh.Ikk2 f/f Sftpc cre mice were a kind gift from Dr. Pasparakis (University of Cologne).Mice were age-and sex-matched and used between 6 and 8 weeks of age.All mice were used and housed in accordance with the National Institutes of Health guidelines for housing and care of laboratory animals.Il1r f/f Sftpc cre mice were obtained by crossing the Il1r f/f strain with Sftpc cre mice.Similarly, Il1r f/f Ikk2 f/f Sftpc cre and Cxcr5 f/f Rorγt cre mice were generated.

Bacterial infection
Mtb W. Beijing strain, HN878 (BEI Resources), were grown to mid-log phase in Proskauer Beck medium containing 0.05% Tween 80 and frozen in at −80°C.Mice were infected using a Glas-Col airborne infection system, with approximately 100 CFU of bacteria as described previously (40).Lungs were collected and homogenized; serial dilutions of tissue homogenates were plated on 7H11 plates; and the CFU was counted.At given time points following infection, lungs were collected and homogenized, and the tissue homogenates were plated following serial dilutions on 7H11 agar (BD Bioscience) to assess bacterial burden.

Generation of single-cell suspensions from tissues
Lung single-cell suspensions from Mtb-infected mice were isolated as previously described (41).Briefly, mice were euthanized with CO 2 , and lungs were perfused with heparin in saline.The lungs were minced and incubated in Collagenase/DNAse (Sigma-Aldrich) for 30 minutes at 37°C.Lung tissue was pushed through a 70-µm nylon screen to obtain a single-cell suspension.Red blood cells were lysed, and the cells were re-suspen ded in suitable media or buffer for further use.

Generation of bone marrow chimera
Recipient Il1r f/f , Il1r f/f Sftpc cre , or B6 mice were exposed to two doses of 450 cGy of X-rays in 48 hours.Following irradiation, mice were rested for 6-8 hours before donor's bone marrow cell transplant.Naïve B6 or Cxcr5 −/− mice euthanized, and cells were isolated from the femur and tibia.Bone marrow cells (1 × 10 7 ) were transplanted into the recipient mice by intravenous route, and the mice were put on antibiotic-water (sulfamethoxazole and trimethoprim oral suspension, NDC 62559-550-16).After 9 weeks, the reconstitution efficacy was checked via flow cytometry, and the chimeric mice were infected with Mtb as described before.

ILC cell sort
ILC populations were sorted from mice infected with HN878 Mtb at 5 and 14 dpi.At the indicated time post-infection, mice were sacrificed, and lung single-cell suspensions were made as previously described (2).CD45 + immune cells populations were first enriched from mice using CD45 + microbeads according to the manufacturer's instruc tion (Miltenyi Biotec).These enriched populations were then stained with the following antibody cocktail: CD45, CD127, NK1.1, and lineage markers (CD3, CD5, CD19, TER119, CD11b, and CD11c).To isolate out the ILC population, the Lin − NK1.1 − CD45 + CD127 + cells were sorted using the BD FACS Jazz flow cytometer with FACS Sortware software (BD).The purity of the sorted population was analyzed and reported to be >98% (Fig. S1A).Cells were collected into phosphate-buffered saline containing 0.04% non-acetylated bovine serum albumin and were subjected to scRNA seq.

scRNA-seq library generation and sequencing
Highly pure ILC populations were sorted as mentioned before and subjected to droplet-based, massively parallel single-cell RNA sequencing using a Chromium Single Cell 3′ (v.3) reagent kit according to the manufacturer's instructions (10× Genomics) as described before (41).Briefly, cell suspensions were loaded at 1,000 cells/µL to capture 10,000 cells/lane.The 10× Chromium Controller generated gel bead-in-emuslion (GEM) droplets, where each cell was labeled with a specific barcode, and each transcript was labeled with a unique molecular identifier (UMI) during reverse transcription.The barcoded cDNA was isolated and removed from the BSL-3 space for library genera tion.The cDNA underwent 11 cycles of amplification, followed by fragmentation, end repair, A-tailing, adapter ligation, and sample index PCR according to the manufacturer's instructions.Libraries were sequenced on a NovaSeq S4 (200 cycle) flow cell, targeting 50,000 read pairs/cell.

scRNA-seq data processing
The raw gene expression matrices were generated by the Cell Ranger software (10× Genomics, v.3) available on the 10× website.After demultiplexing, the resulting fastq files were aligned against the mouse reference genome mm10 with cellranger count.For each sample, the recovered-cell parameter was set to 10,000 cells that we expect to recover for each library.The output filtered gene count matrices were analyzed by R software (v.4.0.4) with the Seurat (42) package (v.4.1.0).Cells that had less than 600 and more than 2,200 detected genes were filtered out.Different thresholds were chosen for mitochondrial genes because of their distinct distribution in each sample.Criteria for filtered cells were as follows: (i) had more than 6% (D5) and (ii) 5% (D14) of mitochon drial genes.Samples were merged and normalized with default parameters, and most variable genes were detected by the FindVariableFeatures function.ScaleData was used to regress out the number of UMIs and mitochondrial content, and principal compo nent analysis (PCA) was performed with RunPCA.The t-distributed stochastic neighbor embedding (tSNE) dimensionality reduction was performed on the scaled matrix using the first 18 PCA.For clustering, the FindNeighbors (18 PCA) and FindClusters (resolution 0.5) functions were used.FindAllMarkers was used to compare a cluster against all other clusters to identify the marker genes.For each cluster, the minimum required average log fold change in gene expression was set to 0.25, and the minimum percentage of cells that must express gene in either cluster was set to 25%.To re-analyze ILC populations, we pooled the clusters that we identified as ILC origin and re-run PCA, tSNE, and clustering (resolution 1) to get a better resolution for further analysis.

Histological analysis
The right upper lobe was collected at the time of harvest for histological analysis of inflammation as previously described.Briefly, the lobes were perfused with 10% neutral buffered formalin and embedded in paraffin.Sections (5 µm) of formalin-fixed and paraffin-embedded (FFPE) lung were cut with a microtome, stained with hema toxylin and eosin, and processed for light microscopy.Images were captured using the automated NanoZoomer digital whole-slide imaging system (Hamamatsu Photon ics).Regions of inflammatory cell infiltration were delineated utilizing the NDP view2 software (Hamamatsu Photonics), and the total area of inflammation was calculated in individual lung lobes by adding all the inflammatory regions of individual lung lobes.All scoring was conducted in a blinded manner utilizing n = 5-10 mice per group.

Immunofluorescence staining
Immunofluorescence staining was done as described before (2).FFPE lung sections were cut, immersed in xylene, and then hydrated in alcohol and PBS.Antigens were unmasked using DakoCytomation target retrieval solution (Dako), and non-specific binding was blocked by adding 5% (vol/vol) normal donkey serum and Fc block (BD).Avidin was used to neutralize endogenous biotin, followed by incubation with biotin (Sigma-Aldrich).Sections were then probed with anti-B220 (clone RA3-6B2, BD) to detect B cells.For analysis of B-cell follicles, follicles were outlined with an automated tool of the Zeiss Axioplan 2 microscope (Zeiss), and total area and average size were calculated in squared microns.

Cytokine and chemokine quantification using Luminex or ELISA
Cytokine and chemokine protein contents in lung homogenates were quantified using Luminex multianalyte technology (Millipore) according to the manufacturer's protocols.

Statistical analysis
The differences between the two groups were analyzed using two-tailed Student's t-test in Prism (v.9, GraphPad).Differences between the means of three or more groups were analyzed using one-way analysis of variance (ANOVA) with Tukey's post-test.For comparisons between two or more groups with two independent variables, two-way ANOVA with Sidak's or Tukey's post-test was used.A P value of <0.05 was considered significant.

FIG 1
FIG 1 scRNA-seq transcriptional profiling of lung ILCs isolated following Mtb infection.B6 mice were aerosol infected with Mtb HN878, and lungs were harvested at 5 and 14 dpi.(A) t-Distributed stochastic neighbor embedding (tSNE) visualization of the different clusters (all samples together).(B) tSNE plot with the expression of known cell markers.The expression of marker genes was used to characterize distinct clusters according to cell identity.(C) tSNE plot of ILC subtypes after re-clustering the ILCs only, colored according to cellular identity, split by days.(D) Dot plot indicating expression of key genes detected for each cell subtype.The dot color represents the expression level, and the dot size represents the percentage of cells in each cluster expressing a particular gene.(E) Distribution of ILC subsets by cluster per days (infected at 5 dpi, pooled from 10 mice; infected at 14 dpi, pooled from 10 mice).(F) Violin plot of genes in ILC3, ILC3/ILC1-like and γδ T-cell subsets by cluster per days (infected at 5 dpi, pooled from 10 mice; infected at 14 dpi, pooled from 10 mice).

FIG 3
FIG 3 IL-1R signaling on lung epithelial cells is important for early accumulation of ILC3s following Mtb infection.Il1r f/f and Il1r f/f Sftpc cre mice were aerosol-infec ted with 100 CFU Mtb HN878.(A) Mice were harvested at 14 dpi, and the lung bacterial burden was determined by plating (n = 5-9 per group).(B-E) Numbers of ILC1s, ILC2s, ILC3s, and NKp46 + ILC3s, (F) AMs, and (G) monocytes were measured by flow cytometry on total lung single-cell suspensions at 14 dpi.In a separate experiment, Il1r f/f and Il1r f/f Sftpc cre mice were sublethally irradiated and reconstituted with B6 bone marrow cells.Chimeric animals were allowed to recover, then infection was done with around 100-CFU Mtb HN878.(H) Lung bacterial burden was determined by plating.(I-N) Numbers of ILC1s, ILC2s and ILC3s, NKp46 + ILC3s, AMs, and monocytes were quantified at 14 dpi using flow cytometry (n = 6-7 mice per group).All data are mean ± SD. *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001 were determined by Student's t-test (A-N).

FIG 4
FIG 4 CXCR5 signaling plays critical role in ILC3 recruitment to the lung during Mtb infection.Il17/22 −/− , Cxcr5 −/− , and B6 mice were aerosol-infected with 100 CFU Mtb HN878.(A) Mice were harvested at different times post-infection, and the number of total ILC3s was measured by flow cytometry on total lung single-cell suspensions (n = 5-9 per group).In a separate experiment, B6 (WT) mice were sublethally irradiated and reconstituted with either B6 or Cxcr5 −/− bone marrow cells.Chimeric animals were allowed to fully reconstitute, then infection was carried out with ~100-CFU Mtb HN878.(B) Mice were harvested at 14 dpi, and the number of lung ILC3s was quantified by flow cytometry (n = 4-5 mice per group).Cxcr5 f/f and Cxcr5 f/f Rorγt cre mice were aerosol-infected with ~100-CFU Mtb HN878.(C and K) Lung bacterial burden at 24 and 30 dpi was determined by plating (n = 5-10 per group).(D-G and L-O) Numbers of ILC1s, ILC2s, ILC3s, and NKp46 + ILC3s, (H and P) AMs and (I and Q) monocytes were measured by flow cytometry on total lung single-cell suspensions at 24 and 30 dpi, respectively.(J) Formalin-fixed paraffin-embedded lung sections from Mtb-infected mice were stained with antibodies against B220, and the average sizes of B cell follicles were quantified.All data are mean ± SD. *P ≤ 0.05, **P ≤ 0.01 were determined by either two-way analysis of variance (ANOVA) (A), one-way ANOVA (B), or Student's t-test (C-Q).ND, not detected.