TNF-α sculpts a maturation process in vivo by pruning tolerogenic dendritic cells.

It remains unclear how the pro-immunogenic maturation of conventional dendritic cells (cDCs) abrogates their tolerogenic functions. Here, we report that the loss of tolerogenic functions depends on the rapid death of BTLAhi cDC1s, which, in the steady state, are present in systemic peripheral lymphoid organs and promote tolerance that limits subsequent immune responses. A canonical inducer of maturation, lipopolysaccharide (LPS), initiates a burst of tumor necrosis factor alpha (TNF-α) production and the resultant acute death of BTLAhi cDC1s mediated by tumor necrosis factor receptor 1. The ablation of these individual tolerogenic cDCs is amplified by TNF-α produced by neighboring cells. This loss of tolerogenic cDCs is transient, accentuating the restoration of homeostatic conditions through biological turnover of cDCs in vivo. Therefore, our results reveal that the abrogation of tolerogenic functions during an acute immunogenic maturation depends on an ablation of the tolerogenic cDC population, resulting in a dynamic remodeling of the cDC functional landscape.

INTRODUCTION

The absence of specific pro-inflammatory signals, referred to as the steady state, is conducive to induction and maintenance of peripheral T cell tolerance, while the disturbance of steady-state conditions by pro-inflammatory stimuli generally results in a change from tolerogenic to immunogenic outcomes of conventional dendritic cell (cDC) and T cell interactions (Cabeza-Cabrerizo et al., 2021; Dalod et al., 2014; Iberg et al., 2017; Iwasaki and Medzhitov, 2015; Mellman, 2013; Steinman et al., 2003a). The recognition of pathogen-associated molecular patterns (PAMPs) through the corresponding pattern recognition receptors (PRRs) triggers activation of cDCs in a process referred to as maturation that results in the acquisition of specific immune functions (Dalod et al., 2014; Iwasaki and Medzhitov, 2015; Mellman, 2013; Steinman, 2012; Steinman et al., 2003a). The archetypical concept of maturation was established early by studying the impact of a canonical PAMP, lipopolysaccharide (LPS), on T cell priming (Cabeza-Cabrerizo et al., 2021; De Smedt et al., 1996). Subsequent studies found that maturation induced by LPS and other PAMPs enhances in individual cell functions of the inflammasome and increases the expression of major histocompatibility complexes (MHCs), multiple costimulatory ligands, cytokines, and other molecules that collectively enable pro-immunogenic effector T cell priming (Dalod et al., 2014; Fitzgerald and Kagan, 2020; Mellman, 2013). After mature cDCs complete the process of T cell priming, their death may be critical to avoid excessive immune activation (Chen et al., 2006; Stranges et al., 2007). Further, long-term sequelae of infections include remodeling of the cDC population (Fonseca et al., 2015). However, cDCs also may undergo a programmed cell death soon after initiation of the maturation process (De Trez et al., 2005; Zanoni et al., 2009, 2016). This is consistent with other studies that observed an LPS-induced death and reduced numbers of cDCs, especially among the CD8 α+ subset, usually attributing it to mechanisms regulating T cell priming while also serving as a potential source of antigenic materials (De Smedt et al., 1996; Fuertes Marraco et al., 2011; Qiu et al., 2009; Sundquist and Wick, 2009; Xu et al., 2017). Further, in a model of sepsis, functions of glucocorticoids regulated the loss of CD8α+ cDCs (Li et al., 2015). However, more recent studies uncovered relevant molecular mechanisms that specifically prevent the pyroptotic death of maturing cDCs, therefore directly preserving their immune functions (McDaniel et al., 2020; Zhivaki et al., 2020).

In addition to stimulating immune functions, some microbial as well as endogenous, including tumor-associated, maturation stimuli also induce specific tolerogenic properties in cDCs (Anderson et al., 2009; Harimoto et al., 2013; Iberg and Hawiger, 2020; Maier et al., 2020; Wang et al., 2019; Zhou et al., 2014). Various types of cDCs with such tolerogenic functions, especially those that migrate to draining lymph nodes (LNs), have important roles for maintaining homeostasis at anatomical barriers exposed to the commensal microbiota and other environmental cues (Arpaia et al., 2013; Cummings et al., 2016; Esterhazy et al., 2019; Guilliams et al., 2010; Iberg and Hawiger, 2020; Kashem et al., 2017; Manicassamy and Pulendran, 2011; Miller et al., 2012; Russler-Germain et al., 2021; Vitali et al., 2012). In contrast to cDCs at anatomical barriers, systemic cDCs are present throughout the peripheral lymphoid organs and constantly survey local and circulating antigens derived from parenchymal, interstitial, and other non-barrier tissues. Further, the systemic cDCs can present antigens from pathogens during infections as well as those introduced by intramuscular vaccinations (Bourque and Hawiger, 2022). However, most systemic cDCs do not remain as immunologically inert bystanders in the steady state and are instead functionally competent in vivo, possibly in response to yet incompletely identified molecular ligands constitutively present in the steady state as well as other specific mechanisms operating intrinsically in cDCs under homeostatic conditions (Ardouin et al., 2016; Baratin et al., 2015; Dalod et al., 2014; Hammer and Ma, 2013; Iberg et al., 2017; Jiang et al., 2007; Manh et al., 2013; Vander Lugt et al., 2017). Such immunologically competent cDCs are responsible for the various effector and regulatory outcomes of T cell activation observed in the steady state (Bourque and Hawiger, 2022; Opejin et al., 2020).

The dominant forms of peripheral tolerance dependent on de novo induced peripheral regulatory T (pTreg) cells limit effector immune responses and crucially complement immunological tolerance first initiated in the thymus (Iberg et al., 2017; Jones and Hawiger, 2017; Jones et al., 2015). The induction of such systemic pTreg cells is efficiently mediated in the steady state by cDCs with inherent tolerogenic functions that belong to the Batf3-dependent, XCR1+ type 1 subset (cDC1), corresponding to previously defined DEC-205+CD8α+ cDCs (Guilliams et al., 2014; Iberg and Hawiger, 2020; Iberg et al., 2017; Steinman et al., 2003b). These tolerogenic cDC1s are further distinguished by their high expression of B- and T-lymphocyte associated/attenuator (BTLA) (Jones et al., 2016). In concert with some other mechanisms present in these cDC1s, BTLA promotes an efficient induction of pTreg cells in response to systemic antigens (Bourque and Hawiger, 2018; Iberg et al., 2017; Jones et al., 2016; Yin et al., 2021). The BTLAhi cDC1s are present in peripheral lymphoid organs, including spleen and LNs, and are therefore ideally positioned to present to T cells self-antigens, including those obtained from apoptotic materials (Albert et al., 1998; Belz et al., 2002; Eisenbarth, 2019; Iberg and Hawiger, 2020; Iyoda et al., 2002; Steinman et al., 2003b). Overall, the tolerance promoted in the steady state blocks subsequent immune responses induced by a pro-immunogenic maturation of cDCs. Crucially, however, it remains unclear how the functional re-programming during a pro-immunogenic maturation induced by PAMPs such as LPS abrogates tolerogenic functions of cDCs.

We now found that, instead of altering the functionality of individual tolerogenic cDCs, pro-immunogenic signals transiently remove the entire population of specialized tolerogenic cDCs. Upon induction of pro-inflammatory conditions in vivo, tolerogenic BTLAhi cDC1s are specifically ablated through the mechanisms of programmed cell death mediated by the functions of tumor necrosis factor alpha (TNF-α) and tumor necrosis factor receptor 1 (TNFR1), resulting in an abrogation of the specific de novo conversion of pTreg cells. Our findings further revealed that the production of TNF-α by neighboring cells potentiates this ablation of BTLAhi cDC1s, therefore amplifying the initial LPS-initiated signals. Consistent with the established rapid turnover and replenishment of cDCs in vivo (Kamath et al., 2000; Liu et al., 2007), the BTLAhi cDC1 tolerogenic population is restored within a few days. Overall, these findings expand the concept of immunogenic activation and maturation to crucially include a dynamic reshaping of the cDC population and ablation of tolerogenic cDCs. By revealing the role of TNF-α and TNFR1 in mediating the loss of tolerogenic cDCs, these results also clarify functions of TNF-α during the maturation process. Further, by elucidating the dynamic cycle of a transient loss and subsequent restoration of tolerogenic cDCs, they help to clarify the biological role of the constant turnover of cDCs in vivo.

RESULTS

Tolerogenic cDCs are ablated under pro-inflammatory conditions

We applied the currently established approach for identification of XCR1+ cDC1s and CD172a+ cDC2s (Guilliams et al., 2016) to rigorously analyze splenic cDCs following administration of LPS in vivo (Figure S1A). Within 1 day after administration of LPS, the proportion of XCR1+ cDC1s among cDCs was reduced by about 60%–70%, whereas the relative proportion of cDC2s among cDCs correspondingly increased (Figures 1A–1C). A terminal deoxynucleotidyl transferase deoxyuridine triphosphate (dUTP) nick end labeling (TUNEL) assay confirmed the death of cDC1s (Figure 1D). Consistent with their multifaceted functions in priming and regulating immune responses, cDC1s are heterogeneous (Durai and Murphy, 2016). The BTLAhi cDC1s are characterized by the highest expression of BTLA among all cDCs (Figure S1B) and have established tolerogenic functions (Jones et al., 2016). The BTLAhi cDC1s were disproportionally decreased by about 3-fold among the splenic cDC1s that remained 24 h after administration of LPS (Figure 1E). In contrast, the BTLAlo cDC1s and cDC2s increased expression of costimulatory molecules, such as CD86, consistent with their maturation (Figures S1C and S1D). We next confirmed the almost complete loss of BTLAhi cDC1s by measuring their proportion among all cDCs as well as leukocytes and by counting the absolute numbers of splenic BTLAhi cDC1s, whereas the number of total splenocytes remained similar in mice that were treated with either PBS or LPS (Figures 1F–1I and S1E). Overall, we established that the relevant loss of BTLAhi cDC1s could be comparably represented as a reduced fraction of either total cDCs or total leukocytes as well as by the absolute numbers (Figure 1J). Importantly, a similar loss of BTLAhi cDC1s was observed in response to different doses (1–10 μg/mouse) of LPS (Figure 1K).

Figure 1.

Ablation of tolerogenic cDCs under pro-inflammatory conditions

(A–C) Wild-type mice were treated with PBS or LPS (1 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. (A) Plots (representative of multiple independent experiments) show anti-XCR1 and anti-CD172a staining intensity in all cDCs (I-Ab+CD11c+—see gating strategy in Figure S1A). cDC1s are identified as XCR1+(CD172aneg), and cDC2s are identified as CD172a+(XCR1neg). Graphs shows percentages of cDC1s (B) or cDC2s (C) among all cDCs in the indicated groups (n = 4–6 mice per group pooled from two independent experiments).

(D) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 6 h before TUNEL analysis of splenocytes by flow cytometry. Graph shows percentages of TUNEL+ cells among cDC1s in the indicated groups (n = 2 samples of three mice pooled together per group from two independent experiments).

(E and F) Wild-type mice were treated with PBS or LPS (1 μg/mouse) 24 h before analysis of splenocytes by flow cytometry.

(E) Graph shows percentages of BTLAhi cells among cDC1s (see Figure S1B for the BTLAhi cutoff applied in all experiments; see also F) in the indicated groups (n = 4–6 mice per group pooled from two independent experiments).

(F) Plots (representative of multiple independent experiments) show anti-XCR1 and anti-BTLA staining intensity in all cDCs. The BTLAhi cDC1s are indicated by the purple shaded regions.

(G–J) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry (n = 3–5 mice per group pooled from two independent experiments). Graphs show percentages of BTLAhi cDC1s among all cDCs (G) or among all leukocytes (H) or the absolute number of BTLAhi cDC1s per spleen (I) in the indicated groups. (J) Graph shows the percent of baseline (average of the corresponding PBS-treated mice) for BTLAhi cDC1s calculated as a percent among all cDCs, among all leukocytes, or as the absolute number of cells per spleen.

(K) Wild-type mice were treated with PBS or the indicated doses of LPS 24 h before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs (in the indicated groups; n = 4–6 mice per group pooled from two independent experiments).

(L) Sorted Foxp3negCD25neg T cells from OT-II TCR tg Foxp3RFP mice were adoptively transferred into wild-type mice that were treated with anti-DEC-OVA only (PBS) or anti-DEC-OVA and 1 μg LPS (LPS) 14 days before analysis of splenocytes by flow cytometry. Graph shows percentages of Foxp3(RFP)+ cells among the Vα2+ Vβ5.1,5.2+ CD4+ T cells in the indicated groups of recipients (n = 5 mice per group pooled from two independent experiments).

(A and F) Numbers next to outlined regions indicate corresponding percentages. (B–E and G–L) Graphs show mean ± standard deviation (SD). **p < 0.01, ***p < 0.001, and ****p < 0.0001 determined by unpaired two-tailed t test (B–E, G–I, and L), two-way ANOVA with Šídák’s multiple comparisons (J), or one-way ANOVA with Tukey’s multiple comparisons (K).

In Batf3−/− mice that lack most splenic cDC1s, the de novo conversion of pTreg cells is reduced (Hildner et al., 2008; Jones et al., 2016; Tussiwand et al., 2012). Consistent with these results, we found a near absence of splenic BTLAhi cDC1s in Batf3−/− mice and a corresponding lack of induction of antigen-specific pTreg cells following the specific in vivo delivery of cognate antigen (Figures S1F–S1H). As expected, and in agreement with a general abrogation of tolerogenic outcomes of cDCs and T cell interactions under acutely induced pro-inflammatory conditions, we observed a comparable absence of such de novo induced pTreg cells in the Batf3+/+ (wild-type [WT]) mice in which BTLAhi cDC1s were ablated by LPS (Figure 1L). Overall, we conclude that, consistent with their established general pro-immunogenic effects, inflammatory conditions acutely induced by LPS result in the loss of tolerogenic BTLAhi cDC1s.

BTLAhi cDC1s are ablated in trans

In addition to its roles in promoting pTreg cell conversion through ligand-receptor interactions with herpesvirus entry mediator (HVEM) expressed by T cells (Bourque and Hawiger, 2019; Henderson et al., 2015; Jones et al., 2016), BTLA also has cell-autonomous functions (Kobayashi et al., 2013; Murphy and Murphy, 2010; Shui et al., 2011). To test the possible impact of such functions on the loss of BTLAhi cDC1s, we analyzed cDCs in Btla−/− mice and found a similar decrease in cDC1s in response to LPS, regardless of the presence or absence of BTLA (Figure S2A). Further, the ablation of BTLAhi cDC1s was independent of Fas and type I interferon receptor (IFNAR), which are involved in mediating certain types of cell death and cDC homeostasis (Fuertes Marraco et al., 2011; Mattei et al., 2009; Schaupp et al., 2020; Stranges et al., 2007; Figures S2B and S2C). The effects of LPS are signaled through several molecular pathways (Fitzgerald and Kagan, 2020). Toll-like receptor 4 (TLR4) is a canonical LPS receptor that is expressed by cDC1s (Edwards et al., 2003), and the loss of BTLAhi cDC1s following LPS administration was completely dependent on the presence of TLR4 (Figure 2A). To further clarify the mechanisms resulting in the loss of BTLAhi cDC1s, we confirmed the roles of pathways known to mediate cell death. Receptor-interacting serine and threonine-protein kinase 3 (RIPK3), caspase 3, and caspase 8 are established executers of programmed death pathways (Kolb et al., 2017; Nagata and Tanaka, 2017). In the steady state, the proportion of BTLAhi cDC1s among cDCs was unaffected by the absence of RIPK3, caspase 3, and caspase 8 (Figures S2D and S2E). Supporting the functions of these programmed death pathways in mediating the ablation of BTLAhi cDC1s, this population was only moderately reduced in the Ripk3−/−Casp3−/−Casp8ΔCD11c mice in response to LPS, remaining at about 50% of its steady-state baseline (Figure 2B). The relationship between TLR4-mediated signals and programmed death mechanisms is complex, and LPS can affect functions of multiple hematopoietic and non-hematopoietic cell populations (Fitzgerald and Kagan, 2020). In agreement with autonomous functions of TLR4 in hematopoietic cells, LPS-induced ablation was prevented in BTLAhi cDC1s developing from the Tlr4−/− bone marrow (BM) that reconstituted chimera mice similarly to Tlr4+/+ BM (Figures 2C and S2F). A conditional deletion of Tlr4 in macrophages or B cells did not prevent a loss of BTLAhi cDC1s in response to LPS (Figures S2G and S2H). In contrast, a specific deletion of Tlr4 using Itgax-Cre that deletes most efficiently in CD11c+ cDCs (Abram et al., 2014; Melillo et al., 2010; Figure S2I) rescued such a loss of BTLAhi cDC1s (Figure 2D), revealing the relevant mechanisms to be dependent on the expression of TLR4 specifically in CD11c+ cDCs. These specific TLR4-mediated functions were also dependent on signaling through myeloid differentiation primary response protein MyD88 (MyD88) and Toll/interleukin-1 (IL-1) receptor (TIR)-domain-containing, adapter-inducing interferon-β (TRIF) (Figure 2E). To distinguish whether the TLR4-dependent death and ablation of BTLAhi cDC1s proceeded cell autonomously in the individual cDCs or whether such ablation depended on additional extrinsic factors, we constructed Trif+/+Myd88+/+:Trif−/−Myd88ΔCD11c mixed BM chimeras and confirmed an equal reconstitution of the cDC populations (Figures 2F and S2J). As expected, we observed a specific loss of the Trif+/+Myd88+/+ BTLAhi cDC1s in response to LPS. However, unexpectedly, we found a similar ablation of the Trif−/−Myd88ΔCD11c BTLAhi cDC1s in these LPS-treated mixed BM chimeras (Figure 2F). This suggested that the ablation of BTLAhi cDC1s is mediated indirectly (in trans) within the cDC population. Further consistent with such an indirect mechanism, we observed a comparable ablation of both Tlr4−/− and Tlr4+/+ BTLAhi cDC1 populations within the equally reconstituted Tlr4−/−:Tlr4+/+ mixed BM chimeras treated with LPS (Figures 2G and S2K). We conclude that the initial TLR4-dependent sensing of LPS by cDCs mediates a secondary, TLR4-independent death and ablation of BTLAhi cDC1s.

Figure 2.

BTLAhi cDC1s are ablated in trans

(A) Tlr4+/+ and Tlr4−/− mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs in the indicated groups (n = 3–6 mice per group pooled from three independent experiments).

(B) Ripk3+/+Casp3+/+Casp8+/+ and Ripk3−/−Casp3−/−Casp8ΔCD11c mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs in LPS-treated mice as a percentage of the baseline (average of the corresponding PBS-treated mice) in the indicated groups (n = 6–8 mice per group pooled from three independent experiments).

(C) Lethally irradiated Tlr4+/+ mice were reconstituted with congenically labeled Tlr4−/− or Tlr4+/+ bone marrow (BM). The reconstituted BM chimera mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. (Left) General experimental outline is shown. (Right) Graph shows percentages of BTLAhi cDC1s among all cDCs in LPS-treated mice as a percentage of the baseline (average of the corresponding PBS-treated mice) in the indicated groups (n = 3–6 mice per group pooled from two independent experiments).

(D) Tlr4+/+ and Tlr4ΔCD11c mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs in LPS-treated mice as a percentage of the baseline (average of corresponding PBS-treated mice) in the indicated groups (n = 4–7 mice per group pooled from two independent experiments).

(E) Trif+/+Myd88+/+ and Trif−/−Myd88ΔCD11c mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs in LPS-treated mice as a percentage of the baseline (average of corresponding PBS-treated mice) in the indicated groups (n = 4–6 per group pooled from two independent experiments).

(F) Lethally irradiated Trif+/+Myd88+/+ mice were reconstituted with a 1:1 mix of congenically labeled Trif−/−Myd88ΔCD11c and Trif+/+Myd88+/+ BM. The reconstituted mixed BM chimera mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. (Left) General experimental outline is shown. (Right) Graph shows percentages of BTLAhi cDC1s among cDCs within the indicated genotype for each treatment (n = 3–8 mice per treatment pooled from two independent experiments).

(G) Lethally irradiated Tlr4+/+ mice were reconstituted with a 1:1 mix of congenically labeled Tlr4−/− and Tlr4+/+ BM. The reconstituted mixed BM chimera mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. (Left) General experimental outline is shown. (Right) Graph shows percentages of BTLAhi cDC1s among cDCs within the indicated genotype for each treatment (n = 3–6 mice per treatment pooled from two independent experiments).

(A–G) Graphs show mean ± SD. ns, not significant, ***p < 0.001, and ****p < 0.0001 determined by unpaired two-tailed t test (B–E) or one-way ANOVA with Tukey’s multiple comparisons (A, F, and G).

Rapid production of TNF-α mediates and amplifies the ablation of BTLAhi cDC1s

To further characterize the ablation of BTLAhi cDC1s, we examined the cDC population at multiple time points after administration of different doses of LPS. We observed an initial, about 25% decrease of BTLAhi cDC1s already by 3 h, a 50% reduction of this population by 6 h, and an almost complete ablation of the BTLAhi cDC1s within 24 h (Figures 3A and S3A). Such an early onset of ablation suggested a role for a rapidly produced mediator. TNF-α, which is encoded by an immediate-early gene, is produced following exposure to specific signals (Falvo et al., 2010). Production of TNF-α is associated with a process of cDC activation (Kaisho et al., 2001). Further, TNF-α is well established to mediate programmed cell death (Wajant and Siegmund, 2019). We found that TNF-α production by cDCs peaked about 1 h after LPS administration, and cDC1s were the main producers of TNF-α with about 60% of cDC1s producing TNF-α at the peak production point, whereas fewer than 10% of cDC2s produced TNF-α (Figures 3B–3E and S3B). This TNF-α production was dependent on the expression of TLR4 in cDCs and the corresponding functions of MyD88/TRIF but under these conditions did not require functions of other signaling pathways involving receptors such as IFNAR (Figures 3F, 3G, and S3C–S3F). Among cDC1s, the major producers of TNF-α were BTLAhi cDC1s, with about 70% of these cells becoming positive for TNF-α, whereas only 30%–40% of BTLAlo cDC1s produced TNF-α (Figure 3H). To test directly whether such acutely produced TNF-α promotes ablation of BTLAhi cDC1s, we pre-treated mice with an anti-TNF-α blocking antibody before administering LPS. LPS-treated mice that received anti-TNF-α blocking antibody had five to eight times more BTLAhi cDC1s among total cDCs as compared with LPS-treated mice without such an anti-TNF-α blockade, indicating a key role for TNF-α in mediating the ablation of BTLAhi cDC1s (Figure 3I). To elucidate whether TNF-α was required to mediate the ablation of BTLAhi cDC1s in trans in the absence of other signals that could be directly induced by LPS in such responding cells, we tested the impact of blocking TNF-α in the Tlr4−/−:Tlr4+/+ mixed BM chimeras. As expected, the administration of LPS resulted in a loss of Tlr4−/− BTLAhi cDC1s in the Tlr4−/−:Tlr4+/+ mixed BM chimeras treated with an isotype control antibody. However, blocking of TNF-α reversed the effects of LPS, completely preventing the loss of these BTLAhi cDC1s (Figure 3J). Therefore, under pro-inflammatory conditions, TNF-α propagates the ablation of BTLAhi cDC1s independently of specific innate signaling in these cells.

Figure 3.

Rapid production of TNF-α mediates and amplifies the ablation of BTLAhi cDC1s

(A) Wild-type mice were treated with PBS or LPS (10 μg/mouse) at the indicated time points before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs as a percentage of the baseline (average of PBS-treated mice for each time point) in the indicated groups (n = 4–6 mice per group pooled from two independent experiments per time point). Significance is shown for the comparison of PBS and LPS groups at each time point.

(B and C) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 1 h before analysis of splenocytes by flow cytometry.

(B) Plots (representative of multiple independent experiments) show anti-XCR1 and anti-TNF-α intracellular staining intensity in all cDCs.

(C) Graph shows percentages of TNF-α+ cells within the cDC1 (XCR1+ CD172aneg) and cDC2 (CD172a+XCR1neg) subsets, as indicated, from mice treated with LPS (n = 6 mice pooled from three independent experiments).

(D and E) Wild-type mice were treated with PBS (0 h) or LPS (10 μg/mouse) at the indicated time points before analysis of splenocytes by flow cytometry. Graphs show percentages of TNF-α+ cells within cDC1 (D) or cDC2 (E) subsets for each time point (n = 2–6 mice per group pooled from three independent experiments). Significance is shown for each time point compared with the baseline (0 h).

(F and G) Tlr4+/+ and Tlr4ΔCD11c were treated with LPS (10 μg) 1 h before analysis of splenocytes by flow cytometry.

(F) Plots (representative of multiple independent experiments) show anti-XCR1 and anti-TNF-α intracellular staining intensity in all cDCs.

(G) Graph shows percentages of TNF-α+ cells within the cDC1 subset in the indicated groups (n = 5–6 mice per group pooled from three independent experiments).

(H) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 1 h before analysis of splenocytes by flow cytometry. Graph shows percentages of TNF-α+ cells within the indicated populations of cDC1s (n = 6 mice pooled from three independent experiments).

(I) Wild-type mice were treated with anti-TNF-α or isotype control antibody as indicated and LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs in the indicated groups (n = 5–6 mice per group pooled from two independent experiments).

(J) Lethally irradiated Tlr4+/+ mice were reconstituted with a 1:1 mix of congenically labeled Tlr4−/− and Tlr4+/+ BM. The reconstituted mixed BM chimera mice were treated with anti-TNF-α or isotype control antibody and PBS or LPS (10 μg/mouse) as indicated 24 h before analysis of splenocytes by flow cytometry. (Left) General experimental outline is shown. (Right) Graph shows percentages of BTLAhi cDC1s among Tlr4−/− cDCs in the indicated groups (n = 3–7 mice per group pooled from two independent experiments).

(K and L) Wild-type mice were treated with PBS or recombinant mouse TNF-α 24 h before analysis of splenocytes by flow cytometry.

(K) (Top) General experimental outline is shown. (Bottom) Representative plots show anti-XCR1 and anti-BTLA staining intensity in all cDCs.

(L) Graph shows percentages of BTLAhi cDC1s among all cDCs in the indicated groups. Results represent one of two similar experiments (n = 3–5 mice per group).

(M) Lethally irradiated Tlr4+/+ mice were reconstituted with the indicated ratios of congenically labeled Tlr4−/− and Tlr4+/+ BM. The reconstituted mixed BM chimera mice were treated with LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. (Left) General experimental outline is shown. (Right) Graph shows percentages of BTLAhi cDC1s among Tlr4−/− cDCs in the indicated groups (n = 4–7 mice per group pooled from two independent experiments).

(B, F, and K) Numbers next to outlined regions indicate corresponding percentages. (A, D, and E) Graphs show mean ± standard error of mean (SEM).

(C, G–J, L, and M) Graphs show mean ± SD. *p < 0.05, **p < 0.01, and ****p < 0.0001 determined by two-way ANOVA with Šídák’s multiple comparisons (A), unpaired two-tailed t test (C, G–I, and L), or one-way ANOVA with Tukey’s multiple comparisons (D, E, J, and M).

Some non-specific signals produced under LPS-induced pro-inflammatory conditions might indirectly facilitate the TNF-α-mediated ablation of BTLAhi cDC1s. Therefore, to test directly whether TNF-α is sufficient to mediate the ablation of BTLAhi cDC1s, we administered purified recombinant TNF-α protein in vivo in the absence of an exposure to LPS. In the absence of LPS treatment, administration of TNF-α induced a robust loss of BTLAhi cDC1s (Figures 3K and 3L). Therefore, we conclude that TNF-α is sufficient to mediate ablation of BTLAhi cDC1s even in the absence of an innate stimulus. This result also suggested that the production of TNF-α in response to LPS can amplify the ablation of BTLAhi cDC1s when the innate stimulus is available only to a minor portion of cells in vivo. To clarify this, we used mixed BM chimeras with different ratios of Tlr4−/−:Tlr4+/+ cDCs (Figure 3M). Upon administration of LPS in these chimeras, we observed a robust ablation of Tlr4−/− BTLAhi cDC1s, even when they outnumbered their LPS-responsive counterparts by 4:1 (Figure 3M). Overall, we conclude that this induced production of TNF-α amplifies the ablation of individual BTLAhi cDC1s independently of a direct sensing by such cells of the original innate trigger of the pro-inflammatory maturation process.

TNF-α-dependent ablation of BTLAhi(CCR7lo) cDC1s throughout peripheral lymphoid organs

CCR7 governs migration of cDCs; therefore, CCR7hi cDCs are considered to have migratory properties, whereas CCR7lo cDCs reside in the lymphoid organs (Ohl et al., 2004). We found almost all splenic BTLAhi cDC1s to have low expression of CCR7, and such BTLAhi(CCR7lo) cDC1s constituted the main population of splenic cDC1s in the steady state (Figure S4A). This is consistent with BTLA being a selective marker of tolerogenic cDCs in the lymphoid organs independently of other markers’ expression (Jones et al., 2016). In agreement with their ablation in response to LPS, the BTLAhiCCR7lo cDC1 population was decreased in the absence of a corresponding increase of a proportion of CCR7-expressing cDC1s (Figures 4A and S4B). Similar to their splenic counterparts, the BTLAhi cDC1s present in LNs are also CCR7lo, but these BTLAhiCCR7lo cDC1s constitute only a minor population among the LN cDC1s in the steady state (Figures S4C and S4D). Nevertheless, treatment with LPS severely reduced this lymphoid BTLAhiCCR7lo cDC1 population, underscoring a curtailed induction of pTreg cells in LNs under the pro-inflammatory conditions following administration of LPS (Figures 4B–4D). To confirm the key roles of TNF-α in mediating the deletion of BTLAhiCCR7lo cDC1s independently of innate pro-inflammatory intrinsic signaling in these cells, we used Tlr4−/−:Tlr4+/+ mixed BM chimeras. We found that, similar to splenic BTLAhiCCR7lo cDC1s (Figure S4E), blocking the TNF-α produced in response to LPS restored the numbers of Tlr4−/− BTLAhiCCR7lo cDC1s in the LNs from the Tlr4−/−:Tlr4+/+ mixed BM chimeras (Figure 4E). To further ascertain the specific impact of TNF-α on BTLAhiCCR7lo cDC1s independently of the conditions induced by an innate pro-inflammatory stimulus, we administered purified recombinant TNF-α protein in vivo in the absence of an exposure to LPS, similar to what we did in Figures 3K and 3L. Administration of TNF-α in the absence of LPS treatment specifically reduced the BTLAhiCCR7lo population without affecting the proportions of other cDC populations present within LNs (Figures 4F, S4F, and S4G). Overall, we conclude that, upon acute induction of pro-inflammatory conditions, TNF-α promotes the loss of tolerogenic BTLAhi(CCR7lo) cDC1s in peripheral lymphoid organs.

Figure 4.

TNF-α ablates BTLAhi(CCR7lo) cDC1s throughout peripheral lymphoid organs

(A) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. Graphs show percentages of BTLAhiCCR7lo cDC1s (gated as in Figure S4A) among all leukocytes in the indicated groups (n = 4–6 mice per group pooled from two independent experiments).

(B and C) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 48 h before analysis of peripheral lymph nodes by flow cytometry. Graph shows percentages of BTLAhiCCR7lo cDC1s (gated as in Figures S4C and S4D) among all cDCs (B) or among all leukocytes (C) in the indicated groups (n = 3 mice per group). Results represent one of two similar experiments.

(D) Sorted Foxp3negCD25neg T cells from OT-II TCR tg Foxp3RFP mice were adoptively transferred into wild-type mice that were treated with anti-DEC-OVA only (PBS) or anti-DEC-OVA and 1 μg LPS (LPS) 14 days before analysis of peripheral lymph nodes by flow cytometry. Graphs show percentages of Foxp3(RFP)+ cells among Vα2+ Vβ5.1,5.2+ CD4+ T cells in the indicated groups (n = 5 mice per group pooled from two independent experiments).

(E) Lethally irradiated Tlr4+/+ mice were reconstituted with a 1:1 mix of congenically labeled Tlr4−/− and Tlr4+/+ BM. The reconstituted mixed BM chimera mice were treated with anti-TNF-α or isotype control antibody and PBS or LPS (10 μg/mouse) 48 h before analysis of peripheral lymph nodes by flow cytometry. (Top) General experimental outline is shown. (Bottom) Graph shows percentages of BTLAhiCCR7lo cDC1s among Tlr4−/− cDCs in the indicated groups (n = 3–7 mice per group pooled from two independent experiments).

(F) Wild-type mice were treated with PBS or recombinant mouse TNF-α 24 h before analysis of peripheral lymph nodes by flow cytometry. (Top) General experimental outline and diagram specifying the populations of cDC1s based on BTLA and CCR7 expression corresponding to populations gated as in Figures S4C and S4D is shown. (Bottom) Graphs show the percentages of the specified populations defined by specific BTLA and CCR7 expression in cDC1s among all cDCs in lymph nodes for the indicated groups. Results represent one of two similar experiments (n = 3–5 mice per group).

(A–F) Graphs show mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 determined by unpaired two-tailed t test (A–D and F) or one-way ANOVA with Tukey’s multiple comparisons (E).

TNFR1 orchestrates the ablation of tolerogenic BTLAhi cDC1s

The cytotoxic effects of TNF-α in various types of cells are crucially mediated by TNFR1, also known as TNF receptor superfamily member 1A (TNFRSF1A) (Wajant and Siegmund, 2019). Consistent with the specific impact of TNF-α on BTLAhi cDC1s, this population exhibited the highest expression of TNFR1 among all other splenic cDCs in the steady state (Figures 5A and 5B). Also, the corresponding BTLAhiCCR7lo population in LNs had a similar high expression of TNFR1, consistent with its sensitivity to TNF-α-mediated ablation (Figure S5A). To directly examine the role of TNFR1 in mediating the ablation of the BTLAhi cDC1s under pro-inflammatory conditions, we constructed Tlr4−/−Tnfrsf1a−/−:Tlr4+/+Tnfrsf1a+/+ and Tlr4−/−Tnfrsf1a+/+:Tlr4+/+Tnfrsf1a+/+ BM chimeras and confirmed the expected equal ratios of the correspondingly reconstituted cDCs both in spleens and LNs (Figures S5B and S5C). The absence of TNFR1 did not alter the proportions of BTLAhi cDC1s in these chimeras in the steady state (Figures 5D and 5E). However, whereas Tlr4−/−Tnfrsf1a+/+ BTLAhi cDC1s were ablated in trans as expected, the ablation of the splenic Tlr4−/−Tnfrsf1a−/− BTLAhi cDC1s was completely prevented under the LPS-induced pro-inflammatory conditions (Figure 5D). Similarly, we observed a comparable lack of ablation of the Tlr4−/−Tnfrsf1a−/− BTLAhiCCR7lo cDC1 population in LNs (Figure 5E). To clarify the impact of TNFR1-mediated effects on the tolerogenic functions of BTLAhi cDC1s, we examined the de novo conversion of pTreg cells in response to a specific antigen. In mice with unaltered responsiveness to LPS and TNF-α, the administration of LPS results in ablation of BTLAhi(CCR7lo) cDC1s and prevents pTreg cell formation as described above. In contrast, such pTreg cells could still be formed in both spleens and LNs by a small portion of Tlr4−/−Tnfrsf1a−/− BTLAhi(CCR7lo) cDC1s whose loss was prevented by their inability to respond to both the original innate stimulus as well as the ablation-amplifying signals mediated by TNF-α (Figures 5F–5H). Overall, we conclude that the abrogation of specific tolerogenic functions under acutely induced pro-inflammatory conditions proceeds through a physical deletion of a specialized cDC population that is amplified by TNFR1-dependent signals.

Figure 5.

TNFR1 orchestrates the ablation of tolerogenic BTLAhi cDC1s

(A and B) Spleens from untreated wild-type mice were analyzed by flow cytometry.

(A) Plot (left) shows anti-XCR1 and anti-BTLA staining intensity in all cDCs. Shaded regions specify gating of populations denoted in the overlaid histograms (right) that show anti-TNFR1 staining intensity of the indicated populations. Results are representative of multiple independent experiments.

(B) Graph shows median fluorescence intensity (MFI) of anti-TNFR1 staining in the indicated populations (gated as in A; n = 4 mice pooled from three independent experiments).

(C–E) Lethally irradiated Tlr4+/+Tnfrsf1a+/+ mice were reconstituted with a 1:1 mix of congenically labeled Tlr4−/−Tnfrsf1a+/+ and Tlr4+/+Tnfrsf1a+/+ or Tlr4−/−Tnfrsf1a−/− and Tlr4+/+Tnfrsf1a+/+ BM.

(C) General experimental outline for mixed BM chimeras.

(D) Reconstituted mixed BM chimera mice were treated with PBS or LPS (10 μg/mouse) 24 h before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among cDCs within the indicated genotype for each treatment (n = 3–8 mice per group pooled from two independent experiments).

(E) Reconstituted mixed BM chimera mice were treated with PBS or LPS (10 μg/mouse) 48 h before analysis of peripheral lymph nodes by flow cytometry. Graph shows percentages of BTLAhiCCR7lo cDC1s among cDCs within the indicated genotype for each treatment (n = 4–7 mice per group pooled from three independent experiments).

(F–H) Lethally irradiated Tlr4+/+Tnfrsf1a+/+ mice were reconstituted with a 2:1 mix of congenically labeled Tlr4−/−Tnfrsf1a+/+ and Tlr4+/+Tnfrsf1a+/+ or Tlr4−/−Tnfrsf1a−/− and Tlr4+/+Tnfrsf1a+/+ BM. Sorted Foxp3negCD25neg T cells from OT-II TCR tg Foxp3RFP mice were adoptively transferred into the reconstituted mixed BM chimera mice that were treated with anti-DEC-OVA and LPS (10 μg/mouse) 14 to 15 days before analysis of spleens and peripheral lymph nodes by flow cytometry.

(F) General experimental outline for adoptive transfers into reconstituted mixed BM chimeras.

(G and H) Graphs show percentages of Foxp3(RFP)+ cells among Vα2+ Vβ5.1,5.2+ CD4+ T cells in the indicated groups of recipients among splenocytes (G) or lymph node cells (H) (n = 10–12 mice per group pooled from three independent experiments).

(B, D, E, G, and H) Graphs show mean ± SD. *p < 0.05, **p < 0.01, and ****p < 0.0001 determined by one-way ANOVA with Tukey’s multiple comparisons (B, D, and E) or unpaired two-tailed t test (G and H).

A return to the homeostatic baseline by BTLAhi cDC1s

A prolonged absence of tolerogenic BTLAhi(CCR7lo) cDC1s and their functions could lead to an aberrant immuno-activation. Therefore, we extended our analysis of the populations of cDCs to later time points following the administration of different doses of LPS. The population of splenic BTLAhi cDC1s remained depressed for about 3 days (Figures 6A and S6A). An overall similar kinetic was also observed in the case of the relevant BTLAhiCCR7lo cDC1 population in the LNs (Figure S6B). However, by 6 days after the initial exposure to LPS, the BTLAhi cDC1 population rebounded, and within about 12 days, it fully stabilized at its initial baseline (Figures 6A, 6B, and S6A–S6C). Therefore, the loss of tolerogenic BTLAhi cDC1s following the acute pro-inflammatory stimulus is transient, accentuating the constant biological turnover of cDCs in vivo from their BM precursors (Kamath et al., 2000; Liu et al., 2007). Further stressing this established compensatory replenishment of cDCs and a return to homeostatic conditions, the expression of TNFR1 in the newly replenished BTLAhi cDC1 population resembled that of the original population prior to the LPS exposure (Figures 6C and 6D). Consistent with a resetting of the sensitivity to the innate stimulus, the newly replenished BTLAhi cDC1 population could be ablated by a subsequent challenge with LPS (Figure 6E). Overall, we conclude that, following the ablation of tolerogenic BTLAhi cDC1s under pro-inflammatory conditions, the return to homeostatic conditions involves a re-establishment of this crucial population in the lymphoid organs and resetting their ability to respond to subsequent pro-inflammatory challenges.

Figure 6.

Return to the homeostatic baseline by BTLAhi cDC1s

(A) Wild-type mice were treated with PBS or LPS (10 μg/mouse) at the indicated time points before analysis of splenocytes by flow cytometry. Graph shows percentages of BTLAhi cDC1s among all cDCs as a percentage of the baseline (average of PBS-treated mice for each time point) at the indicated time points (n = 4–7 mice per group pooled from two independent experiments per time point). Significance is shown for the comparison of PBS and LPS groups at each time point.

(B) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 12 days before analysis of splenocytes by flow cytometry. Plots (representative of multiple independent experiments) show anti-XCR1 and anti-BTLA staining intensity in all cDCs. Numbers next to outlined regions indicate corresponding percentages. Graph shows percentages of BTLAhi cDC1s among all cDCs in the indicated groups (n = 4 mice per group pooled from two independent experiments).

(C and D) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 9 days before analysis of splenocytes by flow cytometry.

(C) Overlaid histograms (representative of two independent experiments) show anti-TNFR1 staining intensity of the indicated populations (gated as in Figure 5A) from PBS-treated mice and 9 days post-LPS treatment. Corresponding graphs show MFI of anti-TNFR1 staining in the indicated populations for the indicated treatments (n = 4–5 mice per group pooled from two independent experiments).

(D) Overlaid histogram (representative of two independent experiments) shows anti-TNFR1 staining intensity of BTLAhi cDC1s from PBS-treated mice and 9 days post-LPS treatment as indicated. Corresponding graph shows MFI of anti-TNFR1 staining in BTLAhi cDC1s for the indicated treatments (n = 4–5 mice per group pooled from two independent experiments).

(E) Wild-type mice were treated with PBS or LPS (10 μg/mouse) 8 days before re-challenge with either PBS or LPS (10 μg/mouse). Splenocytes were analyzed by flow cytometry 1 day later. (Left) General experimental outline is shown. (Right) Graph shows percentages of BTLAhi cells among cDC1s in the indicated groups (n = 3 mice per group).

(A) Graph shows mean ± SEM. (B–E) Graphs show mean ± SD. ***p < 0.001 and ****p < 0.0001 determined by two-way ANOVA with Šídák’s multiple comparisons (A), one-way ANOVA with Tukey’s multiple comparisons (C), or unpaired two-tailed t test (B, D, and E).

DISCUSSION

The pro-immunogenic maturation of cDCs abrogates their tolerogenic functions, but how this happened was unclear (Cabeza-Cabrerizo et al., 2021; Dalod et al., 2014; Iberg et al., 2017; Iwasaki and Medzhitov, 2015; Mellman, 2013; Steinman et al., 2003a). Our studies now expand the concept of the maturation of cDCs resulting from the disturbance of steady-state conditions by the introduction of PAMPs, such as LPS, as a process that crucially involves dynamic changes that reshape the composition of the cDC population. We revealed that the inherently tolerogenic cDCs are physically deleted from the peripheral lymphoid organs at the onset of the pro-inflammatory conditions induced by LPS. Therefore, maturation of cDCs emerges as a two-pronged process during which immunogenic activation of some cDCs is accompanied by a rapid ablation of the inherently tolerogenic cDCs to prevent de novo induction of pTreg cells. Further, and accentuating the dynamic role of biological turnover of cDCs in vivo, the loss of tolerogenic cDCs is transient, therefore avoiding potential long-term risks due to insufficient immunoregulation.

Our results clarify the functions of TNF-α as a key regulator of the fate of tolerogenic cDCs. BMDCs or monocyte-derived DCs (moDCs) cultured in vitro can acquire tolerogenic therapeutic potentials that could possibly be enhanced by TNF-α (Iberg and Hawiger, 2020; Menges et al., 2002; Zhou et al., 2014). However, despite potential translational significance, these results offer only limited insights into the biological characteristics of cDCs found in vivo because the BMDCs differ substantially from endogenous cDCs (Helft et al., 2015). By focusing on bona fide cDCs found in vivo, we instead revealed that, during LPS-induced maturation, TNF-α mediates the death of BTLAhi cDC1s, resulting in a loss of tolerogenic functions. These effects depend on the TNF-α produced in vivo and can be prevented by a specific blockade of TNF-α. Consistent with such effects of TNF-α, BTLAhi cDC1s are key producers of TNF-α in response to an innate signal. Both the production of TNF-α and the ablation of BTLAhi cDC1s in response to LPS depend on intact TLR4 signaling in cDCs. However, the autocrine production of TNF-α is not essential for these specific outcomes, and instead, the TNF-α-mediated ablation is also extended in trans to those BTLAhi cDC1s that could not sense the original innate signal. A comprehensive ablation of the tolerogenic population is robustly propagated even if only a minor proportion of the BTLAhi cDC1s can respond to the initial innate stimulus. Further stressing the specific ablation of the BTLAhi cDC1s mediated by TNF-α, the expression of the death-inducing TNFR1 is highest in the BTLAhi cDC1s among all other cDCs. Therefore, by ablating the BTLAhi cDC1 population independently of the specific innate signals induced in the individual cDCs, functions of TNF-α and TNFR1 amplify the effects of the original innate signal to effectively extinguish the dominant capacity for the de novo conversion of antigen-specific pTreg cells during an acute immunogenic maturation process.

Similar to some other cytokines associated with maturation, the analogous “in trans” functions of TNF-α mediated by its different receptors can increase the activation and survival of some pro-immunogenic cDCs (Bachus et al., 2019; Bardou et al., 2021; Brunner et al., 2000; Le Bon et al., 2001; Maney et al., 2014; Sallusto and Lanzavecchia, 1994; Serbina et al., 2003; Sporri and Reis e Sousa, 2005; Sundquist and Wick, 2005; Trevejo et al., 2001; Winzler et al., 1997). Further, production of TNF-α by cDCs in the context of immune responses was proposed to depend on type I interferon and the relevant functions of its receptor IFNAR (Schaupp et al., 2020). Our findings now revealed that, in response to LPS, TNF-α is produced independently of IFNAR signaling and rapidly amplifies the ablation of tolerogenic BTLAhi cDC1s. Further underscoring these direct functions of TNF-α in mediating the ablation of BTLAhi cDC1s, a loss of BTLAhi cDC1s can be recapitulated by treatment with recombinant TNF-α in vivo in the absence of a specific exposure to innate signals.

Overall, by revealing a rapid death of BTLAhi cDC1s, our studies may also help to reconcile the results of the earlier observations of the LPS-induced death of cDCs and the most recent results that uncovered mechanisms to specifically prevent the inflammasome-mediated pyroptotic death of maturing cDCs with immunogenic functions (De Smedt et al., 1996; McDaniel et al., 2020; Qiu et al., 2009; Sundquist and Wick, 2009; Xu et al., 2017; Zhivaki et al., 2020). We propose that, whereas tolerogenic BTLAhi cDC1s are promptly and specifically ablated via the TNF-α-dependent mechanism, the remaining activated (mature) BTLAlo cDC1s and cDC2s are left to prime immune responses. Because cDC1s are well established to coordinate anti-viral and anti-tumor immune responses (Durai and Murphy, 2016; Eickhoff et al., 2015; Ferris et al., 2020), future studies will clarify the functions of BTLAlo cDC1s under such specific conditions in vivo.

The population of BTLAlo cDC1s that are preserved under pro-inflammatory conditions includes CCR7+ cDC1s, possibly further consistent with the pre-existing specific pro-immunogenic programs in some CCR7+ cDC1s (Ardouin et al., 2016). CCR7 governs migration of cDCs (Ohl et al., 2004). Therefore, CCR7hi cDCs are considered to have migratory properties, and these cDCs play multiple roles in the initiation of effector responses under pro-inflammatory conditions as well as in the orchestration of tolerance, particularly that limiting anti-tumor responses (Bajana et al., 2012; GeurtsvanKessel et al., 2008; Ghiringhelli et al., 2005; Harimoto et al., 2013; Jenkins et al., 2021; Kim et al., 2010; Krishnaswamy et al., 2018; Maier et al., 2020; Scheinecker et al., 2002; Wang et al., 2019). Migratory cDCs transport antigens from parenchymal tissues to the lymph nodes in the steady state (Randolph et al., 2008) and, in the case of some respiratory infections, also to the spleen (Jenkins et al., 2021; Sichien et al., 2017). However, systemic soluble antigens are also delivered directly to cDCs present in the lymphoid organs, such as to spleen via the bloodstream (Bourque and Hawiger, 2022; Eisenbarth, 2019). Particularly, pTreg cells are readily induced in response to antigens that can be delivered either directly or indirectly to lymphoid organs (Iberg et al., 2017; Idoyaga et al., 2013; Jones et al., 2016). The de novo induction of such antigen-specific pTreg cells is compromised in vivo in the absence of specific tolerogenic functions of BTLAhi cDC1s (Iberg et al., 2017; Jones et al., 2016). We now found that BTLAhi cDC1s have a low expression of CCR7, and in the steady state, these BTLAhi(CCR7lo) cells constitute the major population of cDC1s within the spleen, and in the LNs, they correspond to previously characterized cells with “resident” characteristic (Crozat et al., 2011; Idoyaga et al., 2013). However, it may be difficult to unequivocally functionally distinguish migratory and resident systemic cDC populations in vivo. Most crucially, upon a specific pro-inflammatory activation, the BTLAhi cDCs die rapidly in order to limit the mechanisms of tolerance. Therefore, we propose that a high expression of BTLA and the preferential ablation of tolerogenic BTLAhi cDC1s transcend “migratory” versus resident identities of the systemic cDCs that in the steady state induce pTreg cells in the peripheral lymphoid organs.

After they complete the process of T cell priming, individual mature cDCs may undergo programmed death and be eventually replaced by precursors forming in the BM (Chen et al., 2006; Stranges et al., 2007; Winzler et al., 1997). In contrast, our findings uncovered an acute ablation of the entire population of tolerogenic BTLAhi cDC1s immediately after the initiation of an immunogenic maturation process. Such a rapid ablation of tolerogenic cDCs results in a loss of the de novo induction of antigen-specific pTreg cells, consistent with an abrogation of tolerogenic mechanisms during an acute immunogenic maturation (Cabeza-Cabrerizo et al., 2021; Dalod et al., 2014; Iberg and Hawiger, 2020; Iwasaki and Medzhitov, 2015; Mellman, 2013; Steinman et al., 2003a). However, the long-term absence of tolerogenic cDCs and their capacity for the de novo induction of dominant mechanisms of tolerance dependent on antigen-specific pTreg cells would pose a risk for insufficient immunoregulation. Therefore, our results help to clarify the biological benefits of the rapid cDC turnover in vivo by revealing that the tolerogenic BTLAhi cDC1 population is promptly re-established within the lymphoid organs and remains sensitive to a subsequent pro-inflammatory stimulus. Overall, by revealing an ablation of tolerogenic cDCs and the role of TNF-α in this process, the current studies are directly relevant for the design of new vaccines and immunotherapies based on specific abrogation or enhancement of tolerogenic mechanisms as clinically desired.

Limitations of the study

In our experiments, we have not explored the corresponding mechanisms in other tolerogenic cDCs, such as those present at the anatomical barriers, and instead focused on the systemic cDCs relevant for various autoimmune and immune responses. Our studies utilized LPS, an archetypical pro-inflammatory stimulus. The mechanisms involving other adjuvants, specific infections, and other pro-inflammatory states will require future studies.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Daniel Hawiger (daniel.hawiger@health.slu.edu).

Materials availability

This study did not generate new, unique reagents.

All data reported in this paper will be shared by the lead contact upon request.

This paper does not report original code.

Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Mice

All mouse strains were available on a C57BL/6J background. Btla−/− (Sedy et al., 2005), Casp3−/− (Kuida et al., 1996), Casp8fl/fl (Beisner et al., 2005), Cd19-Cre (CD19-Cre) (Rickert et al., 1997), B6.MRL-Faslpr/J (Faslpr) (Watanabe-Fukunaga et al., 1992), Foxp3RFP (Wan and Flavell, 2005), Ifnar1−/− (Hayashi et al., 2002; MÜller et al., 1994; Prigge et al., 2015), Itgax-Cre (CD11c-Cre) (Caton et al., 2007), Lyz2-Cre (LysM-Cre) (Clausen et al., 1999), Myd88fl/fl (Hou et al., 2008), OTII TCR tg (Barnden et al., 1998), Ripk3−/− (JAX stock #025738), Tlr4−/− (Hayashi et al., 2002; McAlees et al., 2015), Tlr4fl/fl (McAlees et al., 2015), Tnfrsf1a−/− (Peschon et al., 1998), and Ticam1−/− (Trif−/−) (Hoebe et al., 2003) mice were previously described and available from Jackson Laboratory. OTII TCR tg (Barnden et al., 1998) mice were crossed with Foxp3RFP (Wan and Flavell, 2005) reporter mice to derive OTII Foxp3RFP mice. Ripk3−/− (JAX stock #025738), Casp3−/− (Kuida et al., 1996), Casp8fl/fl (Beisner et al., 2005), and Itgax-Cre (CD11c-Cre) (Caton et al., 2007) mice were crossed to derive Ripk3−/−Casp3−/−Casp8ΔCD11c mice. Ticam1−/− (Trif−/−) (Hoebe et al., 2003) mice were crossed with Myd88fl/fl (Hou et al., 2008) and Itgax-Cre (CD11c-Cre) (Caton et al., 2007) mice to derive Trif−/−Myd88ΔCD11c mice. Tlr4fl/fl (McAlees et al., 2015) mice were crossed with Itgax-Cre (CD11c-Cre) (Caton et al., 2007), Lyz2-Cre (LysM-Cre) (Clausen et al., 1999), or Cd19-Cre (CD19-Cre) (Rickert et al., 1997) mice to derive Tlr4ΔCD11c, Tlr4ΔLysM, and Tlr4ΔCD19 mice, respectively. Tlr4−/− (Hayashi et al., 2002; McAlees et al., 2015) mice were crossed with Tnfrsf1a−/− (Peschon et al., 1998) mice to derive Tlr4−/−Tnfrsf1a−/− mice. Some C57BL/6J WT mice were also bred onto congenic B6.SJL-Ptprca Pepcb/BoyJ (CD45.1) background. 6- to 9-week-old sex- (both male and female) and age-matched mice were used for experiments. Mice were randomly assigned to experimental groups. All mice were maintained in our facility under specific pathogen free conditions and used in accordance with guidelines of the Saint Louis University Institutional Animal Care and Use Committee.

METHOD DETAILS

cDC preparation

Unless otherwise noted, spleens and peripheral (axial, brachial, and inguinal) lymph nodes were isolated and analyzed separately. To isolate cDCs, spleens and lymph nodes were dissected, individually mechanically fragmented with forceps and 30-gauge needles (BD), and treated with 2.5mg/mL Collagenase D (Roche) in RPMI 1640 media (Hyclone) supplemented with 5% fetal bovine serum (FBS) (Gemini Bio), Penicillin/Streptomycin (100U/mL), HEPES (10mM), Sodium Pyruvate (1mM), and 2-Mercaptoethanol (55μM) (all Gibco) at 37°C for 37 minutes, followed by incubation with EDTA (10mM) for 5 min at 37°C. After incubation, cells were passed through 100μm strainers (VWR) and washed using Hanks’ Balanced Salt Solution (Gibco) supplemented with 2% FBS and 1mM EDTA to obtain single cell suspensions.

Flow cytometry

For analysis of T cells, spleens and lymph nodes were dissected and passed through 70μm strainers (Fisherbrand) to obtain single cell suspensions. cDCs were processed as above. Following lysis of red blood cells (RBCs) with ammonium chloride solution (0.16M) for 5 minutes at room temperature, single cell suspensions were washed with Dulbecco’s phosphate buffered saline (DPBS; referred to as “PBS” throughout the text) (Hyclone) supplemented with 5% FBS and stained with Zombie Aqua Live/Dead viability dye (BioLegend) for 15 minutes at room temperature. Cells were then stained with Fc block (anti-CD16/32, clone 2.4G2; purified in-house using Protein G Sepharose Beads (GE Healthcare) from supernatants of corresponding hybridomas (ATCC) that were concentrated with ammonium sulfate) for 15 minutes at room temperature. For surface staining, cells were then incubated with fluorochrome-conjugated antibodies diluted in PBS supplemented with 2% FBS for 30 minutes on ice. After incubation, cells were washed twice and strained through 35μm strainers into FACS tubes (BD) for acquisition. To conduct staining for surface expression of CCR7, cells were stained with anti-CCR7 fluorochrome-conjugated antibody (BioLegend) diluted in PBS supplemented with 2% FBS for 30 minutes at 37°C prior to incubation with other surface marker antibodies. All samples were acquired on LSRII or BDFortessa (BD) instruments. Data was analyzed using FlowJo 10 software (FlowJo, LLC).

Intracellular cytokine staining

To conduct intracellular staining for TNF-α, spleens were processed and stained for surface markers as described above. Cells were then fixed and permeabilized using Cytofix-Cytoperm buffers (BD) according to the manufacturer’s protocol and stained for intracellular TNF-α using a fluorochrome-conjugated anti-TNF-α antibody (BioLegend) for 30 minutes at 4°C.

Bone marrow chimeras

Wild-type recipient mice were placed on Sulfamethoxazole/Trimethoprim (SulfaTrim) antibiotic water for approximately one week prior to irradiation. Mice were lethally irradiated (10 Gy) approximately 6 hours prior to bone marrow (BM) transfer. Congenically-labeled donor BM was obtained from indicated groups of mice. Briefly, donor mice were sacrificed, and BM was flushed from the femurs and tibias with a 30-gauge needle using PBS supplemented with 5% FBS. After red blood cell lysis, cells were washed twice in PBS and counted. Cells from mice of the indicated genotypes were mixed together at the indicated ratios, and 10–12 million total cells were injected intravenously into the tail vein of each recipient mouse. Recipient mice were kept on SulfaTrim antibiotic water for about 2 weeks post-BM transfer. 5 weeks after BM transfer, chimerism was verified by flow cytometry of blood. Chimera mice were utilized for experiments at least 6 weeks after BM transfer.

In vivo TLR stimulation

Ultrapure Lipopolysaccharide (LPS) from Escherichia coli 0111:B4 (InvivoGen) was diluted in PBS and injected intraperitoneally into each mouse at defined doses and at specific time points as indicated in corresponding figure legends.

In vivo TNF-α stimulation

Carrier-free recombinant mouse TNF-α (Biolegend) was diluted in PBS and injected intraperitoneally into each mouse at 4.5 μg/mouse.

In vivo TNF-α blockade

250 μg blocking anti-TNF-α antibody (clone XT3.11) or 250 μg rat IgG1 isotype control (anti-horseradish peroxidase; clone HRPN) (both BioXcell) was injected intraperitoneally into each mouse 3.5 hours prior to administration of LPS or PBS. For experiments analyzed 48 hours after LPS administration, mice were injected a second time with 250 μg anti-TNFα antibody or isotype control antibody 24 hours after LPS administration.

FACS-based TUNEL assay

The APO-BrdU TUNEL kit (BD) was used to detect DNA strand breaks. cDCs were prepared as described above prior to enrichment by negative magnetic streptavidin microbead selection (Miltenyi) to exclude T cells (CD3ε), B cells (B220), and NK cells (CD49b) (biotinylated antibodies from BioLegend). Cells were then stained for cell surface markers as described above prior to TUNEL analysis according to the manufacturer’s protocol.

T Cell adoptive transfers

Spleens and peripheral lymph nodes of OT-II Foxp3RFP donor mice were pooled together and enriched by depletion. CD4+ T cells were enriched by negative magnetic streptavidin microbead selection (Miltenyi), which excluded cells expressing the surface molecules CD8α, B220, CD11b, CD11c, and CD49b (biotinylated antibodies from BioLegend). Enriched Foxp3(RFP)neg CD25neg cells were sorted on FACSAria Fusion or FACSAriaIII (BD) instruments. After sorting, cells were washed twice in PBS and counted. 3–5 million cells were injected intravenously into the tail vein of each recipient mouse.

In vivo delivery of antigens to cDCs

OT-II TCR tg T cells were activated in vivo by cognate OVA323–339 antigen that was delivered by anti-DEC-OVA, a recombinant chimeric antibody specific for DEC-205 (Hawiger et al., 2010; Iberg and Hawiger, 2019). Chimeric antibodies were expressed in Expi293 cells (ThermoFisher) as previously described (Fang et al., 2017). Cellular supernatants were concentrated using a Vivaflow 50R cassette (Sartorius), and antibodies were purified using Protein G Sepharose Beads (GE Healthcare) as described in (Jones et al., 2016). Chimeric antibodies were injected intraperitoneally at 125 ng/mouse in PBS as previously established (Jones et al., 2016).

Genomic DNA analysis

To check for germline deletion of Tlr4 in Tlr4ΔCD11c mice, genomic DNA was extracted from ear biopsies of mice and analyzed by PCR using the following primers: Tlr4-F: 5′-TGACCACCCATATTGCCTATAC-3′ and Tlr4-R: 5′-TGATGGTGTGAGCAGGAGAG-3′. Products were run on a 2% agarose gel and imaged.

QUANTIFICATION AND STATISTICAL ANALYSES

Sex- and age-matched mice of specified genotypes were randomly assigned into individual experimental groups. Data are presented as mean ± standard error of the mean (SEM) or mean ± standard deviation (SD). Group sizes were determined based on the results of preliminary experiments. No statistical method was used to predetermine sample size. p values were calculated in Prism 9 (GraphPad Software) using unpaired two-tailed t tests, one-way ANOVAs with Tukey’s multiple comparisons, or two-way ANOVAs with Šídák’s multiple comparisons, as indicated in corresponding figure legends. Differences were considered to be statistically significant when p % 0.05.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Anti-Mouse CD3ε, Clone: 145-2C11	BioLegend	Cat#:100308; RRID: AB_312673
Anti-Mouse CD3ε, Clone: 145-2C11	BioLegend	Cat#:100310; RRID: AB_312675
Anti-Mouse CD3ε, Clone: 145-2C11	BioLegend	Cat#: 100304; RRID: AB_312669
Anti-Mouse CD4, Clone GK1.5	BioLegend	Cat#: 100447; RRID: AB_2564586
Anti-Mouse CD8α, Clone: 53-6.7	BioLegend	Cat#:100704; RRID: AB_312743
Anti-Mouse CD11b, Clone: M1/70	BioLegend	Cat#:101204; RRID: AB_312787
Anti-Mouse CD11c, Clone: N418	BioLegend	Cat#:117304; RRID: AB_313773
Anti-Mouse CD11c, Clone: N418	BioLegend	Cat#:117336; RRID: AB_2565268
Anti-Mouse CD120a / TNFR Type I/p55, Clone: 55R-286	BioLegend	Cat#:113003; RRID: AB_313532
Anti-Mouse CD16/32, Clone: 2.4G2	ATCC	Produced in-house from hybridomas obtained from ATCC (Cat#: HB-197™; RRID: CVCL_9148)
Anti-Mouse CD172a, Clone: P84	BioLegend	Cat#:144010; RRID: AB_2563548
Anti-Mouse CD19, Clone: 6D5	BioLegend	Cat#:115508; RRID: AB_313643
Anti-Mouse CD19, Clone: 6D5	BioLegend	Cat#:115510; RRID: AB_313645
Anti-Mouse CD197 (CCR7), Clone: 4B12	BioLegend	Cat#:120105; RRID: AB_389357
Anti-Mouse CD197 (CCR7), Clone: 4B12	BioLegend	Cat#:120110; RRID: AB_492841
Anti-Mouse CD25, Clone: PC61	BioLegend	Cat#:102030; RRID: AB_893288
Anti-Mouse CD272 (BTLA), Clone: 6A6	BioLegend	Cat#:139106; RRID: AB_10613297
Anti-Mouse CD45, Clone: 30-F11	BioLegend	Cat#: 103128; RRID: AB_493715
Anti-Mouse CD45R/B220, Clone: RA3-6B2	BioLegend	Cat#:103208; RRID: AB_312993
Anti-Mouse CD45R/B220, Clone: RA3-6B2	BioLegend	Cat#:103210; RRID: AB_312995
Anti-Mouse CD45R/B220, Clone: RA3-6B2	BioLegend	Cat#:103204; RRID: AB_312989
Anti-Mouse CD45.1, Clone: A20	BioLegend	Cat#:110714; RRID: AB_313503
Anti-Mouse CD45.1, Clone: A20	BioLegend	Cat#:110706; RRID: AB_313495
Anti-Mouse CD45.1, Clone: A20	BioLegend	Cat#: 110722; RRID: AB_492866
Anti-Mouse CD45.2, Clone: 104	BioLegend	Cat#:109830; RRID: AB_1186098
Anti-Mouse CD45.2, Clone: 104	BioLegend	Cat#:109814; RRID: AB_389211
Anti-Mouse CD49b, Clone: DX5	BioLegend	Cat#:108904; RRID: AB_313411
Anti-Mouse CD64, Clone: X54-5/7.1	BioLegend	Cat#:139323; RRID: AB_2629778
Anti-Mouse CD86, Clone GL-1	BioLegend	Cat#: 105007; RRID: AB_313150
Anti-Mouse F4/80, Clone: BM8	BioLegend	Cat#:123118; RRID: AB_893477
Anti-Mouse I-Ab, Clone: AF6-120.1	BioLegend	Cat#:116422; RRID: AB_10613473
Anti-Mouse NK1.1, Clone: PK136	BioLegend	Cat#:108707; RRID: AB_313394
Anti-Mouse NK1.1, Clone: PK136	BioLegend	Cat#:108716; RRID: AB_493590
Anti-Mouse TCR Vα2, Clone: B20.1	BioLegend	Cat#: 127824; RRID: AB_2814019
Anti-Mouse TCR Vβ5.1/5.2, Clone: MR9-4	BioLegend	Cat#:139507; RRID: AB_2566020
Anti-Mouse TNF-α, Clone: MP6-XT22	BioLegend	Cat#: 506307; RRID: AB_315428
Anti-Mouse TNF-α, Clone: MP6-XT22	BioLegend	Cat#:506305; RRID: AB_315426
Anti-Mouse XCR1, Clone: ZET	BioLegend	Cat#:148220; RRID: AB_2566410
Armenian Hamster IgG Isotype Ctrl Antibody, Clone HTK888	Biolegend	Cat#: 400924
InVivoMab Anti-Horseradish Peroxidase (Rat IgG1 Isotype Control), Clone: HRPN	BioXcell	Cat#:BE0088; RRID: AB_1107775
InVivoMab Anti-Mouse TNF-α, Clone: XT3.11	BioXcell	Cat#:BE0058; RRID: AB_1107764
Chemicals, peptides, and recombinant proteins
2-Mercaptoethanol	Gibco	Cat#: 21985-023
Dulbecco’s Phosphate Buffered Saline	Hyclone	Cat#: SH30028FS
Hanks’ Balanced Salt Solution	Gibco	Cat#: 14175095
HEPES	Gibco	Cat#: 15630-080
Penicillin-Streptomycin	Gibco	Cat#: 15140-122
RPMI 1640 medium	Hyclone	Cat#: SH30027FS
Sodium Pyruvate	Gibco	Cat#: 11360-070
Recombinant Mouse TNF-α (carrier-free)	Biolegend	Cat#: 575204
Critical commercial assays
APO-BrdU™ TUNEL Kit	BD	Cat#: 556405
BD Cytofix/Cytoperm™ Fixation/Permeabilization Kit	BD	Cat#: 554714
Zombie Aqua™ Fixable Viability Kit	BioLegend	Cat#: 423102
Experimental models: Cell lines
Cell Line: Expi293F™	ThermoFisher	Cat#: A14527; RRID: CVCL_D615
Experimental models: Organisms/strains
Mouse: C57BL/6J	Jackson Laboratory	Stock# 000664
Mouse: B6.SJL-Ptprca Pepcb/BoyJ	Jackson Laboratory	Stock# 002014
Mouse: Btla−/−	Jackson Laboratory	Stock# 006353; Sedy et al., 2005
Mouse: Casp3−/−	Jackson Laboratory	Stock# 006233; Kuida et al., 1996
Mouse: Casp8fl/fl	Jackson Laboratory	Stock# 027002; Beisner et al., 2005
Mouse: CD19-Cre	Jackson Laboratory	Stock# 006785; Rickert et al., 1997
Mouse: B6.MRL-Faslpr/J	Jackson Laboratory	Stock# 000482; Watanabe-Fukunaga et al., 1992
Mouse: Foxp3RFP reporter	Jackson Laboratory	Stock# 008374; Wan and Flavell, 2005
Mouse: Ifnar1−/−	Jackson Laboratory	Stock# 032045-JAX; Müller et al., 1994
Mouse: Ifnar1−/−	Jackson Laboratory	Stock# 028288; Hayashi et al., 2002; Prigge et al., 2015
Mouse: Itgax-Cre	Jackson Laboratory	Stock# 008068; Caton et al., 2007
Mouse: LysM-Cre	Jackson Laboratory	Stock# 004781; Clausen et al., 1999
Mouse: Myd88fl/fl	Jackson Laboratory	Stock# 008888; Hou et al., 2008
Mouse: OTII TCR tg	Jackson Laboratory	Stock# 004194; Barnden et al., 1998
Mouse: Ripk3−/−	Jackson Laboratory	Stock# 025738; Newton et al., 2004
Mouse: Ticam1−/− (Trif−/−)	Jackson Laboratory	Stock# 005037; Hoebe et al., 2003
Mouse: Tlr4−/−	Jackson Laboratory	Stock# 029015; Hayashi et al., 2002; McAlees et al., 2015
Mouse: Tlr4fl/fl	Jackson Laboratory	Stock# 024872; McAlees et al., 2015
Mouse: Tnfrsf1a−/−	Jackson Laboratory	Stock# 003242; Peschon et al., 1998
Software and algorithms
FlowJo 10	FlowJo, LLC	https://www.flowjo.com; RRID: SCR_008520
GraphPad Prism	GraphPad Software	https://www.graphpad.com; RRID: SCR_002798
Other
BenchMark Fetal Bovine Serum	Gemini Bio-Products	Cat#: 100-106
Roche Collagenase D	MilliporeSigma	Cat#: 11088882001
Ultrapure LPS, E. coli 0111:B4 (LPS-EB Ultrapure)	InvivoGen	Cat#: tlrl-3pelps
Protein G Sepharose Beads	GE Healthcare	Cat#: 17061801
Streptavidin MicroBeads	Miltenyi	Cat#: 130-048-101
LS Columns	Miltenyi	Cat#: 130-042-401