Tissue residency of innate lymphoid cells in lymphoid and non-lymphoid organs

Abstract

Innate lymphoid cells (ILC) contribute to barrier immunity, tissue homeostasis, and immune regulation at various anatomical sites throughout the body. How ILCs maintain their presence in lymphoid and peripheral tissues is currently unknown. We found that in the lymphoid and non-lymphoid organs of adult mice, ILC are tissue-resident cells that were maintained and expanded locally under physiologic conditions, upon systemic perturbation of immune homeostasis, and during acute helminth infection. However, at later time points post-infection, cells from hematogenous sources helped to partially replenish the pool of resident ILCs. Thus, ILC are maintained by self-renewal in broadly different microenvironments and physiological settings. Such an extreme “sedentary” lifestyle is consistent with the proposed roles of ILCs as sentinels and local keepers of tissue function.

Tissue-resident leukocytes can be divided by their cellular origin and means of maintenance into self-renewing cells that seed non-lymphoid organs during ontogeny, or cells that are replenished hematogenously from precursors in the bone marrow (BM) or secondary lymphoid organs (SLO) such as the spleen and lymph nodes (LN). Innate lymphoid cells (ILC) have been identified in embryonic tissues, BM, SLO, peripheral blood, and many non-lymphoid tissues, including mucosal sites like the lung and small intestine where they contribute to tissue immunosurveillance, immunoregulation, repair, and homeostasis (1, 2). How ILC populations in lymphoid and non-lymphoid organs are maintained, and whether ILCs recirculate from lymphoid to non-lymphoid tissues remain controversial. Based on the identification of fetal ILC progenitors that seed the mouse intestine (3) and the proposed development of ILC in human tonsils (4), it is reasonable to expect that ILC may self-renew locally or be generated from immature precursors in these tissues. However, recent studies have identified progenitors to all currently known ILC subsets in adult BM, raising the possibility that ILCs in lymphoid and non-lymphoid organs are continuously replenished through medullary lymphopoiesis and subsequent recruitment to peripheral tissues (5, 6).

To directly test whether hematogenous precursors continuously replenish the pool of peripheral tissue ILCs in adult mice, we generated parabiotic mice, which establish blood chimerism through joint circulation (7, 8). Congenically marked CD45.1⁺ and CD45.2⁺ mice were surgically connected for 30–40 days until complete chimerism (~50:50 CD45.1⁺/CD45.2⁺ ratio) of major lymphocyte populations was established in the peripheral blood and spleens of parabionts (fig. S1). We analyzed the percentages of cells that derived from the donor or host parabiont for currently known populations of “helper-like” ILCs, including Lin⁻RORγt⁻Eomes⁻NK1.1⁺ ILC1, Lin⁻RORγt⁻GATA-3⁺ ILC2, Lin⁻RORγt⁺CD4⁻ ILC3, and Lin⁻RORγt⁺CD4⁺ lymphoid tissue inducer (LTi) cells (fig. S2). We found that >95% of ILC1, ILC2, ILC3, and LTi cells residing in the small intestine lamina propria, salivary gland, lung, and liver (9) were of host origin, indicating that these ILC types are bona fide tissue-resident cells (Fig. 1a–c and fig. S3). One exception to the overwhelming tissue residency of ILCs were Eomes⁻NK1.1⁺ ILC1 in the lung and peripheral blood, which originated from both parabionts (fig. S3). Because cells isolated from enzymatically dissociated lung tissue include cells from both the vasculature and the parenchyma (10), we injected parabiotic mice intravenously with fluorescently labeled anti-CD45 antibody 5 min before sacrifice to identify intravascular leukocytes. This revealed that virtually all ILC2, but only ~20% of Eomes⁻NK1.1⁺ cells, localized to the lung parenchyma (Fig. 1d,e and fig. S3). In contrast to intravascular ILC1 which derived evenly from both parabionts (Fig. 1e and fig. S3), extravascular ILC1 were >90% host-derived. This observation suggested that these cells represent lung parenchymal ILC1 that do not continuously exchange with, and are distinguishable from their intravascular counterparts, in agreement with the observation that Eomes⁻NK1.1⁺ ILC1 in the small intestine, salivary gland, and liver were also overwhelmingly of host origin. Together, these experiments identified all analyzed ILC types as tissue-resident cells in all examined non-lymphoid organs.

Figure 1. Innate lymphoid cells are tissue-resident cells.

Congenically marked CD45.1⁺ and CD45.2⁺ mice underwent parabiosis surgery and were then analyzed on d30–40 (A–E, J) or d104–130 of parabiosis (F–I, K). The percentage of cells derived from the host parabiont was determined for all “helper-like” ILCs including Lin⁻RORγt⁻Eomes⁻NK1.1⁺ ILC1, Lin⁻RORγt⁻GATA-3⁺ ILC2, Lin⁻RORγt⁺CD4⁻ ILC3, and Lin⁻RORγt⁺CD4⁺ LTi cells; Lin⁻RORγt⁻Eomes⁺NK1.1⁺ conventional NK cells (cNK); as well as CD4⁺ and CD8⁺ T cells in the small intestine lamina propria (A, F, G), salivary gland (B, H) lung (C, I) and mesenteric LN (J, K). Intravascular leukocytes were stained by anti-CD45-PE administered intravenously 5 min prior to the isolation and analysis of lung leukocytes on d40 of parabiosis (D, E). Each symbol represents an individual parabiont. The red dotted line at 50% marks complete chimerism (equal contribution from host and donor parabionts). na = not analyzed. The data are shown as mean ± SD and represent 2 or 3 independent experiments (n = 4–8).

Next, we explored whether the long-term maintenance of tissue-resident ILC was accomplished through self-renewal, or through progressive replacement by hematogenous precursors. We analyzed mice which had been in parabiosis for >3 months and found that all analyzed ILC subsets remained >95% host-derived in all non-lymphoid organs tested (Fig. 1f–i and fig. S4). In contrast, conventional natural killer (NK) and T lymphocytes, which in our analysis of d30–40 parabiosis appeared relatively tissue-resident in the small intestine, had equilibrated with donor-derived cells at 3 months (Fig. 1f,g). However, over time, we did detect a minor increase from <2% to ~5% donor-derived cells for ILC2 in the small intestine and ILC1 in the salivary gland, which was not observed for the other ILC subsets (fig. S4). Therefore, while our data are consistent with a minor contribution of hematogenous precursors (5, 6) or circulating mature ILCs (11) to the physiologic renewal of ILC subsets, the majority of ILCs appear to be tissue-resident cells that are locally maintained in peripheral organs. This was in contrast to, for example, the progressive replacement of tissue-resident embryonic macrophages in some adult tissues, such as the intestine, which experience constant immune stimulation and tissue renewal (12).

Our observation of tissue-resident ILC in non-lymphoid tissues raised the question of whether ILCs in SLO are replenished from hematogenous sources, or whether they also represent locally renewing cells. Unexpectedly, our analysis of mesenteric LNs and spleens of parabiotic mice showed that all ILC subsets tested were >95% host-derived at both 1 and 3 months (Fig. 1j,k and fig. S5), demonstrating that ILCs establish tissue-residency in both lymphoid and non-lymphoid organs. Although ILCs that leave peripheral tissues via afferent lymphatics can contribute to ILC populations in draining LNs (13), the identification of tissue-resident ILCs in the spleen, which does not recruit cells via afferent lymphatics, further supports the idea of ILC residency in SLO. The regional maintenance of ILCs in SLO thus distinguishes “helper-like” ILCs from developmentally related NK cells (“killer ILCs”) which, like αβ T cells, are continuously replaced by BM-derived precursors and systemically recirculate through SLO. Nevertheless, some Eomes⁺ cNK cells can also establish tissue residency, e.g. in the salivary gland and small intestine (Fig. 1)(9, 14).

To test whether tissue residency of ILC could be maintained under inflammatory conditions, we generated parabiotic Foxp3^DTR mice. In these mice, Foxp3⁺ regulatory T (Treg) cells express the human diphtheria toxin receptor (DTR) under the control of the endogenous Foxp3 locus (15). Upon administration of diphtheria toxin (DTX), the depletion of Treg cells initiates a systemic lymphoproliferative syndrome and fatal autoimmunity (15, 16). Surprisingly, we failed to detect any contribution of hematogenous cells to the expanded ILC populations in the analyzed organs of parabiotic mice depleted of Treg cells (Fig. 2 and fig. S6). This suggests that the mobilization of ILCs or ILC precursors into peripheral tissues is negligible even during systemic autoimmunity, and that the associated increase in ILC populations in lymphoid and non-lymphoid organs likely results from local expansion despite the influx of hematogenously derived myeloid cells and adaptive lymphocytes (15, 16).

Figure 2. Tissue-residency of ILCs is maintained upon systemic immune activation.

Congenically marked pairs of Foxp3^DTR or wild-type (WT) control mice underwent parabiosis surgery. After 40 days, both mice within Foxp3^DTR or WT pairs were subjected to diphtheria toxin (DTX) treatment. For the indicated populations of lymphocytes, the percentage of host-derived cells was analyzed in the small intestine lamina propria (A) and mesenteric lymph node (B) on d9 DTX. Each symbol represents an individual parabiont. The data are shown as mean ± SD and were pooled from 2 independent experiments (n = 6).

Next, we examined whether local ILC expansion is observed during infection with a helminth, Nippostrongylus brasiliensis (N.b.) (17), that induces strong proliferation and activation of ILC2 in the lung and small intestine. Recent work suggests that N.b. infection triggers the appearance of “inflammatory” ILC2 that differentiate further to “natural” ILC2 to sustain the expansion of ILC2 necessary for worm expulsion (18). To determine whether the ILC response in this setting is due to the mobilization of inflammatory ILC from the BM or elsewhere, or to local differentiation and expansion, we infected parabiotic mice with N.b. and analyzed ILC2 during the acute phase of helminth infection (d7). At this time point, ILC2 in the lung, small intestine, and mesenteric LN remained host-derived, despite their robust proliferation and expansion (Fig. 3a–d). In contrast, we observed a small but statistically significant increase in donor-derived ILC2 during the chronic inflammation phase on day 15 post-infection (Fig. 3c,d) (20, 21). At this time point, ILC2 continued to exhibit increased proliferation (Fig. 3c) and have been shown to serve critical functions in tissue repair (22). Thus, the local expansion of resident ILCs during the acute phase of N.b. infection is followed by moderately increased hematogenous recruitment or redistribution of ILCs during the chronic inflammation and repair phase. The increase in donor-derived ILC2 was specific among innate immune cell types, because we could not detect any increase in donor-derived ILC3 or tissue-resident alveolar macrophages (Fig. 3e and fig. S7). However, even at this later stage, >90% of ILC2 originated from the host parabiont.

Figure 3. Local expansion of tissue-resident ILC2 upon helminth infection.

Congenically marked CD45.1⁺ and CD45.2⁺ mice underwent parabiosis surgery. After 40 days, both mice within each parabiotic pair were infected with Nippostrongylus brasiliensis (N.b.) or mock-infected. The percentages of Ki-67⁺ (A, C) and donor-derived (B, D) ILC2 were analyzed in the lung, small intestine lamina propria, and mesenteric lymph node on d7 and d15 post-infection (pi). (E) Analysis of ILC2 and ILC3 in the small intestine lamina propria on d15 pi. Each symbol represents an individual parabiont. The data are shown as mean ± SD and represent 2–3 independent experiments (n = 6–12).

The local proliferation of tissue-resident ILCs led us to ask whether these cells can access the cytokine interleukin (IL)-2, whose availability in lymphoid organs is controlled by Treg cells (23), but whose physiological relevance to ILC responses in non-lymphoid tissues remains controversial. For this purpose, we generated mixed BM chimeras harboring both wild-type ILC, and ILC deficient for CD25, the high-affinity α-chain of the IL-2 receptor (Fig. 4a) as previously described (24). We found that CD25-sufficient and -deficient ILC2 proliferated to a similar extent and were found at equal ratios in N.b.-infected and mock-infected chimeric mice, indicating that IL-2 did not directly influence the proliferation of ILC2 in response to N.b. infection (Fig. 4b,c). However, we observed that CD25-sufficient ILC2 produced more IL-13 (Fig. 4d), which is required for worm expulsion. These results demonstrate that ILC2 can indeed access physiological levels of IL-2 in non-lymphoid organs and support the notion that one function of locally produced IL-2 may be to modulate ILC effector function (23, 25, 26). We have previously suggested competition for IL-2 as one mechanism by which Foxp3⁺ Treg cells restrain the homeostasis of splenic ILC1 (24). It remained unclear whether Treg cells are able to restrain IL-2–dependent ILC activation in non-lymphoid tissues, where they form co-clusters with ILC2 (27). Transient depletion of Treg cells induced the IL-2 and CD25-dependent proliferation of ILC2 (Fig. 4e–g), suggesting that ILC2 can directly access IL-2 in the lung and the small intestine, and that Treg cells restrain the IL-2 dependent expansion of these tissue-resident ILCs.

Figure 4. Interleukin-2 acts directly on tissue-resident ILC2 to promote cytokine production.

(A–D, G) Mixed chimeras were generated by co-transfer of bone marrow from CD45.2⁺ Il2ra^−/− mice (CD25^−/−) and CD25-sufficient CD45.1⁺ Foxp3^DTR mice (CD25^+/+) into irradiated RAG-γc^KO mice. (B–D) Mixed chimeras were infected with Nippostrongylus brasiliensis, and lung and small intestine lamina propria were analyzed on d7 post-infection. (B) Percentage of Ki-67⁺ ILC2 among CD25^+/+ and CD25^−/− cells. (C) Ratios of CD25^−/− to CD25^+/+ ILC2. (D) Intracellular IL-13 staining of cells stimulated with PMA/ionomycin. (E–G) Foxp3^DTR mice were subjected to diphtheria toxin (DTX) or mock treatment, and additionally received IL-2 neutralizing antibodies JES6-1A and S4B6-1, or isotype control IgG, as indicated. On d9 of DTX, the absolute numbers of ILC2 per lung (E) and the percentages of Ki-67⁺ lung ILC2 (F) were analyzed. (G) Ratios of CD25^−/− to CD25^+/+ lung and small intestine lamina propria lymphocytes were analyzed in mixed chimeras on d9 of DTX. The data are shown as mean ± SD and represent 2–3 independent experiments (n = 4–8).

In conclusion, our experiments identify ILCs in both lymphoid and non-lymphoid organs as tissue-resident cells that are locally renewed and expanded in response to acute environmental challenges. These findings also suggest that observed ILC plasticity, exemplified by the differentiation from iILC2 to nILC2 (18), ILC3 to RORγt⁻Eomes⁻NK1.1⁺ “ex-ILC3” (28), the polarization of ILC1 towards ILC3 (29), and the differentiation of CD4⁺ LTi cells (30), are local processes occurring within peripheral organs. Our studies further demonstrate that the local pool of resident ILCs can be replenished and complemented, albeit only in part, through contributions of hematogenously-derived precursors or mature cells in situations of extended inflammation and tissue repair. Interestingly, ILC subsets are elevated in the peripheral blood of patients suffering from chronic inflammatory psoriasis (32, 33). Peripheral blood ILC2 have been shown to dynamically modulate expression of molecules that regulate tissue homing in mice and humans (21, 31). In addition, we have detected donor-derived lymphoid and ILC progenitors in parabiotic BM (fig. S8), raising the possibility that ILC progenitors can physiologically seed tissues not only during embryonic development (3), but also in adult mice. It remains to be determined whether this observation reflects the physiologic migration of ILC progenitors or the engraftment of donor-derived hematopoietic stem cells (34, 35) giving rise to ILCs. Independent of these considerations, our data support a model in which ILCs are locally maintained and expanded as tissue-resident cells during homeostasis and acute infection. The “sedentary” lifestyle of ILCs in broadly differing microenvironments is consistent with the proposed roles of ILCs as sentinels and local keepers of tissue function.

Material and Methods

Animals

Foxp3^DTR mice have been previously described (15). C57BL/6J mice were purchased from the Jackson Laboratories. All the mice were bred and housed in the specific-pathogen-free animal facility at the Memorial Sloan-Kettering Cancer Center and used in accordance with institutional guidelines.

Parabiosis

Female 6–8 week old congenic CD45.1 and CD45.2 mice were surgically connected in parabiosis as previously described (7,8). After corresponding lateral skin incisions were made from elbow to knee in each mouse, forelimbs and hindlimbs were tied together using nylon suture, and the skin incisions were closed using stainless steel wound clips. After surgery, mice were maintained on a diet supplemented with trimethoprimsulfamethoxazole for prophylaxis of infection.

Cell depletions and IL-2 blockade

To deplete regulatory T cells, Foxp3^DTR mice were injected i.p. with 50µg/kg diphtheria toxin (DT; Sigma Aldrich) every other day (q2d). Chimeras received 15µg/kg DT q2d. To block IL-2 in vivo, mice received 200µg each of JES6-1A12 and S4B6-1 anti-IL-2 antibodies (BioXcell) q2d i.p.

Helminth infection

Mice were injected subcutaneously (s.c.) with 500 L3 Nippostrongylus brasiliensis (Nb) larvae cultured as previously described (17). Briefly, Nb were maintained by passage in 4 to 8 week-old BALB/cJ mice. Mice were injected s.c. with 500–1000 L3 Nb and stool was collected on days 6–8 post infection. Mice were killed on day 8 post infection and intestinal and cecal contents were collected as well. Stool and intestinal contents were mixed with water and a 1:1 mixture of sphagnum moss peat and activated charcoal in petri dishes and cultured at 25°C for up to 4 weeks in a chamber lined with wet paper towels to maintain humidity. Dishes were kept moist with dH₂O and mold-free with 0.5% Nystatin as necessary. L3 Nb were collected as needed using a Baermann apparatus after at least 7 days of culture. L3 larvae were extensively washed with dH₂O to eliminate culture contaminants before infection. Injections were carried out using a 23G needle and a concentration of 500 L3 Nb in 200uL.

Isolation of mononuclear cells

Peripheral blood was drawn from the right ventricle before mice were perfused with PBS. Spleens and lymph nodes were dissociated using ground glass slides and filtered through a 100-µm strainer. To isolate lamina propria lymphocytes (LPL), small intestines were removed and flushed, and Peyer’s patches were excised. The intestines were then opened up lengthwise and sliced transversely into 0.5 cm-long pieces, which were incubated in 1 mM dithiothreitol and 1 mM EDTA for 15 minutes to remove intraepithelial lymphocytes (IEL). After IEL were removed, the pieces of intestine were incubated in digest solution (1 mg/mL collagenase and 1 U/mL DNase I in RPMI supplemented with 5% fetal calf serum, 1% L-glutamine, 1% penicillin-streptomycin, and 10 mM HEPES) for 40 minutes, passed through a 100-µm strainer, and purified using 35% Percoll. To isolate mononuclear cells from salivary glands, lungs, and liver, the tissues were physically dissociated using scissors and incubated for 45 minutes in digest solution before being passed through 100-µm strainers. Liver cells were further purified using a discontinuous gradient of 44% over 67% Percoll. Red blood cells in spleen, blood, liver, and lung were lysed using ACK lysis buffer.

Analysis of phenotype and cytokine production

Cell suspensions were prepared and incubated with Fc block (2.4G2) and fixable live/dead dye (Life Technologies), and were subsequently stained with fluorophore-conjugated antibodies purchased from eBioscience, BioLegend, BD Biosciences, or Tonbo: anti-CD3ε (145-2C11), anti-CD5 (53-7.3), anti-CD19 (1D3), anti-CD11b (M1/70), anti-NK1.1 (PK136), anti-TCRβ (H57-597), anti-TCRγδ (eBioGL3), anti-FcεR1 (MAR-1), anti-CD4 (RM4-5), anti-CD8α (53-6.7), anti-CD49b (DX5), anti-Thy-1.2 (30-H12), anti- CD45.1 (A20), anti-CD45.2 (104), anti-CD45 (30-F11), anti-IL-13 (eBio13A), anti-RORγt (Q31-378), anti-GATA-3 (TWAJ), anti-EOMES (Dan11mag). Lin⁺ cells were stained with a cocktail of antibodies against CD3ε, CD5, CD8α, CD19, TCRγδ, TCRβ and FcεR1. For intracellular cytokine staining, cells were stimulated with 50ng/ml PMA and 500ng/ml ionomycin and incubated for 3h in the presence of 1µg/ml Brefeldin A (Sigma-Aldrich). Intracellular staining of IL-13, Ki-67, Foxp3, GATA-3, RORγt and Eomes was performed using the Foxp3 mouse Treg cell staining buffers (Tonbo). Cells were analyzed on a LSRII cytometer (BD Biosciences) and analyzed with FlowJo software (Treestar).

Statistical analysis

All statistical analyses were performed using GraphPad Prism 6 software. Results are expressed as means +/− standard deviation. Differences between individual groups were analyzed for statistical significance using the unpaired Student’s t-test or one-way ANOVA followed by Tukey’s multiple comparison test (Fig. 3C,D and 4E,F). *=p<0.05, **=p<0.01, ***=p<0.001, ns= not significant.