Migration of Lung Resident Group 2 Innate Lymphoid Cells Link Allergic Lung Inflammation and Liver Immunity

Abstract

Group 2 innate lymphoid cells (ILC2s) are tissue resident in the lung and activated by inhaled allergens via epithelial-derived alarmins including IL-33. Activated ILC2s proliferate, produce IL-5 and IL-13, and induce eosinophilic inflammation. Here, we report that intranasal IL-33 or the protease allergen papain administration resulted in increased numbers of ILC2s not only in the lung but also in peripheral blood and liver. Analyses of IL-33 treated parabiosis mice showed that the increase in lung ILC2s was due to proliferation of lung resident ILC2s, whereas the increase in liver ILC2s was due to the migration of activated lung ILC2s. Lung-derived ILC2s induced eosinophilic hepatitis and expression of fibrosis-related genes. Intranasal IL-33 pre-treatment also attenuated concanavalin A-induced acute hepatitis and cirrhosis. These results suggest that activated lung resident ILC2s emigrate from the lung, circulate, settle in the liver and promote type 2 inflammation and attenuate type 1 inflammation.

Keywords: allergic lung inflammation; group 2 innate lymphoid cells; liver inflammation; migration; parabiosis; tissue resident.

Figure 1

I.n. papain or IL-33 administration causes an increase in ILC2 numbers in the lung, PB and liver. (A) Mice received four daily i.n papain injections and were analyzed on day 6. (B) ILC2s were quantified in the lung, PB and liver. Black bar=naïve, white bar = papain treated. (C) Mice received three daily i.n injections of IL-33 and were analyzed at various time points. d = day, w = week, m = month. (D–G) The number of ILC2s in the lung (D), PB (E), spleen (F) and liver (G) after i.n IL-33 administrations. Day 0 is naïve. Asterisks indicate significant differences from naïve. (H) Mice received three daily i.n injections of IL-33 and ILC2s were quantified in the small intestine (SI) and mLN on day 7 [see (C)]. (I) Mice received three daily i.n. injections of IL-33 and the amount of IL-33 was measured in lung and liver homogenate and serum on day 3 [see (C)]. Black bar = naïve, white bar = IL-33 treated, ND = not detected (H, I). Data shown are mean ± SEM. n = 5-10 (B), n = 4-48 (D–G), n = 6 (H), n = 4 (I). Unpaired two-tailed t test (B, H, I) or Kruskal-Wallis test (D–G) was used to determine statistical significance. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, n.s, not significant (P < 0.05).

Figure 2

ILC2s migrate from the lung to the liver upon i.n. IL-33 treatment. B6 and Pep3b mice were conjoined by parabiosis surgery. B6 mice were given three daily i.n. PBS or IL-33 injections and both B6 and Pep3b mice were analyzed on day 7 (see Figure 1C ). CD45.1⁺ and CD45.2⁺ ILC2s were quantified in the lung (A), PB (B), spleen (C), liver (D), BM (E) and SI (F). X-axes show the strains of mice analyzed (B6 or Pep3b) and treatment groups (PBS or IL-33). IL-33 highlighted in red indicates the mouse that received IL-33 injections. Numbers within graphs indicate the percentages of CD45.2⁺ ILC2s. Black and red asterisks indicate statistical significance of percentages and cell numbers, respectively, of CD45.2⁺ (within blue bars) and CD45.1⁺ (within orange bars) ILC2s compared to PBS treated pairs. Data shown are mean ± SEM. n = 6-7 (except SI, where n = 3-4). Two-way ANOVA with Bonferroni correction was used to determine statistical significance. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 3

Liver ILC2s phenotypically differ from lung ILC2s. (A, B) Mice received three daily i.n IL-33 injections and CXCR6 (A) and CD103 (B) expression was analyzed before and 3 days after IL-33 treatment (day 5 in Figure 1C ). Histograms show representative expression of each marker. Numbers show percentages of CXCR6⁺ (A) or CD103⁺ (B) ILC2s. Histograms: dotted lines/numbers at the bottom = naïve, solid lines/numbers on top = IL-33 treated, shaded = isotype. (C) Mice were administered with i.n. IL-33 on day 0-2 and i.p. FTY720 on day 0-4, and ILC2s were quantified in the lung, liver, mediastinal LN, and PB on day 5. (D) ILC2s were quantified in WT (black bars) and Rorc ^-/- (white bars) lung, liver and PB before and 5 days after three daily i.n. IL-33 injections (day 7 in Figure 1C ). Data shown are mean ± SEM. n = 6-15 (A), n = 5-23 (B), n = 6 (C), n = 5-19 (D). Unpaired two-tailed t test (A–C) or two-way ANOVA with Bonferroni correction (D) was used to determine statistical significance. *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001.

Figure 4

Lung-derived liver ILC2s produce more IL-5, IL-13 and IL-6 than lung ILC2s. (A) Mice received three daily i.n IL-33 injections and lung and liver were analyzed for intracellular cytokine expressions at various time points as shown in Figure 1C . The plots show IL-5 and IL-13 expressions in the lung (top) and liver (bottom) ILC2s at three selected time points. Boxes indicate IL-5⁺IL-13⁺ ILC2s and the numbers show mean ± SEM percentages of IL-5⁺IL-13⁺ ILC2s. The graph shows the IL-5⁺IL-13⁺ lung (filled circle) and liver (cross) ILC2 percentages. d = day, w = week, m = month. Asterisks indicate significant differences compared to naïve. (B–D) Mice were treated as above and ILC2s were purified on day 14 (see Figure 1C ) and cultured for 72 hours. The amounts of IL-5 [(B), top], IL-13 [(B), bottom], amphiregulin (Areg) (C) and various cytokines (D) in conditions indicated below graphs (B, C), P/I [(D), top] or IL-33 + IL-25 + IL-7 + IL-2 [(D), bottom] conditions. P/I = PMA + ionomycin, “+” = present, “-” = absent. Black = lung, white = liver. (E, F) Lung and liver ILC2s were stained for intracellular IL-5, IL-6 and IL-13 before and 5 days after IL-33 treatment (day 7 in Figure 1C ). (E) Histograms (left) show IL-6 expression in IL-33 treated lung (dashed line/number at the bottom) and liver (solid line/number on top) ILC2s. Shaded = isotype. Graphs (right) show IL-6⁺ ILC2 percentages in naïve (black bars) and IL-33 treated (white bars) lung and liver. (F) IL-5, IL-6 and IL-13 expressions in the lung (top) and liver (bottom) ILC2s in naïve (left) and IL-33 treated (right) mice. Numbers indicate percentages of ILC2s in cytokine positive quadrants. Data shown are mean ± SEM. n = 3-22 (A), n = 9-15 (except n = 3-5 for IL-25/IL-33 + TSLP/IL-7 and IL-2/IL-7 only conditions) (B, C), n = 7 (D), n = 6 (E, F). One-way (A) or two-way (E) ANOVA with Bonferroni correction, or unpaired two-tailed t test (B–D) was used to determine statistical significance. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 5

Lung-derived liver ILC2s induce type 2 skewing in the liver. (A–C) WT mice received three daily injections of i.n. IL-33 and the livers were analyzed for various lymphocytes (A), myeloid populations (B), and CD206⁺ macrophages (C) 5 days later (day 7 in Figure 1C ). (D) ILC2 numbers in the lung and liver of naïve WT (black bars) and CD127 cKO (white bars) mice. (E) WT and CD127 cKO mice were treated as above and liver ILC2s were quantified. (F, G) CD127 cKO mice were treated as above and the livers were analyzed for myeloid cells (F) and CD206⁺ macrophages (G). (H, I) WT (H) and CD127 cKO (I) mice were treated as above and the livers were collected 3 days later (day 5 in Figure 1C ) and the expression of indicated genes was analyzed. Black bars = naïve, white bars = IL-33 treated except (D) Numbers in plots show percentages of CD206⁺ macrophages (C, G). Tregs = regulatory T cells, cDC1 = conventional type 1 dendritic cells (CD103⁺), cDC2 = conventional type 2 DCs (CD11b⁺), Mo = monocytes. Data shown are mean ± SEM. n = 6-17 (A–C), n = 17-21 (D), n = 6 (E), n = 6-18 (F), n = 9-10 (G), n = 5-11 (H), n = 5 (I). Unpaired two-tailed t test (A–D, F–I) or two-way ANOVA with Bonferroni correction (E) was used to determine statistical significance. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 6

Repeated i.n. IL-33 injections cause eosinophilic inflammation and expression of fibrosis-related genes in the liver. (A) WT mice received three daily i.n. IL-33 treatment and challenged with another set of IL-33 injections a month later. Mice were analyzed three days after the last injection. (B–F) ILC2s (B) and IL-5⁺, IL-6⁺ and IL-13⁺ ILC2s (C) were quantified in the lung and liver, and various myeloid populations (D) and CD206⁺ macrophages (E) were quantified in the liver. The livers were analyzed for expression of indicated genes by qPCR (F). (G) H&E staining of -IL33/-IL33 (left, scale bar = 100 µm) and +IL33/+IL33 treated liver at low (middle, scale bar = 100 µm) and high (right, scale bar = 20 µm) magnifications. (H) The eosinophilic clusters were quantified in -IL33/-IL33 and +IL33/+IL33 treated liver sections. (I–N) CD127 cKO mice were treated as in (A). ILC2s (I), various myeloid populations (J). CD206⁺ macrophages (K) and expression of indicated genes (L) were quantified in the liver. (M) H&E staining of +IL33/+IL33 treated liver at low (left, scale bar = 100 µm) and high (right, scale bar = 20 µm) magnifications. (N) The eosinophilic clusters were quantified in -IL33/-IL33 and +IL33/+IL33 treated liver sections. “+/-” indicates presence or absence of the first and second treatments, respectively. Data shown are mean ± SEM. n = 6-12 (B–F), n = 5-6 (G, H), n = 5 (I–N). Two-way ANOVA with Bonferroni correction (B–F, I–L) or unpaired two-tailed t test (H, N) was used to determine statistical significance. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 7

I.n. IL-33 treatment attenuates ConA-induced hepatitis. (A) WT mice received three daily i.n. IL-33 treatment and i.v. injected with ConA one month later. Mice were analyzed 1 week after ConA treatment. (B) Mice were weighed daily after ConA injections. Black, blue and red asterisks indicate statistically significant differences between -IL33/-ConA vs. -IL33/+ConA, +IL33/-ConA vs. +IL33/+ConA, and -IL33/+ConA vs. +IL33/+ConA, respectively. ILC2s (C), myeloid cells (D), CD206⁺ macrophages (E) and transcripts of indicated genes (F) were quantified in the liver. (G) Picrosirius red staining of -IL33/+ConA (red) and +IL33/+ConA (black) treated livers at low (left, scale bar = 200 µm) and high (right, scale bar = 50 µm) magnifications. (H) Picrosirius red+ area was quantified in -IL33/+ConA (red circles) and +IL33/+ConA (black squares) treated livers. (I–P) CD127 cKO mice were treated and analyzed as described in (A) The percentage body weights of -IL33/+ConA treated WT (circles) and cKO (squares) mice 24 hours after ConA treatment was calculated with respect to the weights at the time of ConA injections (I). Mice were weighed daily after ConA injections (J). ILC2s (K), myeloid cells (L), CD206⁺ macrophages (M) and transcripts of indicated genes (N) were quantified in the liver. (O) Picrosirius red staining of -IL33/+ConA (red) and +IL33/+ConA (black) treated livers at low (left, scale bar = 200 µm) and high (right, scale bar = 50 µm) magnifications. (P) Picrosirius red+ area was quantified in -IL33/+ConA (red circles) and +IL33/+ConA (black squares) livers. “+/-” indicates presence or absence of i.n. IL-33 and i.v. ConA treatments, respectively. Data shown are mean ± SEM. n = 4-8 (B–H), n = 6-8 (I–P). Two-way ANOVA with Bonferroni correction (B–F, K–N) or unpaired two-tailed t test (H–J, P) was used to determine statistical significance. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.