Damage-induced IL-18 stimulates thymic NK cells limiting endogenous tissue regeneration

Abstract

Interleukin-18 (IL-18) is an acute-phase proinflammatory molecule crucial for mediating viral clearance by activating T helper 1 CD4⁺ T cells, cytotoxic CD8⁺ T cells and natural killer (NK) cells. Here, we show that mature IL-18 is generated in the thymus following numerous distinct forms of tissue damage, all of which cause caspase-1-mediated immunogenic cell death. We report that IL-18-stimulated cytotoxic NK cells limit endogenous thymic regeneration, a critical process that ensures the restoration of immune competence after acute insults such as stress, infection, chemotherapy and radiation. NK cells suppress thymus recovery by aberrantly targeting thymic epithelial cells, which act as the master regulators of organ function and regeneration. Together, our data reveal a new pathway regulating tissue regeneration in the thymus and suggest IL-18 as a potential therapeutic target to boost thymic function. Moreover, given the enthusiasm for IL-18 as a cancer immunotherapy due to its capacity to elicit a type 1 immune response, these findings also offer insight into potential off-target effects.

Fig. 1. Acute thymic damage triggers the cleavage of Cas-1 and the activation of IL-18, which suppresses thymus regeneration.

a,b, Female 1- to 2-month-old C57/BL6 mice were administered SL-TBI (550 cGy), dexamethasone (intraperitoneal (i.p.) injection, 20 mg kg⁻¹), cyclophosphamide (i.p., 200 mg kg⁻¹) or LPS (i.p., 1.5 mg kg⁻¹). a, Thymus cellularity (black) and cl-Cas-1 expression (red) were measured using fluorescently conjugated FAM-YVAD-FMK (a fluorescent probe that irreversibly binds and labels cl-Cas-1) in mice killed at baseline (n = 15), day 0.5 (n = 7), day 1 (n = 8), day 3 (n = 8), day 5 (n = 4) and day 7 (n = 4) after treatment; all statistics are compared to day 0. b, Amount of active IL-1β and active IL-18 in the thymus, measured by ELISA at the indicated time points after SL-TBI (IL-18: day 0, n = 9; day 0.5, n = 6; day 1, n = 6; day 3, n = 5; IL-1β: day 0, n = 4; day 0.5, n = 3; day 1, n = 3; day 3, n = 3), dexamethasone (i.p., 20 mg kg⁻¹) (IL-18: day 0, n = 9; day 0.5, n = 3; day 1, n = 7; day 3, n = 6; IL-1β: day 0, n = 4; day 0.5, n = 3; day 1, n = 3; day 3, n = 3), cyclophosphamide (i.p., 200 mg kg⁻¹) (IL-18: day 0, n = 5; day 1, n = 4; day 3, n = 4; IL-1β: day 0, n = 4; day 0.5, n = 3; day 1, n = 2; day 3, n = 3) or LPS (i.p., 1.5 mg kg⁻¹) (IL-18: day 0, n = 9; day 0.5, n = 6; day 1, n = 6; day 3, n = 6; IL-1β: day 0, n = 3; day 0.5, n = 4; day 1, n = 3; day 3, n = 3); all statistics are compared to day 0. c, Amount of active IL-18 measured by ELISA in thymuses of female 1- to 2-month-old Cas1^Δ10 mice on day 0 (n = 7) and day 1 (n = 8) after SL-TBI. d, Amount of IL-18BP in thymuses of 1- to 2-month-old C57/BL6 WT mice on days 0, 1 and 3 after SL-TBI (n = 3 per group). e, Ratio of active IL-18 to IL-18BP averaged on day 0 (n = 9), day 1 (n = 6) and day 3 (n = 5) after SL-TBI, representing the amount of free active IL-18. f, Female 1- to 2-month-old C57/BL6 WT (n = 18), Il1r1^−/− (n = 3), Il18^−/− (n = 7) and Il18r1^−/− (n = 8) mice were exposed to SL-TBI, and thymus cellularity was measured 7 days later. g, Female 1- to 2-month-old C57/BL6 WT (n = 7) or Cas1^Δ10 (n = 8) mice were exposed to SL-TBI, and thymus cellularity was measured on day 7. h, Female 1- to 2-month-old C57/BL6 WT mice were exposed to SL-TBI and then administered PBS vehicle (n = 12) or rIL-18 (n = 10) on day 3 (subcutaneous (s.c.) injection, 2.5 mg kg⁻¹); thymuses were isolated on day 7. i, Female 1- to 2-month-old C57BL/6 mice were lethally irradiated and transplanted (intravenous (i.v.) injection) with 5 × 10⁶ CD45.1⁺ WT bone marrow hematopoietic cells. Recipient mice were treated with 200 μg of anti-IL-18 mAb (n = 10) or equal-volume control (PBS) (n = 11), and thymus cellularity was measured on day 50 following transplant. Graphs represent mean ± s.e.m.; each dot represents an individual biological replicate; NS, not significant. Statistics were generated for a, b and d–f using one-way analysis of variance (ANOVA) with Dunnet’s correction for multiple comparisons and for c and g–i using unpaired two-tailed t tests. Panel a icons created with BioRender.com. Source data

Fig. 2. Hematopoietic and nonhematopoietic sources of IL-18 after damage.

a,b, scRNAseq was performed on (1) nonthymocyte CD45⁺ stromal cells (CD45⁺ Rag2-GFP⁻ cells isolated from female 1- to 2-month-old Rag2^GFP mice) and (2) CD45⁻ stromal cells isolated from thymuses of female 1- to 2-month-old C57BL/6 mice at baseline and on days 1, 4 and 7 following SL-TBI. Data were previously integrated and published in ref. . a, Integrated UMAP of both hematopoietic and nonhematopoietic cells from datasets, showing undamaged cells, major clusters in the thymus at baseline and annotation. b, Expression of Il18 by population. artEC, arterial endothelial cell; capEC, capillary endothelial cell; venEC, venous endothelial cell; capsFB, capsular fibroblast; intFB, intermediate fibroblast; medFB, medullary fibroblast; vSMC/PC, vascular smooth muscle/pericyte; cTEC, cortical TEC; mTEC1 and mTEC2, medullary TECs; cDC, classical dendritic cell; pDC, plasmacytoid dendritic cell; Mac, macrophage; Eos, eosinophil; B, B cell; NK/ILC1, NK and type 1 ILC; ILC2, type 2 ILC; ILC3, type 3 ILC; γδT, γδ T cell; NKT, NK T cell; T_reg, regulatory T cell; Thy, thymocyte. Colors represent unbiased clusters. c, cl-Cas-1 expression measured using fluorescently conjugated FAM-YVAD-FMK in cDC1s, macrophages, cTECs, mTECs, endothelial cells, fibroblasts and MECs on days 0, 0.5, 1 and 3 after SL-TBI (n = 3–5 per group). Gating and phenotypes can be found in Extended Data Fig. 3. d, Amount of active IL-18 measured by ELISA in female 1- to 2-month-old Il18^fl/fl:Zbtb-Cre⁻ (Il18^WT, n = 15 per group) and Il18^fl/fl:Zbtb-Cre⁺ (Il18^ΔcDC; day 0, n = 10; day 1, n = 13) mice on day 0 or 1 following SL-TBI. e, Female 1- to 2-month-old C57/BL6 WT (WT→WT) or Il18^−/− (WT→Il18^−/−) mice were lethally irradiated (2 × 550 cGy) and transplanted with 5 × 10⁶ CD45.1⁺ WT bone marrow hematopoietic cells. At 10 weeks after transplantation, recipient mice were administered a second dose of SL-TBI (550 cGy), and active IL-18 was measured at baseline and on day 1 after this subsequent damage (WT→WT: n = 6 per group; WT→Il18^−/−: day 0, n = 5; day 1, n = 7). Graphs represent mean ± s.e.m.; each dot represents an individual biological replicate. Statistics were generated for d and e using one-way ANOVA with Tukey’s correction for multiple comparisons. Source data

Fig. 3. IL-18 suppression of thymus function after damage is not mediated directly through TECs or hematopoietic progenitors.

a, Standard scaled dot plot of Il18r1 and Il18rap gene expression by population of cells from thymuses of female 1- to 2-month-old C57BL/6 mice at baseline, taken from the scRNAseq dataset described in Fig. 2a. nmSC, nonmyelinating Schwann cells; mTEC^prol, proliferating mTECs. b, Concatenated flow cytometry plots showing the expression of IL-18R in CD45⁺NK1.1⁺TCRβ⁺ CD1d-αGalCer tetramer⁺ (NKT1) and CD1d-αGalCer tetramer⁻ (NKT2) invariant NK cells, CD45⁺NK1.1⁺TCRβ⁻CD49b⁺ NK cells and CD45⁺NK1.1⁺TCRβ⁻CD49a⁺ ILC1s (n = 5 per group). Gates were based on expression in Il18r1^−/− mice. c, Percentage of IL-18R-expressing cTECs, mTECs, fibroblasts, endothelial cells, other CD45⁻ cells, early thymic progenitors (ETP), thymocytes (DN1–4, double-positive (DP), and single-positive CD4⁺ (SP4) and CD8⁺ (SP8) cells), T_reg cells, γδ T cells, NK cells (n = 9), ILC1s (n = 4), ILC2s (n = 4), ILC3s (n = 4), cDC1s, cDC2s and macrophages (n = 5 per group unless otherwise specified). d, Female 1- to 2-month-old Il18r1^fl/fl:Foxn1-Cre⁻ (Il18r1^WT, n = 11) and Il18r1^fl/fl:Foxn1-Cre⁺ (Il18r1^ΔTEC, n = 8) mice were exposed to SL-TBI, and thymus cellularity was assessed 7 days later. e,f, Bone marrow populations were measured for IL-18R expression (n = 6 per group), shown as flow cytometry plots (e) and percentage of positive cells (f). LSK, lineage (Lin)⁻Sca-1⁺c-Kit⁺ cells; LT-HSC, long-term hematopoietic stem cells; ST-HSC, short-term hematopoietic stem cells; MMP2–4, multipotent progenitors. g, Female 1- to 2-month-old WT CD45.1⁺ mice were lethally irradiated and transplanted (i.v.) with 2.5 × 10⁶ WT CD45.1⁺ bone marrow cells and 2.5 × 10⁶ bone marrow cells from either CD45.2⁺ WT or Il18r1^−/− mice. h, Contribution of CD45.2⁺ cells in the thymus at 2 weeks following transplant (n = 5 per group). i, Contribution of CD45.2⁺ cells to the total CD45⁺ cell (left) or T cell (right) reconstitution in peripheral blood over 17 weeks after transplantation (WT→WT n = 6; Il18r1^−/−→WT n = 8). j,k, At 17 weeks after transplantation, recipient mice were administered a subsequent dose of SL-TBI (550 cGy). Thymuses were collected after 7 days, and the percentage of CD45.2⁺ cells relative to all thymic CD45⁺ cells (j) and the total thymus cellularity (k) were measured (WT→WT: n = 6; Il18r1^−/−→WT: n = 8). Graphs represent mean ± s.e.m.; each dot represents an individual biological replicate. Statistics were generated for d, h, j and k using unpaired two-tailed t tests. Source data

Fig. 4. Damage-resistant IL-18R⁺ NK cells suppress thymus repair.

a, CellChat interaction analysis for IL-18 at baseline and following SL-TBI, taken from the scRNAseq dataset described in Fig. 2a, with quantification of the aggregate signal strength for each IL-18 target cell. b–d, Female 1- to 2-month-old C57BL/6 CD45.2⁺ mice were lethally irradiated and transplanted (i.v.) with 5 × 10⁶ WT CD45.1⁺ bone marrow cells. b, Concatenated flow cytometry plots showing CD45⁺CD45.1⁻CD4⁻CD8⁻ cells (top) and CD45⁺CD45.1⁻CD4⁻CD8⁻NK1.1⁺IL-18R⁺ cells gated on CD3⁺ NKT cells and CD49b⁺ NK cells (bottom) from thymus-recipient mice at the indicated time points after HCT (n = 4–7 per time point). c,d, Proportion (c) and total number (d) of recipient NK or NKT cells before HCT (day 0; n = 7) and on days 1, 3, 7 and 14 after HCT (n = 4 per group). e, Thymuses of female 1- to 2-month-old C57BL/6 mice were visualized at steady state or on day 3 or 7 after SL-TBI at 12×, assessing keratin-14-positive (Krt14⁺) mTECs (green), keratin-8-positive (Krt8⁺) cTECs (pink) and NKp46⁺ NK cells (arrows). The NKp46⁺ NK/ILC1 cell distribution within the thymus cortex or medulla at 0, 3 and 7 days after SL-TBI is shown (n = 3 per group). f, Female 1- to 2-month-old WT C57BL/6 mice were administered 200 μg of anti-NK1.1 mAb or control PBS (i.p.) on days −1, 1 and 3 after SL-TBI, and thymus cellularity was assessed on day 7 (n = 9 per group). g, Female 1- to 2-month-old C57BL/6 WT, Il18^−/− and Il18r1^−/− mice were administered 200 μg of anti-NK1.1 mAb or isotype/PBS (i.p.) as above. The relative change in thymus cellularity is shown, comparing control-treated (WT, n = 17; Il18^−/−, n = 9; Il18r1^−/−, n = 8) and anti-NK1.1 mAb-treated (WT, n = 21; Il18^−/−, n = 8; Il18r1^−/−, n = 9) mice within each strain 7 days after SL-TBI. h, Female 1- to 2-month-old C57BL/6 WT (Cd1d^+/+, n = 9) and Cd1d^−/− (n = 6) mice were exposed to SL-TBI, and thymus cellularity was measured 7 days later. i, Female 1- to 2-month-old Il18r1^fl/fl:Lck-Cre⁻ (Il18r1^WT, n = 5) and Il18r1^fl/fl:Lck-Cre⁺ (Il18r1^ΔT/NKT, n = 5) mice were exposed to SL-TBI and administered rIL-18 (s.c., 2.5 mg kg⁻¹) on day 3. Thymus cellularity was measured on day 7 after SL-TBI. j, Female 1- to 2-month-old Il18r1^fl/fl:Ncr1-Cre⁻ (Il18r1^WT, n = 6) and Il18r1^fl/fl:Ncr1-Cre⁺ (Il18r1^ΔNK/ILC1, n = 7) mice were exposed to SL-TBI and administered rIL-18 (s.c., 2.5 mg kg⁻¹) on day 3. Thymus cellularity was measured on day 7 after SL-TBI. Graphs represent mean ± s.e.m.; each dot represents an individual biological replicate. Statistics were generated for c and d using one-way ANOVA with Dunnet’s correction for multiple comparisons, for e using one-way ANOVA with Tukey’s correction for multiple comparisons, and for f–j using unpaired two-tailed t tests. Source data

Fig. 5. Acute insult activates thymic NK cells.

a, Normalized gene expression in NK/ILC1 or NKT cells of the cytotoxicity factors Ifng, Prf1, Gzma and Gzmb, as well as the activation markers Nkg7, Klrd1, Klrk1, Ncr1, Klrc2 and Klra4, on days 0, 1, 4 and 7 after SL-TBI, taken from the scRNAseq dataset described in Fig. 2a. b, Concatenated flow cytometry plots and corresponding geometric mean fluorescence intensity (gMFI) of CD45⁺NK1.1⁺TCRβ⁻ NK cell expression of Ifng-GFP (day 0, n = 4; day 3, n = 5), perforin (day 0, n = 6; day 3, n = 9) and granzyme B (GZMB; day 0, n = 4; day 3, n = 5) on days 0 and 3 following SL-TBI in female 1- to 2-month-old C57BL/6 WT or Ifng-reporter mice. c, Amount of thymic IFNγ (day 0, n = 8; day 3, n = 3), perforin (n = 8 per group) and granzyme B (n = 8 per group) measured by ELISA in female 1- to 2-month-old C57BL/6 mice on days 0 and 3 after SL-TBI. d, Thymuses were collected from female 1- to 2-month-old C57BL/6 mice on day 3 after SL-TBI. Concatenated flow cytometry plots gated on all CD45⁺ Ifng-GFP (left), CD45⁺perforin⁺ (middle) and CD45⁺GZMB⁺ (right) cells, as well as the total thymus cellularity of Ifng-GFP (left; day 0, n = 3; day 3, n = 4), perforin⁺ (middle; day 0, n = 3; day 3, n = 7) and GZMB⁺ (right; day 0, n = 3; day 3, n = 7) CD45⁺NK1.1⁺TCRβ⁻CD49b⁺ NK cells, CD45⁺NK1.1⁺TCRβ⁻CD49a⁺ ILC1s, CD45⁺NK1.1⁺TCRβ⁺CD49b⁻ NKT cells and CD45⁺TCRβ⁺NK1.1⁻ T cells, are shown. e, Gene expression heat map of thymic NK/ILC1 and NKT cells for Gzma, Gzmb, Prf1, Ifng, Nkg7, Klrk1, Ncr1, Klrd1 and Il18r1 at 1, 4 and 7 days following SL-TBI. Each column represents a cell, with the cells ordered based on the expression of Il18rap (in ascending order from left to right). The time after TBI is indicated by the colors at the bottom. f, NK1.1⁺IL-18R⁺TCRβ⁻CD49b⁺ NK cells from female 1- to 2-month-old C57BL/6 mice were purified using FACS at baseline (d0) or 2 days after SL-TBI (d2) and cocultured with CellTrace-labeled RMA-S target cells at a 2:1 effector-to-target ratio. RMA-S target cell Annexin V expression was measured 5 h after coculture, and cell death was assessed (n = 4 biological replicates per group, representative of three independent experiments). Dashed lines represent RMA-S alone (bottom) or the positive control (top). Graphs represent mean ± s.e.m.; each dot represents an individual biological replicate. Statistics were generated for b–d and f using unpaired two-tailed t tests. Source data

Fig. 6. IL-18 stimulation of cytotoxic NK cells suppresses thymus regeneration.

a,b, Female 1- to 2-month-old C57BL/6 WT or Ifng-reporter mice were exposed to SL-TBI, and thymuses were collected on days 0 and 3 after irradiation. IL-18R⁺ and IL-18R^lo–neg CD45⁺NK1.1⁺TCRβ⁻CD49b⁺ NK cells, CD45⁺NK1.1⁺TCRβ⁻CD49a⁺ ILC1s and CD45⁺NK1.1⁺TCRβ⁺CD49b⁻ NKT cells were compared. a, NK, ILC1 and NKT cellularity on day 0 (n = 3) and day 3 (n = 7) after SL-TBI. b, Fold change in NK cells (n = 10), ILC1s (n = 10) and NKT cells (n = 7) between days 0 and 3 after SL-TBI. c, Female C57BL/6 CD45.1⁺ and CD45.2⁺ mice were surgically conjoined to establish parabiotic pairs when both members of the pair were subjected to SL-TBI. Thymuses were collected, and chimerism was calculated on day 0 (n = 8), day 1 (n = 6), day 4 (n = 4) or day 7 (n = 8) after SL-TBI. Congenic markers (CD45.1 and CD45.2) were used to determine the mouse of origin (that is, cells expressing the same CD45 isoform as the mouse-pair thymus were classified as ‘intrathymically’ derived, while cells expressing the alternate isoform were considered ‘extrathymically’ derived). The numbers of intrathymic or extrathymic CD45⁺NK1.1⁺CD3⁻ NK/ILC1 cells at the indicated time points are shown graphically. d, Concatenated flow cytometry plots showing Ifng-GFP, perforin and granzyme B expression within CD45⁺NK1.1⁺TCRβ⁻CD49b⁺ NK cells. The gMFI of Ifng-GFP, perforin and granzyme B expression in IL-18R^lo–neg and IL-18R⁺ NK cells on day 3 after SL-TBI (n = 8 per group) is shown. e, Female 1- to 2-month-old C57BL/6 WT or Il18r1^−/− mice were exposed to SL-TBI, and thymuses were isolated 3 days later. Concatenated flow cytometry plots and gMFI of perforin expression within CD45⁺NK1.1⁺TCRβ⁻CD49b⁺ NK cells are shown (n = 5 per group). f, Female 1- to 2-month-old Il18r1^fl/fl:Ncr1-Cre⁻ (Il18r1^WT, n = 8) and Il18r1^fl/fl:Ncr1-Cre⁺ (Il18r1^ΔNK/ILC1, n = 9) mice were exposed to SL-TBI, and the total thymic IFNγ and perforin levels were measured 3 days later. g, Ifngr1 and Ifngr2 expression at baseline, taken from the scRNAseq dataset described in Fig. 2a. h, Female 1- to 2-month-old C57BL/6 Ifngr^fl/fl:Foxn1-Cre⁻ (Ifngr^WT, n = 6) and Ifngr^fl/fl:Foxn1-Cre⁺ (Ifngr^ΔTEC, n = 9) mice were exposed to SL-TBI, and thymus cellularity was measured on day 7. i, Female 1- to 2-month-old C57BL/6 WT (Ifngr^+/+, n = 10) or Ifngr1^−/− (n = 7) mice were exposed to SL-TBI, and thymus cellularity was measured on day 7. j, Female 1- to 2-month-old C57BL/6 WT (Prf^+/+, n = 10 and Gzmb^+/+, n = 5), Prf^−/− (n = 7) and Gzmb^−/− (n = 4) mice were exposed to SL-TBI, and thymus cellularity was measured on day 7. Graphs represent mean ± s.e.m.; each dot represents an individual biological replicate. Statistics were generated for a, e, f and h–j using unpaired two-tailed t tests, for b and d using paired two-tailed t tests, and for c using one-way ANOVA with Dunnet’s correction for multiple comparisons. Panel c created with BioRender.com. Source data

Fig. 7. Cytotoxic NK cells aberrantly target TECs.

a, Normalized expression of the MHC-I genes H2-D1, H2-K1 and B2m following SL-TBI, taken from the scRNAseq dataset described in Fig. 2a. Red box highlights epithelial populations. b, Thymuses from female 1- to 2-month-old C57BL/6 mice were collected at baseline (n = 4) or 3 days after SL-TBI (n = 5). Concatenated flow cytometry plots showing H-2K^b expression in stromal subsets (gating and phenotypes are provided in Extended Data Fig. 3) and the CD45⁺ population (n = 10) are presented. c, Thymuses from female 1- to 2-month-old C57BL/6 mice were collected at baseline or 3 days after SL-TBI. Concatenated flow cytometry plots and quantification of RAE-1 expression in stromal subsets (n = 5 per group) are shown. d, Female 1- to 2-month-old C57BL/6 mice were exposed to SL-TBI, and thymuses were collected 3 days later and enriched for nonhematopoietic stromal cells, which were cultured with or without poly(I:C)-stimulated NK cells. The expression of Annexin V (AnnV) and 7-aminoactinomycin D (7-AAD) in CD45⁻EpCAM⁺MHC-II⁺Ly51⁺ cTECs and CD45⁻EpCAM⁺MHC-II⁺UEA-1⁺ mTECs was measured 5 h after coculture (n = 4 per group). e, Female 1- to 2-month-old C57BL/6 WT (n = 14) or Il18r1^−/− (n = 15) mice were exposed to SL-TBI. The expression of Annexin V and 7-AAD in CD45⁻EpCAM⁺MHC-II⁺Ly51⁺ cTECs, CD45⁻EpCAM⁺MHC-II⁺UEA-1⁺ mTECs was measured 5 days later. f, Female 1- to 2-month-old C57BL/6 WT or Il18r1^−/− mice were exposed to SL-TBI, and CD45⁻EpCAM⁺MHC-II⁺Ly51⁺ cTEC and CD45⁻EpCAM⁺MHC-II⁺UEA-1⁺ mTEC cellularity was measured 3 days later (n = 5 biological replicates per group, representative of two independent experiments). g, Data extrapolated from the images in Fig. 4e. The distance between NKp46⁺ cells and either keratin-14⁺ mTECs or keratin-8⁺ cTECs was estimated by nearest-neighbor analysis and shown as a waterfall plot (day 0, n = 278; day 3, n = 1,663; day 7, n = 426). Graphs represent mean ± s.e.m.; each dot represents an individual biological replicate. Statistics were generated for b–f using unpaired two-tailed t tests and for g using one-way ANOVA with Tukey’s correction for multiple comparisons. Source data

Extended Data Fig. 1. Increased pyroptosis after distinct modalities of thymic damage.

a, Thymus supernatants generated from female 1-2 mo C57BL/6 mice given SL-TBI (550 cGy), Dexamethasone (i.p., 20 mg/kg), Cyclophosphamide (i.p., 200 mg/kg) or LPS (i.p., 1.5 mg/kg) and mature Caspase-1 assayed by ELISA on days 0 (n = 4), 0.5 (n = 3), 1 (n = 3) and 3 (n = 3). b, Female 1-2 mo C57BL/6 WT (d0, n = 18; day 3, n = 15) or Il18^−/− (d0, n = 9; day 3, n = 8) and Il18r1^−/− (d0, n = 10; d3, n = 5) mice were given SL-TBI and thymus cellularity was assessed at baseline (day 0) or 3 days post-SL-TBI. c, Thymus supernatants generated from female 1-2 mo C57BL/6 WT (d0, n = 5; d1, n = 8; d3, n = 3) and Il18^−/− (n = 4/group) mice given SL-TBI (550 cGy) and Cortisol assayed by ELISA on days 0, 1 and 3. Graphs represent mean ± SEM; each dot represents an individual biological replicate; ns = not significant. Statistics were generated for a using one-way ANOVA with Dunnet’s correction for multiple comparisons, b using Tukey’s correction for multiple comparisons, and for c using an unpaired two-tailed t-test. Panel a icons created with BioRender.com. Source data

Extended Data Fig. 2. Expression of IL-18 related genes across subsets in the thymus.

a, Relative expression of Il18, Il18r1 and Il18rap on subsets of thymus populations at baseline in 1mo mice (n = 2). Data extracted from Gene Expression Commons using the “Complete thymocyte:stromal interaction model dataset” (https://gexc.riken.jp/models/475/) which is based on GSE56928 (ref. ). b, UMAPs from scRNAseq described in Fig. 2a with color overlays indicating dataset (Left, Rag2-GFP⁻ CD45⁺ or CD45-) or timepoint (right, d0, 1, 4, 7 after SL-TBI). c, Il18 relative expression measured by qPCR within cDC1, cDC2, MECs, and Macrophages FACS isolated (see Extended Data Fig. 3 for gating) from thymuses of female 1-2 mo C57/BL6 WT mice (n = 4/group). Graphs represent mean ± SEM; each dot represents an individual biological replicate. Source data

Extended Data Fig. 3. Gating strategies for cell subsets in bone marrow and thymus.

Bone marrow and thymuses of untreated female 1-2 mo C57BL/6 mice were isolated and digested for follow cytometric identification of the populations described. Concatenated flow plots of gating strategies for mTEC1, mTEC2, mTEc3hi/lo, Tuft, Endothelial, Mesothelial and Fibroblast cells shown above. Gating strategies for NKT1, NKT2, NK, ILC-1, ILC-2, and ILC3 cells shown below. Tables show phenotypes used to gate individual populations.

Extended Data Fig. 4. IL-18R expression across thymus cell subsets.

IL-18R expression on thymic cellular populations (see Extended Data Fig. 3 for gating) taken from female 1-2 mo C57BL/6 WT or Il18r1^−/− mice (n = 3-5/group).

Extended Data Fig. 5. Effect of IL-18 on T cell development in vitro.

a–c, 50,000 lineage depleted bone marrow cells were co-cultured with OP9-DLL1^GFP fibroblasts confluent in a 6-well dish with 5 ng/mL Flt3L plus 1 ng/mL IL-7 and 0, 1 or 10 ng/mL rIL-18 (n = 1/group). a, (Top) CD45⁺ Thy1⁺ CD4⁻ CD8⁻ DN1-4 thymocyte differentiation measured according to CD44 and CD25 expression 2 weeks post co-culture and (Bottom) CD45⁺Thy1⁺ thymocyte differentiation into CD4⁺ CD8⁺ Double Positive population 3 weeks post co-culture. (Right) Ratio of DN2:DN1 thymocyte differentiation 2 weeks post co-culture. b, Thymocyte expansion measured by non-adherent cell count quantified 2 and 3 weeks post co-culture with OP9-DLL1^GFP adherent cells. c, 50,000 bone marrow CD45+ Lineage- cKit+ Sca-1+ LSKs were FACS purified and co-cultured with OP9-DLL1^GFP fibroblasts confluent in a 6-well dish with 5 ng/mL Ftl3L plus 1 ng/mL IL-7 and 0, 1 or 10 ng/mL rIL-18 (n = 4/group) and 10 days later, thymocyte expansion was quantified by measuring non-adherent cell expansion. d, Female 1-2 mo C57BL/6 mice were administered 200μg αNK1.1 monoclonal antibody or control PBS (i.p) at days -1, 1 and 3 days post SL-TBI and thymuses assessed at day 7. Thymus NK1.1⁺IL-18R⁺ (Left) and NKG2D⁺IL-18R⁺ (Right) cells (parent gated on viable CD45⁺CD4⁻CD8⁻ cells) (n = 9/group). e, Female 1-2 mo C57BL/6 WT (Cd1d ^+/+) and Cd1d^−/− mice were given SL-TBI and thymus cellularity was measured 35 days later (n = 5/group). f, Female 1-2 mo Il18r1^fl/fl:Ncr1-Cre⁻ (Il18r1^WT, n = 5) and Il18r1^fl/fl:Ncr1-Cre⁺ (Il18r1^ΔNK/ILC1, n = 6) mice were given SL-TBI and thymus cellularity measured on day 7 post SL-TBI. g, Female 1-2 mo l18r1^WT and Il18r1^ΔNK/ILC1 thymus CD45⁺ NK1.1⁺ TCRβ^—CD49b⁺ NK cells at baseline (n = 5/group). Graphs represent mean ± SEM; each dot represents an individual biological replicate; ns = not significant. Statistics were generated for e, f using unpaired two-tailed t-tests. Source data

Extended Data Fig. 6. Increased NK cells is a function of both extrathymic recruitment and intrathymic expansion.

a, Taken from scRNAseq described in from Fig. 2a: integrated UMAP showing undamaged, baseline thymus major clusters and annotations (left) and heatmap expression Ncr1, Itga1 and Itga1. b, Number of NK cells at day 3 in WT or Il18r1^−/− mice (n = 5/group). c, d, Female C57BL/6 CD45.1⁺ and CD45.2⁺ mice were surgically conjoined to establish parabiotic pairs when both members of the pair received sublethal total body irradiation (SL-TBI; 550 cGy). Intrathymic/extrathymic chimerism was calculated on days 0, 1, 4, or 7 days after SL-TBI using congenic markers (CD45.1 and CD45.2) to determine mouse of origin (that is, cells expressing the same CD45 isoform as assessed mouse pair thymus = “intrathymic”; cells expressing the alternate isoform = “extrathymic”-derived). Absolute chimerism (c) (n = 8/group) and fold-change in NK cell number (d) across timepoints (d1, n = 6; d4, n = 6; d7, n = 8). e, Concatenated flow cytometry plots shown of NKG2D expression on CD45⁺ NK1.1⁺ TCRβ^—CD49b⁺ NK cells and gMFI of NKG2D expression of IL-18R⁻ and IL-18R⁺ NK cells at day 3 post SL-TBI (n = 3/group). f–h, Female 1-2 mo C57BL/6 WT (-rIL-18, n = 6; +rIL-18, n = 4) or Ifng-reporter (-rIL-18, n = 5; +rIL-18, n = 3) mice were given rIL-18 (s.c., 2.5 mg/kg) or PBS. f, Total thymus cellularity assessed on day 3. g–h, Expression of Ifng-GFP (g) or Perforin (h) at day 3. i, Female 1-2mo C57BL/6 mice were given SL-TBI and 3 days later administered with rIL-18 (s.c., 2.5 mg/kg, n = 3) or PBS (n = 4) as in Fig. 4k. On day 7 NK1.1⁺IL-18R⁺TCRβ⁻CD49b⁺ NK cells were then FACS purified and cocultured with cell-dye labeled RMA-S target cells at a 5:1 Effector to Target ratio. RMA-S target cell Annexin V expression was measured 5 h post co-culture. Graphs represent mean ± SEM; each dot represents an individual biological replicate; ns = not significant. Statistics were generated for b–d and f–i using unpaired two-tailed t-tests, e using paired two-tailed t-tests, and for c using one-way ANOVA with Dunnet’s correction for multiple comparisons. Source data

Extended Data Fig. 7. MHC I expression across thymocyte subsets after damage.

a, Female 1-2 mo C57BL/6 mice were given SL-TBI and thymuses collected 0 and 3 days later. Concatenated flow plots showing H2kb expression of early thymic progenitors (ETPs), DN2-4, DP, SP4, SP8 cDC1, cDC2 cells (see Extended Data Fig. 3 for gating) and CD45⁺CD11b⁺Tim-4⁺ Macrophages and their respective gMFI 0 and 3 days post SL-TBI (n = 5/group). b, Female 1-2mo C57BL/6 mice were given SL-TBI and thymuses collected 3 days later and enriched for non-hematopoietic stromal cells which were cultured with or without poly(I:C) stimulated NK cells. Annexin V and 7-AAD expression of CD45⁻ EpCAM⁻ MHC-II⁻ non-TECs was measured at 5 h post co-culture (n = 4/group). c, Female 1-2 mo C57BL/6 WT (n = 5) or Il18r1^−/− (n = 6) thymuses collected at baseline and Annexin V and 7-AAD expression of cTECs and mTECs assessed. d, Female 1-2 mo C57BL/6 WT (n = 10) or Il18r1^−/− (n = 9) were given SL-TBI and thymuses collected at 5 days later. Annexin V and 7-AAD expression of CD45⁻EpCAM⁻MHC-II⁻ non-TECs was measured. Graphs represent mean ± SEM; each dot represents an individual biological replicate; ns = not significant. Statistics were generated for a–d using unpaired two-tailed t-tests. Source data