Supplemental Figures

Supplemental Figure 1. Levels of sIL-15 complexes in MCA-205, MC-38, B16-F10, and B16-OVA tumors. A, B) MCA-205 or MC-38 tumors were isolated from Wt mice (MCA between 5-10 days/MC-38-between 10 and 18 days post-implantation) and levels of sIL-15 complexes in tumors were analyzed by ELISA. Plots show the amount of sIL-15 complexes present in each tumor versus its tumor mass in a representative experiment for each tumor line. C) Levels of sIL-15 complexes in B16-F10 and B16-OVA implanted at the same time and isolated 10 days post-tumor implantation. Average tumor weight (mg): B16-F10= 58± 4; B16-OVA= 41±17.

Supplemental Figure 2. Tumor infiltrating lymphocytes and tumor growth. A) Graph shows levels of sIL-15 complexes (black bar, left axis) and total CD8 T cells/tumor (gray bars, right axis) in MC-38 tumors grouped by tumor size. * indicates p<0.05 compared to 20-50 mg tumors. B) Frequency of Ki67+ cells among indicated TILs isolated from B16-OVA tumors treated systemically with αIL-15 Ab or IgG 4 and 13 days after tumor implantation and analyzed one day later. C) Frequency of TIM3+ and TIM3+PD-1+ cells among CD8 TILs isolated from B16-OVA tumors treated with αIL-15 Ab or IgG. D) MC-38 tumors implanted into Wt mice were treated intratumorally with αIL-15Ab (50 µg) 8 days post implantation. One week later, tumor lymphocytes were analyzed by flow cytometry and total cell numbers were normalized to tumor weight. **p<0.01. E, F) Tumor growth of B16-OVA and MC-38 tumors in Wt mice treated with intratumoral αIL-15 mAb or PBS 8 days post B16 tumor implantation or 14 and 22 days post-MC-38 tumor implantation. All experiments contained 5 mice/group.

Supplemental Figure 3. Production of sIL-15 complexes by various tumor cell lines. A) The following tumor cell lines: BP-1 Melanoma cells, T3M4 pancreatic cancer, A20 Lymphoma, CT26 colon carcinomas cell lines, MB49 bladder carcinoma and 4T1 mammary tumor cells were grown to almost confluent levels in complete media and culture supernatants were collected. Culture supernatants and the respective complete media were analyzed for sIL-15/IL-15Rα complexes by ELISA (3 wells/cell line), *p<0.05. B) In a different experiment, levels of sIL-15 complexes in tumor culture supernatants were measured and normalized to the number of cells recovered.

Supplemental Figure 4. Composition of IL-15-GFP expressing myeloid cells in MCA-205 and MC-38 tumors. A) Flow cytometric plots showing cells in spleens and MCA and MC-38 tumors isolated at the same time from IL-15 translational reporter mice after gating on CD45+, lineage (TCR-β, CD19, NK1.1) negative, and GFP+ cells. Ly6C and Ly6G expression was examined in GFP+ CD11b+cells. B) GFP/IL-15 expression in IL-15 translational and transcriptional reporter mice. Flow cytometric plots showing composition of GFP+ myeloid subsets in similar size MCA tumors isolated from IL-15 translational and transcriptional reporter mice.

Supplemental Figure 5. F4/80 staining by splenic and tumor myeloid cells. Tumor cells were implanted s.c. into IL-15 reporter mice. Tumors and spleens were isolated and stained as described. A) Panels show tumor and spleen cells stained with F4/80-APC and Rat IgG2a-APC after gating on CD45+ lineage- cells. B) Histograms show staining of F4/80-APC and Rat IgG2a-APC after gating on CD45+lineageneg, GFP+ CD11b+ cells in spleens and B16 tumors implanted into IL-15 translational reporter mice.

Supplemental Figure 6. IL-15Rα expression by tumor myeloid cells. B16 tumor cells were implanted s.c. into Wt and IL-15Rα-/- mice 9 and 14 days prior to analysis to allow simultaneous analysis of IL-15Rα of two time points. Tumors were isolated, incubated with BD mouse Fc block, and stained for IL-15Rα using anti-mouse IL-15Rα-biotin (R&D Systems) followed by Streptavidin-APC (Jackson ImmunoResearch). A) Histograms show representative staining for IL-15Rα after gating on CD45+ lineageneg CD11b+ cells isolated from tumors in Wt (grey filled) and IL-15Rα-/- (open histogram) mice. B) Graph shows average MFI of IL-15Rα staining from indicated tumor myeloid cells from IL-15Rα-/- mice (white bars) and from Wt mice 9 days (light grey bars) and 14 days (dark grey bars) post-tumor implantation. N=3-5 mice/group.

Supplemental Figure 7. Neutralizing IL-15 during STING agonists treatment. A) Flow cytometry analysis of CFSE

dilution in OT-I T cells in Spleen and tumor-draining LNs (dLN) of Wt mice bearing B16-OVA tumors. CFSE-labelled

OT-I cells (0.5x106 cells) were injected on day 7 post-implantation, and received one i.t. treatment with c-di-GMP (25µg)

once tumors became palpable. Three days post-treatment, the mice were sacrificed and spleen and draining lymph nodes

were analyzed by flow cytometry. B) Frequency of OT-I (CD45.1+) among CD8 T cells IL-15 Ab (50µg) was delivered

i.p. one day prior to and on the same day as treatment with c-di-GMP. Spleens and tumor-dLNs and were isolated three

days later. C) Representative flow cytometry plots of NK1.1 and CD4 staining in spleens and B16 tumors treated with

αIL-15 mAb or IgG (i.p.) D) Frequency of Granzyme B+ CD8 T cells from B16 tumors from similar experiment shown

in Figure 5B. E) Representative IFN-γ staining by CD8 T cells from B16 tumors from similar experiment shown in

Figure 5B.

Supplemental Figure 8. Dependence of IL-15 in STING-mediated tumor regression of MCA tumors. Wt and IL-15Rα-/- mice bearing MCA-205 tumors on the right flank received three i.t. injections of STING agonist, c-di-GMP (25µg) on days 8, 11, and 14 post-implantation. (n=4-5 per group). Error bars represent SEM, *p<0.05.

Supplemental Figure 9. STING-mediated tumor regression is not dependent on CD11c+ cell-derived IL-15. A, B) Growth of B16-OVA tumors in CD11c-Cre Tg X IL-15Rαfl/fl and Wt mice treated with i.t. injections of c-di-GMP (25µg) or PBS on days 7, 10, and 13 post-implantation (indicated by arrows). Wt mice were treated alongside the CD11c-Cre Tg X IL-15Rαfl/fl mice to confirm the effectiveness of c-di-GMP treatment. (n=4-5 per group). Error bars represent SEM, *p<0.05, ***p<0.001. C) In a separate experiment, tumor growth of B16-OVA cells was compared in CD11c-Cre Tg X IL-15Rαfl/fl and IL-15Rαfl/fl mice (n= 5-7 mice/group).

Supplemental Figure 10. Myeloid cell reconstitution in tumors from bone marrow chimeras (BM) chimeras. A) Flow cytometry plots showing frequency of CD45.2 (BM donor) and CD45.1 cells in B16 tumors after gating on CD11b+ cells in the lymphocyte gate. B) Averaged frequency of CD45.2 BM donor cells among CD11b+ cells in B16 tumors, n=4-5/group.

Supplemental Figure 11. Depletion of myeloid cell subsets in tumors. A) B16-OVA tumors were implanted into either CCR2-DTR- Tg+ and Tg- littermates. Seven days later, mice were treated i.p with DT every 2 days until time of sacrifice. Tumors and spleens were isolated, and frequency of specific myeloid subsets was analyzed alongside the levels of sIL-15 complexes. B) B16 tumor cells were implanted s.c. into Wt mice followed by treatment with αLy6G Ab or Ig control. Tumors and spleens were isolated, analyzed for frequency of specific myeloid subsets and levels of sIL-15 complexes. Both panels show staining of Ly6C and Ly6G after gating on CD45+ Lin- CD11b+ cells.