Renner K , Bruss C , Schnell A , Koehl G , Becker HM , Fante M , Menevse AN , Kauer N , Blazquez R , Hacker L , Decking SM , Bohn T , Faerber S , Evert K , Aigle L , Amslinger S , Landa M , Krijgsman O , Rozeman EA , Brummer C , Siska PJ , Singer K , Pektor S , Miederer M , Peter K , Gottfried E , Herr W , Marchiq I , Pouyssegur J , Roush WR , Ong S , Warren S , Pukrop T , Beckhove P , Lang SA , Bopp T , Blank CU , Cleveland JL , Oefner PJ , Dettmer K , Selby M , Kreutz M .

P30 CA076292 NCI NIH HHS , R01 CA175094 NCI NIH HHS

             glycolytic process
             lactate transmembrane transporter activity

	Graphical Abstract
	Figure 1. Glycolytic Activity Limits T Cell Response (A) Expression of AKT1, HIF1A, SLC2A1, HK2, HK1, TPI1, ENO1, LDHA, PFKFB3, PFKM, GOT1, GOT2, and GLUD1 was analyzed in melanoma biopsies of 47 patients prior to anti-PD1 therapy, and a glycolytic index was calculated. Patients were stratified according to the median index calculated; progression-free survival was plotted as a Kaplan Meier estimation curve. Significance was calculated applying the log-rank (Mantel-Cox) test. (B–G) M579-LUC or PANC-1-LUC cells were transfected either with pools of small interfering RNA (siRNA) either scrambled (siSCR) or siRNAs targeting PD-L1 (siPD-L1). (B) Lactate levels in supernatants were determined after 72 h of transfection in the presence or absence of 0.1 mM diclofenac. The experiment was conducted twice; each time two independent plates with four technical replicates were performed. Supernatants of technical replicates were pooled for analysis (mean, n = 2). (C) 72 h after transfection, tumor cells were pulsed with influenza peptide and influenza-specific T (FluT) cells were added, and after 20 h of coculture, luciferase activity of tumor cells was determined. Tumor cell viability was calculated as the ratio of luciferase activity of tumor cells only to tumor cells cocultured with FluT cells within one treatment condition. The experiment was conducted two times; each time two independent plates with four technical replicates on each were performed. Mean of technical replicates of each plate was calculated (one-way ANOVA, Dunnett’s multiple comparisons test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, mean + SEM, n = 4). (D) PD-L1 mRNA expression was analyzed in wild-type cells (WT), in cells treated with siSCR and cells treated with siPD-L1 (one-way ANOVA paired, Dunnett’s multiple comparisons test, ∗p < 0.05, ∗∗p < 0.01, mean + SEM, n = 4). (E and F) M579 cells were transfected as indicated in the presence or absence of 0.1 mM diclofenac (diclo), 0.1 mM lumiracoxib (lumira), 1 mM aspirin (ASA), or 0.2 mM ketoprofen (keto). (E) Lactate levels were determined in supernatants after 72 h of transfection (mean, n = 2 to 4). (F) 72 h after transfection, tumor cells were pulsed with influenza peptide, peptide was removed, and medium only or FluT was added. After 20 h, luciferase activity of tumor cells was determined. Viability was calculated by dividing siSCR normalized luciferase values of tumor cells cultured with FluT cells to respective medium-only values within each treatment condition (mean + SEM, n = 3). (G) Lactic acid was added during coculture of FluT cells with PANC-1-LUC cells transfected with siPD-L1, and normalized viability was calculated. The experiment was performed once in triplicates; shown is the mean.
	Figure 2. Diclofenac and Lumiracoxib Inhibit MCT1 and MCT4 Transport Activity (A and B) Indicated concentrations of diclofenac were washed in for 10 min, and relative rate of change in [H+]i (Δ[H+]i/Δt) was determined as induced by application of (A) 3 mM lactate in Xenopus oocytes expressing MCT1 and (B) 10 mM lactate in oocytes expressing MCT4 (mean ± SEM, n = 5). (C and D) Indicated concentrations of lumiracoxib were washed in for 10 min, and relative rate of change in [H+]i (Δ[H+]i/Δt) was determined as induced by application of (C) 3 mM lactate in oocytes expressing MCT1 and (D) by application of 10 mM lactate in oocytes expressing MCT4 (mean ± SEM, n = 5). (E and F) Original recording showing change in intracellular H+ concentration in oocytes during application of aspirin. Δ[H+]i/Δt as induced by application of (E) 3 mM lactate in oocytes expressing MCT1 and (F) 10 mM lactate in oocytes expressing MCT4 before (light gray bar) and after (dark gray bar) 10 min wash in of 1 mM aspirin (mean ± SEM, n = 5). (G and H) Impact of diclofenac derivatives (G) on lactate secretion (H) in a human melanoma cell line (MelIM). Lactate levels were normalized to control (mean+SEM, n = 3).
	Figure 3. Effects of Diclofenac Are Independent of Changes in Glycolysis-Associated Proteins (A and B) Expression of MCT1, MCT4, LDHA, and LDHB were analyzed in (A) M579 cells and (B) PANC-1 cells after 72 h of 0.1 or 0.2 mM diclofenac treatment. Protein expression was determined by western blot. One representative blot out of three independent experiments is shown. (C and D) Surface expression of PD-L1 and MHC-I in (C) M579 and (D) PANC-1 cells was analyzed after 72 h of 0.1 and 0.2 mM diclofenac treatment by flow cytometry. One representative out of three independent experiments is shown.
	Figure 4. Diclofenac Treatment Preserves Effector Functions but Induces a Metabolic Shift from Glycolysis to Respiration in Human CD8+ T Cells In Vitro (A–E) CD8+ T cells were freshly isolated and activated by anti-CD3/CD28 dynabeads. (A) MCT1 and MCT4 expression in quiescent (qui.) and 2 day (2 d) and 6 day (6 d) stimulated CD8+ T cells was determined by western blot. One representative blot is shown. (B–F) MCT inhibitors were applied at the following concentrations: 0.1 μM AZD3965 (AZD), 0.2 mM diclofenac (diclo), or 0.2 mM lumiracoxib (lumira), and the combination of 0.2 mM diclofenac and 0.1 μM AZD3965. (B) Lactate levels were measured after 48 h in supernatants and normalized to control (one-way ANOVA, Dunnett’s multiple comparisons test, ∗ shows significant differences between control and treated cells, ∗∗∗p < 0.001; mean + SEM, n = 13 for diclofenac, n = 7 for AZD3965, and n = 3 for the combination thereof, n = 6 for lumiracoxib-treated CD8+ T cells). (C) Proliferation of CD8+ T cells was monitored over 7 days. Cell number was measured by the cell analyzing system (CASY; one-way ANOVA, Dunnett’s multiple comparisons test, ∗ shows significant differences between control and treated cells, ∗∗∗p < 0.001; mean + SEM, n = 17 for control cells, n = 12 for diclofenac, n = 3 for AZD3965, and n = 10 for lumiracoxib-treated CD8+ T cells). (D) Viability of CD8+ T cells was determined after 7 days (mean + SEM, n = 17 for control cells, n = 11 for diclofenac, n = 4 for AZD3965, and, n = 7 for lumiracoxib-treated CD8+ T cells). (E) IFNγ levels secreted by CD8+ T cells were determined after 48 h in supernatants by ELISA (median, each symbol represents an independent donor). (F) Representative flow cytometry blots show expression of CD25 and CD137 in freshly isolated and 48 h stimulated CD8+ T cells. PD-1 expression was analyzed in 6 day expanded and for 48 h re-stimulated T cells. The signal on quiescent T cells as a negative control is shown for each antibody (filled gray). (G–L) CD8+ T cells were isolated, activated with anti-CD3/CD28 dynabeads, and expanded for 6 days. After 6 days, T cells were pooled and re-stimulated. Diclofenac was applied at 0.1 or 0.2 mM, AZD3965 at 0.1 μM, and lumiracoxib and ketoprofen at 0.2 mM. Low glucose (0.5 mM) was achieved by the usage of a glucose-free medium supplemented with 10% human serum. (G) MCT1, MCT4, LDHA, and LDHB protein expression was determined by western blot after 72 h. One representative blot is shown. (H) Lactate levels were measured in 48-h supernatants and normalized to control (one-way ANOVA, Dunnett’s multiple comparisons test, ∗ shows significant differences between control and treated cells, ∗∗p < 0.01, ∗∗∗p < 0.001, mean + SEM, n = 9 for diclofenac, n = 3 for lumiracoxib, AZD, or low glucose). (I) IFNγ levels in 48-h supernatants were determined by ELISA (median, each symbol represents an independent donor). (J) Oxygen consumption was measured by the PreSens technology (mean of three independent donors). (K and L) Glucose flux into (K) intermediates of glycolysis and TCA cycle or (L) amino acids was determined by mass spectrometry after 48 h (one-way ANOVA, Dunnett’s multiple comparisons test, ∗ shows significant differences between control and treated cells, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, mean + SEM, n = 3).
	Figure 5. Combined MCT1/2 and MCT4 Blockade Only Moderately Affects T cell Function In Vitro CD4+ and CD8+ T cell populations were isolated from spleens of WT and MCT4 knockout (MCT4−/−) mice and activated with plate-bound anti-CD3 and soluble anti-CD28 antibodies. (A and B) MCT4 expression in WT and MCT4−/− (A) CD4+ and (B) CD8+ T cells was determined by western blot at indicated time points. One representative western blot is shown. (C–H) MCT inhibitors were applied at the following concentrations: 1 μM SR13800 (SR), 0.1, or 0.2 mM diclofenac. (C and D) Lactate levels were measured in 48-h supernatants of SR-treated WT and MCT4−/− (C) CD4+ and (D) CD8+ T cells (Mann Whitney U test, ∗p < 0.05, mean + SEM, n = 4). (E–H) IFNγ levels secreted by WT and MCT4−/− (E and G) CD4+ or (F and H) CD8+ T cells were determined by ELISA in supernatants of 48-h stimulated T cells (E and F, median, each symbol represents an individual mouse; G and H, mean of two independent experiments).
	Figure 6. Modulating Glycolysis Augments the Efficacy of Checkpoint Inhibition in B16 Tumors (A) Lactate levels were measured in supernatants of B16 WT cells in the presence or absence of 0.1 or 0.2 mM diclofenac as well as lumiracoxib (mean + SEM, n = 4, one-way ANOVA, Dunnett’s multiple comparisons test, ∗p < 0.05, ∗∗∗p < 0.001). (B and C) Medium pH was monitored with the PreSens technology in supernatants of B16 WT cells under diclofenac treatment and B16 LDH–/– cells (B) or lumiracoxib-treated B16 WT cells (C) (mean, n = 3 for B16 WT cells, n = 2 for B16 LDH−/− cells). (D) Cell number was analyzed using the CASY system (one-way ANOVA, Dunnett’s multiple comparisons test, ∗p < 0.05, ∗∗∗p < 0.001, mean+SEM, n = 3). (E) Viability was determined (one-way ANOVA, Dunnett’s multiple comparisons test, ∗∗p < 0.01, mean+SEM, n = 3, except for diclofenac 0.2 mM, where n = 4). (F) MCT1, MCT4, LDHA, and LDHB protein expression was determined by western blot after 72 h. One representative blot is shown. (G) Representative FACS blots showing expression of PD-L1, MHC-I, and MHC-II after 72 h of diclofenac treatment. Isotype staining is shown (filled gray). (H) B16.SIY WT cells and SIY-specific 2C CD8+ T cells were cocultured for 24 h in the presence or absence of 0.1 mM diclofenac, anti-PD-1 antibody, or the respective isotype control. IFNγ levels were determined in supernatants by ELISA (one-way ANOVA, Tukey’s multiple comparisons test, ∗∗p < 0.01, median, each symbol represents an independent experiment). (I) 1.0 × 105 B16 WT cells were injected subcutaneously into the flank of C57BL/6 mice. Treatment was started when tumors were palpable. Diclofenac was applied intraperitoneally (i.p.) daily (7.5 mg/kg); anti-PD-1 antibody (20 mg/kg body weight) was administered i.p. every third day. Tumor volume was monitored over time (median, n = 7 for vehicle and anti-PD-1, n = 8 for diclofenac and diclofenac + anti-PD-1). (J) 0.3 × 105 B16 WT cells were injected subcutaneously into the flank of C57BL/6 mice, and treatment was started when tumors were palpable. Diclofenac and lumiracoxib were applied i.p. daily (both 15 mg/kg); anti-PD-1 and anti-CTLA-4 (CP, both 10 mg/kg body weight) were administered i.p. Tumor volume was monitored over time (median, n = 9 for vehicle, n = 10 for checkpoint inhibitors combined with NSAIDs, n = 4 for diclofenac, n = 5 for lumiracoxib). (K and L) Tumor pH was measured using a pH meter either at 1 or 2 mm depth (median, each symbol represents one individual tumor). (K) pH of B16 WT tumors in vehicle-, diclofenac-, or lumiracoxib-treated mice was determined. (L) pH in B16 LDH−/− tumors was measured. (M) Glycolytic activity was determined in vivo by PET scan in B16 WT and LDH−/− tumors and injected into the right and left flank (shown is one representative image). (N) 0.3 × 105 B16 LDH−/− cells were injected subcutaneously into the flank of C57BL/6 mice, and treatment was started when tumors were palpable. Anti-PD-1 and anti-CTLA-4 antibodies (CP, 10 mg/kg body weight) were administered i.p. every third day (median, n = 10). (O) Percentage of CD3+IFNγ+ T cells in B16 WT or LDH−/− tumors under indicated conditions (V, vehicle;, CP, anti-PD-1 and anti-CTLA-4; D, diclofenac; L, lumiracoxib) was measured by flow cytometry (median, each symbol represents one individual tumor).
	Figure 7. Diclofenac Augments the Efficacy of Checkpoint Inhibition in 4T1 Tumors (A) Lactate levels were measured in supernatants of diclofenac- or aspirin-treated 4T1 cells (mean + SEM, n = 4, one-way ANOVA, Dunnet’s multiple comparisons, ∗p < 0.05). (B) Medium pH was monitored with the PreSens technology in supernatants of 4T1 cells under indicated conditions (mean of n = 3). (C) Cell number was analyzed using the CASY system (one-way ANOVA, Dunnett’s multiple comparisons test, ∗∗∗p < 0.001, mean + SEM, n = 3). (D) Viability was determined (one-way ANOVA, Dunnett’s multiple comparisons test, ∗∗∗p < 0.001, mean + SEM, n = 3). (E) MCT1, MCT4, LDHA, and LDHB protein expression was determined by western blot after 72 h. One representative blot is shown. (F) Representative FACS blots showing expression of PD-L1, MHC-I, and MHC-II after 72 h of treatment. Isotype staining is shown (filled gray). (G–M) 1 × 106 4T1 cells were injected subcutaneously into BALB/c mice. NSAID and checkpoint inhibitor treatment was started on day 6. Diclofenac was applied i.p. twice a day (7.5 mg/kg) and aspirin (ASA) by addition to the drinking water at 600 μg/mL for 14 d; anti-PD-1 and anti-CTLA-4 antibodies were administered i.p. at a concentration of 10 mg/kg every 3–4 day for 1 week. (G) Growth curves of vehicle-, diclofenac-, or aspirin-treated tumors are shown (median, n = 14). (H) pH in 4T1 tumors was measured either at 1 or 2 mm depth (median, each symbol represents an individual tumor). (I and J) Percentage of IFNγ+ (I) and IL-2+ (J) cells among NK cells derived from 4T1 tumors was determined by flow cytometry (median, each symbol represents one individual tumor). (K–M) Individual growth curves of 4T1 tumors treated either with (K) anti-PD1 and anti-CTLA-4 antibodies, (L) aspirin combined with anti-PD1 and anti-CTLA-4 antibodies, or (M) diclofenac in combination with anti-PD-1 and anti-CTLA-4 antibodies are shown.
	Supplementary Figure 1. Diclofenac lowers lactate secretion independent of effects on proliferation and does not affect viability which is compromised by lactic acid. Related to Figure 1. A, Lactate levels were measured in 72 h culture supernatants of M579-LUC and PANC-1- LUC cells and normalized to cell number applying the CORE method (Mann-Whitney U test, **p<0.01, mean+SEM, n=5). B, Luciferase activity of M579-LUC and PANC-1-LUC cells non- transfected or transfected either with scrambled siRNA (siSCR) or a pool of siRNA against PD-L1 (siPD-L1) in the presence or absence of 0.1 mM diclofenac. The experiment was conducted twice, each time two independent plates with four technical replicates on each plate were performed. The mean of the technical replicates on each plate was calculated (mean+SEM, n=4). C,D, Cell death was evaluated by monitoring the number of YOYO™-1 Iodide positive cells using the Incucyte ZOOM live-cell imager after 72 h of 0.1 mM diclofenac treatment or after adding varying lactic acid (LA) concentrations for 24 h in C, M579-LUC and PANC-1-LUC cells as well as in D, FluT cells.
	Supplementary Figure 2. Diclofenac, but not ketoprofen, inhibits MCT1 and MCT4 and augments intracellular acidification. Related to Figure 2. A, Original recording of the intracellular H+ concentration ([H+]i) induced by application of 3 mM lactate in Xenopus oocytes expressing MCT1 after wash in of 0.1, 0.3, 1, 3, 10 and 30 µM diclofenac. B, Original recording of [H+]i in MCT4-expressing oocytes induced by application of 10 mM lactate in t e presence of 0.01, 0.03, 0.1, 0.3, 1 and 3 µM diclofenac. C, Original recordings showing the change in intracellular H+ concentration in oocytes expressing MCT1 before, in the presence of diclofenac and 10, 20, and 30 min after washout of the drug. D, Rate of change in intracellular H+ concentration (Δ[H+]/Δt) induced by application of lactate in oocytes expressing either MCT1 (white circles) or MCT4 (black circles) in the pres nce of diclofenac (0 min) and 10, 20 and 30 min after washout of diclofenac (mean±SEM, n=5). E Rate of change in intracellular H+ concentration Δ[H+]i/Δt as induced by application of 3 mM lactate in oocytes expressing MCT1 after 10 min wash in of 3, 10, 30, 100, 300 and 1000 µM ketoprofen (mean±SEM, n=5). F, Δ[H+]i/Δt as induced by application of 10 mM lactate in oocytes expressing MCT4 after 10 min wash in of 3, 10, 30, 100, 300 and 1000 µM ketoprofen (mean±SEM, n=5).
	Supplementary Figure 3. Impact of diclofenac and a specific MCT1/2 inhibitor on lactate secretion in human tumor cells with different MCT1 and MCT4 expression profiles. Related to Figure 3. A, MCT1 and MCT4 expression in OC-316 and IGROV-1 cells as determined by western blot analysis. B, Lactate levels were measured in culture supernatants of OC-316 and IGROV-1 cells treated with 0.1 µM AZD3965 (AZD), 0.1, 0.2 mM diclofenac (diclo) and the combination of 0.2 mM diclo and 0.1 µM AZD3965 after 24 h. Lactate levels were normalized to control levels (one-way ANOVA paired, Tukey’s multiple comparisons, **p<0.01, ***p<0.001, mean+SEM, n=4). C, MCT1 and MCT4 expression in LS147T wild type cells (wt) and LS147T knockout clones for MCT1 (MCT1-/-), MCT4 (MCT4-/-) and MCT1 and 4 (MCT1/4-/-) as determined by western blot analysis. One representative blot is shown. D, Lactate levels were measured in supernatants of LS147T clones treated with 0.1 µM AZD3965 (AZD), 0.1 or 0.2 mM diclofenac (diclo) and the combination thereof. Lactate levels were normalized to control levels (one-way ANOVA paired, Tukey’s multiple comparisons, *p<0.05, **p<0.01, mean+SEM, n=4).
	Supplementary Figure 4. Diclofenac treatment preserves effector functions but induces a metabolic shift from glycolysis to respiration in human CD4+ T cells in vitro. Related to Figure 4. A-E, CD4+ T cells were freshly isolated and activated with anti-CD3/CD28 dynabeads. A, MCT1 and MCT4 expression in quiescent (qui.) and 2 days (2 d) and 6 days (6 d) stimulated CD4+ T cells was determined by western blot. One representative blot is shown. B-F, MCT inhibitors were applied at the following concentrations: 0.1 µM AZD3965 (AZD), 0.2 mM diclofenac (diclo) or 0.2 mM lumiracoxib (lumira) and the combination of 0.2 mM diclo and 0.1 µM AZD3965. B, Lactate levels were measured in 48 h supernatants and normalized to control cells (one-way ANOVA, Dunnett’s multiple comparisons test, * shows significant differences between control and treated cells, ***p<0.001, mean+SEM, n=20 for diclofenac, n=5 for AZD3965 and n=4 for the combination thereof, n=12 for lumiracoxib treated CD4+ T cells). C, Proliferation of CD4+ T cells was monitored over 7 d. Cell number was measured by the CASY system (one-way ANOVA, Dunnett’s multiple comparisons test, * shows significant differences between control and treated cells, ***p<0.001, mean+SEM, n=27 for control cells, n=18 for diclofenac, n=4 for AZD3965 and, n=9 for lumiracoxib treated CD4+ T cells). D, Viability of CD4+ T cells was determined after 7 d (mean+SEM, n=16 for control cells, n=16 for diclofenac, n=4 for AZD3965 and n=9 for lumiracoxib treated CD4+ T cells). E, IFNγ levels secreted by CD4+ T cells were determined after 48 h in supernatants by ELISA (median, each symbol represents an independent donor). F, Representative flow cytometry blots show expression of CD25 and CD137 in freshly isolated and 48 h stimulated CD4+ T cells. PD-1 expression was analyzed in 6 day expanded and for 48 h re-stimulated T cells. The signal on quiescent T cells as a negative control is shown for each antibody (filled grey). G-L, CD4+ T cells were isolated, activated with anti-CD3/CD28 dynabeads and expanded for 6 d. After 6 d T cells were pooled and re-stimulated. Diclofenac was applied at 0.1 or 0.2 mM, AZD3965 at 0.1 µM (AZD), lumiracoxib at 0.2 mM and ketoprofen at 0.2 mM. Low glucose (0.5 mM) was achieved by the usage of a glucose free medium supplemented with 10% human serum. G, MCT1, MCT4, LDHA and LDHB protein expression was determined by western blot after 72 h. One representative blot is shown. H, Lactate levels were measured in 48 h supernatants and normalized to control (one-way ANOVA, Dunnett’s multiple comparisons test, * shows significant differences between control and treated cells, ***p<0.001, mean+SEM, n=10 for diclofenac, n=3 for lumiracoxib, AZD, low glucose). I, IFNγ levels in 48 h supernatants were determined (median, each symbol represents an independent donor). J, Oxygen consumption was measured by the PreSens technology (mean of n=3 independent donors). K,L, Glucose flux into K, intermediates of glycolysis and TCA cycle or L, amino acids was determined by mass spectrometry after 48 h (one-way ANOVA, Dunnett’s multiple comparisons test, * shows significant differences between control and treated cells, *p<0.05, **p<0.01, ***p<0.001, mean+SEM, n=3).
	Supplementary Figure 5. Effect of diclofenac in healthy and tumor bearing mice. Related to Figure 5 and 6. A-C, Healthy C57BL/6 mice were treated with 15 mg/kg body weight diclofenac injected i. p for one week. Percentage of A, CD3+, B, CD8+ T cells and C, NK1.1+ cells among CD45+ leukocytes determined in blood, spleen and lymph nodes (l.n.) are shown determined by specific surface marker staining and subsequent analysis by flow cytometry (median, each symbol represents one mouse). D-F, B16 wt tumor bearing C57BL/6 mice were treated with 15 mg/kg body weight diclofenac injected i. p. every day for 7-10 d. Percentage of D, CD3+, E, CD8+ T cells and F, NK1.1+ cells among CD45+ leukocytes was determined in blood, spleen and lymph nodes (l.n.) by specific surface marker staining and subsequent analysis by flow cytometry (median, each symbol represents one mouse; Mann-Whitney U test, *p<0.05, **p<0.01).
	Supplementary Figure 6. Checkpoint therapy and NSAID treatment augments effector cell infiltration. Related to Figure 6. Immune cell infiltration in B16 wt or LDH-/- tumors under indicated treatment conditions (V = vehicle, CP = anti-PD-1 and anti-CTLA-4, D = diclofenac, L = lumiracoxib) was measured by flow cytometry. Shown is the percentage of A, CD45+ leukocytes among the entire cell population; B, CD3+ T cells and C, CD3+CD8+ T cells among CD45+ cells. D, Percentage of IL-2+ cells among CD3+ T cells and E, percentage of PD1+ cells among the CD8+ T cell population are displayed. F, Percentage of NK1.1+ cells among CD45+ leukocytes is presented. A-F, Median is displayed, each symbol represents one individual tumor (one-way ANOVA, Tukey’s multiple comparisons test, *p<0.05, **p<0.01).
	Supplementary Figure 7. Diclofenac alters characteristics of tumor-infiltrating effector cells in T1 and B16 tumors. Related to Figure 6 and Figure 7. A-J, Immune cell infiltration in tumors was measured by flow cytometry. A,B, Percentage of CD45+ leukocytes within tumors is shown. C,D, Percentage of CD8+ T cells among CD45+ leukocyt s is shown. E,F, Percentage of CD25+ cells among CD8+ T cells is displayed. G,H, NK cells among CD45+ leukocytes, I,J, CD11b+ cells among CD45+ leukocytes are displayed (A-J, median, each symbol represents one individual tumor). K,L, Representative histological sections of organs are shown. K, Sections of lung (1), pancreas (2), liver (3), kidney (4), bone marrow (5) and stomach (6) without signs of pathological changes are presented. L, Sections of lung (1,2) and large bowel (3,4) showing local inflammation (indicated by red arrows) are displayed in two different magnifications, scale bars are displayed in the upper left angle (scale bar K1, K3, K4 500 µm, K2 200 µm, K5, K6 100 µm; L1 200 µm, L2, L4 50 µm, L3 2000 µm).