Article Figures & Data
Figures
- Fig. 1.
Immunohistochemical staining of human pancreatic cancer tissue for PD-L1 and PD-L2. Representative case of PD-L1–positive (A) and PD-L2–positive (C) tumor. B and D, negative controls stained with mouse immunoglobulin G. Original magnification, ×400.
- Fig. 2.
A, overall survival of 51 patients with pancreatic cancer in relation to tumor PD-L1 status. PD-L1–positive patients had a poorer prognosis than the PD-L1–negative patients (P = 0.016). B, overall survival of 51 patients with pancreatic cancer in relation to tumor PD-L2 status. There was no significant difference in prognosis between PD-L2–positive and PD-L2–negative patients (P = 0.297).
- Fig. 3.
Therapeutic efficacy of PD-L1/PD-1 blockade in murine pancreatic cancer. A, expression of PD-L1 on murine pancreatic cancer cells, PAN 02. IFN-γ stimulation for 3 d in vitro induced its high expression (>95%) as estimated by flow cytometric analysis. Dashed line, negative control; thin line, naïve PAN 02; bold line, PAN 02 stimulated with IFN-γ. B, PD-L1 or PD-1 blockade induced significant antitumor effect and inhibited tumor growth. Mice were treated with anti–PD-L1 or anti–PD-1 mAb at a dose of 0.3 mg thrice per week for 4 wk. Tumor size at 4 wk after tumor establishment: anti–PD-L1 mAb, 6.3 ± 1.0 mm, n = 5; anti–PD-1 mAb, 6.0 ± 0.8 mm, n = 5; P = 0.0087 and P = 0.0025, respectively, compared with controls (9.0 ± 0.4 mm, n = 10). Points, mean maximal tumor size; bars, SE. C, immunohistochemical staining of CD8+ T cells in tumor tissue at 4 wk after tumor establishment. Representative pictures of three mice for each treatment. PD-L1 blockade induced massive CD8+ T-cell infiltration (magnification: top right, ×100; bottom right, ×400) compared with controls (top left, ×100; bottom left, ×400). D, induction of local immune activation by PD-L1 blockade. The expressions of IFN-γ, granzyme B, and perforin were significantly higher in tumors treated with anti–PD-L1 mAb than in controls (P = 0.030, P = 0.015, and P = 0.012, respectively). Points, mean of real-time PCR data from three individual mice for each group; bars, SE.
- Fig. 4.
Combination of PD-L1 blockade and gemcitabine. A, gemcitabine and delayed PD-L1 blockade significantly inhibited tumor growth of pancreatic cancer. Gemcitabine was given i.p. on days 1, 4, 7, 10, and 13 at a dose of 60 μg/g. Anti–PD-L1 mAb was given at a dose of 0.3 mg on days 15, 17, 19, 22, 24, and 26. Tumor size at 4 wk after tumor establishment: gemcitabine + anti–PD-L1 mAb, 5.8 ± 0.2 mm; P = 0.0077, versus gemcitabine alone, 6.8 ± 0.2 mm; P < 0.0001, versus anti–PD-L1 mAb alone, 9.0 ± 0.3 mm. B, gemcitabine and simultaneous PD-L1 blockade had a synergistic antitumor effect and resulted in complete response in treated mice. Anti–PD-L1 mAb was given at a dose of 0.3 mg for thrice per week for 4 wk. Tumor size at 4 wk after tumor establishment: gemcitabine + anti–PD-L1 mAb, 1.2 ± 0.2 mm; P < 0.0001, versus gemcitabine alone, 6.8 ± 0.2 mm; P = 0.001, versus anti–PD-L1 mAb alone, 6.2 ± 1.0 mm. Points, mean tumor size of five mice; bars, SE.
Tables
- Table 1.
One-year survival rate of 51 patients with pancreatic cancer according to clinicopathologic characteristics and tumor PD-L1 status
Variable n (%) PD-L1 positive, n (%) PD-L1 negative, n (%) P Tumor status T1 1 (100) 1 (ND) 0 (—) ND T2 7 (42.9) 3 (0) 4 (75.0) 0.093 T3 40 (54.7) 15 (40.8) 25 (62.9) 0.019 T4 3 (54.7) 1 (ND) 2 (ND) ND Nodal status N0 15 (61.9) 6 (20.0) 9 (87.5) 0.019 N1 36 (46.2) 14 (40.8) 22 (50.0) 0.252 Metastatic status M0 45 (52.8) 17 (33.0) 28 (63.2) 0.035 M1 6 (33.3) 3 (33.3) 3 (33.3) 0.486 Pathologic stage Ia, Ib 2 50.0 (0) — (2) ND (ND) IIa 13 (64.2) 6 (20.0) 7 (100) 0.006 IIb 29 (50.6) 11 (42.4) 18 (55.6) 0.426 III 1 (0) 0 (—) 1 (ND) ND IV 6 (33.3) 3 (33.3) 3 (33.3) 0.486 Total 51 (50.4) 20 (33.5) 31 (60.3) 0.016 Abbreviation: ND, not determined.
- Table 2.
Inverse correlation between tumor PD-L1 status and TILs
PD-L1 CD4 CD8 Positive Negative Positive Negative Positive 9 11 5 15 Negative 24 7 25 6
Volume 13, Issue 7
