In acute myeloid leukemia, B7-H1 (PD-L1) protection of blasts from cytotoxic T cells is induced by TLR ligands and interferon-gamma and can be reversed using MEK inhibitors

Abstract

B7-H1 (PD-L1) is a B7-related protein that inhibits T-cell responses. B7-H1 participates in the immunoescape of cancer cells and is also involved in the long-term persistence of leukemic cells in a mouse model of leukemia. B7-H1 can be constitutively expressed by cancer cells, but is also induced by various stimuli. Therefore, we examined the constitutive and inducible expression of B7-H1 and the consequences of this expression in human acute myeloid leukemia (AML). We analyzed B7-H1 expression in a cohort of 79 patients with AML. In addition, we studied blast cells after incubation with interferon-gamma or toll-like receptors (TLR) ligands. Finally, we evaluated functionality of cytotoxic T-cell activity against blast cells. Expression of B7-H1 upon diagnosis was high in 18% of patients. Expression of TLR2, 4 and 9 was detected in one-third of AML samples. Expression of TLR2 and TLR4 ligands or IFN-γ induced by B7-H1 was found to protect AML cells from CTL-mediated lysis. Spontaneous B7-H1 expression was also found to be enhanced upon relapse in some patients. MEK inhibitors, including UO126 and AZD6244, reduced B7-H1 expression and restored CTL-mediated lysis of blast cells. In AML, B7-H1 expression by blasts represents a possible immune escape mechanism. The inducibility of B7-H1 expression by IFN-γ or TLR ligands suggests that various stimuli, either produced during the immune response against leukemia cells or released by infectious microorganisms, could protect leukemic cells from T cells. The efficacy of MEK inhibitors against B7-H1-mediated inhibition of CTLs suggests a possible cancer immunotherapy strategy using targeted drugs.

Fig. 1

B7-H1 and TLR expression in leukemic cell lines and blast cells from AML. a Flow cytometry analysis of B7-H1 expression in myeloid leukemic cell lines (LAMA-84, HL60, K562, U937, KG1a, THP-1) and lymphoid lines (Raji, Jurkat) with or without incubation with 500 IU/ml IFN-γ for 24 h. Data represent the mean and SD of three experiments. b B7-H1, B7-DC, B7.1, B7.2 and B7-H4 expression measured by flow cytometry in 79 blast samples from AML patients collected at the time of diagnosis. c Same as in (b), but this is for TLR2, TLR4 and TLR9 expression. d B7-H1 expression in 60 blast samples collected upon diagnosis with or without incubation with either 500 IU/mL IFN-γ or TLR ligands (500 ng/ml LPS, 2.5 μg/ml PGN or 5 μM ODN) for 24 h. *P < 1 × 10⁻⁶, **P < 1 × 10⁻⁵, ***P < 6 × 10⁻⁵, paired t test. e Correlation between TLR2 or TLR4 expression and percent of B7-H1-positive cells (Pearson’s correlation)

Fig. 2

B7-H1 at relapse. a Histograms of B7-H1 flow cytometric analysis of blast cells collected from a patient with AML at the time of diagnosis and relapse. Thick line anti-B7-H1 mAb. Thin line control isotype. b B7-H1 expression upon diagnosis and relapse in nine AML samples. c Same patients as in (b), but cells were incubated for 24 h with 500 IU/mL IFN-γ

Fig. 3

T-cell inhibition by blast cells. a CTL activity of CD8-sorted T cells expanded for 2 weeks against blast cells from six patients with AML. Before the CTL assay, blast cells were pre-incubated for 1 h with or without 500 ng/ml Ultrapure LPS, anti-B7-H1 mAb or isotype control. Data represent the mean and SD of experiments performed in quadruplicate. *indicates significant differences with LPS (Student’s t test). E:T effector:target ratio. b CTL activity of CD8-sorted T cells expanded for 2 weeks against blast cells from one patient with AML collected upon diagnosis and upon first relapse. Target cells were pre-incubated for 1 h with anti-B7-H1 or isotype control before the CTL assay. Data represent the mean and SD of experiments performed in quadruplicate

Fig. 4

B7-H1 expression and signal transduction inhibitors. a B7-H1 mRNA levels (relative to rRNA) in THP-1 (left panel) and U937 (right panel) cells measured by RQ-PCR following 24 h incubation with IFN-γ with or without 1-h pre-treatment with signal transduction inhibitors. Inhibitors included the following: 20 μM UO126 (MEK1/2), 25 μM PD98059 (MEK1), 25 μM SP600125 (JNK), 25 μM AG490 (JAK2), 25 μM LY294002 (PI3K), 3 μM SB203850 (p38MAPK) and 10 nM rapamycin (mTOR). *indicates significant differences from cells treated with IFN-γ alone (Student’s t test). b B7-H1 expression in 60 blast samples with or without incubation with 25 μM of the PI3K inhibitor LY294002. c Same as in (b), but with incubation with the mTOR inhibitor rapamycin

Fig. 5

B7-H1 protein expression and MEK inhibitors. a Western blot analysis of B7-H1 levels with anti-mouse B7-H1 mAb (clone 179711) in DA1-3b mouse leukemic cells exposed to 500 IU/ml IFN-γ with or without 1-h pre-incubation with 20 μM UO126 (right panel) or 25 μM rapamycin (left panel). b Western blot analysis of B7H1 levels with anti-human B7-H1 mAb (clone 130002) in blast cells from one patient with AML. Cells were exposed to 500 ng/ml Ultrapure LPS for 24 h with or without pre-incubation with 20 μM UO126 or 1 μM AZD6244. c Same as in (b), but in another AML patient and with exposure to 500 IU/ml IFN-γ instead of LPS

Fig. 6

CTL-mediated killing of blast cells and MEK inhibitors. a CTL activity of CD8-sorted T cells expanded for 2 weeks against blast cells from seven patients with AML; the cells were chosen for their expression of B7-H1 after pre-incubation for 1 h with 500 ng/ml Ultrapure LPS. Blast cells were exposed to 20 μM UO126 or 1 μM AZD6244. Data represent the mean and SD of seven patients performed in quadruplicate. *indicates significant differences with LPS (Student’s t test). b B7-H1 expression in blast cells from (a) after exposure to LPS and 1 μM AZD6244. *indicates significant differences between control and LPS; **indicates significant differences between LPS and LPS plus AZD6244 (Student’s t test). c B7-H1 expression in one representative sample from (b). d Same as in (a), but with three patients who did not express B7-H1 after exposure to LPS