STAT3 in acute myeloid leukemia facilitates natural killer cell-mediated surveillance

Abstract

Acute myeloid leukemia (AML) is a heterogenous disease characterized by the clonal expansion of myeloid progenitor cells. Despite recent advancements in the treatment of AML, relapse still remains a significant challenge, necessitating the development of innovative therapies to eliminate minimal residual disease. One promising approach to address these unmet clinical needs is natural killer (NK) cell immunotherapy. To implement such treatments effectively, it is vital to comprehend how AML cells escape the NK-cell surveillance. Signal transducer and activator of transcription 3 (STAT3), a component of the Janus kinase (JAK)-STAT signaling pathway, is well-known for its role in driving immune evasion in various cancer types. Nevertheless, the specific function of STAT3 in AML cell escape from NK cells has not been deeply investigated. In this study, we unravel a novel role of STAT3 in sensitizing AML cells to NK-cell surveillance. We demonstrate that STAT3-deficient AML cell lines are inefficiently eliminated by NK cells. Mechanistically, AML cells lacking STAT3 fail to form an immune synapse as efficiently as their wild-type counterparts due to significantly reduced surface expression of intercellular adhesion molecule 1 (ICAM-1). The impaired killing of STAT3-deficient cells can be rescued by ICAM-1 overexpression proving its central role in the observed phenotype. Importantly, analysis of our AML patient cohort revealed a positive correlation between ICAM1 and STAT3 expression suggesting a predominant role of STAT3 in ICAM-1 regulation in this disease. In line, high ICAM1 expression correlates with better survival of AML patients underscoring the translational relevance of our findings. Taken together, our data unveil a novel role of STAT3 in preventing AML cells from escaping NK-cell surveillance and highlight the STAT3/ICAM-1 axis as a potential biomarker for NK-cell therapies in AML.

Keywords: AML; ICAM-1; NK cells; STAT3; immunotherapy.

Figure 1

THP-1 and HEL cell line xenografts are killed by NK cells in vivo. (A–D) NSG-Tg(Hu-IL15) mice were injected via the tail vein with 0.5x10⁶ THP-1 cells. In addition, the mice were injected i.v. with 10⁶ human NK cells 24 h prior to leukemic cell inoculation or left untreated. (A) Schematic experimental setup. (B) hCD45^+(high) NKp46⁺ NK cells were analyzed by flow cytometry in the splenic single cell suspensions 28 days post injection. (C, D) Infiltration of hCD45^+(low) leukemic cells in the bone marrow 28 days post injection was analyzed by flow cytometry. (B, C) Bar graphs represent mean +/- SD from two independent experiments (n=5-6 per group). Statistical analysis was performed using unpaired t-test. (D) Representative flow cytometry plots. (E, F) NSG-Tg(Hu-IL15) mice were injected with 0.5x10⁶ HEL cells i.v. In addition, the mice were injected i.v. with 10⁶ human NK cells 24 h prior to leukemic cell inoculation or left untreated. (E) Kaplan-Meier plot showing survival of mice injected with HEL cells only (n=6) or additionally with NK cells (n=8) as pool of two independent experiments. Kaplan-Meier estimates were utilized for survival analysis, with differences assessed through log-rank (Mantel-Cox) test. (F) hCD45^+(low) leukemic cells were analyzed by flow cytometry in the splenic single cell suspensions of terminally diseased animals. Bar graphs represent mean +/- SD from two independent experiments (n=6 per group). Statistical analysis was performed using unpaired t-test. (B, C, F) *p < 0.05; **p<0.01; ****p<0.0001.

Figure 2

STAT3-deficient AML cells escape NK-cell recognition. (A, B) CFSE-stained HEL STAT3^WT, STAT3^KO6 and STAT3^KO8 cells were mixed at indicated effector: target (E:T) ratios with (A) NK92 for 4 h or (B) expanded primary human NK cells for 2 h. The specific lysis of target cells was assessed by flow cytometry. One representative out of at least two independent experiments is shown. (C, D) CFSE-stained THP-1 STAT3^WT and STAT3^KO8 cells were mixed at indicated effector: target (E:T) ratios with (C) NK92 for 4 h or (D) expanded primary human NK cells for 2 (h) The specific lysis of target cells was assessed by flow cytometry. One representative out of at least two independent experiments is shown. (A–D) Symbols and bars represent mean of technical duplicates +/- SD. Statistical analysis for each ratio was performed using (A, B) one-way ANOVA with Tukey post-test (STAT3^WT vs ^KO6 – indicated on the left; STAT3^WT vs ^KO8 indicated on the right side) or (C, D) unpaired t-test. (E, F) CFSE stained THP-1 STAT3^WT and STAT3^KO6 and STAT3^KO8 cells were mixed at 1:1 ratio with human primary NK cells from three different healthy donors and the percentage of CD107a⁺ NK cells (CFSE-; in contrast to CFSE+ target cells) was analyzed by flow cytometry. (E) Representative flow cytometry plots pre-gated on CFSE- cells (NK cells only). (F) Bar graphs represent mean +/- SD of n= 4 per group (3 donors; each donor 1-2 experiments). Statistical analysis was performed using unpaired t-test. (A–D, F) *p < 0.05, **p < 0.01.

Figure 3

STAT3-deficient AML cells show reduced expression of ICAM-1 (A, B) Expression levels of CD48, CD58, ICAM1, HLA-A, ULBP1, ULBP2, MICA, B7-H6 were analyzed in (A) HEL and (B) THP-1 STAT3^WT and STAT3^KO8 cells using RT-qPCR (n=6-8 per group, out of two or three independent experiments). The relative values to the respective control cell line are shown as mean +/- SEM in the bar graphs. Statistical analysis was performed using unpaired t-test for each gene. (C, D) THP-1 and HEL STAT3^WT and STAT3^KO8 were analyzed for surface expression of ICAM-1 via flow cytometry. Representative histograms (C) and bar graphs showing mean percentage of ICAM-1⁺ cells +/- SD (D) are shown for each cell line (n=2 per group from two independent experiments). Statistical analysis was performed using unpaired t-test. (A, B, D) *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 4

STAT3-deficient AML cells form fewer immune synapses with NK cells. (A, B) THP-1 STAT3^WT and STAT3^KO8 were incubated with NK92 at 1:1 ratio and the number of NK92-THP-1 doublets was analyzed after indicated timepoints. (A) Representative flow cytometry plots. (B) Symbols and bars represent the mean percentage of doublets +/- SD (n=3-4 per group from 2 independent experiments). Statistical analysis was performed using unpaired t-test for each timepoint. (C) Representative fluorescence microscopy pictures showing THP-1 cells in green and NK92 in blue and stained for CD56 in red. (D) The synapse area from (C) was quantified for timepoint 10 min as ratio of mean fluorescence intensity (MFI) of CD56 at the NK92-THP-1 contact site to mean fluorescence intensity of CD56 at NK92 surface. (B) **p < 0.01, ****p < 0.0001.

Figure 5

The impaired killing of STAT3-deficient AML cells is caused by low expression of ICAM-1. (A) THP-1 STAT3^WT and STAT3^KO8 cells were incubated with isotype control (IgG) or blocking antibody against ICAM-1 followed by surface expression analysis of ICAM-1 via flow cytometry. Bar graphs show the mean percentage of ICAM-1⁺ cells +/- SD (n=2 per group from two independent experiments). Statistical analysis was performed using one-way ANOVA with Tukey post-test. (B) CFSE-stained THP-1 STAT3^WT and STAT3^KO8 cells were preincubated with isotype control (IgG) or ICAM-1 blocking antibody and mixed at the indicated effector: target (E:T) ratios with NK92 cells for 4 h. The specific lysis of target cells was assessed by flow cytometry. One representative out of two independent experiments is shown. Symbols and bars represent the mean of technical duplicates +/- SD. Statistical analysis was performed using one-way ANOVA with Tukey post-test for each ratio. Parentheses indicate that the same significance level refers to all conditions covered. (C) THP-1 STAT3^WT and STAT3^KO8 and HEL STAT3^WT and STAT3^KO8 cells transduced with lentivirus encoding for GFP or ICAM-1 were analyzed for surface expression of ICAM-1 via flow cytometry. Representative flow cytometry plots out of 2 independent experiments. (D, E) CFSE-stained (D) THP-1 or (E) HEL STAT3^WT_GFP, STAT3^KO_GFP, STAT3^WT_ICAM1 and STAT3^KO_ICAM1 cells were mixed at indicated effector: target (E:T) ratios with (D) NK92 cells for 4 h or (E) primary human NK cells for 2 h. The specific lysis of target cells was assessed by flow cytometry. One representative out of two independent experiments is shown. Symbols and bars represent the mean of technical duplicates +/- SD. Statistical analysis was performed using one-way ANOVA with Tukey post-test for each ratio. Parentheses indicate that the same significance level refers to all conditions covered. (A, B, D, E) *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 6

STAT3 correlates with ICAM1 expression in AML patient samples. (A–C) 79 AML patient samples were analyzed for expression of STAT3 and ICAM1 using RT-qPCR and (A) ICAM1 expression was compared between STAT3 high and STAT3 low group. Statistical analysis was performed using unpaired t-test; **p < 0.01. (B) The correlation between STAT3 and ICAM1 expression was analyzed using Pearson test. (C) Kaplan-Meier plot shows the probability of survival of AML patients expressing high or low ICAM1. (D) Kaplan Meier plot shows the probability of survival of AML patients from publicly available datasets (kmplot.com) stratified into ICAM1 high and low groups. (C, D) Statistical analysis was performed using log-rank (Mantel-Cox) test.