Abstract
Background: Possible correlations between the expression of immune checkpoint molecules and prognosis in childhood acute leukemia were investigated. Materials and Methods: The expression of programmed-death 1 (PD1), cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), and B- and T-lymphocyte attenuator (BTLA) was determined by flow cytometry on peripheral αβ+ and γδ+ T-cells from patients with newly diagnosed acute lymphoblastic leukemia (ALL) (n=9) or acute myeloid leukemia (AML) (n=12), and from healthy volunteers (n=7). The expression of programmed-death ligand 1 (PD-L1), B7-1, B7-2, human leukocyte antigen-ABC (HLA-ABC), and herpesvirus-entry mediator (HVEM) ligands was determined on leukemia blasts. Results: PD1 expression on αβ+ and γδ+ T-cells was significantly higher in patients with ALL than in those with AML (p=0.0019 and 0.0239, respectively). CTLA-4 expression was moderately higher on αβ+ and γδ+ T-cells in ALL (p=0.077 and 0.077, respectively), whereas HLA-ABC expression was significantly higher in AML blast cells (p=0.0182). The expression of CTLA-4 on γδ+ T-cells and the B7-2 ligand on blasts was higher in patients with high-risk ALL (p=0.02 and 0.02, respectively). In AML, PD1 expression on αβ+ T-cells was higher in the intermediate-risk group (p=0.05), whereas HVEM expression was significantly higher in the low-risk group (p=0.02). Expression of CTLA-4 on γδ+ T-cells and PD-L1 on blasts were both associated with poor relapse-free survival outcomes in ALL (p=0.049). Conclusion: The higher expression of immune checkpoint molecules, in particular, CTLA-4 and PD-L1 are associated with a poorer prognosis in ALL, suggesting that selective use of the immune checkpoint blockade might improve the clinical outcomes in patients with ALL.
Cancer acquires several biological capabilities during multistage development, including proliferative signaling, evasion of growth suppressors, resistance to cell death, replicative immortality, induction of angiogenesis, and metastasis (1). The cellular processes that are specifically activated in cancer cells during tumorigenesis accumulate genetic alterations that lead to the expression of tumor neoantigens (2). Ideally, immune cells would recognize these neoantigens, which are not present in normal cells, and kill the cancer cells. The process by which cancer cells are killed by normal immune cells can be explained as a cancer-immunity cycle of multiple steps (3). After initial capture and process by antigen-presenting cells (APCs), the following steps are involved in the priming and activation of effector T-cells, for which two separate signals are required (4). In the first, the antigen-dependent T-cell receptor (TCR) binds to the major histocompatibility complex (MHC) molecule of an APC. The second signal is an antigen-independent, co-stimulatory or inhibitory signal which is delivered by the APCs. This second signal determines and modulates TCR signaling and the fate of a T-cell (5).
Within the cancer-immunity cycle, immune checkpoint (IC) or inhibitory molecules function to reset or reinstate effector T-cells. Since the identification of checkpoint inhibitors, numerous clinical studies have reported a significant association between these factors and the clinical outcomes in multiple cancer types, such as a poorer prognosis associated with the overexpression of tumor programmed death-ligand 1 (PD-L1) (5-9). On the other hand, in terms of targeting immunotherapeutics, some clinical reports have revealed that patients with higher levels of PD-L1 expression show improvements in their response rates, and in progression-free and overall survival [reviewed in (9)]. These findings have included various cancer types such as Hodgkin's lymphoma, melanoma, multiple myeloma, lung cancer, and acute leukemia (9-11).
Mansour et al. reported that expression of soluble cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) is associated with a poorer prognosis in both adult and pediatric patients (12). Additionally, a phase I trial of nivolumab, which causes programmed death 1 (PD1) blockade, in children with relapsed or refractory solid tumors is currently ongoing through the Children's Oncology Group (13). However, despite the dramatic advances in our understanding of IC pathways and the success of some clinical trials, few studies to date have assessed the targeting of these pathways in children. Notably, the roles of checkpoint molecules and their clinical implications in pediatric acute leukemia, which is the most common malignancy in children, have not been investigated.
The purpose of our present study was to investigate the expression of IC receptors on T-cells and their cognate ligands in leukemia blasts in childhood acute leukemia and assess their possible correlations with the prognosis of these patients. We compared children with B-cell acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML), along with healthy volunteers, to identify potential biomarkers or targets for future cancer immunotherapies.
Materials and Methods
Patient characteristics. A prospective cohort of 21 children newly diagnosed with de novo acute leukemia prior to chemotherapy was enrolled. All of the participants or their legal guardians provided written informed consent (IRB approval no. 2018-0445). Nine patients were diagnosed with ALL, and 12 with AML. The B-cell ALL group included three boys and six girls, with a median age of 8.8 (range=0.8 to 12.3) years and a National Cancer Institute (NCI) risk group classification (14) of standard risk in four and high risk in five cases. The patients with AML included seven boys and five girls, with a median age of 10.3 (range=1.6 to 17.7) years, with eight with favorable risk and four with intermediate risk determined by the National Comprehensive Cancer Network (NCCN) risk group classification (15). All the patients were provided standard therapy and care. The healthy volunteers (HV) were three men and four women, with a median age of 40 (range=28.0 to 46.0) years. The age and sex ratios between the ALL and AML groups were not significantly different (Table I).
Isolation of peripheral blood mononuclear cells (PBMCs). PBMCs from patients with AML and B-cell ALL prior to treatment and HV were isolated by density-gradient centrifugation using Ficoll-Paque™ Plus (GE Healthcare, Milwaukee, WI, USA). Isolated cells were resuspended in heat-inactivated fetal bovine serum (FBS) (Welgene, Daegu, South Korea) containing 10% dimethyl sulfoxide (Sigma-Aldrich, St. Louis, MO, USA) and stored in liquid nitrogen until use.
Flow cytometry. After thawing and washing in phosphate-buffered saline (PBS), PBMCs were resuspended in PBS containing 2% FBS and treated with a human Fc gamma receptor (FcγRc) blocking antibody (Miltenyi Biotec, Bergisch Gladbach, Germany) for 10 min on ice. The cells were then stained with the following fluorescence-conjugated monoclonal antibodies against: Cluster of differentiation (CD)3 (clone SK7), TCRγδ (clone MOPC-21), PD1 (CD279; clone EH12.2H7), CTLA-4 (CD152; clone BNI3), B- and T-lymphocyte attenuator (BTLA) (CD272; clone MIH26), CD19 (clone 4G7), CD33 (clone WM53), PD-L1 (CD274; clone 29E.2A3), B7-1 (CD80; clone 2D10), B7-2 (CD86; clone IT2.2), and human leukocyte antigen (HLA)-ABC (clone G46-2.6). All antibodies and their appropriate isotype controls were purchased from BD Biosciences, eBioscience (San Diego, CA, USA) or BioLegend (San Diego, CA, USA). The results were obtained with a FACScanto II flow cytometer (BD Biosciences) and analyzed using FlowJo software (TreeStar, Ashland, OR, USA).
Statistical analysis. Differences between mean values were evaluated using the Mann-Whitney U-test and Kruskal–Wallis nonparametric test. Comparisons between survival outcomes were made using the log-rank test. All statistical analysis was performed using SPSS software version 21.0 (IBM Corp., Armonk, NY, USA) and GraphPad Prism version 8 (GraphPad Prism Software Inc., San Diego, CA, USA). p-Values of less than 0.05 were considered statistically significant.
Results
PD1 is more highly expressed in T-cells from patients with ALL. PBMC samples from nine patients with ALL and nine with AML were analyzed to determine the expression level of IC receptors. The gating strategies are shown in Figure 1A and C. The percentage of PD1+ cells among CD3+ T-cells was significantly higher in the ALL compared to the AML cases (median: 32.6% vs. 5.0%; p=0.0005) (Figure 1B). T-Cells were then gated as γδ+ T-cells and αβ+ T-cells (Figure 1C). For the γδ+ T-cells, the frequency of PD1-expressing cells was significantly different among the ALL, AML and HV groups. γδ+ T-Cells from patients with ALL showed a significantly higher expression of PD1 compared with those with AML and HV (p=0.0028 and 0.0239, respectively) (Figure 1D). The PD1 expression on αβ+ T-cells was also significantly higher in patients with ALL than in the AML and HV groups (p<0.0001 and 0.0019, respectively; Figure 1E).
CTLA-4 is more highly expressed in T-cells from patients with ALL. The samples and gating strategies used to evaluate CTLA-4 expression on T-cells were as described above. The percentage of CTLA-4-expressing cells was significantly higher in the T-cells from the children with ALL compared with AML (median: 14.4% vs. 1.3%; p=0.0039) (Figure 2A). The CTLA-4 level on γδ+ T-cells were also slightly higher in ALL compared to AML (p=0.077) and HV (p=0.5043) (Figure 2B), but this was not a statistically significant difference. CTLA-4 expression on αβ+ T-cells also showed a trend for marginally higher expression in ALL than AML (p=0.077; Figure 2C).
BTLA is expressed at similar levels in ALL and AML. The expression of BTLA on T-cells sampled from our patients with ALL and AML did not significantly differ (median: 57.5% vs. 27.6%, p=0.4283; Figure 3A). The BTLA expression on the γδ+ and αβ+ T-cells was also not significantly different among the ALL, AML and HV groups either (Figure 3B and C).
HLA class I is more highly expressed by AML blasts. AML blasts were gated as CD33+, and B-cell ALL blasts as CD19+. PD-L1 is a ligand specific to the PD1 receptor, whereas B7-1 (CD80) and B7-2 (CD86) are ligands that bind to CTLA-4. The expression of PD-L1, B7-1 and B7-2 in blast cells were not significantly different between the ALL and AML groups (p=0.4555, 0.2136, and 0.3028, respectively). However, HLA-ABC was expressed at a significantly higher level by AML blast cells (p=0.0182; Figure 4).
Expression of IC receptors and ligands stratified by the NCI risk grouping of patients with ALL. The expression of CTLA-4 on γδ+ T-cells was significantly higher in patients with high-risk ALL compared with those with standard-risk (median: 48.6% vs. 3.5%, p=0.02). B7-2 expression was significantly higher in leukemia blasts from patients with high-risk ALL (median: 46.6% vs. 18.0%, p=0.02). The median expression levels of the IC receptors and ligands in T-cells and blast cells were higher in high-risk cases although mostly not to a statistically significant degree (Table II).
Expression of IC receptors and ligands stratified by NCCN risk grouping of patients with AML. The expression of PD1 on αβ+ T-cells was significantly higher in the intermediate-risk AML cases compared to the low-risk ones (median: 6.0% vs. 2.6%, p=0.05). The expression of HVEM, a ligand for BTLA, was significantly higher on the blast cells from the low risk group (median: 31.9% vs. 75.6%, p=0.02). The differences in the expression of other receptors and ligands between these AML risk groups were not significant (Table III).
Expression of IC receptors and ligands stratified by ALL relapse. Among patients with ALL, PBMCs in 55.6% (5/9) expressed low levels of CTLA-4 (≤38%), whereas in the remaining 44.4% (4/9) expressed high levels of CTLA-4 (>38%), on γδ+ T-cells. The 4-year relapse-free survival rate was significantly higher for the patients with low CTLA-4 expression on γδ+ T-cells (100% vs. 25%, p=0.049). In terms of the PD-L1 expression on leukemia blasts, 55.6% (5/9) of ALL cases expressed a low level (≤24.5%), and the remaining 44.4% (4/9) expressed a high level (>24.5%). Relapse-free survival at 4 years was again significantly higher in the patients with low PD-L1 expression on leukemia blasts (100% vs. 25%, p=0.049; Figure 5). In patients with AML, the expression of IC receptors and ligands did not contribute to relapse-free survival outcomes (Table IV).
Discussion
We investigated IC receptor and ligand associations with acute leukemia in children, which has been less characterized than in adults with these types of cancer. Despite the clinical successes of using IC inhibitors in different forms of cancer in adults, pediatric studies of these therapies are rare (16). In our current investigation, we compared the expression of IC receptors and ligands in pediatric hematological malignancies including ALL and AML. PD1 and CTLA-4 receptor expression were found to be higher in ALL whereas HLA-ABC expression was significantly higher in AML. This indicates the downregulation of MHC I molecules on ALL blasts and may explain the low alloreactivity of allogeneic hematopoietic stem cell transplantation in this type of leukemia. The expression of a higher level of CTLA-4 on γδ+ T-cells and the B7-2 ligand on blasts was related to NCI high-risk group in ALL, whilst higher PD1 level on αβ+ T-cells and lower HVEM expression on blasts were associated with NCCN intermediate-risk group in AML. In children with ALL, a higher CTLA-4 expression level on γδ+ T-cells and PD-L1 expression level on blasts were significantly associated with lower relapse-free survival.
The differences in the expression of IC receptors on αβ+ and γδ+ T-cells and the relationship between this and the pathogenesis of acute leukemia has important clinical implications. This is because of the possibility that enhancing immunotherapeutic effect may be achieved by stimulating specific T-cell subtypes. We were particularly interested in γδ+ T-cells as we practice the depletion of these cells prior to haploidentical hematopoietic stem cell transplantation (17). The role of BTLA has been elucidated in γδ+ T-cells as an inhibitory receptor that regulates proliferation (18, 19). HLA-ABC, MHC class I molecules are well known ligands for killer cell-inhibitory receptors on natural killer cells, as well as γδ+ T-cells (20). However, the higher expression of HVEM in low-risk AML implies a contradictory role of γδ+ T-cells.
The expression patterns of the receptors and ligands between the two types of hematological malignancies tested in the present study were quite different. It is not yet clear whether these differences are related to the pathogenesis of specific malignancies. From the perspective of tumorigenesis, previous studies have reported effects from the interaction between PD1 and γδ+ T-cells. The PD1 pathway is a potentially important mechanism by which γδ+ T-cells are either functionally impaired or otherwise exhausted in tumor-bearing mice (21). Additionally, Bhat et al. reported that checkpoint blockade rescues the repressive effects of histone deacetylase inhibitors on γδ+ T-cell function (22).
CTLA-4 is an IC receptor that prevents autoimmunity in regulatory and memory T-cells through their deactivation when B7-1/2 ligands on APCs are conjugated (16, 23-25). A CTLA-4-blocking agent has been approved by the United States for treating adult and pediatric melanoma, but several preclinical results have revealed that other solid tumors also have a high expression of CTLA-4 (26, 27). In hematological malignancies, CTLA-4 polymorphisms are involved in relapse in adult patients with AML following a first complete remission after standard chemotherapy (28). A prior phase 1/1b multicenter study of ipilimumab, which induces a CTLA-4 blockade, in persisting leukemia after allogeneic hematopoietic stem cell transplantation reported a feasible response rate (32%) in adult patients (29). There have however been no pediatric clinical studies of these therapies to date. A previous study reported that the circulating soluble form of CTLA-4 was elevated in 70% of pediatric patients with B-cell ALL and positively correlated with CD1d expression, which is considered a negative prognostic marker (30). Our current study also found that there are clinical implications of CTLA-4 in pediatric ALL. The positive correlations between CTLA-4 expression and disease risk as well as relapse-free survival, implicate the possibility of using blocking agents as a therapeutic approach in pediatric ALL.
PD1 is an immune-inhibitory receptor expressed on activated T-cells. The PD1 and PD-L1 interaction inhibits T-cell proliferation, survival and effector functions (13). PD-L1 is highly expressed in myeloid leukemia cells including myelodysplastic syndrome, chronic myelomonocytic leukemia, and AML (13, 31). In our study series, PD-L1 expression was slightly higher in AML than in ALL, but not statistically significantly. The expression of PD1 on αβ+ T-cells was, however, significantly higher in NCCN intermediate-risk group comparing to the low-risk group in AML. These results are consistent with the findings of previous reports implicating the PD1/PD-L1 pathway as an immunotherapeutic target. In this context, nivolumab is currently being tested in a clinical trial for pediatric refractory AML, which is the first trial of an IC receptor inhibitor in children with a myeloid malignancy (13). Notably, PD1 was found in our current study to be highly expressed on T-cells of patients with ALL, and PD-L1 expression on blasts was also correlated with shorter relapse-free survival in ALL in our present analyses. Taken together, the evidence to date suggests that blocking this interaction may be translated to the clinic to improve the survival outcomes in refractory ALL cases as well as in patients with AML (13, 32).
In summary, IC ligands and their receptors may have differential roles in the pathogenesis of pediatric ALL and AML.
The relatively small number of patients and short period of evaluation are major limitations of our study. Follow-up analyses including treatments were also not performed. Therefore, further studies are required to expand on our current findings, as well as explore the therapeutic potential of IC inhibitors in pediatric acute leukemia.
Acknowledgements
This study was supported by a grant from the National R&D Program for Cancer Control, Ministry of Health & Welfare, Republic of Korea (1520060). This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2018R1C1B6008852).
Footnotes
Authors' Contributions
SH Kang, HJ Hwang performed the experiments, analyzed the data and wrote the first draft of the article. SH Hwang, YU Cho, SS Jang, Cj Park performed the experiments and edited the article. JW Yoo, ES Choi, H Kim, HJ Im, JJ Seo analyzed, quality controlled the data and edited the article. N Kim, and KN Koh conceptualized the project, oversaw the experiments and edited the article.
Conflicts of Interest
The Authors have no conflicts of interest to declare regarding this study.
- Received August 21, 2019.
- Revision received September 13, 2019.
- Accepted September 17, 2019.
- Copyright© 2019, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved