Background
Novel immunotherapeutic strategies are increasingly evolving for the treatment of acute myeloid leukemia (AML) [
1,
2]. Many of these strategies inherently rely on the efficiency and functionality of autologous T cells; a detailed understanding of T cell function at different phases of the disease, e.g., at diagnosis and at relapse, is therefore of highest importance for their optimal application. These treatment options include multispecific antibody constructs such as CD33/CD3-bispecific T cell engaging (BiTE) antibodies [
3‐
5] or other bispecific antibodies [
6] that bring the CD3
+ T cells in close contact with leukemic cells. Chimeric antigen receptors (CARs) or transgenic T cell receptors are introduced into patients’ T cells [
7,
8]. And finally, immune checkpoint inhibitors such as PD-1/PD-L1-blocking antibodies unleash spontaneously pre-existing tumor- or leukemia-specific T cells [
9‐
11].
However, different degrees of T cell dysfunctionality have been described in various hematologic malignancies, including adult T cell leukemia/lymphoma [
12,
13], chronic myeloid leukemia [
14,
15], and chronic lymphoid leukemia (CLL) [
16‐
18]. These observations have recently been put into the context of T cell exhaustion, a state of T cell dysfunction that is defined by increased expression of several inhibitory receptors (CD244, PD-1, CD160, TIM-3, LAG-3) in combination with poor effector function (hypoproliferation, diminished cytokine production, impaired cytotoxicity) and finally apoptosis [
19]. It was first described for antigen-specific T cells in chronic viral infection in mice [
20‐
22] but has since been demonstrated in several human chronic infections, among others in patients with human immunodeficiency virus (HIV) [
23‐
25]. Most of the data has been gathered on CD8
+ T cells, but loss of effector function has also been described in virus-specific CD4
+ T cells [
26,
27].
CD244, PD-1, CD160, and TIM-3 are expressed on T cells and interact with their ligands on antigen-presenting cells upon TCR ligation, resulting in modulation of the T cell response. CD244, also known as 2B4, is a dual-function receptor that mediates activating or inhibitory signals depending on its expression level, extent of ligation, and relative amounts of certain adaptor molecules [
28]. PD-1 limits T cell responses in infection [
29] and autoimmunity [
30]. Tumors can exploit this axis to escape the immune system by constitutive or inducible expression of PD-1 ligand [
31]. CD160 coinhibits CD4
+ T cells upon ligand binding [
32], and when coexpressed with PD-1, it also inhibits CD8
+ T cells [
33]. TIM-3 inhibits Th1 T cells by binding to galectin-9 [
34] and has been shown to promote T cell exhaustion during chronic viral infection [
35] and in cancer [
36,
37].
So far, data on T cell function or potential T cell exhaustion in AML is mainly based on the analysis of murine models. In a syngeneic AML model, it was reported that coexpression of PD-1 and TIM-3 defined a subset of CD8
+ T cells deficient in cytokine production. During AML progression, the number of these cells increased [
38].
In our study, we set out to analyze the phenotype (with a particular focus on the inhibitory molecules CD244, PD-1, CD160, and TIM-3) and function (proliferation, cytokine production) of peripheral blood (PB) and bone marrow (BM) T cells in AML patients at different stages of the disease (diagnosis, relapse after intensive chemotherapy, relapse after allogeneic stem cell transplantation (allo-SCT)) in comparison to healthy controls (HC) and untreated HIV-infected patients.
Discussion
Many novel immunotherapeutic strategies in AML including bispecific antibodies, CAR T cells, checkpoint inhibitors, and vaccinations rely on T cell function. For the optimal application and timing of these therapies, it is therefore of utmost importance to have detailed knowledge about the characteristics of T cells during the disease. Phenotypic changes and functional defects associated with T cell exhaustion have been described in patients with solid cancers [
40,
41] and hematologic malignancies [
12‐
15]. The most detailed studies of T cell status have been conducted in patients with CLL. An increased expression of CD244, PD-1, and CD160 was described on T cells of untreated CLL patients, preferentially on the CD8
+ effector cells, accompanied by defects in proliferation and cytotoxicity, but at the same time, increased production of IFN-γ and TNF-α [
17]. Within CLL patients in an early stage of disease, higher PD-1 positivity among CD8
+ T cells was shown to be associated with worse prognosis [
16]. Chemotherapy seemed to increase the expression of inhibitory receptors (CD244, PD-1) on T cells of CLL patients, while lenalidomide reversed this effect [
18].
In comparison, little data is available on different aspects of T cell number, phenotype, and function in AML patients. An early study described an increase of some activation markers (HLA-DR, CD69, CD71, CD57) on T cells at diagnosis [
42]. This was in line with data from gene expression profiling of T cells providing some hints at aberrant T cell activation in AML patients [
43]. Patients in complete remission after intensive chemotherapy had normal CD8
+ T cell counts but reduced numbers of CD4
+ T cells and Tregs, and the proliferation of CD4
+ T cells was not impaired [
44]. During chemotherapy-induced leukopenia, however, when T cell counts are very low, the remaining T cells were shown to be functionally impaired and to need optimal costimulation in order to proliferate [
45]. Antigen-specific CD8
+ T cell responses are generally very rare in AML and were only studied after allo-SCT. In this setting, increased PD-1 levels on MiHA-specific CD8
+ T cells [
46] and the existence of a special subset of TNF-α
+/IFN-γ
− T cells without further characterization [
47] were described.
However, expression levels of inhibitory molecules associated with T cell exhaustion and the functional status of T cells at different phases of the disease have not been studied. We report here that CD8+ and CD4+ T cells in PB as well as in BM of AML patients at relapse after allo-SCT showed increased expression of PD-1, in contrast to T cells of AML patients at diagnosis.
Inhibitory molecules on CD4
+ T cells have not been studied as broadly as on CD8
+ T cells. However, the same molecules also seem to play a role in CD4
+ T cells. Similarly to CD8
+ T cells, persistent antigen exposure can induce a dysfunctional state in CD4
+ T cells [
48], which correlates with PD-1 expression [
49,
50]. Tregs are unlikely to account for the observed increased expression of PD-1 on CD4
+ T cells, as freshly isolated Tregs from healthy volunteers have been reported to express PD-1 solely intracellularly [
51].
We were able to demonstrate that the inhibitory molecule expression pattern was independent of age and CMV status. Instead, high expression of CD244 and PD-1 was associated with T cell memory subset distribution. As had been described before [
39], we found that the numbers of naïve T cells decline with age. However, we saw that patients with a relapse had a significantly higher proportion of effector memory T cells, independent of age. This shift was clearly detectable as an increase in the CCR7
−/CD45RA
− subset as well as in CD27
− T cells, which have been described to be more differentiated [
39,
52]. We could show that the subset of effector memory T cells inherently expressed higher levels of CD244 and PD-1, in healthy controls as well as across all patient cohorts. Therefore, we concluded that the increase in inhibitory molecule expression was most likely a surrogate for a shift towards differentiated effector T cells instead of a sign for T cell exhaustion. This is in line with the emerging notion that T cells with the characteristics defining exhaustion might rather be chronically activated [
53].
Supporting our results, it has recently been shown in a study of healthy controls that inhibitory receptor expression depends more dominantly on differentiation and activation than on exhaustion of CD8
+ T cells [
53]. The increased expression of PD-1 on T cells of CLL patients was also found to be accompanied by a shift towards effector memory cells [
54]. Similarly, in a very detailed study on T cell status in CLL, it was described that inhibitory molecule expression correlated with a skewing of T cells towards effector differentiation, although this study did not analyze patients after allogeneic SCT [
17]. Supposedly, chronic antigen stimulation accounts for this shift in subsets. Importantly, however, functional defects were described for the T cells in CLL, particularly concerning proliferation and cytotoxicity [
17], although it was demonstrated, on the other side, that CMV-specific CD8
+ T cell function was not impaired [
55].
In our study, we did not detect relevant functional impairment with respect to proliferation or cytokine production in T cells from AML patients at diagnosis or at relapse. Cytokine production was measured after antigen-unspecific T cell stimulation based on PMA and ionomycin, which bypasses TCR signaling. Unfortunately, antileukemic T cell responses were too rare to be measured. Even CD3/CD28 stimulation only resulted in very small T cell responses in such a setting. After stimulation with PMA and ionomycin, however, we observed compromised cytokine production in HIV patients. We therefore conclude that we can at least detect a functional defect downstream of TCR signaling this way. Our data are perfectly in line with a recent publication showing that the expression of PD-1 and CD244 on CD8
+ T cells marks differentiated cells that intrinsically produce more cytokines [
53]. The sole deficit found was a significantly decreased IFN-γ production of CD4
+ T cells in diagnosis patients, but not in patients with an AML relapse. To our knowledge, this has not been described before and could constitute an important observation with respect to the evolving immunotherapeutic options, as the reversal of this defect by Th1-polarizing therapies could be part of the therapeutic effects. More unspecifically, serum levels of cytokines and chemokines have been measured before, and different groups found either identical [
56] or reduced levels [
57] of IFN-γ for untreated AML patients compared to healthy controls. Of course, our data applies to bulk T cells; therefore, antigen-specific T cells occurring at low frequencies could potentially vary from this pattern.
Methods
Sample collection
After written informed consent in accordance with the Declaration of Helsinki and approval by the Institutional Review Board of the Ludwig-Maximilians-Universität (Munich, Germany), PB or BM samples were collected from HC and AML patients at diagnosis or relapse before the start of treatment. Patients were treated according to the German AML-CG treatment recommendations. Samples of HIV patients before start of highly active antiretroviral therapy (HAART) were collected at the Division of Clinical Infectiology at the Klinikum der Universität München. The cytomegalovirus (CMV) serostatus was determined at the Virology Department of the Max-von-Pettenkofer-Institute of the Ludwig-Maximilians-Universität.
Immunophenotyping of lymphocytes
Mononuclear cells (MCs) were isolated from PB or BM by Ficoll density gradient centrifugation. Immunofluorescent staining of cell surface antigens was performed using the following fluorescence-conjugated monoclonal antibodies: CD244 (PE or APC, C1.7), PD-1 (APC or Brilliant Violet 421, EH12.7H7), CD3 (AlexaFluor 488, UCHT1), CD45RA (Brilliant Violet 421, HI100), CCR7 (PE, G043H7), CD27 (APC, O323; all BioLegend, San Diego, CA, USA), CD160 (APC, 688327), TIM-3 (PE, 344823; both R&D Systems, Minneapolis, MN, USA), CD8 (PerCP-eFluor 710, SK1; eBioscience, San Diego, CA, USA), and CD4 (APC-H7, RPA-T4; BD Biosciences, San Jose, CA, USA). Corresponding isotype controls were used. PB and BM cells were analyzed using LSR II (BD Biosciences, Heidelberg, Germany) and Navios (Beckman Coulter, Krefeld, Germany) instruments, respectively. Post-acquisition analysis was performed using FlowJo software (version 9.6; Tree Star, Ashland, OR, USA).
CFSE proliferation assay
CD3+ T cells were isolated from fresh or frozen PBMCs of HC and AML patients by MACS (Miltenyi Biotec, Bergisch Gladbach, Germany) labeled with 0.625 μm CFSE (Life Technologies, Carlsbad, CA, USA), and cultured in the presence of Dynabeads Human T-Activator CD3/CD28 (Life Technologies GmbH, Darmstadt, Germany) at a bead-to-cell ratio of 1:3 for 7 days. Unstimulated cells served as negative control. Harvested cells were then stained with antibodies for CD3 (APC, UCHT1; BioLegend), CD4 (APC-H7, RPA-T4; BD Biosciences), and CD8 (PerCP-eFluor 710, SK1; eBioscience).
Cytokine production assay
Isolated CD3+ T cells were stimulated with PMA (20 μg/ml) and Ionomycin (750 ng/ml; both Sigma-Aldrich, St. Louis, MO, USA) after overnight resting. After 1 h, Golgi stop solution consisting of Monensin at 25 μM and Brefeldin A at 10 μg/ml (both Sigma-Aldrich) was added for additional 5 h. Harvested cells were surface stained for CD3 (AlexaFluor 488, UCHT1; BioLegend), CD4 (APC-H7, RPA-T4; BD Biosciences), and CD8 (PerCP-eFluor 710, SK1; eBioscience) and intracellularly with the BD Cytofix/Cytoperm Kit (BD Biosciences) and antibodies for IFN-γ (PE, B27), IL-2 (Brilliant Violet 421, MQ1-17H12), and TNF-α (APC, MAb11; all BioLegend).
Statistical analysis
Data was analyzed using Prism 6 (GraphPad Software, La Jolla, CA, USA) and is reported in scatter plots. Statistical significance of differences was determined using the Mann-Whitney U test. p ≤ 0.05 was considered statistically significant (* in all figures), p ≤ 0.01 is designated with **, p ≤ 0.001 with ***, and p ≤ 0.0001 with ****.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
FMS, FSL, WH, and MS conceived and designed the experiments. FMS, KE, MS, and JSN performed the experiments. RD provided the HIV samples. FMS, FSL, and MS analyzed the data and designed the figures. FSL and MS wrote the manuscript. All authors read and approved the final manuscript.