Chronic myeloid leukemia (CML) is a clonal hematopoietic stem-cell malignancy characterized by the presence of the chimeric BCR-ABL oncoprotein with deregulated tyrosine-kinase (TK) activity. Although conventional T cells are acknowledged as important players in the control of CML, a possible modification of invariant NKT (iNKT) cells, known for their antitumoral activity, has not been established as yet. Here, we showed that the expression of perforin, CD95L, and promyelocytic leukemia zinc finger, a transcription factor required for maintenance of iNKT cell functions, was reduced or suppressed in CML patients at diagnosis, as compared with healthy individuals. The proliferation rate of blood iNKT cells in response to their cognate ligand was likewise diminished. These functional deficiencies were corrected in patients having achieved complete cytogenetic remission following TK inhibitor or IFN-α therapy. iNKT cells from CML patients in the chronic phase did not display increased TK activity, which argued against a direct autonomous action of BCR-ABL. Instead, we found that their anergic status originated from both intrinsic and APC-dependent dysfunctions. Our data demonstrate that chronic phase CML is associated with functional deficiencies of iNKT cells that are restored upon remission. These results suggest a possible contribution to disease control by TK inhibitor therapies.
Chronic myeloid leukemia (CML) is a well-characterized myeloproliferative disorder, initiated by the presence of the Philadelphia chromosome (Ph) generating the BCR-ABL oncoprotein, which gives rise to deregulated tyrosine-kinase (TK) activity in all leukemic cells . For many years, allogeneic stem cell transplantation has been regarded as the standard curative therapy and was offered front line in patients in whom an HLA-compatible donor was available, with IFN-α therapy as a possible alternative . However, based on the results of the International Randomized Study of Interferon and STI571 (IRIS) trial, imatinib mesylate (IM), a competitive inhibitor of the BCR-ABL TK activity, is currently used as a first line therapy of CML [2, 3].
A critical role of the immune system in the control of CML [4-8] is supported by several reports, even though the mechanisms of this antileukemic response are poorly understood. Invariant NKT (iNKT) cells constitute a distinct lymphocyte population sharing a conserved semi-restricted TCR that recognizes glycolipidic antigens in the context of CD1d , with the unusual ability to rapidly secrete both Th1 and Th2 cytokines upon primary stimulation, together with a complete cytotoxic arsenal. The transcription factor promyelocytic leukemia zinc finger (PLZF) directs the effector program of the iNKT cell lineage [10, 11] which is believed to play a key role in the regulation of various immune responses, including antitumor responses in experimental mouse models [12-15], as well as in human solid tumors [16-18] and hematopoietic malignancies [19, 20]. However, although the role of conventional T cells, particularly their tumor Ag-specific cytotoxic subset [4, 7, 8], in the control of CML has been largely documented, the implication of iNKT cells has not been clearly established and has only been investigated in a single study comprising exclusively CML patients having undergone treatment with IM .
Here, we addressed this issue by comparing frequencies, perforin, CD95L, and PLZF expression, as well as proliferative responses between blood iNKT cells from healthy donors (HDs) and CML patients, either at diagnosis or after induction of complete remission (CR) by IM or IFN-α therapy. Based on our results, it can be concluded that iNKT cell functions are impaired during chronic phase (CP) CML, but return to normal after therapy.
Results and discussion
Functional deficiencies of iNKT cells in CML-CP patients are reversed by IM or IFN-α therapy
Significantly reduced blood iNKT cell frequencies have been reported in several forms of cancer, including hematologic malignancies [16-20]. Such a decrease in cell counts could not be evidenced in CML patients as assessed by flow cytometry analysis of PBMCs (Supporting Information Fig. 1), either in CP at the time of diagnosis or after molecular remission induced by IM (Supporting Information Fig. 2A). iNKT cells from CML-CP patients were both CD4+ and CD4− at ratios similar to those found in HDs (Supporting Information Fig. 2B and C). Only CML patients who achieved CR upon treatment with IFN-α had less iNKT cells than normal with a higher proportion of CD4+ T cells.
Knowing that iNKT cells are endowed with cytolytic activity enabling them to exert an antileukemic effect [19, 20], we examined whether their expression of perforin and CD95L was modified in CML patients relative to HDs (Fig. 1A and B, Supporting Information Fig. 3). This was clearly the case, since these molecules were either undetectable or expressed at very low levels during disease, while they were present in almost 40% iNKT cells in HDs, as measured by intracytoplasmic staining. However, in patients having achieved CR following IM or IFN-α therapy we observed a remarkable reversal of this deficiency, back to proportions occurring in HDs.
Because iNKT cell proliferation is reportedly arrested in several malignant diseases [17-20], we addressed the question whether it is also the case in CML-CP patients. As shown in Fig. 1C and in Supporting Information Fig. 4A, iNKT cells from CML-CP patients proliferated much less in response to their cognate ligand α-GalCer than their counterpart from HDs (mean ± SEM: 2.2 ± 0.8-fold versus 83.1 ± 16.9-fold, respectively). This result was in accordance with the reduced proportion of cycling cells among the CML-CP iNKT cell population in the same culture conditions (Fig. 1D, and Supporting Information Fig. 4B). Once again, the defect was corrected in CML-CR patients having received IM and IFN-α therapy, in which iNKT cells increased 100-fold or 20-fold, respectively, upon exposure to α-GalCer. In contrast with iNKT cells, conventional T cells conserved their proliferative potential in CML-CP patients. Indeed, Vβ2+ polyclonal MHC class II-restricted T cells responded normally to the super-antigen TSST1 (Supporting Information Fig. 5A and B), and PBMCs stimulation with anti-CD3/CD28 beads revealed also no significant difference between CML-CP patients and HDs in terms of cell growth (Supporting information Fig. 5C).
The anergic status of iNKT cells is partially reversed by IM in vitro implicating BCR-ABL TK activity
The fact that iNKT cells from CML patients emerge from their anergic status after IM therapy suggests that BCR-ABL activity contributes to the functional defect of iNKT cells. In accordance with this notion, in vitro culture of PBMCs with the TK inhibitor IM led to a significant improvement of the proliferation rate of iNKT cells from CML-CP patients with a 2- to 50-fold increase induced by IM (Fig. 2A). The fact that in the same experimental set up the number of iNKT cells recovered remained virtually unchanged in HDs reflects the selective sensitivity of CP patients to IM. However, it should be noted that exposure to IM in vitro restored the proliferative capacity of iNKT cells from CML-CP patients only partially. This result, together with our demonstration that both IM therapy and IFN-α therapy reversed the anergic status of iNKT cells suggests the involvement of complex immunoregulatory mechanisms rather than the sole direct inhibition of BCR-ABL activity in hematopoietic cells. Consistent with this view, the expression of Ph chromosome in T cells is commonly considered a very rare event. Also, we found that TK activity assessed by cytometry analysis of intra-cytoplasmic phospho-tyrosine levels was modified neither in iNKT cells nor in T cells from CML-CP patients, relative to their counterpart from HDs, while it was clearly enhanced among the CD34+ myeloid progenitor subset (Fig. 2B and C). Moreover, in vitro treatment with IM had no effect on phospho-tyrosine MFI in iNKT cells and T cells from both CML-CP and HDs, while it led to a substantial decrease in CD34+ cells from CML-CP patients. These results argue against a direct autonomous action of BCR-ABL in iNKT cells.
Anergy is associated withintrinsic and APC-dependent functional deficiencies of iNKT cells during CML-CP
The anergic status of iNKT cells during CP CML might be an indirect effect resulting from BCR-ABL expression in some APCs required for activation. Indeed, several functional defects (including inefficient actin polymerization and antigen processing) have previously been reported in myeloid DCs from CML patients . Consistent with this hypothesis, we found that α-GalCer-loaded ex vivo-generated monocyte-derived DCs obtained from CML-CP patients were unable to ensure optimal expansion of a HD-derived iNKT cell line, conversely to their counterpart from HDs (Supporting Information Fig. 6). These findings along with the fact that in vitro exposure to IM did not completely restore normal iNKT cell proliferation led us to assume that total reversal of the anergic state requires activation and/or maturation steps for which APCs are critical. Furthermore, we showed that intrinsic modifications underlie the BCR-ABL-dependent anergy of iNKT cells from CML-CP patients (Fig. 3, and Supporting Information Fig. 7). Indeed, these cells proliferated less in response to anti-CD3/CD28 beads, which provide TCR-dependent stimulation that bypasses APCs, than their HDs counterpart. In the same line of evidence, PLZF, a transcription factor requisite for the differentiation and maintenance of functional iNKT cells failed to be expressed in cells from CML-CP patients (Fig. 3B). Lastly, in accordance with recent studies in mice showing that PLZF-deficient iNKT cells do not acquire the capacity to secrete IL-4 [10, 11], we observed the same modification in iNKT cells from CML-CP patients upon stimulation with PMA/ionomycin (Fig. 3C). These findings are consistent with the observation that the proliferative response of iNKT cells from CML-CP patients could not be restored in vitro by adding APCs from HDs (data not shown). Importantly, proliferation in response to anti-CD3/CD28 beads (Fig. 3A) as well as expression of PLZF and IL-4 (Fig. 3B and C) returned to normal when iNKT cells were recovered from patients who had achieved CR after IM or IFN-α therapy, confirming that the functional impairment during CP CML is reversible.
Taken together, our results show that iNKT cells in CML-CP patients have specifically lost their ability to proliferate in response to their cognate ligand and to exert their cytolytic functions through perforin and CD95L expression. Further investigations are required to elucidate the exact mechanism accounting for the effects of IM and IFN-α on these parameters. Nonetheless, our results suggest that the dysfunctions of iNKT cells are partly caused by defective APCs that are probably rescued by IM or IFN-α therapy, a view that is in agreement with a previous study reporting that ex vivo-generated monocyte-derived DCs from IM-treated CML-CP patients can activate iNKT cell lines with similar efficiency than their HDs counterpart . As α-GalCer has already been used in clinical trials , our findings lead to propose that combination of IM with an immunotherapy using α-GalCer-pulsed auto DCs might become a new and useful strategy for targeting iNKT cells to restore their functions in CML-CP patients. However, the possibility that IM and/or IFN-α act(s) in part by enabling iNKT cells to exert their antileukemic effects should be investigated before using their ligands as a therapeutic strategy in CML.
The present study provides evidence that CP CML is associated with an acquired but potentially reversible defect in iNKT cells. Given the critical role of this innate cell population in tumor surveillance, our data reveal a new mechanism allowing leukemic cells to escape from an anti-CML immune response, and support the notion that iNKT cells play a role in controlling the disease in patients undergoing IM or IFN-α therapy.
Materials and methods
CML patients treated in the Oncology-Hematology and Cell Therapy Department and INSERM CIC-P-802 were included in this study. Patients were divided into three groups according to their treatment and to their clinical status as follows: (i) 22 untreated patients studied at diagnosis (100% Ph+ mitosis in cytogenetic analysis); (ii) 23 patients currently treated with IM and who have achieved either complete cytogenetic remission (0% Ph+ mitosis in cytogenetic analysis of at least 20 mitosis) and major molecular response (BCR-ABL/ABL ratio at 0.1% in International Scale or the status of undetectable BCR-ABL in peripheral blood by RT-qPCR (undetectable molecular residual disease, UMRD). These cytogenetic and molecular responses are referred to as CR, and (iii) 14 patients who were initially treated with IFN-α and developed a sustained CR maintained at least 3 years after treatment discontinuation. The 24 HDs were volunteers from the Pôle Biologie Santé (Poitiers, France). An informed consent was obtained from patients and HDs. The study was approved by the local Institutional Review Board (CHU La Milétrie). Blood samples were collected on heparin and PBMCs were isolated and cultured as previously described .
A total of 10 ng/mL IL-2 and 10 ng/mL IL-15 (R&D systems, Abingdon, UK), 100 ng/mL α-GalCer (Kirin Brewery, Gunma, Japan), 10 ng/mL TSST1 superantigen (Sigma-Aldrich, Lyon, France), and 0.5 or 1 μM IM (Glivec™, Novartis, Rueil Malmaison, France) were used in this study. Anti-IFN-γ-FITC, anti-IL-4-PE, anti-perforin-FITC, anti-CD34-PE, anti-CD3-PE-Cy5.5, anti-CD3-PerCP-Cy5.5, anti-CD4-FITC mAb, anti-iNKT-PE 6B11 clonotype, and anti-CD95L-allophycocyanin mAb were purchased from BD Biosciences (Le Pont de Claix, France) and Miltenyi Biotec SAS (Paris, France), respectively. Anti-CD8-PE-Cy7 was purchased from eBioscience (Paris, France). Anti-Vα24-FITC, anti-Vβ2-FITC and anti-Vα24-PE, and anti-phosphotyrosine mAb (mouse IgG2b) and FITC polyclonal Ab anti-mouse IgG2b were purchased from Beckman Coulter (Villepinte, France) and Southern Biotechnology (Birmingham, AL), respectively. Anti-PLZF mAb (clone D-9, Santa Cruz Biotechnology, Santa Cruz, CA, USA) was coupled in our laboratory with Alexa Fluor 647 with a kit from Invitrogen (Cergy Pontoise, France). Allophycocyanin-conjugated CD1d-tetramer loaded with α-GalCer (α-GalCer-loaded CD1d TT-allophycocyanin) was provided by the NIH tetramer facility (Atlanta, GA, USA).
Flow cytometry analysis
Cells were analyzed by four-color flow cytometry (FACScalibur™ and CellQuest™ software, BD Biosciences) or six-color flow cytometry (FacsCanto II™ and FacsDiva™ software, BD Biosciences), and data were reanalyzed with FlowJo™ (Treestar, Ashland, OR). For in vitro experiments, dead cells were excluded by propidium iodide staining. At least 105 viable cell events were acquired in the peripheral blood lymphocyte gate (assessed by forward and side scatter parameters). Positive staining for each marker was determined by comparison to appropriate isotype-matched negative controls or nonloaded CD1d-TT.
We gratefully acknowledge the technical assistance of S. Noel. The authors are especially indebted to Elke Schneider and Michael Melkus for critically reviewing the manuscript and thank the NIH tetramer core facility for the α-GalCer-loaded CD1d-TT. Financial support came from INSERM, CHU de Poitiers, Université Paris Sud 11 (AAP « Attractivité » 2011), Université de Poitiers, Ligue contre le Cancer (Comité de la Vienne et Comité du Val-de-Marne), ARI-PC (Association pour la Recherche en Immunologie-Poitou-Charentes), and Ministère de la Recherche.
Conflict of interest
The authors declare no financial or commercial conflict of interest.