|Home | About | Journals | Submit | Contact Us | Français|
Immune effector cells such as T and NK cells can efficiently eliminate tumor cells. However, when activating oncogenic signaling pathways or protective mechanisms against cell death are active, immune cells can also confer therapy resistance. Here, we analyzed the role of activated T and NK cells and released cytokines on tyrosine kinase inhibitors imatinib and nilotinib – mediated apoptosis induction and proliferation of chronic myelogenous leukemia (CML) cells. Incubation of CML cells with activated, but not with resting CD3+ T cells or with activated NK cells significantly inhibited TKI-induced apoptosis induction in CML cells as quantified by nuclear fragmentation assays. Transwell experiments revealed a critical role for T or NK cell-derived cytokines for CML cell protection. Accordingly, CML cells treated with IFNγ also showed a clearly reduced sensitivity to TKI-mediated cell death induction and inhibition of proliferation. In contrast, IFNα or other pro-inflammatory mediators and cytokines, such as TNFα and GM-CSF did not impair TKI-induced apoptosis in CML cells. On a molecular level, IFNγ-exposed CML cells showed a significantly reduced caspase-3 activation and PARP-1 cleavage as well as an increased expression of anti-apoptotic molecule xIAP. Finally, IFNγ diminished TKI-induced downregulation of Jak-2 and STAT-5 phosphorylation and increased nuclear expression of RUNX-1, which may at least in part contribute to the reduced sensitivity to TKI effects. Our results demonstrate that IFNγ released by activated T or NK cells may interfere with the therapeutic effects of TKI in CML. Our findings may have important implications for the understanding of inflammation-mediated BCR-ABL independent resistance mechanisms.
The interaction of malignant cells with the infiltrating immune cells and released pro-inflammatory mediators can either mediate tumor destruction with subsequent elimination or induce cancer cell resistance to various treatment approaches. In several pre-clinical animal models as well as in clinical studies it was reproducibly demonstrated that the composition of infiltrating T lymphocyte and macrophage subpopulations1,2 as well as the repertoire of released soluble immunomodulatory mediators may dramatically impact the clinical outcome. This may at least in part be explained by the effects of cytokines on a plethora of tumor-biologically relevant processes, such as tumorigenesis, angiogenesis and induction of resistance to anticancer agents.3,4
Chronic myelogenous leukemia (CML) is a myeloproliferative disease, characterized by the recurrent genetic abnormality (i.e. the Philadelphia chromosome (Ph)) leading to the constitutively active BCR-ABL tyrosine kinase. The latter drives the Ph-positive clone by conferring growth advantage over normal hematopoiesis, reduced apoptosis and an altered interaction with the stromal compartment of the bone marrow. Since the introduction of the prototypic tyrosine kinase inhibitor (TKI) imatinib mesylate (IM), the overall survival of patients with Ph-positive CML or ALL was substantially improved.5 Despite the high clinical activity of IM, there is a considerable proportion of patients developing resistance to the drug. In case of BCR-ABL point mutations, with the exception of the T315I mutation, second generation TKIs such as nilotinib (NI) or dasatinib are available. Recently, these compounds also proved to be superior to IM in randomized phase III clinical trials as first-line treatment, thus leading to approval of the drugs for up-front therapy of CML in chronic phase.6,7 In addition to point mutations as resistance conferring events, other molecular mechanisms leading to BCR-ABL independent IM- or NI resistance are increasingly acknowledged.8,9 As an example, reduced drug influx by impaired Oct-1 expression10 or increased drug efflux by PgP11 may contribute to resistance. However, BCR-ABL independent resistance mechanisms are poorly understood so far. Notably, BCR-ABL independent resistance is predominantly seen in newly diagnosed patients without detectable point mutations indicating that the interaction of CML cells with non-malignant cells of the microenvironment as well as with the soluble mediators and extracellular components may contribute to treatment failure. Previously, it was demonstrated that autocrine and paracrine secretion of GM-CSF can reduce the anti-leukemic activity of TKIs in CML patients.12 In this study, we analyzed the impact of T lymphocyte-derived and NK cell-derived mediators on the CML cell sensitivity toward TKI effects. Using CML cell lines and primary CML cells, we found that IFNγ released by activated T and NK cells impedes the therapeutic activity of TKIs by reducing TKI-induced induction of apoptotic cell death and inhibition of TKI-mediated growth inhibition.
Imatinib and Nilotinib were obtained from Novartis (Basel, Switzerland). Imatinib was dissolved in dimethylsulfoxid (DMSO) and distilled water (AMPUWA) in a 1:1 ratio and aliquots with 2 mg/ml stock solution were stored at −20 °C. Nilotinib was dissolved in DMSO as a 10 mM stock solution and stored at −20 °C. IFNγ was purchased from R&D systems, Minneapolis, USA and solved in PBS/BSA as a 10 µg/mL stock solution and stored at −20 °C. Phorbol myristate acetate and Ionomycin (PMA/Iono) (both purchased from Sigma Aldrich, Hamburg, Germany) for the activation of T cells were dissolved in DMSO as stock solutions of 0.1 mg/mL and 1 mg/mL and stored at −20 °C. Cells were cultured in RPMI1640 Glutamax Media supplemented with 10% heat-inactivated fetal calf serum, 100 units/mL penicillin, and 100 µg/mL streptomycin, all purchased from Invitrogen, Karlsruhe, Germany. Freshly isolated NK cells were cultivated overnight with 1000U/mL IL-2 (IL-2, Proleukin, Novartis, Basel, Switzerland). In addition, to induce IFNγ production NK cells were stimulated with IL-12 (Immunotools, Friesoythe, Germany) and IL-18 (Biozol, Eching, Germany) during transwell assays. IFNγ production was analyzed by intracellular cytokine staining and subsequent flow cytometry (CD3-APC (biolegend, San Diego, USA) CD56-PE, IFNy-FITC (both Miltenyi Biotec, Bergisch-Gladbach, Germany)).
Primary CML cells were acquired from patients with diagnosed CML after written informed consent and approval by the local ethics committee at the department of hematology and oncology, University of Bonn (265/08). Peripheral blood samples were collected from CML patients and stored at −80 °C. The obtained blood samples were then used for preparation of PBMNCs by Biocoll (Biochrome, Berlin, Germany) density gradient centrifugation as described previously.13,14 All samples were collected at time of diagnosis and had no TKI treatment before. Patient sample 4 was collected after 2 d treated with hydroxyurea (see Table 1).
The tumor cell lines K-562 (CML in blast crisis; American Type Culture Collection), Meg01 (CML in megakaryocytic blast crisis; Deutsche Sammlung von Mikroorganismen und Zellkulturen), BV173 (CML in blast crisis, human B cell precursor leukemia; Deutsche Sammlung von Mikroorganismen und Zellkulturen), Kyo-1 (CML in myeloid blast crisis; Deutsche Sammlung von Mikroorganismen und Zellkulturen) and the HLA-*A0201 transfected cell line K-562-A2 (kindly provided from Dr. T. Wölfl, Universitätsklinikum Mainz, Germany) were all cultured in RPMI1640 Glutamax. All used CML cell lines are BCR-ABL positive and are highly sensitive to imatinib and nilotinib. All cell lines are tested twice a year by PCR for BCR-ABL expression. The K-562-A2 cells were additionally tested for the expression of HLA-A2 by flow cytometry.
Cell isolation from PBMNCs Isolation of CD3+ cells was performed using MACS CD3 MicroBeads (Miltenyi Biotech, Bergisch-Gladbach, Germany) strictly according to the protocol provided by the manufacturer. NK cells were isolated by magnetic bead separation (NK Cell Enrichment Kit; StemCell Technologies, Vancouver, Canada).
DNA fragmentation in apoptotic nuclei was measured by the method of Nicoletti et al.15 Cells were lysed with hypotonic buffer (1% sodium citrate, 0.1% Triton X-100, 50 µg/mL propidium iodide) and apoptotic nuclei were released. Stained apoptotic nuclei were subsequently analyzed by flow cytometry on a Cytomics FC 500 (Beckman Coulter, Krefeld, Germany) using CXP analysis software. Nuclei to the left of the 2N peak containing hypodiploid DNA were considered apoptotic.
CML cell lines or primary CML cells were seeded (105) into a 24-well plate for 24 h respectively 6 h and treated once with TKIs in different concentrations. In addition, either PMA/ionomycin- or unstimulated CD3+ T cells (in an E:T ration of 2:1) as well as naïve or cytokine activated NK cells (in an E:T ratio of 1:2) were seeded into a transwell chamber (4 µm pore size excluding cell transmigration and direct interaction of T or NK and leukemic cells, BD Falcon, BD Biosciences, Heidelberg, Germany) and added into the 24-well plate. CML cell lines or primary CML cells were pre-incubated for 8 h with the CD3+ T or NK cells before subsequent exposure for 24 h with TKIs. After 24 h, nuclear fragmentation of cell lines or primary CML cells was analyzed via the method of Nicoletti by flow cytometry.
HLA-ABC expression was measured by staining HLA-A2 transfected K-562 cells, Meg01 cells, Kyo-1 cells as well as primary patient samples with FITC-conjugated mouse antibody (anti-HLA-ABC-FITC Clone B9.12.1, Beckman Coulter) raised against MHC-class I. Cells were harvested after incubation, washed once in PBS, stained with the antibodies and incubated for at least 15 min at room temperature in the dark. Measurement was done by flow cytometry on a Cytomics FC 500 (Beckman Coulter) using CXP analysis software. Percentages of HLA-class I positive cells were measured.
CML cells (1 × 105 responding cells) were cultured in 96-well flat-bottomed microplates (Nunc, Wiesbaden, Germany) in the presence of IM or NI with or without IFNγ for 2 d. On day 3, H3+-thymidine (0,0148 MBq/well; GE Healthcare, München, Germany) was added for 18 h to each well. After incubation, the incorporation was measured by MicroBeta® TriLux counter (Perkin-Elmer, Rodgau-Jügesheim, Germany).
Nuclear extracts were prepared from CML cell lines or from primary CML cells as described previously(Appel et al., 2005b). For whole-cell lysates, a buffer containing 0.1 M phosphate, 0.1 mM ethylenediaminetetraaceticacid (EDTA), 1% Triton X-100, 2 mg/mL aprotinin, and 1 mM sodium orthovanadate was used. Protein concentration was determined using a bicinchoninic acid (BCA) assay (Pierce, Perbio Science, Bonn, Germany). Western blot analysis for protein detection was performed as described before.16-18 For detection of nuclear localization of Runx1, at least 20 µg of protein were separated on a polyacrylamide gel and transferred to a nitrocellulose membrane (GE Healthcare). The membrane was blocked with TBST containing 4% Slimfast solution for 1 h. Subsequently the blot was probed with a monoclonal antibody against Runx1 (H-65, Santa Cruz Biotechnology, Santa Cruz, CA, USA). Ponceau S staining of the membrane was performed to confirm that equal amounts of proteins had been loaded onto the gel.
For analysis of the activation and expression status of pro-Caspase-3, PARP-1, JAK2, STAT5, ERK1 (all purchased from Santa Cruz Biotechnology, Santa Cruz, CA, USA) as well as p38, pERK1/2 (Cell Signaling, Frankfurt a.M., Germany) and hILP/xIAP (BD Transduction Laboratories, Heidelberg, Germany), 20–40 μg whole-cell lysates were separated on a polyacrylamide gel and subsequently transferred on a nitrocellulose membrane. The blot was probed with monoclonal antibodies against Caspase-3, PARP-1, hILP/xIAP, phosphoJak2, JAK2, p38, phosphoSTAT5 and STAT5. GAPDH (10B8, Santa Cruz Biotechnology) or p38 was used as a loading control. Protein bands were detected using an enhanced chemiluminescence (ECL) kit (GE Healthcare).
CML cell lines and primary CML cells were seeded into a 6-well plate for up to 24 h. IFNγ or IFNα were incubated for 1 h before TKI treatment. After 24 h CML cells were harvested an directly added into Trizol→ reagent (Invitrogen) for the isolation of total RNA. RNA preparation was performed as described by the manufacturer. A quantity of 1 µg total RNA was transcribed into cDNA using the Transcriptor First Strand cDNA Synthesis kit (Roche) with random hexamer primer. Subsequently, Real Time quantitative PCR was performed using Runx1 and GAPDH primer set from RealTimePrimers.com (Real Time Primers). PCR reaction was conducted on the LightCycler 480 system (Roche) for 50 cycles with the following program: 1 hold 95 °C (10 min); 50 cylces of 95 °C (10 s), 85 °C (45 s), 72 °C (30 s); 1 hold 72 °C (5 min).
Analyses of significance of different treatment groups were calculated with the one way analysis of variance (ANOVA) and Bonferroni´s multiple comparison test using Prism 5 software (Graphpad Software). Transwell assays, apoptosis measurement, and proliferation were performed at least 3 times. Western blot analyses and analyses of HLA-molecules expression were performed at least 3 times. Representative experiments are shown.
To evaluate the effect of T cells on the anti-leukemic activity of TKIs, transwell assays were performed (Fig. 1). Using PBMNCs from healthy donors, CD3+ cells were sorted and purified by MACS and either stimulated up to 24 h with PMA/ionomycin or left untreated. CML cells were seeded in 24-well plates and treated with different concentrations of IM or NI. The addition of activated CD3+ T cells in the upper transwell chambers to the cell line K-562 led to a significant inhibition of TKI-induced cell death. In contrast, the unstimulated CD3+ cells had no effect on the TKI-mediated nuclei fragmentation in K-562 cells (Fig. 1A and B). Similarly, TKI-induced cell death of primary CML cells was not affected by the unstimulated CD3+ cells, but significantly decreased upon co-culture with activated CD3+ cells (Fig. 1C and D). As demonstrated in Fig. 1E activated T cells produced high amounts of IFNγ.
To further expand our findings on activated T cells, we investigated the effects of IL12/IL18 –activated NK cells on TKI treated CML cells. NK cells were seeded in the upper transwell chambers as described above. CML cells were seeded in 24-well plates and treated with different concentrations of IM or NI. The addition of IL12/IL18-activated NK cells in the transwell chamber to the TKI-treated CML cell lines Meg01 (Fig. 2A and B), Kyo-1 (Fig. 2C and D) or K-562 (Fig. 2E and F) as well as to primary CML cells (Fig. 2G and H) led to a highly significant downregulation of TKI-mediated apoptosis induction in these cells, comparable to the results obtained with the activated T cells. In line with the results from unstimulated CD3+ T cells, non-activated NK cells had no effect on the TKI-mediated nuclear fragmentation.
In the next set of experiments, we investigated the potentially secreted factors mediating protective effects of activated T or NK cell-derived soluble factors. IFNγ is among the major cytokines secreted by activated T and NK cells. To analyze a potential functional role of IFNγ in protection of CML cells from TKI-induced cell death, the cytokine was titrated in vitro to K-562 (Fig. 3A), Meg01 (Fig. 3B), or primary CML cells (Fig. 3C, D and E), which were then further exposed to IM or NI. In line with the results from the transwell experiments, addition of IFNγ resulted in a significantly reduced induction of apoptosis by either IM or NI (Fig. 3).
Interestingly, in contrast to IFNγ, IFNα showed no effects on the rate of cell death compared to IM treatment alone (Fig. 4E). Other pro-inflammatory cytokines tested, such as TNFα, IL-6, IL-12, GM-CSF or soluble CD40 ligand did not affect the sensitivity of CML cells to the pro-apoptotic effect of TKIs (data not shown).
In addition to nuclear fragmentation, cleavage of PARP-1 as a product of Caspase-3 activation also indicates apoptosis induction. In line with the results above, pre-incubation of CML cells with IFNγ resulted in a reduced PARP-1 cleavage and diminished pro-caspase-3 expression (Fig. 3F).
As a potential mechanism mediating the anti-apoptotic effects of IFNγ, we found that the TKI-mediated down-regulation of the anti-apoptotic protein xIAP (also known as hILP, Fig. 3G) was markedly inhibited by IFNγ, while no significant regulation of either survivin, BCLxL or BCL2 expression could be detected (data not shown). In summary, these findings indicate that the addition of IFNγ results in a reduced TKI-induced apoptosis in CML cell lines.
An increased rate of proliferation in CML cells may also lead to reduced TKI effectiveness. The addition of IFNγ to the TKI-exposed K-562 (Fig. 4A and B), BV173 (Fig. 4C) or Meg01 (Fig. 4D) cells resulted in a significant reduction of the anti-proliferative TKI-effects. We next analyzed the JAK/STAT-pathway, as it is constitutively activated in leukemic cells.19-22 As shown in Fig. 5, pre-treatment with IFNγ, in comparison to treatment with TKIs alone, resulted in an increased expression and phosphorylation of JAK2 protein (Fig. 5A and B) and phosphorylation of STAT5 (Fig. 5C). In addition, we found that pre-exposure of K-562 cells to IFNγ resulted in increased activation of the MAP-kinase ERK1/2 when compared to TKI only exposed cells (Fig. 5D). In contrast, we could not detect any effect of IFNγ on p38 expression and phosphorylation levels (data not shown). As previously described, the transcription factor Runx1 (also known as AML-1) acts as a modulator of the cellular response toward IM in vitro and in vivo.23 In addition, it supports the generation of trisomy 21 in Ph+ leukemia.24 IFNγ markedly affected Runx1 expression in TKI-exposed CML cells, as a clear up-regulation of Runx1 was seen in all tested CML cell lines as well as in primary patient sample preincubated with IFNγ when compared to sole TKI-exposure (Fig. 5E-I). Of note, we could not detect any effect of IFNγ on nuclear expression of the transcription factors IRF-3, IRF-8 or RelB (data not shown).
Taken together, these results indicate that IFNγ released by activated T cells can reduce the inhibitory effects of IM and NI on proliferation and viability of CML cells.
In the next set of experiments, we aimed to analyze whether activated T cells not only affect TKI sensitivity of CML cells, but may also potentially modify immunological accessibility of the leukemic cells via regulation of MHC class I expression. To address this, we used a HLA-A2-transfected K-562 cell line (Fig. 6A), the CML cell lines Kyo1 (Fig. 6B and C) and Meg01 (Fig. 6D and E) as well as primary leukemic cells (Fig. 6F and G). Treatment of CML cell lines with IM or NI showed a concentration-dependent downregulation of the HLA-class I expression levels. Addition of IFNγ inhibited the effect of the TKIs on the MHC-class I expression (Fig. 6). In contrast, IFNα could not antagonize the activity of TKIs (Fig. 6A–E). Similarly, the incubation of primary cells from a patient in chronic phase CML with IFNγ but not with IFNα resulted in an upregulation of HLA-class I molecules which was not affected by the addition of TKIs (Fig. 6F and G).
CML is characterized by the constitutive activity of the BCR-ABL fusion protein which causes malignant transformation, deregulation of proliferation, apoptosis and adhesion control pathways.25,26 Since the introduction of IM in the last decade, the overall survival of patients with CML was significantly prolonged.5,27,28 Despite this breakthrough in therapeutic options and the development of second and third generation TKIs such as NI, dasatinib or ponatinib,29,30 the CML is still uncurable, with the exception of allogeneic stem cell transplantation. In addition, some patients develop resistance to TKIs which remains until now a great challenge in CML therapy. There are several known mechanisms mediating TKI-resistance, most of them tending to different BCR-ABL point mutations. On the other hand, there are several reports showing that newly diagnosed CML patients or patients without any point mutations are resistant to IM or other TKIs, suggesting that there may be other BCR-ABL independent mechanisms mediating TKI-resistance. GM-CSF has been described by Wang et al.12 as one possible cytokine mediated mechanism of resistance development. In addition, it was shown that the src-related Lyn kinase is able to mediate BCR-ABL-independent resistance.9 Both described pathways are mediated via the JAK2/STAT5 activation. Furthermore, the JAK/STAT signaling in CML, mainly the constitutive activation of STAT5, can obviously be activated by both, BCR-ABL dependent and independent pathways, as previously shown by Chai et al. for some BCR-ABL expressing cell lines and all tested peripheral blood samples of CML patients.31 Another important factor in the constitutive activation of this pathway is the induction via the IL-3 receptor (IL-3R).32 The transduction of signals via this receptor requires conformational changes in the cytoplasmic domain of the β-subunit which then promotes binding and activation among others of JAK2 and STAT5.33
In our study we found that IFNγ, a major cytokine secreted by activated T lymphocytes and NK cells, can affect the viability and survival of CML cells. Exposure of CML cells to activated T or NK cells resulted in a reduced TKI-induced nuclei fragmentation. In further experiments we were able to show that IFNγ secreted by activated T and NK cells is contributing to this effect and the addition of pure IFNγ to the TKI-treated cells decreased the nuclei fragmentation.
Similarly, IFNγ caused a significant inhibition of the TKI-mediated anti-proliferative effect in CML cell lines.
Using western blot analyses, we found that the reduced apoptosis induction and the increased proliferation after IFNγ exposure is mediated via the JAK2/STAT5 and ERK1/2 pathways. Interestingly, we found that the antiapoptotic protein xIAP seems to be involved in the effects of IFNγ while other antiapoptotic proteins such as survivin or Bcl2 remained unaffected. Another interesting and importing finding was the affection of the transcription factor Runx1 by IFNγ. We could show that IFNγ upregulates the nuclear expression of RUNX1 which then mediates proliferation signals and survival signals that further cause reduction of apoptosis induction. In addition, it has been described by Miething et al.23 that RUNX1-expressing cells are protected from IM-induced apoptosis. Beyond that, it was shown that IM treatment selected for RUNX1 expressing cells suggesting an important role of RUNX1 in CML disease persistence.
These results represent a novel mechanistic loop of resistance induction due to a paracrine release of cytokines by immune cells and confirm the important role of the interaction of malignant cells with the microenvironment and soluble mediators on the efficiency of therapeutic approaches.
In the last set of experiments, we examined the role of IFNγ on the HLA class I molecule expression. Interestingly, we found that the addition of IFNγ led to an increased upregulation of HLA class I molecules on the surface of TKI-treated CML cells, in contrast to IFNα.
Our results on the importance of IFNγ on CML biology treatment resistance are further supported by a recent observation demonstrating that IFNγ released by transferred antigen specific CTL can promote the proliferation of chronic myeloid leukemia stem cells and induce resistance to the killing of leukemic cells in a mice model of CML.34 In line with these results, we could detect a significant effect on the TKI-induced anti-proliferative effect in human CML cell lines as well as in primary CML cells derived from patients in a chronic phase of the disease after IFNγ treatment in our experiments, despite these cells probably do not contain leukemic progenitors.
T or NK cell derived cytokines such as IFNγ were shown to have a direct and critical effect on the survival and maintenance of hematopoietic stem cells and to induce the differentiation and proliferation of the granulocyte-monocyte and erythropoietic lineages.35,36 Other studies have shown that IFNγ directly promotes the expansion of lin- Sca-1+ c-kit+ cells.37 Thus, future studies will have to address the possible role of IFNγ on the resistance induction to different treatment approaches in other hematological malignancies.
In summary, our results demonstrate that cytokines such as IFNγ released by activated T lymphocytes or NK cells can interfere with the therapeutic effects of IM or NI and reduce the activity of these drugs. Furthermore, these findings are important for the design of clinical protocols combining TKI treatment with immunotherapeutic approaches.
No potential conflicts of interest were disclosed.
We thank Kati Riethausen and Solveig Daecke for their excellent technical assistance.
This work was supported by grants from DFG (PB), BONFOR (SAEH, AH, KS and ARK), and Deutsche Krebshilfe (PB).