c-Myc regulates expression of NKG2D ligands ULBP1/2/3 in AML and modulates their susceptibility to NK-mediated lysis

Acute myeloid leukemia (AML) is a hematologic malignancy characterized by proliferation of malignant precursors of the myeloid lineage coupled with impaired differentiation of normal hematopoietic progenitors.¹  Chemotherapy is the first line treatment against most leukemia disorders, and cytarabine (cytosine arabinoside) has been one of the most widely used chemotherapy agents against AML blasts for more than 30 years.^2-6  Although cytarabine is an efficient antileukemic agent for AML and other leukemias,⁷  emergence of drug resistance due to prolonged chemotherapy in most patients is a major obstacle.^8,9  Accumulating evidence indicates that the acquisition of drug resistance enhances the sensitivity of leukemic blasts to cytotoxic cells of the immune system. However, other reports indicate decreased susceptibility of leukemic cells to cytotoxic cells.^10-18 

Allogeneic bone marrow transplantation is the only curative treatment of many intermediate and high-risk leukemias. Recent studies suggest that immunotherapy may continue to be an effective approach for patients with leukemia,^19-21  and emerging strategies are currently under investigation based on adoptive transfer of natural killer (NK) cells. NK cells are a component of an innate immune system that play important roles as first line-defenders in the host response to tumors and infections, as well as in transplant rejection and in the development of tolerance.^22-27  Due to their strong ability to target tumor cells, NK cells have been described as promising effectors for adoptive immunotherapy of cancer.²⁸  It is well established that NK cell activity is regulated by a balance between inhibitory and stimulatory signals that are transmitted by cell-surface receptors after interaction with their respective ligands on target cells.^29,30  NK group 2, member D (NKG2D) is one of the activating receptors expressed by NK cells, γ/δ T cells, and activated CD8⁺ T cells in humans.^31-33  Several ligands for this receptor have been identified in humans, including major histocompatibility complex (MHC) class I-related chain A (MICA), MICB, and UL16-binding proteins (ULBP) 1/2/3/4/5. These ligands are abundantly expressed by tumor cells, rendering these cells susceptible to NK-cell–mediated cytotoxicity.^32,34-36 

While the functional role of NKG2D is well established,³⁷  the regulation of its ligands (NKG2DL) remains only partially understood. Various molecular pathways, including extracellular signal-regulated kinase (ERK), AKT, p53, and signal transducer and activator of transcription 3 have been reported to play a regulatory, both at the transcriptional or posttranscriptional level.^38-48 

In this study, we investigated the molecular basis of cytarabine resistance in AML cells. We found that these cells exhibited increased susceptibility to NK lysis that correlates with an increase in c-Myc induction and the subsequent upregulation of ULBPs. Therefore, this study reveals a new regulatory mechanism of ULBPs in AML involving the c-Myc pathway. This knowledge could help predict the efficacy and response to NK-cell–based therapy, and allow for better designing of NK-based immunotherapy.

Human AML cell lines (KG-1 and HL-60) were grown in RPMI 1640 medium supplemented with 10% fetal calf serum (Seromed) and 1% penicillin-streptomycin. Human NK cell lines were grown in RPMI 1640 medium supplemented with 10% fetal calf serum, 1% penicillin-streptomycin, and 300 IU/mL IL-2. Cytarabine-resistant KG-1 and HL-60 cell sub lines were established by exposing parental cells to increasing concentrations of the drug. All experiments were performed using cytarabine-resistant cells subcultured at 7 day intervals without further addition of the drug. Approval for these studies was obtained from the Gustave Roussy Cancer Campus Institutional Review Board. Informed consent was provided in accordance with the Declaration of Helsinki.

Monoclonal antibodies (mAbs) directed against ULBP1/2/3 were purchased from R&D Systems. Anti-CD107a coupled to CyChrome mAb was provided by Becton Dickinson. mAbs-recognizing MICA, MICB, and MHC-I was obtained from BioLegend. Small-molecular-weight c-Myc inhibitor (10058-F4; [Z,E]-5-[4- ethylbenzylidine]-2-thioxothiazolidin-4-one) was purchased from Calbiochem (San Diego, CA). The ERK inhibitor (PD98059) was obtained from GIBCO (Invitrogen, Cergy Pontoise, France) and the AKT inhibitor (MK-2206) was from Selleck Chemicals.

Abs against c-Myc was purchased from Santa Cruz Biotechnology. Abs against total ERK, phospho-ERK (Thr202/Tyr204), total AKT, and phospho-AKT (ser-473) were from Cell Signaling Technology. Abs against β-actin was from Sigma-Aldrich.

Specific c-Myc small interfering RNA (siRNA) (sc-29226) and negative control siRNA (sc-37007) were purchased from Santa Cruz Biotechnology. siRNA was transfected into HL-60 or KG-1 cells by electroporation (Amaxa, Gaithersburg, MD) following the manufacturer’s instructions. Briefly, 2 × 10⁶ cells were electroporated in 100 μL nucleofector solution (Amaxa Reagent V for HL-60 cells and Amaxa Reagent R for KG-1 cells) containing 50 pmol of each siRNA using the preselected Amaxa program T-019 for HL-60 cells and Amaxa program V-001 for KG-1 cells. siRNA transfected cells were plated in a 6-well plate with 5 mL supplemented RPMI 1640 medium for 72 hours and harvested for further experiments.

Total RNA was extracted from cell samples with TRIzol solution (Invitrogen). DNase I-treated 1 μg of total RNA was converted into complementary DNA by using TaqMan Reverse Transcription Reagent (Applied Biosystems), and messenger RNA (mRNA) levels were quantified by SYBR Green qPCR method (Applied Biosystems). Relative expression was calculated by using the comparative C_t method (2-ΔC_t). Primer sequences are available upon request.

Leukemia cells seeded onto flat-bottom 96-well plates were exposed to different concentrations of the drug for 48 hours. Viability of cell was measured with a modified 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide assay (Sigma-Aldrich, Saint-Quentin, France). The percentage of cell viability was calculated as follows: percentage of viability = (A1/A0) × 100, where A1 and A0 represent absorbance obtained, respectively, for treated and untreated cells.

Tumor cells were washed in phosphate-buffered saline and lysed in plates with lysis buffer (62.5 mM Tris-HCl [pH 6.8], 2% weight/volume sodium dodecyl sulfate, 10% glycerol, 1 mM sodium orthovanadate, 2 mM phenylmethylsulfonyl fluoride, 25 mM leupeptin, 5 mM benzamidine, 1 mM pepstatin, and 25 mM aprotinin). Lysates were sonicated on ice, resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (60 mg/lane), and transferred onto nitrocellulose membranes. After incubation in blocking buffer, the membranes were probed overnight at 4°C with the indicated Abs. The labeling was visualized using peroxidase-conjugated secondary Abs and an enhanced chemiluminescence kit (Amersham International). Blots were scanned and processed by Adobe Photoshop 7.0 software.

The cytotoxic activity of the NK cells was measured by a conventional 4-hour ⁵¹Cr-release assay by using triplicate cultures in round-bottom 96-well plates. Different effector to target (E:T) ratios were used as previously described.⁴⁹ 

NK cells were cocultured with laser-assisted microdissection target cells for 4 hours at 2:1 E:T ratio. Degranulation of NK cells was analyzed by flow cytometric analysis of CD107a expression. For c-Myc inhibition experiments, the leukemia cells were treated with c-Myc inhibitor overnight and then assayed by flow cytometry.

Flow cytometry analysis was performed using a FACSCalibur and LSRII flow cytometer. Data were processed using CellQuest software (BD Biosciences).

NK cells and leukemia cells were incubated for 1 hour, respectively, with red cell tracer and green cell tracer. Cells were then washed and put in coculture (at 2:1 E:T ratio) for 1 hour, and then immediately analyzed by flow cytometry.

For ChIP experiments, 3.10⁷ untreated or treated KG-1 cells and primary blasts were harvested and processed with the EpiTect ChiP OneDay Kit (Qiagen) according to the manufacturer’s instructions. The amount of input chromatin per ChIP was 10 µg. For c-Myc–specific ChIP, chromatin was immunoprecipitated with 2 µg monoclonal antibody (sc-40; Santa Cruz Biotechnology) or 2 µg of a respective isotype control mAb (sc-2031; Santa Cruz Biotechnology). The relative amounts of chromatin immunoprecipitated by either isotype control or c-Myc mAb was determined by real-time PCR with specific primers for c-Myc binding site (c-Myc BS) in ULBP1/3 gene promoter, respectively (TGCTATGTC CATGGGACCTG, GCAGGCACCACGTGCCATAGC) purchased from Qiagen.

Data were analyzed with GraphPad Prism. A Student t test was used for single comparisons. A P value of .05 was considered statistically significant.

The cytarabine-resistant AML cell lines, KG-1 and HL-60 cell sublines, were established by exposing parental cells to increasing concentrations of the drug. Results depicted in Figure 1A show that cell viability is correlated to drug concentration in sensitive cell lines. When cytarabine was used at the different indicated concentrations, it had no effect on the viability of resistant cells. To examine the susceptibility of cytarabine-resistant KG-1(R) and HL-60(R) cells to NK-mediated cytotoxicity, we used the NK-cell line (NKL) that selectively expresses NKG2D but does not express most other receptors. Cytotoxicity by NK cells was determined by a conventional 4-hour ⁵¹Cr release assay at different E:T ratios in resistant and parental cells. As shown in Figure 1B, both cytarabine-resistant cell lines were more susceptible to NK cell lysis compared with parental (ie, cytarabine-sensitive) cells, in all tested ratios.

To elucidate the mechanism underlying the increased susceptibility to NK cells by cytarabine resistant targets, we examined the expression pattern of NK cell receptor ligands by parental and cytarabine-resistant cell lines. Our results showed that the expression of ULBP1, 2, and 3 was up-regulated on the surface of KG-1(R) and HL-60 (R) cell lines, while the expression of MICA/B, poliovirus receptor, and MHC-I was not affected (Figure 2A). We next performed the degranulation assay by measuring CD107a expression on NK cells after 4-hour coculture with leukemia cells at 2:1 effector/target ratio. The results showed that the expression of CD107a increased on NK cells after coculture with resistant cells as compared to sensitive cells (Figure 2B). To examine the role of NKG2D and its ligands in the acquisition of leukemic cell resistance to cytarabine, we first incubated leukemic cells with anti-NKG2D mAb. Data revealed that blocking this receptor resulted in the inhibition of the cytotoxic activity of NK cells (Figure 2C). Taken together, these results suggest that induction of NKG2D ligand expression on cytarabine-resistant AML cells renders them susceptible to NK-cell–mediated killing.

To investigate the molecular mechanisms associated with the enhancement of target cell susceptibility to NK cells following acquisition of cytarabine resistance, we analyzed the expression of a series of antiapoptotic and proapoptotic genes using an apoptosis-related array of 94 genes. In the course of these studies, the gene differentiation analysis was performed using noncloned cell lines in order to be closer to what is observed in patient blasts. The results revealed no common expression changes in proapoptotic or antiapoptotic genes between cytarabine-resistant cell lines and their sensitive equivalents (Figure 3A).

It has been shown that ERK and AKT pathways can modulate NKG2D ligand expression. Furthermore, as cytarabine induces apoptosis in AML cells through cell-cycle arrest, we investigated the putative involvement of c-Myc, which is a major cell-cycle modulator in leukemia disorders. Therefore, we assessed the expression of AKT, ERK, and c-Myc. Western blot analysis showed a stronger phosphorylated ERK and AKT, and induced expression of c-Myc in KG-1(R) and HL-60(R) cells as compared with KG-1 and HL-60 cells (Figure 3B).

Next, we asked whether the inhibition of the expression of ERK, AKT, and c-Myc affects the expression of ULBP1/2/3. The expression at the transcriptional level by real-time qPCR and also at the cell surface by flow cytometry was examined. Results depicted in Figure 4A-B revealed that inhibition of c-Myc by chemical inhibitor reduced the mRNA level and cell-surface expression of ULBP1/2/3, while inhibition ERK or AKT by pharmacologic inhibitors had no effect. These results indicate that c-Myc is involved in the regulation of ULBP1/2/3 in cancer cells.

To delineate the role of c-Myc in NK-cell susceptibility of cytarabine-resistant cells, we investigated the effect of c-Myc inhibition in KG-1(R) and HL-60(R) cells using specific siRNA or a c-Myc pharmacologic inhibitor. c-Myc neutralization, whether by siRNA or pharmacological inhibitor, was associated with sensitization to NK-mediated lysis (Figure 5A,C). The inhibition was confirmed by western blot assay, which showed that there was no effect on c-Myc protein level when luciferase siRNA was used as a control (Figure 5B,D). To examine whether c-Myc inhibition in cytarabine-resistant cells would alter the stimulation of cytolytic activity of NK cells, CD107a assay was performed (as described earlier with NK cells) in the presence or absence of c-Myc inhibitor. The results shown in Figure 5E demonstrate that CD107a expression on NK cells decreased dramatically when cocultured with leukemic cells treated with c-Myc inhibitor as compared with nontreated cells (Figure 5E). We also performed the conjugate formation assay to analyze the capacity of NK cells to form immunologic synapse with leukemia cells treated or untreated with c-Myc inhibitor. The results showed that the c-Myc treated KG-1(R) and HL-60(R) cells form much less conjugates with NK cells than with nontreated cells (Figure 5F).

To validate the role of c-Myc in regulation of NKG2D ligands expression in primary AML cells, blasts from 9 different patients in a refractory stage (resistant to chemotherapy) were isolated and treated with c-Myc inhibitor for 24 hours. ULBP expression was measured at the transcriptional level using real-time qPCR and also examined at the cell surface by flow cytometry. The results indicate that ULBP1/2/3 mRNA level in blasts treated with c-Myc inhibitor was decreased in all patients, with varying degrees of down-regulations (Figure 6A).

To further confirm the role of c-Myc in the regulation of primary AML susceptibility to NK-mediated cell lysis, the blasts were incubated with c-Myc inhibitor for 24 hours at 100 µM concentration. Surprisingly, we observed that the refractory blasts of all of the tested patients were resistant to lysis mediated by NKL cells. Since NKL cells express a much lower level of granzyme B as compared with healthy donor NK cells, as well as not expressing most of the stimulatory receptors involved in recognition and elimination of primary blasts, we hypothesized that these NK cells are not cytotoxic enough to eliminate leukemia cells. However, when we used peripheral NK cells of a healthy donor (activated with IL-2, 300 U/mL for 2 days), we observed that NK-cell cytotoxic activity against AML blasts, which had been treated with c-Myc inhibitor, was abolished as compared with nontreated cells (Figure 6B). We also observed that CD107a expression on NK cells decreased dramatically when cocultured with cells treated with c-Myc inhibitor as compared with nontreated cells (Figure 6C). The analysis of ULBP expression at the cell surface revealed that c-Myc inhibition resulted in a dramatic decrease in expression of ULBP1/2/3 (Figure 6D).

To get more insight into the mechanism of ULBP gene regulation by c-Myc, we asked whether the transcriptional upregulation of ULBP1/2/3 involved direct c-Myc binding to c-Myc BS in the promoter region of ULBP1/2/3 genes (based on the SABiosciences database). Data depicted in Figure 7A indicate that potential c-Myc BS was only found for the ULBP1/3 genes. ChIP with a c-Myc–specific mAb revealed the binding of c-Myc to c-Myc BS of the ULBP1/3 in KG-1(R) cells (Figure 7B). We also observed that the binding of c-Myc to c-Myc BS of ULBP1/3 genes was decreased dramatically when cells were treated with c-Myc inhibitor (Figure 7B). In addition, we also performed a ChIP experiment with AML drug-resistant primary blasts, and similar results were obtained (Figure 7C). Together, these data indicate that c-Myc modulates ULBP1/3 expression directly by interacting with c-Myc BS at promoter region of ULBP1/3 genes.

There is renewed interest in exploiting the antitumor effect of NK cells due to data from preclinical and clinical work on the potential of alloreactive NK cells to kill tumor cells. Potential approaches include not only enhancing the generation of NK cells, but also manipulating the susceptibility of blast cells to NK-mediated killing. Pursuing these approaches will require a more thorough understanding of the different mechanisms of resistance and their relationship with susceptibility to NK-mediated killing. For example, following allogeneic hematopoietic stem-cell transplantation, activated donor NK cells can kill residual AML target cells in the recipient and contribute to protection from relapse. In this regard, manipulating NK-cell alloreactivity might improve outcomes after hematopoietic stem-cell transplantation; however, cross-resistance among blasts remains a drawback. In this study, we attempted to elucidate how acquisition of drug resistance in AML cells influences NK-cell recognition and the killing of drug-resistant blasts. We showed that the in vitro acquisition of AML cell resistance to cytarabine resulted in an increase in their susceptibility to NK-mediated cell cytotoxicity. Meanwhile, some studies report decreased sensitivity of drug-resistant leukemic cells to cellular cytotoxicity; others showed that chemotherapeutic drugs, including ara-C, sensitize acute lymphoblastic leukemia cells for NK-cell–mediated apoptosis.^10-18 

Even if the susceptibility of blasts in refractory patients is increased, it is essential to take into account the alteration of autologous NK cells in these patients. In fact, leukemic cells may escape from NK-cell surveillance through different mechanisms, including NK-cell qualitative deficiency (due to expression of HLA class I molecules, impaired activation, downregulation of ligands relevant in natural cytotoxicity receptors, and impaired differentiation signaling), as well as NK quantitative defects. Given the impairment of NK-cell cytotoxicity in almost hematologic malignancies, the restoration of normal NK function or adoptive NK cell therapy remains an attractive option to treat leukemic patients. In this context, NK-based cell therapy approaches should be based on the administration of highly cytotoxic NK cells in the context of target susceptibility to lysis.

The in vitro acquisition of resistance of AML cells to cytarabine was associated with an enhanced expression of NKG2D ligands by a mechanism involving c-Myc induction. More importantly, using AML blasts isolated from patients at refractory stage, c-Myc inhibitor was found to be efficient in downregulating NKG2D ligands expression. Future experiments will focus on the elucidation of c-Myc level, as well as NKG2DL expression in AML patients before and after drug resistance acquisition. In previous reports, it has been shown that ERK or AKT pathways could modulate NKG2DL expression; however, in our experimental model, despite the induction of these pathways, the pharmacologic inhibition of these proteins did not result in downregulation of ULBP1/2/3. We hypothesized that these pathways are switched on in a response to a hostile cell culture condition, which allows cells to survive and overcome drug-induced apoptosis.^50,51  Nevertheless, while inhibition of AKT or ERK had no effect on ULBP1/2/3, inhibition of c-Myc resulted in a decrease in the transcriptional level expression of ULBP1/2/3 and inhibition of NK cytotoxic response. It is assumed that this transcriptional factor plays a critical role in cell growth, proliferation, and apoptosis, by its capacity to orchestrate the expression of more than 15% of all cellular genes. c-Myc is frequently activated in AML and plays an important role in the induction of leukemogenesis, regulating several genes both positively and negatively. Unni et al have provided evidence that NKG2D ligands are induced on spontaneously arising tumors in a murine model of lymphomagenesis and that c-Myc is involved in their regulation.⁵² 

We attempted therefore, to shed light on the correlation between those 2 important genes, NKG2D ligands and c-Myc, respectively, involved in immune response and leukemogenesis in the context of a drug resistance.

ChIP studies revealed a direct correlation between c-Myc activation and ULBP1/3 expression, although no c-Myc BS for ULBP2 was identified. This indicates not only that c-Myc can regulate ULBP expression directly, but also that it may modify its expression indirectly. These data would suggest that quantification of c-Myc in AML blast cells may be useful as a prognostic indicator in AML for NK-cell response in bone marrow transplantation cell therapy trials.

Recently, it has been reported that p53 regulates ULBP1/2 expression at the transcriptional level.⁴⁴  This mechanism is ruled out in our cell models since the one we used displays a mutated p53. Recent studies also indicate that DNA damage could regulate NKG2D ligand expression.³⁹  This mechanism is also unlikely since no modification for MICA/B or poliovirus receptor expressions, which are regulated by DNA damage pathway, was observed. In addition, it is well established that leukemic stem cells play a central role in the relapse of acute AML and the resistance to NK-cell–mediated cytotoxicity.⁵³ 

Nevertheless, under our experimental conditions, the acquisition of resistance cells to cytarabine did not correlate with the induction of stem-cell markers (not shown).

Collectively, these studies suggest that the acquisition of AML resistance to cytarabine does not confer cross-resistance to NK-mediated cell lysis, and that the increased susceptibility resulting from the acquisition of drug resistance may show promise in NK-cell therapy following allogeneic stem-cell transplantation for drug refractory patients with AML.

This work was supported by the Institut National du Cancer, Association Laurette Fugain, and the Qatar Foundation.

Contribution: A.N., C.P., and A.M. performed the experiments; S.C. designed the study; A.N. and S.C. wrote the manuscript; and all authors critically reviewed the manuscript and gave their final approval.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Salem Chouaib, U753 INSERM, Institut Gustave Roussy, 114 rue Edouard Vaillant, 94805 Villejuif cedex, France; e-mail: chouaib@igr.fr.