IL2RA Promotes Aggressiveness and Stem Cell–Related Properties of Acute Myeloid Leukemia

Acute myeloid leukemia (AML) is a hematopoietic malignancy with an annual incidence of 5–8 cases per 100,000 population and a median age of 67 years at diagnosis (1–3). It is organized in a hierarchical manner, that is, bulk leukemic cells are derived from leukemic stem cells (LSC), which are thought to represent the source of disease emergence, therapy resistance, and relapse (4, 5). Standard treatment is based on chemotherapy with the deoxycytidine analogue cytarabine (AraC) and an anthracycline, but both primary and secondary resistance are frequent, so that only 20%–30% of patients achieve long-term disease-free survival (6, 7).

AML is a genetically and prognostically heterogeneous disease. It is caused by various recurrent chromosomal abnormalities and gene mutations, which accumulate in hematopoietic stem and progenitor cells (HSPC) and transform them into LSCs (8–12). In addition, gene expression changes contribute to AML formation (12–15). The identification and characterization of such recurrent molecular and genetic abnormalities has significantly improved our understanding of the mechanisms of leukemogenesis, as well as prognostication. Furthermore, certain aberrations represent potential targets for rationally designed novel therapeutics for the treatment of AML (16–18). Some of these have recently been approved for clinical use, among them Gemtuzumab-Ozogamicin, an antibody–drug conjugate directed against the cell surface molecule CD33 (16, 17). Despite of these advances, AML remains a deadly disease, and additional potential therapeutic targets are actively being searched for (16, 19, 20). Among the identified candidates is the cell surface molecule IL2RA (19).

IL2RA (CD25) represents a low affinity receptor for its ligand IL2. Together with IL2RB (CD122) and IL2RG (CD133), it forms the high-affinity IL2 receptor (21, 22). IL2 binding to its receptor causes activation of JAK1 and JAK3, which, in turn, activate several downstream pathways regulating cell survival and proliferation, namely, the PI3K/AKT, RAS/RAF/MEK/ERK, and STAT5 pathways (21, 23).

In healthy tissues, IL2RA is mainly expressed on activated T cells and regulatory T cells, as well as some other (activated) hematopoietic cell types, but notably not on hematopoietic stem cells (HSC; refs. 19, 21, 24). IL2RA expression was elevated in a variety of cancers, mostly of the hematopoietic system (19, 21, 22, 24, 25). It is a diagnostic marker, and also used for detection of minimal residual disease, in hairy cell leukemia (26). In chronic myelogenous leukemia (CML), IL2RA was specifically upregulated in LSCs (24), but reports on its function were controversial, describing IL2RA as either a promoter or an inhibitor of CML cell proliferation and disease aggressiveness (24, 25). In AML, IL2RA overexpression was consistently associated with poor therapy response and adverse outcome (27–36) and scored as the second top hit in an unbiased search for genes whose expression predicted survival (37). IL2RA expression was also increased upon development of AML from myelodysplastic syndrome or myeloproliferative disease and at relapse of AML (29). Despite this compelling evidence for the role of IL2RA in disease aggressiveness, and even though phase I studies exploring the possible use of IL2RA-directed therapy in AML have been incepted (NCT02588092, NCT00085150; ref. 21), the biological functions of IL2RA in AML and the underlying molecular mechanisms have not been addressed so far.

In this study, we used human AML cell lines, Flt3-ITD/Npm1c and MLL-AF9–driven mouse models of AML, and primary AML samples to show that IL2RA promoted proliferation and inhibited apoptosis and differentiation of AML cells. Moreover, it augmented LSC-related properties and was required for leukemogenesis. IL2RA antibodies inhibited leukemic, but not normal hematopoietic cells, and synergized with AraC and with BCL2 and CDK4/CDK6 inhibitors in this regard, suggesting effective therapeutic strategies for patients with AML with IL2RA overexpression.

Animal experiments were approved by the Animal Ethics Committee of the Medical University of Vienna and the Austrian Federal Ministry of Science, Research, and Economy (GZ66.009/0308-WF/V/3b/2015). Federation of European Laboratory Animal Science Associations and Austrian guidelines to minimize animal distress and suffering were followed. Experiments with primary AML samples were approved by the Ethics Committee of the Medical University of Vienna (EK 1394/2019) and conducted in accordance with the Declaration of Helsinki.

The human myeloid cell lines HL60 and UCSD/AML1 were kindly provided by Dr. Peter Valent (Department of Medicine I, Medical University of Vienna, Austria) in 2013, and by Dr. Frank Speleman (Centre for Medical Genetics, Ghent University Hospital, Ghent, Belgium) in 2012, respectively. The viral packing cell line Phoenix-GP was a gift from Dr. Hannes Stockinger (Institute for Applied Immunology, Medical University of Vienna, Austria) in 2009. Upon receipt, stocks of the cell lines were prepared and stored in liquid nitrogen. New aliquots from these stocks were thawed every 1–3 months, and cultured as described in the Supplementary Methods. Authentication was performed via the specific growth characteristics (adherent or suspension, doubling times), morphology, functional properties (ability to produce infectious particles in the case of Phoenix-GP cells), and expression of specific genes (EVI1, IL2RA, etc.). All cell lines were tested regularly (last test, July, 2020) for Mycoplasma contamination using MycoAlert Mycoplasma Detection Kit (Lonza), and showed negative results.

Preparation of protein lysates, SDS-PAGE, transfer to polyvinylidene difluoride membranes (Hybond-P; Amersham), and incubations with antibodies (Supplementary Table S1) were performed using standard procedures. Blots were developed using SuperSignal West Femto or Pico Chemiluminescent Substrates (both from Thermo Scientific) and scanned using a ChemiDoc Touch Imaging System (Bio Rad). Densitometric analysis was performed with ImageJ software (NIH, Bethesda, MD).

Spleen cells from C57BL/6 mice that had succumbed to AML following transplantation with Flt3-ITD/Npm1c–transformed hematopoietic cells (38) were cultured in Iscove's modified Dulbecco medium containing 10% FBS, 1% l-Glutamine (all from Thermo Fisher Scientific), 50 ng/mL mSCF (PeproTech), 10 ng/mL mIL3 (PeproTech), and 10 ng/mL mIL6 (BioLegend). To knock down Il2ra, they were transduced with the lentiviral vector pRRL_SFFV_GFP_ mirE_PGK_NeoR containing shIl2ra-1 (5′-CACAGCAGTTCTAAAGCTTTA-3′), shIl2ra-2 (5′-AAGAGAGGTTTCCGAAGACTA-3′), or shRen (5′-AAGGAGGAAAGTTAAATAGAAT-3′) as a control. Generation of shRNA constructs and transduction of leukemic cells are described in Supplementary Methods. Sorted GFP⁺ cells were expanded in the above described medium and used for in vitro assays. Proliferation, cell cycle, and apoptosis were measured as described for human cell lines (Supplementary Methods). To assess myeloid differentiation, cells were stained with the respective antibodies (Supplementary Table S1) and subjected to flow cytometry (LSRFortessa, Becton Dickinson). For serial replating assays, 2,000 cells per well of a 6-well plate were seeded into methylcellulose (MethoCult GF M3434; StemCell Technologies). Technical duplicates were performed. Colonies were counted after 7 days and 2,000 cells per condition were used for replating.

For transplantation, 6- to 8-week-old female C57BL/6 recipient mice were sub-lethally irradiated (5 Gy), anesthesized on the next day, and injected retro-orbitally with GFP⁺ shRen- or shIl2ra-transduced spleen cells from mice with Flt3-ITD/Npm1c–driven AML (600,000 cells/mouse). Mice were sacrificed when terminally ill. The significance of differences in survival was probed using the log-rank test. Peripheral blood cell counts were determined using a hematology analyzer Sysmex XN-350 (Sysmex), and the proportion of GFP⁺ cells in bone marrow (BM) and spleen was assessed by flow cytometry.

Cryopreserved primary human AML and healthy BM samples were provided by the Hanusch Hospital (Vienna, Austria), and the General Hospital (Vienna, Austria), respectively. BM CD34⁺ cells from healthy donors were purchased from Lonza. Primary cells were thawed and cultured as described previously (39). Briefly, after thawing in a 37° water bath, cells were washed and incubated with RPMI medium containing 50 μg/mL DNAse (Sigma-Aldrich) for 60 minutes to prevent cell clumping. Cells were cultured in RPMI medium supplemented with 10% FBS, 1% glutamine, 1% penicillin/streptomycin, and 100 ng/mL each of SCF, IL3, and G-CSF (all from PeproTech). For antibody treatment, cells were seeded at a density of 2 × 10⁵/mL and incubated with 3 μg/mL monoclonal human IL2RA (hIL2RA) antibody (clone BC 96, BioLegend), 6 μg/mL basiliximab (Szabo Scandic), or the corresponding amounts of isotype control antibodies (BioLegend or Szabo Scandic) for 72 hours. For combination treatment, cells were incubated for 48 hours with the indicated concentrations of hIL2RA antibody, AraC (provided by the dispensary of the General Hospital of Vienna), the BCL2 inhibitor Navitoclax (MedChemExpress), and/or the CDK4/CDK6 inhibitor Abemaciclib (MedChemExpress). Incubation with human recombinant IL2 (Sigma-Aldrich) employed a concentration of 100 ng/mL for 72 hours. Cell viability, cell-cycle distribution, and apoptosis were measured as described for cell lines (Supplementary Methods). For colony formation assays, cells were treated as described above and transferred to methyl cellulose (MethoCult H4434, StemCell Technologies) at a concentration of 1 × 10⁵ cells per well of a 6-well plate. Technical duplicates were performed, and total colonies were counted after 14 days. To determine total viable cell numbers, colonies were harvested and cells were washed, resuspended in PBS, and counted using a CASY Cell Counter (Roche Innovatis AG).

For experiments with cell lines and primary mouse cells, at least three independent biological replicates were performed; results are displayed as means ± SEM. For experiments with primary AML samples, technical replicates were performed whenever feasible in terms of cell numbers; results are displayed as means ± SD. Apoptosis and cell-cycle assays with primary samples were performed as single measurements. Significance of differences between two independent groups was calculated using Student two-tailed t test; significance of differences between multiple groups was determined by two-way ANOVA followed by Bonferroni post hoc test. The log-rank test was used to evaluate survival differences between groups of mice. P values <0.05 were considered statistically significant. Analyses were performed using GraphPad Prism 6 software (GraphPad Software).

Additional and more detailed methods are available in Supplementary Methods.

Aberrant expression of IL2RA protein or mRNA was associated with unfavorable outcome in AML (27–34, 37). To confirm and extend these findings, we analyzed the expression and potential prognostic relevance of IL2RA in several publicly available gene expression datasets. IL2RA expression was significantly upregulated in AML compared with healthy BM cells (GSE13159; Fig. 1A). Its expression was also significantly higher in LSC-enriched cell populations as compared with LSC-depleted cell populations (GSE76008; Fig. 1B) or to healthy HSCs and progenitor cells (GSE63270, GSE30029; Fig. 1C). In addition, high IL2RA expression was an independent prognostic parameter for decreased overall survival (OS) in 7 AML patient cohorts, contained in datasets GSE37642, GSE12417 (2 cohorts), GSE6891 (2 cohorts), GSE71014, and TCGA_LAML and comprising a total of 1,272 patients [Fig. 1D; Table 1; Supplementary Fig. S1, the value was adjusted for multiple testing (P_adj) according to Altman and colleagues (40)]. European LeukemiaNet (ELN) risk classification data were available for GSE37642. The favorable, intermediate, and adverse risk groups contained 107, 174, and 86 patients, respectively. High IL2RA expression had no impact on OS in the adverse risk group, but was significantly associated with shorter OS in the favorable (P = 0.034) and intermediate (P = 0.006) risk groups (Fig. 1E). We therefore asked whether IL2RA expression could refine the ELN risk classification. On the basis of median survival times, ELN-favorable/IL2RA^low patients were assigned to a redefined favorable risk group, ELN-favorable/IL2RA^high and ELN-intermediate/IL2RA^low patients were combined into an intermediate risk group, and ELN-intermediate/IL2RA^high, and ELN-adverse patients constituted the adverse risk group. The resulting ELN + IL2RA classification substantially improved the ELN risk score (median survival, 4,029, 292, and 215 days according to ELN, and not reached, 493, and 113 days according to the ELN + IL2RA classification; Fig. 1F).

In summary, IL2RA expression was upregulated in AML versus normal BM, and in LSCs versus bulk leukemic cells and versus HSCs. Moreover, its expression was an independent prognostic parameter for poor outcome of AML and could further refine the ELN risk classification.

To address the functional role of IL2RA in AML, we transduced UCSD/AML1 cells, which expressed high levels of IL2RA (Supplementary Fig. S2A), with lentiviral vectors containing two different shRNAs against IL2RA (shIL2RA-1, shIL2RA-2), or a nontargeting shRNA as control (shCtrl). Downregulation of IL2RA, confirmed by flow cytometry (Fig. 2A), significantly reduced cell proliferation (Fig. 2B). Accordingly, it decreased the proportion of actively cycling cells (cells in S/G₂–M), and increased the proportion of cells in G₀–G₁ as well as in sub-G₁ (apoptotic cells; Fig. 2C; Supplementary Fig. S2B). The increase in apoptosis was confirmed by the Annexin V assay (Fig. 2D; Supplementary Fig. S2C; higher proportions of apoptotic cells than in Fig. 2C are most likely due to the fact that Annexin V labels also early apoptotic cells). Corroborating the specificity of these effects, reexpression of a codon-optimized (i.e., shRNA insensitive) version of IL2RA in UCSD/AML_shIL2RA-1 and UCSD/AML_ shIL2RA-2 cells counteracted the effects of the knockdown on cell proliferation, cell-cycle distribution, and apoptosis (Supplementary Fig. S2D–S2G).

Next, we asked whether antibodies specifically targeting IL2RA would also reduce AML cell proliferation. Indeed, treatment of UCSD/AML1 cells with two different hIL2RA antibodies, including basiliximab, which was approved for the prevention of renal transplant rejection (41), resulted in dose- and time-dependent inhibition of cell proliferation (Supplementary Fig. S2H and S2I). In contrast, hIL2RA antibodies had no effect on the proliferation of the IL2RA^low cell line HL60 (Supplementary Fig. S2J).

HL60 cells were also used to investigate the consequences of lentiviral overexpression of IL2RA. Expression of IL2RA on the surface of transduced cells was confirmed by flow cytometry (Fig. 2E). It promoted cell proliferation (Fig. 2F), increased the proportion of cells in S/G₂–M, decreased the proportion of cells in G₀–G₁ and in sub-G₁, and inhibited apoptosis as determined by Annexin V staining (Fig. 2G and H; Supplementary Fig. S2K and S2L).

In summary, our data show that IL2RA promoted proliferation and cell-cycle progression and inhibited apoptosis of human AML cell lines.

To explore the mechanisms by which IL2RA exerts its above described effects, we determined the expression and activation status of key regulators of proliferation, survival, cell-cycle progression, and apoptosis by immunoblot analysis. Knockdown of IL2RA in UCSD/AML1 cells decreased the activating phosphorylations on the proliferation and survival promoting kinases, AKT and ERK (Fig. 3A; Supplementary Fig. S3A). It also resulted in reduced phosphorylation of STAT5, another central regulator of survival and proliferation in AML (Fig. 3A; Supplementary Fig. S3A). Accordingly, JAK–STAT signaling was among the pathways that were enriched in the list of genes coexpressed with IL2RA in the TCGA_LAML dataset (Supplementary Table S2A and S2B). Regarding the core cell-cycle machinery, the positive regulators of the G₁–S transition, CDK6, CDK2, and cyclin E, were downregulated, while the cell-cycle inhibitors p21 and p27 were upregulated in shIL2RA-transduced UCSD/AML1 cells (Fig. 3A; Supplementary Fig. S3A). Furthermore, IL2RA knockdown reduced the level of the antiapoptotic BCL2 protein, and increased expression of the proapoptotic BAX protein (Fig. 3A; Supplementary Fig. S3A).

Overexpression of IL2RA in HL60 cells elicited molecular effects opposite to those observed upon IL2RA knockdown in UCSD/AML1 cells (Fig. 3B; Supplementary Fig. S3B). Taken together, the pro-proliferative effects of IL2RA on AML cells were associated with corresponding effects on key regulators of proliferation, survival, cell-cycle progression, and apoptosis.

In human AML, aberrant IL2RA expression correlated positively with the presence of activating internal tandem duplications in the FLT3 gene (FLT3-ITD) and with nucleophosmin (NPM1) gene mutations (NPM1c; ref. 32). A Flt3-ITD/Npm1c–driven mouse model of AML (38) was, therefore, employed to investigate the effects of Il2ra on leukemia cell proliferation, differentiation, and apoptosis, LSC activity, and leukemogenesis.

BM cells from mice that had succumbed to AML after transplantation with Flt3-ITD/Npm1c–transformed hematopoietic cells contained higher levels of Il2ra mRNA than their healthy counterparts (Supplementary Fig. S4A). Monoclonal mouse IL2RA (mIL2RA) antibody inhibited proliferation, promoted apoptosis, and decreased serial replating activity (considered a proxy of LSC activity) of spleen cells from these mice (Supplementary Fig. S4B–S4D). To further investigate the role of Il2ra in Flt3-ITD/Npm1c–driven AML, leukemic spleen cells were transduced with lentiviral vectors expressing shRNAs against Il2ra (shIl2ra-1, shIl2ra-2), or an shRNA against Renilla luciferase (shRen) as a control (Fig. 4A). Downregulation of IL2RA in shIl2ra- versus shRen-transduced cells was confirmed by flow cytometry (Fig. 4B). Sorted GFP⁺ cells were used for in vitro assays and for transplantation of recipient mice. By analogy to the results obtained with the human AML cell lines, knockdown of Il2ra in murine leukemic cells significantly inhibited cell proliferation (Fig. 4C), decreased the proportion of cells in S/G₂–M, and increased apoptosis (Fig. 4D and E; Supplementary Fig. S4E and S4F). Knockdown of Il2ra also induced myeloid differentiation, as shown by increases in the proportions of Gr-1⁺ among CD11b⁺ GFP⁺, and cKit⁻ among GFP⁺ cells (Fig. 4F and G; Supplementary Fig. S4G and S4H). Furthermore, Il2ra downregulation reduced the serial replating activity of murine AML cells (Fig. 4H).

Upon transplantation into recipient mice, both shRen- and shIl2ra-expressing cells caused AML-like disease with increased white blood cell counts and spleen weight, and decreased red blood cell counts and platelet numbers (Supplementary Fig. S4I), yet downregulation of Il2ra significantly increased disease latency (median survival, shRen: 56 days, shIl2ra-2: 76 days, shIl2ra-1: 113 days; Fig. 4I). Moreover, while large proportions of BM and spleen cells from moribund mice maintained shRen expression (as measured by the proportion of GFP⁺ cells), cells carrying shIl2ra were entirely outcompeted by shRNA-negative cells (Fig. 4J; Supplementary Fig. S4J).

To confirm the role of Il2ra on the background of a different genetic driver lesion, a mouse model based on the AML-associated fusion oncogene MLL-AF9 was employed (39, 42, 43). As with the Flt3-ITD/Npm1c model, Il2ra mRNA expression was upregulated in BM cells from mice that succumbed to AML after transplantation with MLL-AF9–transformed hematopoietic cells as compared with healthy BM cells (Supplementary Fig. S5A). mIL2RA antibody significantly inhibited cell proliferation, and induced differentiation and apoptosis (Supplementary Fig. S5B–S5D). Furthermore, mIL2RA antibody diminished LSC-related properties: it reduced the abundance and quiescence of a cell population enriched for LSCs (defined by the marker combination Venus⁺ lin⁻ Sca1⁻ c-Kit⁺ CD34⁺ CD16/CD32^high in the MLL-AF9 model; refs. 37, 39, 42), and decreased the serial replating activity of MLL-AF9–expressing leukemic BM cells (Supplementary Fig. S5E–S5G). Likewise, shRNA-mediated knockdown of Il2ra in MLL-AF9 cells decreased proliferation, led to a strong depletion of transduced cells over time, increased differentiation, and apoptosis, and reduced LSC-related properties (Supplementary Fig. S5H–S5N).

In summary, these data show that Il2ra inhibits differentiation and apoptosis, and augments proliferation, LSC-related properties, and leukemogenesis in murine AML.

We next asked whether our findings in AML cell lines and murine AML models could be translated to primary AML cells. Flow cytometry confirmed high and low IL2RA expression, respectively, in four and two AML samples selected on the basis of qRT-PCR data (Supplementary Fig. S6A). Clinical characteristics of the patients are summarized in Supplementary Table S3. Exposure of IL2RA^high, but not IL2RA^low, primary AML cells to hIL2RA antibody or basiliximab decreased viability, proliferation, and the proportion of cells in S/G₂–M, and enhanced apoptosis (Fig. 5A–D; Supplementary Fig. S6B–S6D). Pretreatment with hIL2RA antibodies also strongly reduced colony numbers and the total number of viable cells in colonies upon plating in methylcellulose, reflecting reduced stem cell/progenitor activity, in a manner dependent on IL2RA expression (Fig. 5E; Supplementary Fig. S6E and S6F). Furthermore, hIL2RA antibodies led to alterations of the levels of p-ERK, p-AKT, pSTAT5, CDK6, CDK2, BCL2, and BAX in a manner consistent with their biological effects (Fig. 5F). In support of the potential therapeutic utility of hIL2RA antibodies in IL2RA^high AML, these agents had much weaker effects on the viability, apoptosis, and clonogenic growth of CD34⁺ HSPCs from healthy donor BM (Supplementary Fig. S6G–S6I). Also, treatment with hIL2RA antibody had no impact on the relative abundance of different progenitor populations in healthy BM cells (Supplementary Table S4).

In agreement with a previous report (44), AraC treatment induced IL2RA expression both in primary AML samples (Fig. 6A) and in UCSD/AML1 cells (Supplementary Fig. S7A). Therefore, we asked whether the hIL2RA antibody would enhance the antileukemic activity of AraC. Indeed, the combination of these two agents inhibited viability of primary AML samples (Fig. 6B) and of UCSD/AML1 cells (Supplementary Fig. S7B) in a synergistic manner in the majority of dose combinations. These results were confirmed using the Annexin V assay (Fig. 6C; Supplementary Fig. S7E).

To also explore potential combination therapies suggested by our molecular data (Figs. 3 and 5F; Supplementary Fig. S3), human AML cells were cotreated with hIL2RA antibody and either the BCL2 inhibitor navitoclax or the CDK4/CDK6 inhibitor abemaciclib. Again, synergistic inhibition of AML cell viability was observed (Supplementary Fig. S7C–S7G).

In summary, hIL2RA antibody efficiently inhibited the proliferation and clonogenic activity of primary IL2RA^high AML samples, while healthy CD34⁺ BM cells were substantially less sensitive. hIL2RA antibody synergized both with AraC and with molecularly targeted inhibitors to cause cytotoxicity in AML cells, suggesting that these combinations have therapeutic potential for patients with IL2RA^high AML.

Several studies have highlighted the prognostic importance of IL2RA in AML (27–36), and IL2RA is actively pursued as a therapeutic target in this disease (21). Nevertheless, surprisingly little is known about the pathophysiologic role of this gene in AML cells. Here, we confirmed IL2RA as an independent prognostic parameter in several publicly available gene expression datasets, which comprise AML patient populations with different genetic and age compositions and a total of almost 1,300 patients. Moreover, IL2RA expression was able to further refine the ELN classification. These analyses strengthen and extend previous findings, and further emphasize the need to understand the function of IL2RA in AML.

Using gene overexpression and knockdown approaches in, and/or antibody treatment of, human AML cell lines, AML mouse models, and primary AML cells, we report here for the first time that IL2RA promoted proliferation and cell-cycle activity, and inhibited differentiation and apoptosis, of AML cells. Interestingly, these effects likely occurred independently of the IL2RA ligand, IL2. First, neither the primary AML samples nor the AML cell lines used in our study secreted detectable levels of IL2 (Supplementary Fig. S8A). Second, in line with a previous report that suggested that IL2 acted on AML cells in a non–cell-autonomous, indirect manner (45), incubation with human recombinant IL2 had no effect on proliferation, apoptosis, or colony formation of either IL2RA^high or IL2RA^low AML cells (Supplementary Fig. S8B–S8H).

The phenotypes elicited by manipulation of IL2RA expression were associated with corresponding changes in proteins associated with proliferation/survival (p-ERK, p-AKT, p-STAT5), cell-cycle regulation (CDK6, CDK2, Cyclin E, p21, and p27), and apoptosis (BCL2 and BAX). Notably, IL2RA was previously described both as a downstream target and an upstream regulator of STAT5 (24, 46, 47), suggesting a positive feedback loop between these two signaling components. This pivotal role in the JAK–STAT pathway, as well as its effect on several other central signaling pathways, explain the strong impact of alterations in IL2RA levels on numerous key features of AML. Most compellingly, knockdown of Il2ra completely abolished the ability of Flt3-ITD/Npm1c–transformed cells to give rise to AML in vivo.

Beyond its roles in the proliferation and survival of AML blasts, IL2RA was also relevant to AML stem cells. We found that IL2RA expression was higher in AML LSCs compared with bulk leukemic cells or healthy HSCs. Moreover, Il2ra augmented LSC-related properties in two independent AML mouse models, and inhibition of IL2RA by specific antibodies inhibited stem cell/progenitor activity in primary human AML samples. The inability of shIl2ra-transduced murine Flt3-ITD/Npm1c AML cells to contribute to leukemogenesis in vivo precluded bona fide stem cell assays like serial transplantation or in vivo limited dilution assays, but per se suggested a requirement for Il2ra to maintain leukemia initiating potential. In line with our data, IL2RA has previously been reported as an AML LSC–associated gene (19), and the gene expression signature of IL2RA^high AML blasts was enriched for gene expression profiles associated with AML LSCs (32). Similarly, studies on human CML found that IL2RA was predominantly expressed in LSCs, but not in more mature cell fractions or in normal HSCs (24, 25, 48). In a CML mouse model, IL2RA^high cells were enriched for LSCs, and therapeutic targeting of IL2RA reduced LSC numbers and improved animal survival (25).

Taken together, our results provide, for the first time, strong experimental support for the potential of IL2RA as a novel therapeutic target in AML. The fact that knockdown of Il2ra in IL2RA^high murine AML cells diminished LSC-related properties and entirely prevented any contribution of the knockdown cells to leukemia formation suggests that IL2RA^high AML may be “addicted” to this oncogene. Thus, even though numerous different driver lesions are able to cause AML (8–15), a subset of patients may exhibit an actionable dependency on the activity of IL2RA. Correspondingly, IL2RA^high AML cells were sensitive to inhibition by hIL2RA antibodies. In contrast, healthy CD34⁺ BM HSPCs were hardly responsive to this treatment. In line with this, transplantation of healthy IL2RA^low HSCs into NOD-SCID mice resulted in multilineage hematopoietic reconstitution, including the generation of IL2RA^high myeloid and lymphoid progeny (19). These data indicate that normal HSC function is independent of IL2RA, and point toward the possible existence of a therapeutic window. Furthermore, even though Il2ra knockout mice suffered from a lymphoproliferative disorder and autoimmune disease in adulthood, they were healthy and indistinguishable from wild-type littermates up to 4 weeks of age (49), suggesting that even the complete absence of Il2ra function is tolerable for several weeks. Inhibition of IL2RA through specific antibodies, or elimination of IL2RA^high AML cells through antibody–drug or antibody–radioisotope conjugates, may therefore represent an effective and safe treatment strategy for patients with IL2RA^high AML. Indeed, several early-phase clinical studies addressed the potential utility of IL2RA antibodies in hematologic malignancies (21, 50–54). Efficacy and acceptable toxicity of IL2RA antibodies conjugated to radioisotopes or toxins were reported in various types of lymphoma (50, 51, 53). Also, a phase I study of the pyrrolobenzodiazepine-conjugated hIL2RA antibody ADCT-301 in patients with relapsed/refractory IL2RA^high AML or ALL (NCT02588092) showed an acceptable safety profile (54).

Because efficacies of targeted therapies can be augmented by combination with chemotherapeutic drugs (16), we queried the antileukemic activity of hIL2RA antibody and of AraC alone and in combination. Indeed, these two agents inhibited the viability of an AML cell line and of primary AML samples in a moderately synergistic manner. This synergy may be due to the fact that both drugs have different mechanisms of action, that is, hIL2RA antibody enhances antileukemic signaling pathways while AraC causes DNA damage. As an alternative or additional explanation, AraC treatment resulted in higher levels of IL2RA on AML cells (Fig. 6A; Supplementary Fig. S7A; ref. 44). This could be a defense mechanism providing some protection from AraC induced apoptosis that might be alleviated by IL2RA inhibition. In line with our observations, the combined application of the anti-IL2RA immunotoxin LMB-2 and the antimetabolite gemcitabine resulted in increased cytotoxicity and antitumor activity both in vitro and in a mouse xenograft model of a human epidermoid carcinoma cell line, proposed as a model for adult T-cell leukemia (55). IL2RA antibodies also enhanced the activity of targeted drugs. Here, we show synergistic effects with inhibitors of the IL2RA downstream effectors BCL2 and CDK6 both in an AML cell line and in primary AML samples. In a previous report, an IL2RA antibody augmented the reduction in leukemic burden effected by the tyrosine kinase inhibitor nilotinib in a mouse model of CML (25).

In summary, we report here the first functional characterization of the previously proposed therapeutic target, IL2RA, in AML. IL2RA promoted proliferation and stem cell–related properties, and inhibited differentiation and apoptosis in different AML model systems. AML cells with experimental downregulation of Il2ra were unable to contribute to leukemia formation in recipient mice, possibly indicating oncogene addiction. IL2RA inhibition reduced the viability of AML cells alone and in synergy with AraC and with targeted drugs, but had little effect on normal HSPCs. Together, these results provide a strong rationale for further development of therapeutic strategies directed at IL2RA for the treatment of AML.

G.S. Vassiliou reports personal fees and other compensation from Kymab Ltd. (minor stockholder) and grants from Celgene outside the submitted work. P.B. Staber reports grants from Austrian Science Fund during the conduct of the study. R. Wieser reports grants from Austrian Science Fund during the conduct of the study. No potential conflicts of interest were disclosed by the other authors.

C.H. Nguyen: Conceptualization, methodology, writing-original draft, writing-review and editing. A. Schlerka: Methodology. A.M. Grandits: Methodology. E. Koller: Methodology. E. van der Kouwe: Methodology. G.S. Vassiliou: Conceptualization, methodology. P.B. Staber: Conceptualization, methodology. G. Heller: Conceptualization, supervision, methodology, writing-original draft, writing-review and editing. R. Wieser: Conceptualization, supervision, funding acquisition, writing-original draft, writing-review and editing.

This work was funded by the Austrian Science Fund (FWF), projects no. P28256-B28 and P28013-B28 (to R. Wieser). The authors gratefully acknowledge Dr. Andreas Spittler of the Core Facility for Flow Cytometry, Medical University of Vienna, Vienna, Austria, who performed cell sorts and gave invaluable advice for flow cytometry applications. We thank Dr. Klaus Schmetterer and Marlene Gerner, MSc (Department of Laboratory Medicine, Medical University of Vienna, Vienna, Austria) for performing IL2 ELISAs. Dr. Gregor Hoermann (Clinical Department of Medical and Chemical Laboratory Diagnostics, Medical University of Vienna, Vienna, Austria) kindly provided human IL2RA shRNA (pLKO.1_shIL2RA_mCherry and pLKO.1_shCtrl_mCherry) and lentiviral overexpression constructs (pLOC_IL2RA_IRES_GFP, pLOC_IRES_GFP). Dr. Johannes Zuber (Research Institute of Molecular Pathology, Vienna, Austria) is acknowledged for providing pRRL_SFFV_GFP_mirE_shRen713 and pMSCV_MLL-AF9_IRES_Venus. Dr. Karin Nowikovsky, Erwin Tomasich, Msc, and Dr. Peter Valent (all from the Department of Medicine I, Medical University of Vienna, Austria) shared BCL2, BAX, p21, ERK, p-STAT5, and STAT5 antibodies. The dispensary of the General Hospital, Vienna, is gratefully acknowledged for supplying AraC. We thank Dr. Tobias Herold from the Acute Myeloid Leukemia Cooperative Group (AMLCG) Munich for providing clinical data to dataset GSE37642.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.