The bone marrow niche has a pivotal role in progression, survival, and drug resistance of multiple myeloma cells. Therefore, it is important to develop means for targeting the multiple myeloma bone marrow microenvironment. Myeloma-associated macrophages (MAM) in the bone marrow niche are M2 like. They provide nurturing signals to multiple myeloma cells and promote immune escape. Reprogramming M2-like macrophages toward a tumoricidal M1 phenotype represents an intriguing therapeutic strategy. This is especially interesting in view of the successful use of mAbs against multiple myeloma cells, as these therapies hold the potential to trigger macrophage-mediated phagocytosis and cytotoxicity. In this study, we observed that MAMs derived from patients treated with the immunomodulatory drug (IMiD) lenalidomide skewed phenotypically and functionally toward an M1 phenotype. Lenalidomide is known to exert its beneficial effects by modulating the CRBN-CRL4 E3 ligase to ubiquitinate and degrade the transcription factor IKAROS family zinc finger 1 (IKZF1). In M2-like MAMs, we observed enhanced IKZF1 levels that vanished through treatment with lenalidomide, yielding MAMs with a bioenergetic profile, T-cell stimulatory properties, and loss of tumor-promoting capabilities that resemble M1 cells. We also provide evidence that IMiDs interfere epigenetically, via degradation of IKZF1, with IFN regulatory factors 4 and 5, which in turn alters the balance of M1/M2 polarization. We validated our observations in vivo using the CrbnI391V mouse model that recapitulates the IMiD-triggered IKZF1 degradation. These data show a role for IKZF1 in macrophage polarization and can provide explanations for the clinical benefits observed when combining IMiDs with therapeutic antibodies.
See related Spotlight on p. 254
Multiple myeloma is a hematologic malignancy characterized by clonally expanded plasma cells. Despite the introduction of novel therapeutic agents, including immunomodulating drugs (IMiD), proteasome inhibitors, and mAbs, multiple myeloma remains largely incurable (1). One hallmark of multiple myeloma is the close interaction between the malignant plasma cells and the bone marrow (BM) microenvironment (2). Macrophages are a major component of the multiple myeloma-BM microenvironment (3), they provide multiple myeloma cells protection from spontaneous and chemotherapy-induced apoptosis (4), promote multiple myeloma cell proliferation (5), and facilitate immune evasion by multiple myeloma cells (6), thereby laying the foundation for poor patient outcomes (7).
Consequently, macrophages have emerged as a potential therapeutic target in multiple myeloma, as they have in other types of cancer (8). Several targeting strategies have been proposed, including depletion of tumor-associated macrophages (TAM) and reprogramming of TAMs. The latter strategy is based on knowledge about the plasticity of macrophages whose phenotypes form a continuum between an immunoreactive, antitumoral M1 phenotype and a tolerogenic, protumoral M2-like phenotype. Microenvironmental cues, such as macrophage migration inhibitory factor, drive M2-like polarization of myeloma-associated macrophages (MAM; ref. 9). However, these M2-like MAMs retain intrinsic tumoricidal capacity, and this can be reactivated in preclinical models by disrupting M2-like–promoting signals (9) or by interfering with immunologic checkpoints, such as “don't eat me” signals, on multiple myeloma cells (10).
In addition, macrophages are critical mediators of antibody-based multiple myeloma therapies targeting CD38 (11) and SLAMF7 (12), further highlighting the therapeutic importance of these cells in multiple myeloma. This rationale is further corroborated by our in vitro observations that the IMiD, lenalidomide, substantially enhances the anti-multiple myeloma activity of macrophages in the presence of a CD38-specific antibody (13) and by clinical data that support the synergistic effects of lenalidomide and anti-CD38 therapy (14).
On the basis of our observation that MAMs derived from patients treated with lenalidomide display phenotypical and functional shifts toward M1 polarization, we sought out to determine whether lenalidomide directly modulates macrophage polarization and to unravel the underlying molecular mechanisms.
Materials and Methods
Cell culture reagents
Cells were cultured in complete medium, which comprised RPMI1640 Medium (catalog no. #11875085, Biochrom) supplemented with Glutamine (2 mmol/L, catalog no. #G8540, Sigma), 10 mmol/L HEPES (catalog no. #P05–0110), 13 mmol/L NaHCO3 (catalog no. #P06-22100), 100 μg/mL streptomycin (catalog no. #P06-11100P), 60 μg/mL penicillin (catalog no. #P06-08100P; all from Biochrom), and 10% Human AB Serum (catalog no. #P30-2401, PAN Biotech).
The multiple myeloma cell lines, RPMI-8226 (catalog no. #ACC 402), LP-1 (catalog no. #ACC 41), IM-9 (catalog no. #ACC 117), NCI-H929 (catalog no. #ACC 163), and OPM-2 (catalog no. #ACC 50), were bought from the German Collection of Microorganisms and Cell Cultures (DSMZ, Braunschweig, Germany). Cell lines were received in 2019 and were not further authenticated. The multiple myeloma cell lines, U-266 and MM.1S, were a kind gift from J. Krönke. The murine multiple myeloma cell line 5TGM1 was a kind gift from Jens Nolting (Department of Oncology, Hematology and Rheumatology, University Hospital Bonn, Bonn, Germany). The Burkitt lymphoma cell line 291PC was established from the human λ-c-myc transgenic mouse, as described previously (15). The cells were used at <16 passages. Mycoplasma detection was performed monthly by PCR at our institution on all cell lines currently used in this study.
Human sample studies
In accordance with the Declaration of Helsinki and upon approval by the institutional ethics committee, each patient gave informed written consent prior to surgery, blood donation, or BM biopsy (multiple myeloma and healthy donors; Aachen: EK206/09; Erlangen: Ref. #3555, 36_12 B, 219_14B, 200_12B). For patients' characteristics and numbers, please refer to Supplementary Tables S1–S3. Upon single-cell isolation (see below), all cells were cryopreserved in culture medium containing 10% DMSO and stored in liquid nitrogen.
Antibodies and reagents
The following antibodies were used for immunofluorescence, flow cytometry, or Western blotting.
Antibodies specific for IKAROS family zinc finger 1 (IKZF1; 14859S, Cell Signaling Technology), CD68 (PG-M1, Dako Cytomation), CD163 (clone 10D6, Novocastra/Leica), rabbit IgG F(ab')2 Fragment 488 conjugate, mouse IgG F(ab')2 Fragment Alexa 488, rabbit IgG F(ab')2 Fragment 674 conjugate, and mouse IgG F(ab')2 Fragment 647 conjugate (all from Cell Signaling Technology) were used for immunofluorescence.
Antibodies specific for IKZF-AlexaFluor-647 (16B5C71, BioLegend), IKZF1-AlexaFluor-647 (2A9/Ikaros, BioLegend), IRF4-PE (3E4, eBioscience), IRF4-Pacific Blue (IRF4.3E4, BioLegend), IRF5-AlexaFluor-488 (IC4508G, R&D Systems), IRF5-Alexa Fluor488 (903430, R&D Systems), CD163-PE (GHI/61, eBioscience), CD163-BV421 (GHI/61, BioLegend), CD15-BV510 (W6D3, BioLegend), CD11b-FITC (M1/184.108.40.206, Miltenyi Biotec), CD11b-APC (M1/220.127.116.11, Miltenyi Biotec), CD80 (2D10, BioLegend), MHCII (L243, BD Pharmingen), CD86-PE (IT2.2, BioLegend), CD40-FITC (5C3, BioLegend), CD36-APC/Fire750 (5-271, BioLegend), PD-L1-PE (MIH1, eBioscience), PD-L1-APC (10F.9G2, BioLegend), CD204-APC (351615, R&D Systems), Annexin-V-APC (BD), 7-AAD (Sigma), CD14-PerCP (MϕP9), F4/80-PE (BM8, BioLegend), Ly6G-BV510 (1A8, BioLegend), CD115-APC-Cy7 (AFS98, BioLegend), Ly6C-BV421 (HK1.4, BioLegend), Siglec-F-PE-Vio770 (ES22-10D8, BioLegend), Glut1-APC (202915, R&D Systems), Glut2-PE (199017, R&D Systems), Glut3-Alexa Fluor488 (202017, R&D Systems), CD3-BV510 (UCHT1, BD Pharmingen), CD4-PE (SK3, BD Pharmingen), CD8-BV450 (SK1, BioLegend), IFNγ-FITC (B27, BD Pharmingen), and IL17A-AlexaFluro647 (BL168, BioLegend) were used for flow cytometry.
Western blotting and chromatin immunoprecipitation
Antibodies specific for IKZF1 (14859S, Cell Signaling Technology), IRF3 (11904S, Cell Signaling Technology), IRF4 (4299S, Cell Signaling Technology), IRF5 (13496S, Cell Signaling Technology), c-Jun (60A8, Cell Signaling Technology), NF-kB p65 (D14E12, Cell Signaling Technology), Tri-Methyl-Histone H3 (K4) (9751P, Cell Signaling Technology), Tri-Methyl-Histone H3 (K27) (9733S, Cell Signaling Technology), IgG (2729P, Cell Signaling Technology), Tubulin (Abcam, ab7291), β-actin (AC-15, Abcam), goat anti-mouse IgG-HRP (Dako, P0447), and goat anti-rabbit IgG-HRP (Cell Signaling Technology) were used for Western blotting and chromatin immunoprecipitation (ChIP).
6-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-6-deoxyglucose (6-NBDG, catalog no. #N23106) was purchased from Thermo Fisher Scientific. Lenalidomide was provided by Celgene Corporation and solubilized in DMSO (Sigma-Aldrich), which was used as a vehicle control in all experiments. Between 0.1 and 10 μmol/L of lenalidomide was added every 2 days to long-term cultures, unless otherwise indicated. Pomalidomide was purchased from Selleckchem (catalog no. #S1567).
To detect cell surface markers by flow cytometry, 1 × 105 cells were resuspended in 200 μL PBS and stained with antibodies (20 minutes at 4°C), followed by two washes with PBS. To detect intracellular proteins, cells were processed with the BD Transcription Factor Phospho Buffer Set (BD Biosciences, catalog no. #563239) according to the manufacturer's instructions. For some experiments, 123count eBeads (eBioscience, catalog no. #01-1234-42) were used for cell quantification. Flow cytometry was performed on a BD FACS Canto II (BD Biosciences). Flow cytometry data were analyzed by FlowJo v10 (BD Biosciences) or Kaluza (Beckman Coulter).
Isolation of TAMs, macrophages, and myeloma cells from human samples
BM from patients with multiple myeloma or healthy controls was stained with antibodies specific for CD163, CD15, and CD138. Macrophages (defined as CD163+CD15−) and multiple myeloma cells (defined as CD138+) were isolated by flow cytometry. Purity was greater than 95%.
Preparation of human macrophages
Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation of buffy coat preparations from the peripheral blood of healthy donors (DRK). Briefly, blood was layered over 25 mL of the Ficoll-Paque PLUS (GE Healthcare, catalog no. #17-1440-03). Gradients were centrifuged at 400 × g for 30 minutes at room temperature in a swinging-bucket rotor without the brake applied. The PBMC interface was carefully removed by pipetting and washed with PBS by centrifugation at 250 × g for 10 minutes. A total of 5 × 106 PBMCs per well were plated in a 6-well plate for 2 hours, and monocytes were isolated by adherence on plastic and cultured in complete medium in the presence of GM-CSF (50 ng/mL, Berlex) to generate M1 macrophages, or in the presence of macrophage colony-stimulating factor (M-CSF; 50 ng/mL, R&D Systems, catalog no. #216-MC) to obtain M2-like macrophages. Macrophages were detached with EDTA (1 mmol/L, Sigma) after 6 days of culture. In some experiments, M-CSF–generated macrophages were treated with IFNγ and lipopolysaccharide (LPS, 100 ng/mL), IL4 (10 ng/mL, R&D Systems, catalog no. #204-IL), or IL10 (10 ng/mL, R&D Systems, catalog no. #217-IL/CF) for a further 24 hours to generate other macrophage phenotypes. Phenotype was evaluated by flow cytometric analysis of expression of surface CD163, CD86, CD204, and CD11b. Purity of generated macrophages was greater than 90% and minimally contaminated by CD3+ and CD14+ cells.
Formalin-fixed, paraffin-embedded (FFPE) BM specimens were retrieved from the archive of the Institute of Pathology, FAU Erlangen-Nuremberg (Erlangen, Germany), including bone biopsies with plasma cell myeloma before (n = 7) and after lenalidomide therapy (n = 7; for patients' characteristics, please refer to Supplementary Table S1).
Immunocytochemistry of FFPE samples
Sections (2-μm-thick) of FFPE BM biopsies of patients with multiple myeloma were deparaffinized in xylene and rehydrated with graded ethanol (100%–70%). Antigen retrieval was performed in a steam cooker with citrate buffer (pH 6) for 2 minutes at 120°C. Primary antibodies (specific for CD68, 1:200, clone PG-M1, Zytomed Systems and CD163, 1:500, NCL-CD163, clone 10D6, Novocastra/Leica) were added overnight at room temperature. An alkaline phosphatase-labeled polymer kit was used according to the manufacturer's instructions (Zytochem-Plus AP-Polymer Kit, Zytomed Systems, catalog no. #POLAP-006) for detection. Fast Red (Sigma-Aldrich, catalog no. #F4648) was applied as chromogen.
Quantitative evaluation of histologic sections
A computer-assisted analysis was performed on whole-block sections. For image acquisition, a standard light microscope and a CCD camera or a Mirax Midi System (Zeiss) for digital slide scanning and Mirax Viewer Software (Zeiss) were applied. Numbers of positive cells (defined as brown spots with a minimum size of >5 μm2) per mm² in the area with the densest infiltration were evaluated using the image analysis software BZ-9000 (Keyence).
To analyze the effect of lenalidomide on M2-like macrophages, we generated M2-like macrophages by stimulating monocytes from healthy donors with M-CSF (50 ng/mL, 6 days; as described in preparation of human macrophages section). The M2-like macrophages were incubated in complete medium with or without 1 μmol/L lenalidomide for 3 days. Cells from five different donors were analyzed by RNA sequencing (RNA-seq; 100 bp, 30,000,000 Single Reads, GATC Biotech). RNA was isolated from the cells using an RNeasy Mini Kit (Qiagen, catalog no. #74104) according to the manufacturer's recommended protocol. A total of 100 ng RNA of each total RNA sample was used as template. RNA-seq was performed on Illumina HiSeq 2500. Approximately 1,100 genes were significantly (FDR < 0.05) differentially expressed comparing medium versus lenalidomide treatment. Differentially expressed genes were compared with previously published M1 and M2 gene signatures (16, 17). Heatmaps based on the RNA-seq data depict log2 ratios between the FPKM values of two different conditions and were produced in R with libraries gplots or pheatmap. RNA-seq data are deposited in the NCBI Gene Expression Omnibus (GEO) database under GEO accession number GSE165332.
Reanalysis of publicly available datasets
GEO dataset GSE5099, recorded with Affymetrix Human Genome U133A array, was used to illustrate expression differences in macrophages due to polarization. The analysis was performed in R, with preprocessed matrix files obtained through the download functionality of the library GEOquery. We selected the human monocyte–derived M-CSF–matured macrophage samples, either unstimulated at 7 days (n = 3), activated with IFNγ and LPS (n = 3), or activated with IL4 (n = 3). The expression and clustering tendency of several preselected transcription factors were visualized in a heatmap with the library pheatmap using Euclidean distance and complete linkage.
Macrophages from patients (with or without lenalidomide treatment) or after in vitro treatment (lenalidomide, 1 μmol/L, 72 hours or siRNA) were incubated in complete media for 24 hours with no stimulus or with LPS (100 ng/mL). ELISA, using the Quantikine Immunoassays Kit manufactured by R&D Systems, determined IL12 (catalog no. #D1200) and IL10 (catalog no. #D1000B) concentrations in the culture supernatants. Assays were performed according to the instructions provided with each kit following the manufacturer's instructions. Color development was assessed utilizing the Microplate Autoreader (EL309) from Bio-Tech Instruments.
IKZF1 DNA-binding activity was quantitatively assessed using the TransAM FLEXI Kit (ActiveMotif, catalog no. #40098) per the manufacturer's protocol. The 58-bp oligonucleotides containing a 5-bp core consensus sequence of IKZF1 binding motif (IKZF1: 5′-TCAGACATCAGAAAAAGGGAATTCCGTCACTCAGACACTTTTGGTACTGTCACTTGCT-3′-biotin) were custom synthesized and biotinylated (Metabion Inc.). Oligonucleotides were duplexed by incubating 1 pmol/μL of each oligonucleotide in DEPC Water (Thermo Fisher Scientific, catalog no. #AM9916) together at 95°C for 5 minutes and ramping back to 4°C by decreasing 1°C/minute. Twenty micrograms of nuclear extract or cytoplasmic fraction prepared from generated macrophages (as described in preparation of human macrophages section), 50 μL binding buffer, and 1 pmol of biotinylated oligonucleotide were incubated for 30 minutes at room temperature prior to placement in a well on streptavidin-coated 96-microtiter plates (ActiveMotif, catalog no. #40098). After washing, wells were incubated sequentially with a primary antibody for IKZF1, followed by anti-rabbit peroxidase-conjugated antibody (see antibodies and reagents section). After substrate addition, peroxidase activity was measured by reading at 450 nm utilizing the Microplate Autoreader (EL309) from Bio-Tech Instruments.
qPCR-based gene expression analysis
To detect mRNA expression in macrophages, cells were lysed and mRNA was isolated from the cells using an RNeasy Mini Kit (Qiagen, catalog no. #74106) according to the manufacturer's recommended protocol. Contaminating DNA was removed by DNase I treatment. cDNA was prepared by using Random Decamers (Applied Biosystems, catalog no. #AM5722G) as suggested by the manufacturer. Expression of mRNA was quantified by quantitative RT-PCR (qPCR) using a StepOnePlus Real-Time PCR System (Thermo Fisher Scientific) or a Rotor Gene Q (Qiagen). The mRNA levels were determined by normalizing expression of mRNA of interest [using the following Quantitec primer assays from Qiagen: CCL5 (catalog no. #QT00090083), CXCL10 (catalog no. #QT01003065), and CXCL12 (catalog no. #QT00087591) in human macrophages and Tnf (catalog no. #QT00104006), Nos2 (catalog no. #QT00100275), and Arg1 (catalog no. #QT00134288) in mouse macrophages] to β2 microglobulin (mouse, catalog no. #QT01149547 and human, catalog no. #QT00088935) and 18S (mouse, catalog no. #QT02448075 and human, catalog no. #QT00199367).
Macrophages treated with or without lenalidomide (1 μmol/L, 72 hours) were adhered to poly-Lysine slides and fixed with paraformaldehyde (4%), treated with Triton X-100 (Sigma, 0.3%, 10 minutes on ice), and incubated with anti-IKZF1 overnight at 4°C in a humid chamber. After washing with PBS, the slides were incubated with anti-rabbit IgG F(ab')2 Fragment 488 conjugate, anti-mouse IgG F(ab')2 Fragment Alexa 488, or anti-rabbit IgG F(ab')2 Fragment 674 conjugate for 2 hours at 4°C in a humid chamber. Cell membranes and cell nuclei were counterstained using 10 μg/mL WGA (green) and 1 μg/mL DAPI (blue). Slides were analyzed by z-stack sections creating up to 10 optical slices (0.5-μm-thick each) using Confocal Microscope (LSM700, Zeiss) at 630× magnification.
Protein extraction and Western blotting
Modified RIPA buffer [50 mmol/L Tris-HCl (Roth), pH 7.4; 1% NP-40 (Fluka); 0.25% sodium deoxycholate (Fluka); 150 mmol/L NaCl (Roth); and 1 mmol/L EDTA] and Protease Inhibitor Cocktail Tablets (Sigma-Aldrich, catalog no., #5892970001) were added to macrophages (1 × 106/well) after removing culture medium. Macrophages were lysed on ice for 10 minutes. Cell debris were removed by centrifugation (14,000 rpm, 10 minutes) and protein content in the supernatant was determined using a BCA Protein Assay per the manufacturer's protocol (Pierce, catalog no. #23225). Lysates were boiled for 5 minutes at 95°C in 1 × Laemmli sample buffer [62.5 mmol/L Tris HCl, pH 6.8; 10% (w/v) glycerin, 2% (w/v) SDS, 50 mmol/L DTT, and 0.01% (w/v) bromophenol blue] and analyzed by SDS-PAGE (12%, 10 μg protein/lane) and semi-dry Western Blot (TransBlotTurbo, Bio-Rad). The membranes were blocked for 2 hours at room temperature with 5% BSA in TBS-T (TBS with 0.1% Tween20) and incubated overnight with the primary antibodies (IKZF1, 1:1,000; IKZF3, 1:1,000; IRF3, 1:1,000; IRF4, 1:500; IRF5, 1:500; β-actin, 1:1,000; and Tubulin, 1:5,000 in 5% BSA/TBS-T) at 4°C. The secondary antibody was applied subsequently for 6 hours at 4°C. Proteins were detected by chemiluminescence (SignalFire ECL Reagent, Cell Signaling Technology) following the manufacturer's protocol on an Amersham Imager 600.
Flow cytometry–based killing assay
Multiple myeloma cells were labeled with Cell Proliferation Dye eFluor 670 (CPD; Thermo Fisher Scientific, catalog no. #65-0840-85) and cocultured with macrophages [effector to target ratio (E:T) = 1:1] in sterile FACS tubes in complete medium with 10% FCS in the presence or absence of the CD38-specific antibody, daratumumab (1 μg/mL), for 24 hours at 37°C, 5% CO2. An irrelevant IgG1 antibody (BioLegend, catalog no. #400123) was used as an isotype control. To distinguish between phagocytosed CPD+ multiple myeloma cells and free multiple myeloma cells, macrophages were counterstained with anti-CD11b-FITC. Absolute numbers of surviving multiple myeloma cells were determined by flow cytometric analysis of CD11b−CPD+ cells using 123count eBeads according to the manufacturer's protocol (eBioscience, catalog no. #01-1234-42). The percentage of daratumumab-mediated killing was then calculated by using the following formula: daratumumab-mediated killing = 100 – [(absolute number of surviving multiple myeloma cells in the presence of daratumumab/absolute number of surviving multiple myeloma cells in the presence of IgG1) × 100].
Viability of multiple myeloma cells after coculture with M1 and M2-like macrophages
M1 and M2-like macrophages (treated with or without lenalidomide, 0.1–10 μmol/L, 72 hours) were incubated in complete medium in round bottom microplates (96-well) for 24 hours together with different multiple myeloma cell lines (at various E:T, 1:1, 5:1, and 10:1). Nonadherent cells were harvested and stained with a CD11b antibody (to exclude macrophage contamination). Viability of CD11b− cells was measured by Annexin V and 7AAD Staining (BD, catalog no. #559925) according to the manufacturer's instructions. Multiple myeloma cells were immediately analyzed by flow cytometry (FACS Canto II, BD).
TH1 and TH17 assay
M2-like macrophages were generated by adherence on plastic (as described in preparation of human macrophages section) and nonadherent cells from the same donor were cryoconserved in culture medium containing 10% DMSO. After this, M2-like macrophages were cultured in complete medium with lenalidomide (1 μmol/L) for 72 hours. In some experiments, siRNA-transfected macrophages (directed to IKZF1) were used. Pan T cells were isolated from thawed PBMCs (Human Pan T-cell Isolation Kit, Miltenyi Biotec, catalog no. #130-096-535) according to the manufacturer's instructions, and coincubated in fresh complete medium (lenalidomide free) with autologous lenalidomide-pretreated macrophages (macrophages:T cells ratio = 1:10). After 24 hours, cells were stimulated with phorbol-12-myristat-13-acetat (20 ng/mL) and ionomycin (500 ng/mL) for 5 hours. Brefeldin (10 μg/mL) was added for the last 2 hours to the culture and T cells were analyzed for IFNγ and IL17A by flow cytometry.
Extracellular flux analysis and glucose uptake
Bioenergetics of M1/M2-like–polarized macrophages and of lenalidomide-treated macrophages (1 μmol/L, 72 hours) were determined using an XFe96 Extracellular Flux Analyzer (Seahorse Bioscience) using Seahorse XF Assay Medium (Agilent Technologies) supplemented with 2 mmol/L Sodium Pyruvate (Sigma-Aldrich, catalog no. #S8636) and 12.5 mmol/L Glucose (Sigma-Aldrich, catalog no. #G8270). Extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) were recorded under baseline conditions and after port injection of LPS (100 ng/mL, Sigma-Aldrich, catalog no. L3024) and recombinant CD40L (1 μg/mL, Novus Biologicals, catalog no. #NBP2-26514).
Influx of glucose was semiquantified by flow cytometry based on the uptake of the fluorescence glucose analogue, 6-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (6-NBDG, Thermo Fisher Scientific, catalog no. #N23106), by culturing cells in glucose-free RPMI Medium (Thermo Fisher Scientific, catalog no. #11879-020) containing 0.3 mmol/L 6-NBDG for 15 minutes at 37°C, 5% CO2.
siRNA transfection of macrophages
Freshly generated macrophages were cultured for 2 hours in prewarmed serum-free media (AIM V, Thermo Fisher Scientific, catalog no. #12055091) in 24-well plates. Adherent macrophages were transfected with a nontargeted control (catalog no. #4390843) siRNA or on-target–specific siRNA directed to IKZF1 (assay IDs: s20180 and s20181) or IKZF3 (assay ID: s22146) sequences (silencer select siRNA, Thermo Fisher Scientific) using Lipofectamine RNAiMax according to the manufacturer's protocol (Life Technologies, catalog no. #13778030, 15 pmol siRNA:1.5 μL Lipofectamine/well) 24 hours prior to LPS incubation (100 nmol/L, 24 hours).
Macrophages (10 × 106) were cultured in complete medium in a 15-cm dish and treated with lenalidomide (1 μmol/L, 24 hours). Cells were fixed with 37% formaldehyde for cross-linking histones to DNA directly (final formaldehyde concentration was 1%). Fixation step was stopped after 10 minutes with glycine. After cross-linking, nucleosomal DNA was prepared from 3 × 107 cells and processed using Simple ChIP Enzymatic Chromatin IP Kit (with magnetic beads, from Cell Signaling Technology, catalog no. # 9003) according to the manufacturer's instructions. In addition to the digestion step with the micrococcal nuclease, chromatin was sonicated using a TOMY-UR-20P to obtain DNA fragments. Fragment lengths peaked between 150 and 300 bp. ChIP with antibodies (antibody against tri-methyl-histone H3 (K4), tri-methyl-histone H3 (K27), or normal IgG) was also performed using Simple ChIP Enzymatic Chromatin IP Kit (with magnetic beads, from Cell Signaling Technology, catalog no., #9003) according to manufacturer's instruction. Enrichment in the proximal region of the IRF4 promoter was quantified by quantitative PCR (qPCR) using the following primers: forward 5′-ACTCTCAGTTTCACCGCTCG-3′ and reverse 5′-CTCCGGGTCCTCTCTGGTAT-3′, and enrichment of the IRF5 promoter was quantified by qPCR using the following primers: forward 5′-CTCTGAGGGAGGCCTGCAAT-3′ and reverse 5′GAGCCTCAGTTTCCCCTGAA-3′. Data were normalized to input DNA.
For analyzing published human IKZF1 ChIP data, we extracted UCSC's genome browser IKZF1 signal track provided by the ENCODE project (GM12878). For analyzing public murine IKF1 ChIP data, signal data were acquired through a custom track in the UCSC genome browser for the ChIP antibody AB_2614961 manufactured by Active Motif. The genomic loci were manually extracted and inspected through the Integrative Genomics Viewer (18).
Macrophages were labeled with 5 μmol/L carboxyfluorescein diacetate succinimidyl ester (CFSE) according to the manufacturer's instructions (catalog no. #C34554, Thermo Fisher Scientific). CFSE-labeled murine macrophages were plated in 8-chamber slides (Nalge Nunc International) and coincubated for 2 hours with CPD-labeled 5T33M cells or 291PC lymphoma cells (E:T = 1:1, 37°C) in the absence or presence of anti-CD38 (NIMR-5, Abcam) or anti-CD20 (SA271G2, BioLegend, 10 μg/mL). An irrelevant IgG1 antibody was used as an isotype control. Cells were washed and fixed with paraformaldehyde (4%). Fixed cells were analyzed by z-stack sections creating up to 10 optical slices (0.5-μm thick each) using Confocal Microscope (LSM700, Zeiss) at 630× magnification.
B6.CrbnI391V mice (C57BL/6-Crbntm1.1Ble/J; stock no., 032487) were a kind gift from Benjamin Ebert (Dana-Farber Cancer Institute, Boston, MA). C57BL/6 (B6) mice were purchased from Charles River Laboratories. Mice were maintained in a pathogen-free facility and fed a standard diet. Experiments were approved by and performed in accordance with the local authorities of Germany (approval no. #1302, Regierungspräsidium Tübingen).
Treatment of B6.CrbnI391V mice and isolation of mouse BM macrophages
For mouse experiments, lenalidomide was purchased from Tocris Bio-Techne (catalog no. #6305). Lenalidomide was dissolved in DMSO (Sigma-Aldrich) to a concentration of 100 mmol/L and stored at −80°C. According to the individual weight of each mouse, lenalidomide was administered daily at 10 mg/kg i.p. in a total volume of 100 μL PBS containing 10% DMSO. Mice were treated for 7 days before being euthanized to obtain femur, tibia, and hip bones.
BM from the femur and tibia was harvested by flushing with PBS. The resulting single-cell suspension was incubated with Mouse CD115 MicroBeads (Miltenyi Biotec, catalog no. #130-096-354) for 15 minutes on ice. MACS Bead Separation (Miltenyi Biotec, catalog no. #130-042-501) was performed to obtain CD115+ myeloid cells following the manufacturer's protocol. Purity was analyzed by expression of surface markers F4/80, Ly6C, Siglec-F, and CD11b by flow cytometry and was greater than 90%.
GraphPad Prism 8 and Microsoft Excel software were used for statistical analysis. Comparisons between patients and controls were done with nonparametric Mann–Whitney U test. Two-tailed Student t tests were used for all other figures. The results are presented as mean ± SEM. Differences were considered significant if P < 0.05.
Patient-derived MAMs skew from an M2-like to an M1-like phenotype during IMiD treatment
First, we performed IHC staining for CD68 (as a pan-macrophage marker) and CD163 (as an M2-like macrophage marker) of BM biopsies from patients newly diagnosed with multiple myeloma (see Supplementary Table S1 for the patients' characteristics). As displayed in the representative IHC image in Fig. 1A, we observed a universally dispersed BM infiltration with CD68+CD163+ MAMs with M2-like properties (19). When analyzing paired follow-up samples during IMiD therapy, we found a diminished density of CD163+ macrophages. However, density of total CD68+ macrophages remained stable (Fig. 1A; Supplementary Fig. S1), suggesting a shift in the composition of macrophage subsets rather than an overall depletion of macrophages.
Next, we analyzed the cell surface expression of molecules linked to immune activation (CD40, CD80, CD86, and MHCII), immune suppression (PD-L1), and an M2-like phenotype (CD36, CD136, CD204, and CD206) on macrophages of untreated versus IMiD-treated patients in BM-derived single-cell suspensions. Expression of CD40, CD86, and MHCII was found increased and expression of CD36, CD163, CD204, CD206, and PD-L1 decreased on CD163+CD15− macrophages during IMiD therapy (Fig. 1B; Supplementary Fig. S2A and S2B; for patients' characteristics, please refer to Supplementary Table S2). In addition, in macrophages from proteasome inhibitor–treated patients, the marker profile remained, on average, unchanged before and during treatment, reinforcing the notion that the observed shift toward a more tumoricidal phenotype is (co-)mediated by the IMiDs and is not an epiphenomenon of a multiple myeloma–directed intervention. M1-polarized macrophages typically produce IL12 and their M2-like–polarized counterparts IL10, which is associated with a poor prognosis in multiple myeloma (20). We observed substantially skewed IL10/IL12 cytokine production by MAMs toward IL12 during lenalidomide treatment, further pointing to an M1-like shift (Fig. 1C).
Given that IMiDs modulate the CRBN-CRL4 E3 ligase to ubiquitinate and degrade the transcription factor IKZF1 (21), and that IKZF1 is a transcriptional repressor of lymphoid cells (22), we hypothesized that IKZF1 might regulate polarization of human macrophages. When we measured IKZF1 protein levels in macrophages from untreated and IMiD-treated patients with multiple myeloma, we found that macrophages from IMiD-treated patients had markedly reduced IKZF1 expression in comparison with macrophages from untreated patients (Fig. 1D).
IKZF1 expression is upregulated during M2-like polarization
To further establish a connection between IKZF1 and an M2-like phenotype, we reassessed publicly available transcriptome data (accession no. GSE5099) from monocyte-derived macrophages generated under M1- or M2-like–polarizing conditions (16). Monocytes cultured in the presence of the M2-like–promoting cytokines, M-CSF and IL4, had enhanced IKZF expression (Fig. 2A). We confirmed higher IKZF1 protein levels in M2-like macrophages by FACS and Western blotting when testing different M1- and M2-like–promoting conditions (M1-like: GM-CSF or IFNγ + LPS and M2-like: M-CSF, IL4, or IL10; Fig. 2B and C; Supplementary Fig. S3; refs. 23, 24). Nuclear localization and DNA binding are prerequisites for an optimal transcription function of IKZF1. We visualized the subcellular distribution of IKZF1 in M1 and M2-like macrophages using confocal microscopy. We detected a prominent nuclear localization of IKZF1 in M2-like macrophages (in contrast to M1 macrophages, which had only limited nuclear IKZF1; Fig. 2D; Supplementary Fig. S4). Furthermore, sequence-specific DNA binding of IKZF1 was significantly higher in nuclear extracts from M2-like–polarized macrophages as compared with M1-polarized macrophages (Fig. 2E), suggesting increased functional activity.
IKZF1 loss promotes repolarization to a proinflammatory, tumoricidal phenotype
To further elucidate the role of IKZF1 in macrophage polarization, we degraded IKZF1 in M2-like macrophages through exposure to noncytotoxic doses of lenalidomide (Fig. 3A; Supplementary Fig. S5A–S5C), and then performed RNA-seq. Upon IMiD exposure, genes that are characteristic for the M1 signature were upregulated and genes typically enriched in M2-like macrophages were downregulated (Fig. 3B; Supplementary Fig. S6; ref. 16). Enhanced expression of CD40, CD86, CCL5, and CXCL10 (M1 markers) paralleled by a reduction of CD163, CD204, and CXCL12 (M2-like markers) as assessed by flow cytometry and qPCR, supporting the hypothesis that IMiD triggers polarization to an M1-like phenotype from M2-like phenotype (Supplementary Fig. S7A and S7B). In addition, secretion of multiple myeloma–promoting (25) B-cell–activating factor (BAFF) and a proliferation-inducing ligand (APRIL) was significantly diminished (Supplementary Fig. S7C). Genetic ablation of IKZF1 using siRNA mimicked lenalidomide-mediated effects in shifting the cytokine profile away from IL10 (indicative for an M2-like polarization) toward IL12 (indicative for an M1 polarization), further pointing toward the direct involvement of IKZF1 in macrophage polarization to an M1 phenotype (Fig. 3C). Similar to lenalidomide, pomalidomide (another potent second-generation IMiD) increased expression of CD40, CD86, and IL12 on M-CSF–generated macrophages, while CD163, CD204, and IL10 were decreased (Supplementary Fig. S8). In line with our own observations (Supplementary Fig. S9A), and the previously reported abundance of M2-like macrophages within the multiple myeloma-BM niche (9), we measured higher IKZF1 levels in BM macrophages isolated from untreated patients with multiple myeloma in comparison with BM macrophages from healthy donors (Supplementary Fig. S9B; for patients' characteristics, please refer to Supplementary Table S3).
Next, we investigated the functional properties of IMiD-treated MAMs and macrophages to allow a more definitive categorization. Activity against primary multiple myeloma cells and multiple myeloma cell lines (RPMI-8226, OPM-2, and MM.1S) was potentiated by lenalidomide in the presence and absence of the therapeutic monoclonal CD38-specific antibody, daratumumab (Fig. 3D; Supplementary Figs. S10 and S11). Consistent with these data, RNA-seq analyses revealed upregulation of genes encoding molecules that activate antibody-dependent cellular phagocytosis (ADCP), such as SLAMF7 (26) and CD44 (27), and downregulation of ADCP-inhibiting molecules, such as CD300a (28) and SIRPa (Supplementary Fig. S12; ref. 29). Interestingly, enhanced phagocytosis is considered a prominent feature of M2-like macrophages (30), which highlights the shortcomings of the current M1/M2 paradigm (31).
Recent evidence suggests that IKZF1 acts as a metabolic gatekeeper by repressing glycolysis (32), which in turn is critical for M1 macrophages to exercise their tumoricidal activity (33). We observed a significantly higher ECAR, which is indicative of aerobic glycolysis, in macrophages generated under M1-promoting conditions. The overall ratio between ECAR and the OCR, which is indicative of oxidative phosphorylation, was skewed toward ECAR (Supplementary Fig. S13). Application of lenalidomide (and the accompanying IKZF degradation) significantly boosted the M2-like macrophages' ECAR (but not OCR) under steady-state conditions and upon stimulation by CD40L or LPS (Fig. 3E; Supplementary Fig. S13). This is consistent with the increased expression of key glycolytic molecules, glucose transporters, and elevated glucose uptake that we observed in lenalidomide-treated M2-like macrophages (Supplementary Fig. S14). Next, we tested the TH1- and TH17-promoting capacity of lenalidomide-treated M2-like macrophages, as these are key features of M1 macrophages. Production of both IFNγ and IL17A was significantly enhanced in T cells cocultured with IMiD-treated M2-like macrophages, and we recapitulated this finding using siRNA against IKZF1 (Fig. 3F; Supplementary Figs. S15–S17).
IKZF1 epigenetically controls the balance between IRF4 and IRF5 expression
IFN regulatory factors 3, 4, and 5 (IRF3–5) are potential transcriptional targets of IKZF1, as identified within the ENCODE project database (Fig. 4A; ref. 34). The transcription factor IRF4 is downregulated in multiple myeloma cells after lenalidomide-induced IKZF1 degradation (35). Because IRF4 and IRF5 signaling balances between M1/M2-like macrophage polarization (17, 36), we investigated whether this was the mechanistic link between IKZF1 and the lenalidomide-induced M1 phenotype. IKZF1 degradation by lenalidomide, or by genetic ablation, yielded increased IRF5 protein and gene expression and decreased IRF4 protein and gene expression, whereas IRF3, another member of the IRF family, remained unaffected (Fig. 4B–D; Supplementary Fig. S18A). We ruled out a lenalidomide-mediated change in the level of IRF7, which has been linked previously to an IFN response triggered by the novel pleiotropic pathway–modifier, CC-122 (Supplementary Fig. S18B; ref. 37).
Previous studies have highlighted the importance of TNF in promoting the M1 phenotype (38, 39). Here, we did not observe any effects of lenalidomide on TNF secretion or signaling (Supplementary Fig. S19). Beyond transcription factors, epigenetic regulation controls gene expression. IKZF1 holds the potential to regulate chromatin remodeling and target gene expression alone and in complex with histone deacetylase 1 (HDAC1; refs. 40, 41). In the case of histone modifications, trimethylation of H3K4 is associated with active gene transcription, whereas trimethylation of H3K27 is linked to gene silencing. When performing ChIP analyses in M2-like macrophages, we noticed a high H3K4 together with a low H3K27 trimethylation signal in the promoter region close to the transcription start site of IRF4 and vice versa for IRF5 (Fig. 4E). This is consistent with the high IRF4 and low IRF5 mRNA expression levels that we detected in M2-like macrophages (Supplementary Fig. S18A). After IMiD treatment, the trimethylation state of H3K4 and H3K27 in the promotor regions of IRF4 and IRF5 was reversed, suggesting that lenalidomide regulates gene expression of IRF4 (i.e., downregulation) and IRF5 (i.e., upregulation) by histone modification (Fig. 4E). Finally, we found also a shift in the balance of the IRF4/IRF5 axis in macrophages retrieved from lenalidomide-treated patients (Fig. 4F).
Lenalidomide treatment of CrbnI391V mice promotes macrophage tumoricidal activity
A single amino acid at murine Crbn position 391, which is in the IMiD binding site, shields mouse CRBN from the degradative effects of IMiDs (42), thereby preventing extrapolation of IMiD activity in murine models (43) to the human system. To validate our data in vivo, we utilized a humanized CrbnI391V knock-in mouse model in which the critical amino acid of the mouse CRBN is replaced by the relevant human residue (44). Mice underwent daily treatment with either the solvent or 10 mg/kg lenalidomide for 7 consecutive days and BM-resident macrophages were isolated for further analyses (Fig. 5A).
In line with our ex vivo and in vitro studies using patient- and healthy donor–derived macrophages, we observed diminished expression of IKZF1 and IRF4 and an increase in IRF5 in BM-derived macrophages upon treatment with lenalidomide (Fig. 5B; Supplementary Fig. S20). Differential control of IRF4 and IRF5 by IKZF1 in murine macrophages (in analogy to our findings in human macrophages) was further supported by public ChIP sequencing (ChIP-seq) data (custom signal track provided by ActiveMotif, mAb_AM39355), revealing IRF4 and IRF5 as potential IKZF targets in mice (Supplementary Fig. S21). The frequency of BM-residing macrophages was not affected by lenalidomide treatment (Supplementary Fig. S22).
Next, we examined whether IKZF1 degradation promotes macrophage tumoricidal activity in CrbnI391V mice. Expression of inducible nitric oxide synthase (iNOS), which has been shown previously to mediate anti-myeloma activity of macrophages (45), was significantly enhanced by lenalidomide application, whereas levels of arginase-1 and PD-L1, molecules being linked to myeloid cell–mediated T-cell suppression in multiple myeloma (Fig. 5C; refs. 6, 46). In contrast, when treating BM-derived macrophages isolated from wild-type mice, we did not observe any effects on IKZF1, IRF4, and IRF5 or on iNOS, arginase-1, and PD-L1 (Supplementary Fig. S23), thus further corroborating the notion that the effects on macrophage polarization elicited by IMiDs are CRBN dependent.
Finally, we confirmed using this model that, in vitro M2-like–polarizing conditions promote IKZF1 expression and an increased IRF4/IRF5 ratio in BM-derived macrophages, which can be efficiently antagonized by lenalidomide treatment (Supplementary Fig. S24).
ADCP represents a principal tumoricidal effector mechanism of macrophages. On the basis of our observations, we reasoned that lenalidomide treatment enhances the efficacy of mAbs by increasing ADCP. To test this, we coincubated macrophages from lenalidomide-treated and -untreated CrbnI391V mice with multiple myeloma (5TGM1) or lymphoma (291PC) cell lines in the absence or presence of monoclonal CD38- or CD20-specific antibodies, respectively. We assessed ADCP by confocal microscopy. Treatment with lenalidomide alone showed only a small and variable effect on ADCP capacity, but the combination of lenalidomide and either antibody resulted in a significant enhancement of ADCP (Fig. 5D and E). These data are consistent with observations from clinical studies in multiple myeloma and lymphoma showing that combining IMiDs with therapeutic antibodies yields beneficial effects (47, 48). Overall, these experiments indicate that lenalidomide treatment promotes the tumoricidal activity of macrophages in vivo.
Macrophages possess pleiotropic functions in health and disease, which are mirrored by their diverse functional states. Transcriptional control of macrophage differentiation is currently a subject of intense investigation. To date, control of macrophage polarization has largely been attributed to a small group of molecules, including NF-κB, STTA proteins, members of the Krüppel-like factor family, and peroxisome proliferator-activated receptor-γ (49). However, the specific transcriptional program that regulates TAMs, such as the MAMs, remains poorly understood. TAMs populate the tumor tissue and acquire an M2-like phenotype in response to tumor microenvironmental stimuli (50). In multiple myeloma, macrophages are present in abundance, subvert immune responses, directly promote tumor growth and survival, and are linked to an inferior outcome (3–7). However, MAMs retain their capacity to elicit antitumor functions, and observations from preclinical models suggest that macrophage-repolarizing approaches could be a promising strategy for controlling/eradicating multiple myeloma (9, 45). Therefore, it is essential to identify the underlying molecular mechanisms by which they retain their antitumor capacity.
Here, we revealed a role for IKZF1 as a regulator of macrophage polarization. Selective degradation of IKZF1 by lenalidomide (21) or its genetic ablation with siRNA promoted a functional shift in macrophages away from an anti-inflammatory M2-like phenotype toward a tumoricidal M1 phenotype. ADCP efficacy was significantly augmented, which is consistent with emerging clinical reports on the beneficial action of combining IMiDs with therapeutic mAbs in disorders such as multiple myeloma and lymphoma (47, 48). Lenalidomide-mediated degradation of IKZF1 was very rapid, occurring in ≤3 hours, and could allow macrophages a fast repolarization from M2 like to M1, which warrants further analyses, especially in vivo. Nevertheless, the molecular mechanism(s) by which IKZF1 is upregulated in MAMs remains to be elucidated. Several potential regulatory regions have been identified in the proximity of the Ikzf1 locus, suggesting a complex regulatory network that is required to drive IKZF1Ikaros expression.
We are aware of the limitations of the M1/M2-like paradigm (24). Moreover, TAMs represent a unique polarized population in which M1 and M2-like signatures can coexist. The resulting TAM phenotype is determined by the composition of activating or inhibitory stimuli within the tumor microenvironment. To study the role of IKZF1 in macrophage reeducation, we used an in vitro polarization model, in which we generated M2-like macrophages that recapitulate a variety of the phenotypical MAM features, including being BAFF+, CD36high, CD163high, CD204high, IL10+, IL12−, and iNOSlow. These cells also supported the growth of multiple myeloma cells and protected malignant cells from chemotherapy-induced cell death, also indicating that they have a functional resemblance to MAMs.
To date, IKZF1 has been described as a central regulator of lymphocyte differentiation (51). During the early stages of lymphopoiesis, IKZF1 contributes to transcriptional priming of hematopoietic stem cells and their direct progeny, the lympho-myeloid primed multipotent progenitors (52). The transcription factor also holds a prominent role in macrophage chromatin regulation, supporting sustained transcription of numerous innate immune response genes (53). IKZF1 mainly regulates gene expression through association with components of the nucleosome remodeling and deacetylase (NuRD) complex (54), including HDAC1/2 and the ATP-dependent chromatin remodeling proteins, CHD3/4. The NuRD complex is an epigenetic regulator of chromatin structure and gene expression that controls both transcriptional repression, as well as activation. Degradation, or genetic ablation with siRNA, of IKZF1 in macrophages led to a repression of IRF4, which is crucial for M2-like macrophage polarization (36). This is in line with findings in multiple myeloma cells, suggesting that IRF4 is a direct, positively regulated target gene of IKZF1 (21), and we saw that IMiD treatment inhibited IRF4 transcription in multiple myeloma cells. Simultaneously, we noticed increased transcription of IRF5. CHIP-seq data from human and murine cells, together with our methylation analyses, indicate that IRF5 represents a negatively regulated IKZF1 target gene. High IRF5 expression is characteristic of proinflammatory M1 macrophages and has been previously shown to repress IL10 and activate IL12, thereby leading to a potent TH1 and TH17 response (17); effects that we could emulate by targeting IKZF1. Here, we have shown that IKZF1 controls the balance of the IRF4/IRF5 regulatory axis, which is critical for establishing M1/M2-like macrophage phenotypes. Loss of IKZF1 skews this balance toward IRF5 and an M1 phenotype. However, given the complexity of macrophage polarization and plasticity, involvement of other factors and even other IKZF1-mediated effects (beyond and in parallel to IRF4/IRF5 regulation) cannot be ruled out and will require further investigation.
Our findings could have implications beyond tumor immunology, including for inflammatory diseases. IKZF1 expression is found to be diminished in several proinflammatory conditions, such as systemic lupus erythematosus (SLE). SLE patient–derived PMBCs show reduced IKZF1 transcript levels (55) and a genome-wide association study identified IKZF1 as an SLE susceptibility gene (56). In addition, IKZF1 deficiency in host hematopoietic cells exacerbates graft-versus-host disease (GvHD) following allogeneic hematopoietic stem cell transplantation (57). The fact that M1-polarized macrophages can be detrimental in both SLE and GvHD (42, 58), leads us to speculate that lack of IKZF contributes to pathogenesis in these conditions through an M1/M2-like disequilibrium.
Taken together, we revealed what we believe to be a previously unknown role of IKZF1 in macrophage differentiation. IZKF shifts the IRF4/IRF5 balance toward IRF4, thus favoring differentiation of an anti-inflammatory M2-like type macrophage that closely resembles TAMs (such as MAMs). Consequently, in vitro and ex vivo IKZF1 targeting unleashed the macrophages' tumoricidal potential, including ADPC, explaining observations from recent clinical trials showing that combining IKZF1-degrading IMiDs and therapeutic antibodies is an effective treatment in multiple myeloma and lymphoma. Future development of strategies to up- or downregulate IKZF1 might be useful for controlling macrophage polarization in accordance to the underlying biology of the disease.
L. Röhner reports other from International Graduate School in Molecular Medicine Ulm (course fees) outside the submitted work. C. Lischer reports grants from Manfred-Roth Stiftung during the conduct of the study. M. Eberhardt reports grants from BMBF (German Federal Ministry of Education and Research) during the conduct of the study. J. Vera reports grants from BMBF and Manfred-Roth Stiftung during the conduct of the study. C. Schütz reports grants from German Cancer Foundation during the conduct of the study. J. Krönke reports grants from Deutsche Forschungsgemeinschaft and Kasprzak Foundation during the conduct of the study, as well as personal fees from Celgene and Takeda outside the submitted work. H. Bruns reports grants from Celgene, Wilhelm-Sander Foundation, and German Research Foundation during the conduct of the study. No disclosures were reported by the other authors.
D. Mougiakakos: Conceptualization, resources, investigation, writing–original draft, project administration. C. Bach: Investigation, methodology, writing–original draft, writing–review and editing. M. Böttcher: Investigation, methodology. F. Beier: Resources, investigation. L. Röhner: Investigation, methodology. A. Stoll: Data curation, software, formal analysis, investigation. M. Rehli: Formal analysis, methodology. C. Gebhard: Data curation, software, formal analysis. C. Lischer: Software, formal analysis, investigation. M. Eberhardt: Software, formal analysis, investigation. J. Vera: Resources, software, formal analysis. M. Büttner-Herold: Investigation, visualization, methodology, writing–review and editing. K. Bitterer: Investigation. H. Balzer: Investigation. M. Leffler: Investigation. S. Jitschin: Investigation. M. Hundemer: Resources, supervision, investigation. M.H.S. Awwad: Investigation, methodology. M. Busch: Investigation, methodology. S. Stenger: Supervision, investigation, methodology. S. Völkl: Resources, investigation, writing–review and editing. C. Schütz: Formal analysis, investigation, methodology. J. Krönke: Resources, supervision, investigation, methodology. A. Mackensen: Conceptualization, supervision. H. Bruns: Supervision, funding acquisition, validation, investigation, visualization, writing–original draft, project administration.
H. Bruns was supported by the Wilhelm-Sander Foundation, the German Research Foundation (DFG BR 4775/2-1), and the SFB TR 221 (project B12) from the German Research Foundation. This research was also funded by Celgene. D. Mougiakakos was supported by the Else Kröner-Fresenius Foundation and by the SFB TR 221 (project B06). C. Schütz was supported by German Cancer Foundation (grant no. 70112372). J. Krönke was supported through the Emmy Noether Program by the German Research Foundation.
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.