IREM-1 is an inhibitory cell surface receptor with an unknown function and is expressed on myeloid cell lineages, including cell lines derived from acute myeloid leukemia (AML) patients. We have generated a series of monoclonal antibodies (mAbs) against the extracellular domain of IREM-1 and further assessed its expression in normal and AML cells. IREM-1 was restricted to cells from myeloid origin and extensive expression analysis in primary cells obtained from AML patients showed IREM-1 expression in leukemic blasts of 72% (39/54) of samples. We therefore searched for specific IREM-1 mAbs with activity in functional complement-dependent cytotoxicity (CDC) and antibody-dependent cellular cytotoxicity (ADCC). Lead mAbs against IREM-1 showed specific cytotoxic activity against a variety of AML-derived cell lines and freshly isolated blasts from AML patients. Internalization of mAbs upon IREM-1 binding was also shown. In vivo anticancer activity of lead mAbs was observed in an established HL-60 xenograft model with a tumor growth delay of up to 40% and in a model using primary human AML cells, where treatment with anti-IREM-1 mAb resulted in a significant reduction of engrafted human cells. These results demonstrate IREM-1 as a potential novel target for immunotherapy of AML.
Acute myeloid leukemia (AML) has a poor prognosis, primarily due to relapses with conventional chemotherapy regimens. High-dose cytarabine and anthracycline combination regimens induce remission in 70–80% of patients younger than 60 years of age, but with intensive postremission therapy, the overall 5-year leukemia-free survival rate is only 25–35%.1 Several alternative therapeutic strategies have made significant progress in the clinic in recent years, but novel treatment approaches are needed.1, 12 Although FLT3 tyrosine kinase inhibitors are not effective as single agents, they are currently being evaluated in combination with chemotherapy in patients positive for the FLT3 mutation.2, 3
Antibody-based therapies provide a promising targeted approach to eliminate leukemic cells. Gemtuzumab ozogamicin (Mylotarg, Wyeth, Madison, NJ, USA) is an mAb, which binds to CD33 on the surface of myeloid leukemia blasts and is conjugated with a cytotoxic antitumor antibiotic, chalicheamicin. It has been approved for use in patients older than 60 years with relapsed AML and has an overall response rate of 30% in clinical studies.4, 5
Human IREM-1 (CD300LF) is a type I transmembrane protein in the CD300 family. The mature protein consists of an extracellular domain containing an immunoglobulin (V-type) domain, a transmembrane region and an intracellular domain containing three tyrosine signaling motifs (two immunoreceptor tyrosine-based inhibition motifs (ITIM) and one immunoreceptor tyrosine-based switch motif).6, 7 The crystal structure of the IREM-1 ectodomain has been described by Marquez et al.8
IREM-1 was identified as a novel receptor from dendritic cells,7 and as an inhibitory receptor expressed in myeloid cells in a search for SHP-1-interacting molecules in a three-hybrid screen.6 The interaction with SHP-1 in myeloid cells is dependent on phosphorylation. Mutagenesis studies of tyrosine residues identified Tyr 205, which resides in a classic ITIM, to be the main docking site for SHP-1. Interestingly, IREM-1 has also been reported to interact with the p85 subunit of PI3K, revealing a potential dual role for this receptor as an inhibitory and activating molecule on myeloid cells.9
In this study, we report the generation of specific monoclonal antibodies (mAbs) against the extracellular domain of IREM-1, provide a detailed expression analysis of normal and AML samples, and show the therapeutic potential of these mAbs for myeloid leukemia.
Materials and methods
Recombinant protein production
HEK-293 cells (ATCC) were cultured in Dulbecco's modified Eagle's medium+10% fetal bovine serum +Pen/Strep+L-Glu and were transfected with a plasmid construct containing the sequence of IREM-1 extracellular domain fused to a C-terminal tag of V5His (IREM-1-ECD-V5His) using FuGene 6 (Roche, Indianapolis, IN, USA). G418-resistant cells were selected for 2 weeks with selection medium (as above with 1 mg/ml G418). Clonal cell lines expressing IREM-1-ECD-V5His recombinant protein were screened with both HisSorb-based ELISA (Qiagen, Valencia, CA, USA) and western blot analysis detected with an anti-V5mAb (Invitrogen, Carlsbad, CA, USA). Stable clones with highest protein expression level were selected, expanded, and suspension-adapted.
Purification of IREM-1-V5His protein from HEK 293 cells
Culture supernatant from stably transfected HEK 293 cells was adjusted to 1 mM EDTA and 0.4 mM Pefabloc (Roche) protease inhibitors and 0.22 μm filtered. The supernatant was concentrated and diafiltered with 20 mM sodium phosphate, 0.5 M NaCl, pH 7 buffer using a Tangential flow filtration system with 10 kDa cut-off membrane (Pall Filtron, East Hills, NY, USA). The diafiltered medium was loaded on a 5 ml Ni-chelating affinity column (GE healthcare, Piscataway, NJ, USA), which was then washed with buffer containing 20 mM imidazole and eluted with a gradient of 20 mM to 220 mM imidazole. Fractions containing IREM-1 were pooled and buffer exchanged to phosphate-buffered saline (PBS).
Hybridomas were generated using the standard protocols.22 In brief, 6-week-old Balb/c mice were immunized with 100 μg of purified recombinant protein in Freund's adjuvant. After 3 biweekly boosts in incomplete Freund's, titers were assessed and a fusion of spleen cells was performed 3 days after a last boost in saline. Sp2/0 cells were used as fusion partner. Hybridomas were selected and supernatants from the resulting clones were screened by ELISA and fluorescence-activated cell sorting (FACS). mAbs were purified using standard protein A columns (GE Healthcare).
ELISA and FACS screen of hybridoma supernatants for binding to IREM-1
Enzyme-linked immunosorbent assay for binding to IREM-1 was performed using the standard techniques.23 The secondary goat anti-mouse Ig-HRP antibody was from Bio-Rad (Hercules, CA, USA) (no. 170-6516) and TMB substrate from KPL (Gaithersburg, VA, USA) (no. 50–76-03). Plates were read on a Spectramax plate reader (Molecular Devices, Sunnyvale, CA, USA).
Flow cytometry was performed using the standard protocols.24 Secondary antibody was from BD Pharmingen (San Diego, CA, USA) (goat anti-mouse phycoerythrin conjugated, no. 550589). Cells were analyzed using an Automated Microsampler (Cytek, Fremont, CA, USA) attached to a FACScalibur (Becton Dickinson, San Jose, CA, USA).
Real-time binding analysis
Surface plasmon resonance was carried out on a Biacore (Piscataway, NJ, USA) system. mAbs were diluted to 2 μg/ml and then captured on the biosensor surface using an anti-mouse mAb. Antigen was diluted to a starting concentration of 46 nM and tested for binding to the mAb samples using a threefold dilution series. Each of five concentrations was tested twice except the highest concentration that was tested five times in total, two times with a short dissociation of 300 s and then three times with a dissociation of 60 min. The data sets from the long dissociation experiments were globally fit with the shorter association experiments to determine binding constants for the interactions. The analysis was carried out in Hank's buffered saline, pH 7.4 buffer at 25 °C.25
Expression of IREM-1 by quantitative flow cytometry
QIFIKIT (K0078 from Dako, Carpinteria, CA, USA) was used for the quantitative determination of receptor number per cell, according to the manufacturer's instructions. Before use, cells were blocked for 10 min in 100 μl of 10% heat-inactivated human serum.
Cell lines and western blotting
All cells were purchased from ATCC (Manassas, VA, USA) and maintained in their recommended medium. Lysates were prepared and 25 μg of total lysate was analyzed by western blotting using the standard protocols under non-reduced conditions.26 Secondary antibody (goat anti-mouse Ig-HRP, BioRad, no;170–6516) was used and membranes were developed using Pierce's (Rockford, IL, USA) ECL western blot substrate (no. 32209).
Flow cytometry staining of normal human peripheral blood, BM from AML, MDS and all patient samples
Clinical samples were obtained from patients diagnosed at the Cleveland Clinic and the Stanford Cancer Center after informed consent. For the expression of IREM-1 on patients’ samples or normal BM aspirate, four colors staining with IREM-1 Alexa488/CD34PE-CD45PerCP/CD38APC was performed. For normal blood specimen, Immunophenotyping was performed with IREM-1 Alexa488/CD14-16PE/CD45PerCP/CD33APC panel. IgG-Alexa488 was used as control for all of the above tests. Expression of CD33 was analyzed with CD33FITC/CD34PE/CD45PerCP/CD38APC panel and IgG-FITC was used as control. Detailed staining process was described earlier.27 Blasts were gated based on low side scatter versus CD45 dim expression. A sample was considered positive for IREM-1 or CD33 if the ratio of the geometric mean fluorescence intensity of stained sample and that of isotype IgG control (medium fluorescence intensity (MFI)) >2 and >20% of the cells expressed the antigen compared with the control sample. All antibody used for flow cytometry were from BD Biosciences (San Jose, CA, USA) except for IREM-1 Alexa488 and IgG Alexa488 which were self-prepared using Invitrogen's Alexa Fluor 488 mAb labeling kit, according to the instructions.
Fresh peripheral blood or BM aspirate from AML patients was used for blast preparation by ‘Ficoll-Paque gradient separation.’ For most AML and all acute lymphoblastic leukemia cases, enriched blasts accounted for more than 85% of total population based on flow cytometry assay. Cell lines were used at log phase of growth. Cells (105) were suspended in 50 μl complete RPMI media and plated in a 96-well plate, 50 μl of 2 × antibody/IgG isotype, made up in same medium was added to each well and the plates left at room temperature for 15 min. A volume of 2.5–15 μl (for CA46 and HL60, respectively) of freshly prepared baby rabbit complement (CL3441; Cedarlane labs, Burlington, NC, USA) was added to respective wells followed by incubation at 37 °C for 1 h. After equilibrating plates to room temperature, cell viability was analyzed using ‘Cell Titer Glo’ (G7571; Promega, Madison, WI, USA) and luminescence was measured using ‘Victor 1420 Multilabel Counter’ (Perkin Elmer Life Science, Waltham, MA, USA). Triple test was conducted for each group and data were normalized to complement+isotype.
Antibody-dependent cell cytotoxicity
Human natural killer cells were isolated from buffy coat (purchased from Stanford blood center, Palo Alto, CA, USA) by negative selection using Rosette Sep natural killer cell enrichment cocktail from Stem Cell Technologies (Vancouver, BC, Canada), according to the manufacturer's instructions.
Specific lysis of target cells was determined by using a standard 4 h 51Cr release assay in a 96-well plate format as described earlier.28 Target/effector ratio of 1:40 was typically used. No preincubation step of effector cells and antibody was performed. Percent lysis was calculated using the following standard equation:
((TEST−BGD)/(Max−BGD)) × 100, where TEST is sample release of 51Cr, BGD is spontaneous release and Max is Triton-X-mediated release. % Specific lysis has effector control subtracted
Internalization by immunofluorescence microscopy and flow cytometry
For immunofluorescence microscopy experiments, IREM-1-expressing HEK 293 cells were seeded onto 12 mm glass coverslips coated with poly-L-lysine (Sigma-Aldrich, St Louis, MO, USA). The next day, the cells were preincubated in PBS-10% human serum (HS) for 15 min before incubation for 1 h at 4 °C with 10 μg/ml of Alexa-488-conjugated D12 or Alexa-488-chimeric control isotype antibody in PBS-2.5% HS. After three washes with ice-cold FACS buffer (PBS-1% bovine serum albumin), cells were incubated at 37 °C (in 5% CO2 and air) in PBS-5% fetal bovine serum for up to 2 h to allow internalization. Cells were then washed with ice-cold PBS, fixed with 4% paraformaldehyde and mounted onto slides with Prolong Gold antifade reagent with DAPI (4,6-diamidino-2-phenylindole) (Invitrogen). The stained specimens were examined using a Zeiss Axiovert 200 microscope with × 40 oil objective (Zeiss, Thornwood, NY, USA).
For flow cytometry experiments, HL-60 cells were preincubated with ice-cold PBS-10% HS, followed by incubation for 40 min at 4 °C with saturating amounts of Alexa-488-conjugated D12, PE-labeled anti-CD44 antibody (negative control, clone 515, BD Biosciences) or their corresponding isotype control antibodies. After three washes in ice-cold FACS buffer, cells were resuspended in 100 μl of RPMI-10% fetal bovine serum-PS and incubated either at 4 °C (no internalization) or at 37 °C for different time points. After washing, cell-surface-bound antibody complexes were stripped from the cells using 50 μl of Qiagen protease (QP, Qiagen no. 19 155) at 5 AU/ml in cold PBS for 30 min at 4 °C. Cells were washed three times with ice-cold FACS buffer with 0.02% NaN3, fixed in PBS-1% formaldehyde and analyzed by flow cytometry using the FACS Calibur system (Becton Dickinson, San Jose, CA, USA). The background MFI determined with cells incubated with the labeled isotype control antibodies was subtracted for each time point. The specific MFI at T0 in the absence of protease treatment was normalized to 100% and the relative specific MFI for each time point was calculated using the formula: 100 × (specific MFI Ab +QP Tx)/(specific MFI Ab−QP Tx).
Toxin conjugated IREM-1-mediated toxicity
Saporin-conjugated chimeric D12 IREM-1 and chimeric isotype were generated by Advanced Targeting System (San Diego, CA, USA). Cells were blocked in 10% heat-inactivated human serum for 10 min at room temperature and then plated at 5000 cells per well in 50 μl complete media. 50 μl of 2 × conjugated antibody/isotype made up in complete media was added to each well, in triplicate, and plates left for 72 h at 37 °C. Cell viability was measured by relative ATP levels using CellTiter Glo (Promega no. G7571). Wells were read using a Victor 1420 Multilabel Counter (Perkin Elmer Life Science/Veritas software) and data represented as a percent of control. Data analysis and EC50 values were generated using PRISM (GraphPad Software, La Jolla, CA, USA).
The xenograft HL-60 model in BALB/c scid mice was carried out as described.29, 30 The animal study protocols were reviewed and approved by the Institutional Animal Care and Use Committee according to the governmental guidelines for animal welfare. All of the mice were acclimated before use. Xenografts were allowed to establish to an average size of 50–100 mm3, after which mice were randomized into various conditional groups. mAbs were given to each mouse at designated dose through intraperitoneal injection at a frequency of twice a week. Each mouse was measured for tumor size using a caliper on alternate days. Animal body weight and any sign of morbidity were also closed monitored. The mAb treatment lasted for 2 weeks at which point mice were harvested, tumor xenografts were extirpated, weighed, and correlated with the tumor size measurement. Statistical significance for median time to tumor end point values for treatment group comparisons was determined by the log-rank test.
Nonobese diabetic (NOD)/severe combined immunodeficient (SCID) mouse assays
NOD/SCID (NOD.CB17-prdkdc. scid/J) mice (Jackson Laboratories, Bar Harbor, ME, USA) were sub-lethally irradiated with 250 rad using a RadSource X-ray irradiator (RadSource Imaging Technologies, Parkville, MO, USA) 1 day before transplantation. Primary AML cells were injected through tail vein (5 million cells) in a final volume of 0.2 ml of PBS with 0.5% albumin. After 2 weeks, animals were injected intraperitoneally with 50 μg of either huIgG1 isotype control or IREM-1 D12 chimeric antibody twice a week for 4 consecutive weeks. Animals were killed and BM was analyzed for the presence of human cells by flow cytometry. Percent of human cells was determined using anti-human CD45 antibody, cells were also stained with anti-human CD33 and IREM-1-specific antibodies.
Generation of mAbs against IREM-1
Through comprehensive expression analysis we identified several genes encoding for cell surface receptors that are differentially expressed in AML disease samples. This includes a member of the CD300 gene family, IREM-1, which showed a myeloid-restricted expression pattern. To further explore the expression of IREM-1, we generated recombinant protein spanning the extra-cellular domain of IREM-1 and used it to immunize mice and generate mAbs. We selected ∼1000 hybridoma clones recognizing IREM-1 protein by enzyme-linked immunosorbent assay (ELISA) and 200 of these recognize cell surface expression of IREM-1 on the myeloid cell line HL-60 as determined by flow cytometry analysis. We also selected mAbs with cytotoxic activity and identified two lead mAbs (871.1 and D12), which were further characterized and subsequently used in detailed expression analysis and efficacy studies.
Affinity of lead mAbs
Besides the cell surface binding to IREM-1 and cytototxic activity, our lead mAbs also had a high avidity to IREM-1. Surface plasmon resonance was used to determine the affinity of the lead mAbs 871.1 and D12. Anti-IREM-1 mAbs were captured using anti-mouse Fc Ab and recombinant extra-cellular domain of IREM-1 was allowed to bind to the mAb at various concentrations. The kinetic rate constants (ka and kd) of this reaction were determined and used to calculate the KD values. Table 1 shows the KD values of lead mAbs 871.1 and D12.
Expression of IREM-1 in myeloid cells
It was shown earlier that IREM-1 is expressed on the cell surface of myeloid lineages.6 We used our specific IREM-1 mAbs to further analyze the expression of IREM-1 in peripheral blood. A fluorescent-labeled anti-IREM mAb was incubated with whole blood from normal donors and a series of markers found on subsets of peripheral blood cells. As shown in Figure 1, IREM-1 expression was detected on cells from myeloid origin including monocytes (highest expression), dendritic cells and granulocytes (lower levels) as determined by CD14/CD16 and CD33 co-staining (Figure 1a). In contrast, IREM-1 was not detected on lymphocytes and platelets. Analysis of samples from morphological normal bone marrow (BM) confirmed these results, with expression observed on monocytes and granulocytes, but not on lymphocytes or erythroid precursors (Figure 1b). Data from peripheral blood analysis from 10 normal donors are summarized in Figure 1c.
To determine IREM-1 expression in myeloid leukemia, we analyzed blast cells derived from the peripheral blood and BM from 54 AML patients by flow cytometry (example in Figure 1d). AML samples were obtained from patients resembling a variety of classifications as summarized in Table 2. IREM-1 expression was detected in the majority of AML cases (39 of 54, 72.2%) with a range of positive blasts from 20 to 99% (mean 72.1%), with a high incidence in more mature and monocytic AML. IREM-1 was not detected in blasts from acute lymphoblastic leukemia patients (n=5, Figure 1e). Sufficient events of CD34+/CD38− leukemia stem cells were collected from 24 AML cases and stem cells from 13 of these cases were IREM-1 positive (54.2%, example in Figure 1f). CD34+ progenitor cells from negative lymphoma staging BM samples (n=8) showed that only one case expressed IREM-1 (21% positive cells). We also tested CD34+/CD38− BM-derived stem cells from 11 morphological normal samples and all were negative for IREM-1. In addition, we surveyed IREM-1 expression in BM aspirate from 13 cases myelodysplastic syndrome (MDS) patients, and blasts of 7 cases were positive (54%) with a range of positive blasts from 20–78% (mean 43.6%). Among the 13 MDS cases, there is one refractory anemia which was negative for IREM-1, 2 of 5 RAEB-1 cases were positive, 2 of 3 RAEB-2 cases were positive; one MDS/MPD, one therapy-related MDS and one unclassifiable case were all positive. One MDS 5q(−) syndrome was negative. We were able to analyze CD34+/CD38− cells in four patients with MDS of which two showed positivity for IREM-1.
As gemtuzumab ozogamicin (calicheamicin-conjugated humanized anti-CD33 antibody) is approved for treatment of CD33-positive AML in relapsed patients >60 years of age, we next compared the expression pattern of IREM-1 with CD33. Monocytes expressed the highest levels of CD33, while lower expression was observed in granulocytes, and no expression in lymphocytes regardless whether the specimen was obtained from healthy control or AML patients (Figure 2a). Nevertheless, the overall expression pattern of CD33 is similar to that of IREM-1 (Figure 2b). Similar to IREM-1, CD33 was detected in AML blasts of 48/54 (88.9%) of cases tested in this study. Both IREM-1 and CD33 also have similar expression patterns in AML blasts and CD34+/CD38− cells (n=9) (Figure 2c). We analyzed CD34+/CD38− cells from five cases, and CD33 was found in two of five cases, whereas IREM-1 was positive in three of five cases.
We also surveyed the expression of IREM-1 in a wide variety of AML cell lines using flow cytometry and western blotting analysis (Figure 3). Various expression levels of IREM-1 were detected in myeloid cells by quantitative flow cytometry. Receptor number per cell ranged between 0 and 27 000 IREM-1 molecules present on AML cell lines (Figure 3a), with highest expression levels found on HL-60 and U937 cells. These data were in accordance with immunoblotting results obtained using cell lysates from the same cell lines (Figure 3b). IREM-1 was not detected in cells from lymphoid origin (results not shown).
These results confirmed the restricted tissue expression of IREM-1 exclusively in cells from myeloid origin and highlight the potential of IREM-1 as a target for myeloid leukemia therapeutics.
Specific cytotoxic activity of mAbs against IREM-1
To assess the therapeutic potential of IREM-1 mAbs, we characterized their cytotoxic activity against AML-derived cell lines. Lead mAbs were tested in cellular cytotoxicity assays, including CDC activity by incubating HL-60 cells with varying concentrations of mAb 871.1 and fixed amounts of rabbit complement. Dose-dependent CDC activity was detected with mAb 871.1 (with lysis up to 100% of target cells), while no CDC activity was seen with an isotype control mAb (Figure 4a). Moreover, mAb 871.1 was effective in CDC activity against a variety of AML cell lines, including OCI-AML5, AML193 and Kasumi-1 (results not shown). In contrast, no cytotoxic activity was observed against a Burkitt's lymphoma cell line lacking IREM-1, which was used as a control cell line (CA46, Figure 4a), showing the specific effect of IREM-1-mediated cytotoxic activity.
To determine whether the cytotoxic activity detected with anti-IREM-1 mAbs was mediated through a direct interaction with IREM-1 on the cell membrane, CDC assays were performed against a surrogate cell line. HEK293 cells were transfected with IREM-1 plasmid and stably expressing clones were selected. CDC activity with IREM-1 mAb was detected against the HEK293-IREM-1 cells but not against the HEK293 parental line (Figure 4b), indicating that the CDC activity is mediated specifically through IREM-1.
To further explore the therapeutic potential of IREM-1 mAbs, we performed CDC assays on fresh AML blasts obtained from AML patients. In these ex vivo experiments, cytotoxicity was observed in a dose-dependent manner in 10 of 11 (91%) samples tested, which expressed IREM-1 (example in Figure 4c), with lysed cells ranging between 20 and 60%. In contrast, AML blasts not expressing IREM-1 and acute lymphoblastic leukemia blasts were insensitive to CDC mediated through IREM-1 mAbs (Figure 4c).
Antibody-dependent cellular cytotoxicity (ADCC) is an important mechanism of cancer cell cytotoxicity in vivo. For testing lead IREM-1 mAbs in ADCC and subsequent in vivo evaluation, chimeric mAb constructs were generated by fusing the D12 and 871.1 variable sequences to human IgG1 constant region sequences. The recombinant chimeric mAbs were expressed and purified from CHO cells. To determine the ability of anti-IREM-1 mAbs to elicit ADCC activity, HL-60 cells were labeled with 51Cr and subjected to cytotoxic activity of freshly isolated human natural killer cells in the presence of chimeric mAb. A dose-dependent ADCC-mediated lysis was observed using chimeric mAb D12, whereas an isotype control mAb had no ADCC activity (Figure 5a). Up to 40% of cells were lysed with an EC50 of 1.2 μg/ml. CA46, a B-cell line not expressing IREM-1 was insensitive to anti-IREM-1 mAbs (Figure 5a).
To show the specificity of IREM-1 mAb, HEK293 target cells stably expressing IREM-1 were also labeled and used in ADCC assays and appeared to be sensitive to anti-IREM-1 mAbs at low concentrations. In contrast, an isotype control mAb had no cytotoxic effect on these cells (Figure 5b). Potent activity (EC50 of 66 ng/ml) was observed against up to 35% of the cells.
In addition to the effector mechanisms CDC and ADCC, we also tested whether our mAbs were capable of inducing direct cell death through apoptosis upon cell binding. Similar to what has been described for anti-CD33 mAbs,10 anti-IREM-1 mAbs were capable of inducing cell death through apoptosis in primary AML cells (Supplementary Figure 1), but not in AML-derived cell lines (results not shown). After 40 h of incubation with the antibodies we detected Annexin V binding in up to 25% of cells.
These results show that IREM-1 mAbs are potent and selective cytotoxic agents in vitro and ex vivo against IREM-1-expressing cells and exhibit potential as targeted anticancer agents through a combination of effector mechanisms.
Internalization of IREM-1
Current mAb therapy for AML includes a toxin-conjugated anti-CD33 mAb, gemtuzumab ozogamicin. To investigate the potential for an antibody-drug conjugate we first determined whether IREM-1 mAbs trigger receptor internalization. Therefore, 293 cells stably transfected with IREM-1 were stained with fluorescent-labeled chimeric mAbs 871.1 and D12, followed by incubation at 37 °C for 30 min or 2 h (Figure 6a, left panels). An increasingly intracellular staining pattern was observed in cells stained with IREM-1 mAb, indicating rapid internalization of the receptor–mAb complex, with most of the antibody inside the cells after 2 h (Figure 6a, middle and right panels). Similar results were obtained using mAb 871.1 (not shown).
A flow cytometry-based assay was also developed to measure internalization rates in AML-derived cell lines. After incubation with fluorescent-labeled mAbs, cells were subjected to protease treatment and the signal associated with a protease-resistant compartment was measured over time using flow cytometry. Approximately 45% of anti-IREM-1 chimeric mAb D12 became internalized within 3 h post-treatment, whereas an anti-CD44 control mAb remained protease sensitive (Figure 6b).
We next investigated whether drug conjugation to anti-IREM-1 antibodies could improve cytotoxic activity. D12 and an isotype control mAb were conjugated to the ribosome-inactivating protein saporin and were used in direct cytotoxic assays against HL-60 and CA46 cells. A potent (EC50=34 ng/ml), dose-dependent cytotoxic activity was observed against HL-60 cells, but not against control CA46 cells (Figure 6c).
These results show the internalization of IREM-1 upon antibody binding and the potential of designing an antibody-drug conjugate against this target.
Xenograft antitumor activity
The in vivo anticancer effects of chimeric anti-IREM-1 mAbs were evaluated in immunodeficient mice bearing HL-60 xenografts. In the first study, HL-60 cells were inoculated into the left flank of severe combined immunodeficient (SCID) mice and treated with mAb D12 after tumors had established. The tumors were measured bi-weekly for 4 weeks or until the tumors reached a size of 2000 mm3. A dose-dependent reduction in tumor burden was observed in mice treated with D12 (with doses ranging from 0.05 mg to 0.5 mg/mouse, or about 2 mg/kg to 20 mg/kg) compared with mice treated with saline or isotype control mAb (P0.001, Figure 7a and b).
A follow-up experiment included dosing at lower levels with both mAbs 871 and D12 (0.5 and 5 mg/kg), tumor growth delay was calculated relative to the saline control group using the end point set at 2000 mm3 and was found to be up to 40% (P<0.05) in the D12-treated groups, whereas mAb 871.1 reduced tumor growth by about 25% at 5 mg/kg dosing (Table 3).
Moreover, an in vivo assay was also performed using primary cells injected intravenously into sub-lethally irradiated nonobese diabetic (NOD)/SCID mice. Two weeks after the injection, once the disease was established in the BM, treatment was started using 50 μg of IREM-1 mAb in a biweekly schedule for a total of 4 weeks. Tumor burden in the BM was determined 6 weeks after the injection of human cells in the mice. A significant decrease was observed in the percent of human cells (P=0.0024, Figure 7c). Specifically, when the percent of CD33 and IREM expression was assayed in the human cells recovered from the mice, a significant decrease in CD33+/IREM+ cells was observed (P<0.0001, Figure 7d).
These results validate anti-IREM-1 mAbs as inhibitors of myeloid cancer cell growth in vivo.
Selective expression of a cell surface protein on target cells provides an antibody-based therapeutic opportunity for treating leukemia. Using a genomic approach we identified IREM-1 as a potential receptor expressed on normal myeloid and leukemic blast cells. We generated a series of mAbs against IREM-1 and selected two lead mAbs with high affinity and cytotoxic activity. Flow cytometry and western blot analysis of IREM-1 in normal and diseased samples revealed an expression pattern restricted to cells from myeloid origin. We generated murine and chimeric mAbs against IREM-1, showed their specificity and evaluated their anticancer activity against AML cell lines and primary blasts in various in vitro, ex/in vivo models.
To our knowledge, this is the first report to show that IREM-1 is expressed in the majority of malignant AML blasts (39 of 54, 72.2%), with a high incidence in more mature acute myelomonocytic/monocytic leukemia (21 of 22, 95.5%). Our preliminary studies also indicated that IREM-1 is expressed in blasts of >60% of MDS cases, a disease with few effective treatment options. The CD34+/CD38− phenotype identifies early hematopoietic progenitor stem cells and may also identify leukemia stem cells.11 Targeting the latter may be beneficial, while the former may impact BM recovery after therapy. We therefore analyzed the expression pattern of IREM-1 in CD34+/CD38− leukemia stem cells and found expression in 13 of 24 (54.2%) AML cases. In contrast, we failed to identify an IREM-1-positive CD34+/CD38− cell population in a parallel test of non-neoplastic BM (n=11). Although these results are encouraging as a first step toward analyzing stem cell reactivity, a thorough analysis of IREM-1 in normal versus malignant precursors in a large number of samples is warranted. Our functional studies of IREM-1 mouse monoclonal and chimeric antibodies in vitro, ex vivo and in vivo models indicate its ability to induce AML cell killing and delay tumor growth. Taken together, IREM-1 has potential not only as a novel therapeutic target for antibody-mediated immunotherapy but also to serve as possible diagnostic marker in detection of minimal residual disease and relapse prediction.
Our data indicate that IREM-1 has a similar distribution and expression pattern as CD33, either in normal or AML blood. CD33 was positive in 48 of 54 (88.9%) AML cases tested in this study, which is comparable with other published studies.12, 13, 14 Among the six negative CD33 cases, IREM-1 was positive for five cases (83.3%), suggesting it might be a useful target in CD33-negative AML cases as well. Although CD33 density might be a limiting factor prohibiting significant induction of ADCC and CDC activity,15 humanized anti-CD33 mAbs have shown biological activity in in vitro CDC and ADCC assays against CD33-expressing cell lines.12 The novel IREM-1 antibodies described here appear capable of mediating dose-dependent ADCC and also show CDC activity in 10 of 11 freshly isolated IREM-1+ AML blasts. It should be noted that the 11 cases used for CDC are either acute myelomonocytic leukemia or acute monocytic leukemia, all of which have higher IREM-1 expression, as described above.
The anti-CD33 mAb lintuzumab has shown only modest single-agent activity against AML in the clinic16, 17 and native anti-IREM-1 mAbs may display similar potency in patients with active AML. Although IREM-1-mediated CDC and ADCC activity were described here, it is possible that IREM-1-mediated signaling pathways can contribute to IREM-1 antibody-mediated killing activity in malignant blasts. Several tyrosine-based signaling motifs reside in the intracellular domain of IREM-1. As IREM-1 has been identified as an inhibitory receptor in myeloid cells by interaction through its ITIM with SHP-1 phosphatase,6 an interesting avenue to explore would be to screen mAbs for their potential to alter the phosphorylation status of the receptor, thereby inducing inhibitory signals to malignant cells. A recent study showed that orexin-induced apoptosis in cancer cells is driven by ITIM present in the orexin receptor.18 Another group reported that cross-linking of CD300LF with its mAb induced cell death in murine peritoneal macrophages as well as in transfected murine cell lines. Such a CD300LF-mediated cell death was dependent on the cytoplasmic region but did not require an ITIM or immunoreceptor tyrosine-based switch motif.19 These studies suggest that immunotyrosine signaling motifs-bearing receptors may have a different impact on the signal pathway(s) based on the experimental model. Our data indicate that anti-IREM-1 mAbs are able to directly induce apoptosis in primary AML cells (Supplementary Figure 1).
Internalization of IREM-1 receptors was observed within 2 h of exposure of cells to IREM-1 mAbs. Therefore, in addition to CDC/ADCC activity, another potential strategy to explore is the development of this antibody as a toxin or drug conjugate given its ability to be internalized. ITIMs in CD33 have been identified as required for internalization20 and YXXM motifs have been shown to be involved in lysosomal targeting of receptors,21 and therefore, these sequences could be responsible for the internalization of IREM-1 that we have observed upon mAb binding.
In the in vivo xenograft model described here we treated rapidly growing, established HL-60 tumors with anti-IREM-1 chimeric mAbs. We observed a significant tumor growth delay up to 40% compared with control treatments. In vivo xenograft assays performed with primary AML cells showed a significant decrease in tumor burden in the BM (P=0.0024) using 50 μg of IREM-1 mAb, suggesting that the efficacy observed in the subcutaneous model is not specific to the cell line used. Interestingly, when the phenotype of the human cells was determined, a significant decrease (P<0.0001) was observed for CD33+/IREM+ cells, suggesting specificity to this population of cells for IREM-1 mAb. Although treatment did not eradicate the tumors completely in these mice, the results are encouraging and open up the potential for combination therapy with chemotherapeutic agents.
In summary, we have developed mAbs against the myeloid cell surface receptor IREM-1 and have shown their therapeutic potential in preclinical AML models.
Conflict of interest
WK, SS, CP, JZ, SY, SL, XZ, DG, NT, CZ and AA are employed by and own stock in Nuvelo Inc. XZ, JG and EDH are consultants for Nuvelo. We declare that the work is not considered for publication elsewhere.
WK, XZ, SS, CP, JZ, MLG, SS, SY, SL, XZ, NT, CZ and DG performed experiments; WK, XZ, MLG, CTJ, JG, EDH and AA designed the studies, wrote and critically revised the manuscript.
About this article
Annals of Hematology (2011)