Abstract
B7-H3 acts as a checkpoint inhibitor, influences the tumor microenvironment, and is necessary for osteoclast (OC) development. Given its dual roles in promoting cancer survival and bone breakdown, we looked at its expression pattern in >200 multiple myeloma (MM) patients and association with cytogenetic abnormalities and outcomes. We found that most newly diagnosed patients had B7-H3+ plasma cells or stroma regardless of cytogenetic risk and expression on plasma cells increased with subsequent relapses. In marrow aspirates and core biopsies we identified B7-H3+ myeloid derived suppressor cells (MDSC) and (OC). We found that high B7-H3 expression was associated with pathologic fractures, bone lesions, and worse progression free survival (PFS). To confirm the utility of B7-H3 as a target in MM, we tested a tri-specific killer engager (TriKE) that binds CD16 on natural killer (NK) cells and B7-H3 while also delivering a recombinant IL-15 molecule to promote NK cell activity. The B7-H3 TriKE restored the ability of patient-derived NK cells to recognize and kill MM and enhanced NK cell-mediated killing of tumor, MDSC, OCs, and fibroblasts. These data suggest that targeting B7-H3 may positively impact outcomes and prevent the development of new bone lesions. In addition to its ability to directly target MM and immunosuppressive cells in the tumor microenvironment, the B7-H3 TriKE improved cytokine secretion and polyfunctionality of NK cells in a single cell assay. These findings advance our understanding of B7-H3’s role in MM pathogenesis and validate the use of the B7-H3 TriKE to improve therapeutic outcomes in MM.
Keywords: B7-H3, multiple myeloma, natural killer engager, TriKE, pathologic fracture, bone lesion, osteoclast, MDSC
Introduction
In multiple myeloma (MM), natural killer (NK) cells become dysfunctional [1] and resistance is enhanced by suppressive cells in the tumor microenvironment [2, 3]. Antibodies against CD38 and SLAMF7 are approved to treat MM and work, at least in part, by targeting NK cells to kill MM through engagement of CD16 receptor, causing release of cytotoxic granules [4]. One limitation of currently available MM therapies is NK cell depletion through fratricide since NK cells also express CD38 and SLAMF7 [5]. In contrast, B7-H3 (CD276), is a member of the inhibitory B7 family that is highly expressed on numerous cancers, including MM, but has low to absent expression on healthy tissues [6, 7]. Our previous work in immunocompromised mice treated with B7-H3 TriKE showed no off target toxicity [8].
B7-H3 inhibits NK-mediated killing [9] but the receptor it binds has not yet been identified [10]. In addition to its role in immune escape, B7-H3 modulates tumor growth and invasion [11, 12] despite having a short intracellular tail with no known signaling motif [13]. In MM, higher B7-H3 expression was associated with activation of the JAK2/STAT3 pathway and worse overall survival [7]. B7-H3 expression was independent of the targetable antigen BCMA and a chimeric antigen receptor (CAR)-T cell showed efficacy in preclinical models of MM [14]. B7-H3 is also expressed by tumor stroma including endothelial cells [15] and fibroblasts [16] and is associated with increased secretion of vascular endothelial growth factor (VEGF) and resistance to chemotherapy [17].
Myeloid derived suppressor cells (MDSC) can also express B7-H3 [18] and promote MM survival [3]. MDSC from cancer patients can differentiate into OCs which break down bone to form lytic lesions [19] and have directly suppressive effects on NK cells through TGF-β signaling [20]. B7-H3 expression is crucial for OC differentiation [21]. Therefore, modulation of the tumor microenvironment by targeting B7-H3-expressing immunosuppressive and stromal cells might improve outcomes.
In this manuscript we describe the expression characteristics of >200 patients of B7-H3. We then tested a tri-specific killer engager (TriKE) targeting B7-H3 with donor and patient-derived NK cells against tumor, immunosuppressive, and stromal cells. The TriKE possesses a variable heavy chain (VHH) camelid nanobodies targeting CD16 on NK cells and B7-H3 on MM cells. The nanobodies are connected by a linker containing a recombinant human interleukin-15 (IL-15). The B7-H3 TriKE promotes proliferation and enhanced NK cell efficacy against tumors, even in hypoxic environments [22].
Materials and Methods
Patient data.
Patients were included who had their diagnostic bone marrow biopsy done at University of Minnesota, had remaining tissue, and consented to future research. Of 141 potential patients, seven were excluded because their official pathology showed no clonal plasma cells on the biopsy specimen. Another 18 patients were excluded because they did not have full body imaging (PET, CT, or X-ray) within +/− 30 days of diagnosis. All imaging within +/− 30 days of diagnosis were reviewed by an oncologist as well as any notes from oncology or orthopedics to determine if pathologic fracture had occurred. Charts were reviewed to collect laboratory data, fluorescence in situ hybridization (FISH), and cytogenetics within +/− 30 days of diagnosis. Cytogenetic risk was determined by mSMART guidelines [23]. International Staging Score (ISS) was determined using beta-2 microglobulin and albumin as previously described [24]. Calcium was corrected for serum albumin [25].
A second cohort of MM patients were collected sequentially (at diagnosis, first relapse, and at triple class refractory relapse) at Karolinska Institute (Sweden) with aspirates undergoing Ficoll centrifugation to collect mononuclear cells.
Patient samples were collected with IRB approval at University of Minnesota and Karolinska Institutet, and this study was conducted in accordance with the Declaration of Helsinki.
IHC staining of bone marrow.
Unstained sections (4 μm) were de-paraffinized and rehydrated using standard methods. Antigen retrieval is outlined in supplemental materials and methods. Slides were then dehydrated and cover slipped. Scanning was performed with Zeiss axio scan Z1. B7-H3 staining was quantified by estimating the percentage of plasma cells staining on low power (4X objective) and by manual count of the B7-H3 positive plasma cells on high power (40× objective) by author BG and graded 0–3. Stromal staining was performed by examining endothelia and fibroblasts at low power (4x objective) by author BG and graded as 0–3 intensity.
Bone imaging and fractures:
Radiology reports +/− 30 days from diagnosis and physician notes were reviewed by authors AM and PK independently to determine the number of lytic lesions and if pathologic fracture existed. Any discrepancies were reviewed jointly. Lytic lesions at presentation were determined by whole body imaging (bone survey or positron emission tomography or low dose CT).
Blood and bone marrow donors.
Peripheral blood was obtained from the Memorial Blood Centers (Minneapolis, Minnesota, USA). All blood products were deidentified. Patient bone marrow and blood samples were collected fresh.
Cell lines.
MM1S, H929, U266, RPMI-8226, and Hs5 cell lines were obtained from American Type Culture Collection and maintained per their recommendations with quarterly testing for Mycoplasma.
Primary cell isolation and tissue culture.
Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque (GE Healthcare) density gradient centrifugation. NK cell enrichment was performed with EasySep™ Human NK cell enrichment kit (STEMCELL Technologies). CD33 were isolated with EasySep™ Human CD33 positive selection kit II (STEMCELL Technologies). CD14 were isolated with EasySep™ Human CD positive selection kit II (STEMCELL Technologies). MDSC were generated as outlined in supplemental methods.
To test the ability of healthy donor-derived MDSC to suppress NK cell proliferation, NK were co-cultured with 10ng/mL of IL-15 alone or with MDSC (2:1) for 7 days and analyzed by flow cytometry.
OC assays:
Marrow from MM patients was depleted of tumor using Human Plasma Cell Isolation kit (Miltenyi). The remaining cells were cultured for six days with M-CSF (20 ng/mL) and RANKL (40 ng/mL, Peprotech) to differentiate OC. OC were detached with 0.01% EDTA, and 1×105 cells were seeded onto bovine femur bone chips (ThermoFisher) or tissue culture plastic in 96-well plates and allowed to adhere for four hours. NK cells were added to the OC cultures at an effector-to-target ratio of 2:1. Co-cultures were maintained overnight with M-CSF and RANKL before TriKE antibody (3nM) was added and stained for TRAP activity. To assess resorptive activity, bone chips were reacted with NHS-AF647 (Molecular Probes) prior to OC plating, and OC cultures were maintained for 72 hours. OC resorptive activity was detected through the proteolytic release of AF647. Imaging was performed on the EVOS system and TRAP positive, multinucleated OCs or resorptive area (AF647 removal) was quantified using Fiji. Each experiment was performed with at least three MM patient cells.
CRISPR-Cas9 Knockout cell lines.
Chemically synthesized B7-H3 sgRNA (TGCTGCTACCCTCACCTGTG) was resuspended at 120 μM in IDTE buffer (pH 7.5), and stored at −80C. CRISPR ribonucleoprotein (RNP) complexes were assembled by mixing 1 μL sgRNA with 1.5 μL of 40 μM recombinant Cas9. After nucleofection, 80 μL of complete growth medium was added and cells recovered at 37C for 15 minutes. Complete process is outlined in supplemental materials and methods.
Flow cytometry.
Staining was performed with fluorochrome-conjugated Abs against the surface epitopes CD3 (OKT4; BioLegend), CD56 (HDCD56; Biolegend), CD11b (ICRF44; BD Pharmingen) S100A8/A9 (900028; R&D Systems), CD138 (MI15; BD Pharmingen), CD14 (63D3; BDP harmingen). Cells were surface stained with a dead cell stain (Invitrogen, ThermoFisher Scientific) in FACS buffer (PBS supplemented with 2% FBS and 2 mM EDTA) and fixed in 2% paraformaldehyde.
All antibodies were directly conjugated. CD138 (MI15, BD), CD38 (HIT2, BD), CD45 (HI30, BD), CD276 (MIH42, Biolegend), CD3 (SK7, BD), CD14 (63D3, BD), BCMA (REA315, Miltenyi), FcRH5 (REA391, Miltenyi). Flow cytometry data was acquired on a Beckman Coulter CytoFlex LX and collected using CytExpert v2.6. dDta was analyzed with FlowJo software, version 10 (Tree Star).
Live imaging killing assays.
Cell lines were stained with CellTrace Far Red (Invitrogen) and seeded into poly-L-ornithine coated plates with caspase 3/7 dye (Sartorius). Target cells were incubated for one hour before addition of NK cells and/or 3 nM B7-H3 TriKE or 3 nM null TriKE. Target cell survival was assessed using an IncuCyte Live Cell Analysis System (Essen BioScience). Live cell numbers were quantified with IncuCyte S3 software (Essen BioScience) and normalized to the number of live cells remaining in the target cell-only control group.
Single cell polyfunctionality assays.
NK cells were enriched from a blood donor and incubated with 10 ng/mL IL-15; or 10 ng/mL IL-15, 100 ng/mL IL-18 and 10 ng/mL IL-12; or 3nM B7-H3 TriKE, for 48 hours. After 48 hours, cells were stained with dye and loaded on single cell multiplex ELISA chips per the manufacturer’s instructions (Isoplexis Human Adaptive Immune Chips).
Statistical Analysis.
Numerical clinical characteristics were compared by Mann-Whitney U test. Categorical variables were compared by Fisher’s exact test. PFS was estimated by the Kaplan-Meier method and compared using the log-rank test. B7-H3 expression in sequential samples and bone erosion was compared by Friedman test followed by a Dunn’s post hoc test. In vitro killing assays were compared between groups by Wilcoxon matched pairs signed-rank test. All P values were two-sided. GraphPad Prism v7.0 was used for statistical analysis of flow cytometric parameters and killing assays. PFS was analyzed with R version 4.3.2.
Results
B7-H3 is associated with bone pathology and progression free survival.
A total of 116 newly diagnosed patients contributed marrow cores for immunohistochemical (IHC) staining. Plasma cells were positive for B7-H3 in 61 patients (53%) (Figure 1 A–B) and the stroma was positive for B7-H3 in 96 patients (83%) (Figure 1 C–E). Eighteen patients were negative for B7-H3 staining on both plasma cells and stroma (16%). We saw no difference in common genetic abnormalities associated with myeloma pathogenesis, cytogenetic risk, or monoclonal protein type between groups with positive or negative B7-H3 staining. Gain of chromosome 15 (which carries the gene for B7-H3) was not significantly associated with positive stromal or plasma cell staining (Table 1). In the cohort, there were 5 patients with pathologically confirmed AL-amyloidosis; all of whom had B7-H3+ plasma cells and all but one with B7-H3+ stromal staining (p=0.567). None of the patients with double negative staining (in stroma and plasma cells) had concomitant AL-amyloidosis.
Figure 1. B7-H3 is expressed on plasma cells and/or marrow stroma in the majority of newly diagnosed multiple myeloma patients.
A. Representative bone marrow sample with strong B7-H3 staining on plasma cells (grade 3) and stroma (grade 3). B. Bone marrow with negative B7-H3 staining on plasma cells (ISS-0) but moderate staining of stroma and vasculature (grade 2). C. Stromal staining in the area of myeloma involvement (grade 2) with weak staining on plasma cells (grade 1) but negative in areas uninvolved by plasma cells. D. Pleomorphic plasma cells with variable B7-H3 staining (grade 2) and strong stromal staining (grade 3). E. Strong stromal staining with membranous staining of three OCs (marked with yellow circles). F. Membranous staining of B7-H3 on an OC that is degrading a bone spicule. Sample was taken from a patient who presented with multiple vertebral fractures from MM.
Table 1.
Patient demographics by B7-H3 immunohistochemical staining on plasma cells or stroma.
| B7-H3 negative PC (n=55) | B7-H3 positive PC (n=61) | B7-H3 negative stroma (n=20) | B7-H3 positive stroma (n=96) | Double Negative (n=18) | |
|---|---|---|---|---|---|
| Age (median, IQR) | 62 (56,68) | 64 (55,72) | 62 (56,71) | 64 (56,71) | 63 (57,68) |
| Sex Female | 27 (49%) | 21 (34%) | 10 (50%) | 38 (40%) | 10 (56%) |
| ISS-1 | 21 (38%) | 19 (31%) | 10 (50%) | 30 (31%) | 10 (56%) |
| ISS-2 | 22 (40%) | 26 (43%) | 9 (45%) | 39 (41%) | 7 (39%) |
| ISS-3 | 9 (16%) | 13 (21%) | 0 (0%) | 22 (23%) | 0 (0%) |
| ISS-unknown | 3 (1%) | 3 (5%) | 1 (5%) | 5 (5%) | 1 (6%) |
| Pathologic Fx | 15 (27%) | 33 (54%) | 3 (15%) | 45 (47%) | 3 (17%) |
| PC % (median, IQR) | 50% (20,70) | 40% (20,65) | 30% (10,68) | 45% (20,70) | 30% (9,62) |
| High Cytogenetic Risk | 26 (47%) | 25 (41%) | 10 (50%) | 67 (70%) | 10 (56%) |
| 15 gain | 29 (53%) | 39 (64%) | 8 (40%) | 54 (56%) | 7 (39%) |
| t(11:14) | 14 (25%) | 12 (20%) | 8 (40%) | 18 (19%) | 7 (39%) |
| Tp53 loss/mut | 10 (18%) | 10 (16%) | 5 (25%) | 15 (16%) | 5 (28%) |
| t(4:14) | 4 (7%) | 4 (7%) | 1 (5%) | 7 (7%) | 0 (0%) |
| t(14:16) | 0 (0%) | 1 (2%) | 0 (0%) | 1 (1%) | 0 (0%) |
| t(14:20) | 1 (2%) | 0 (0%) | 0 (0%) | 1 (1%) | 0 (0%) |
| 1q dup/gain | 14 (25%) | 19 (31%) | 4 (20%) | 29 (30%) | 4 (22%) |
| IgG | 30 (55%) | 36 (59%) | 13 (65%) | 53 (55%) | 11 (61%) |
| IgA | 9 (16%) | 11 (18%) | 2 (10%) | 18 (19%) | 2 (11%) |
| FLC only | 16 (29%) | 14 (23%) | 5 (25%) | 25 (26%) | 5 (28%) |
| Kappa | 32 (58%) | 34 (56%) | 10 (50%) | 56 (58%) | 10 (56%) |
| Lambda | 23 (42%) | 26 (43%) | 10 (50%) | 39 (41%) | 8 (44%) |
| Heavy chain | 0 (0%) | 1 (2%) | 0 (0%) | 1 (1%) | 0 (0%) |
| AL-Amyloidosis | 0 (0%) | 5 (8%) | 1 (5%) | 4 (4%) | 0 (0%) |
| FLC mg/dL (median, IQR) | 54 (12,195) | 30 (10,140) | 34 (5,149) | 48 (14,173) | 80 (8,168) |
| M protein g/dL (median, IQR) | 1.4 (0.2,3.7) | 1.5 (0.2,2.8) | 1.0 (0.1,4.4) | 1.5 (0.2,2.9) | 1.0 (0.1,3.8) |
| Beta 2 Microglobulin mg/L (median, IQR) | 3.2 (2.2,4,4) | 3.4 (2.5,5.4) | 3.0 (2.5,3.6) | 3.5 (2.3,5.4) | 2.9 (2.2,3.5) |
| Albumin (g/dL) (median, IQR) | 3.7 (3.2, 4.0) | 3.4 (3.0, 3.8) | 3.6 (3.3, 3.9) | 3.5 (3.0, 3.9) | 3.6 (3.4, 3.9) |
| Corrected Calcium mg/dL (median, IQR) | 9.5 (9.2, 10.3) | 9.8 (9.2, 10.2) | 9.6 (9.1, 10.3) | 9.8 (9.2, 10.2) | 9.7 (9.0, 10.3) |
PC=plasma cells, IQR= intraquartile range, ISS= International Staging System, Fx= fracture, M=monoclonal, AL-amyloidosis= confirmed staining of light chain amyloid in tissue
Bolded cells indicate p<0.05 for comparisons between 560 column groups.
Patients with B7-H3 stromal staining had higher International Staging Scores (ISS) by Fischer’s exact test (p=0.001) than patients with no stromal staining (Figure 2 A–B). ISS is based on beta-2 microglobulin (B2M) and albumin levels. While there were no statistically significant differences in either B2M or albumin by B7-H3 staining (Figure 2 C–D), there were several outliers with high B2M in the B7-H3+ stromal staining group (p=0.144) that drove higher ISS.
Figure 2. B7-H3 is associated with higher stage, lytic lesions and pathologic fractures in MM.
PC=plasma cell, Fx=Fracture
A. Number of patients with ISS by B7-H3 staining. B. Proportion of patients by ISS in each B7-H3 staging group. C. Violin plot of beta 2 microglobulin in each group at diagnosis. D. Violin plot of albumin in each group at diagnosis. E. Violin plot of immunohistochemical (IHC) staining of plasma cells with 0–2 lytic lesions within +/− 30 days of diagnosis versus 3 or more lytic lesions. F. Violin plot of IHC staining in the marrow stroma in patients with 0–2 lytic lesions versus 3 or more lytic lesions. G. Number of lytic lesions (0-ε5) with no B7-H3 staining versus positive B7-H3 staining on plasma cells and/or stroma. H. Intensity of B7-H3 staining (0–3) on plasma cells by the presence or absence of pathologic fracture at diagnosis. I. Intensity of B7-H3 staining (0–3) on stroma by the presence or absence of pathologic fracture at diagnosis. J. Number of pathologic fractures in patients with no B7-H3 staining versus staining in plasma cells or stroma versus double positive patients.
Lytic bone lesions were divided into 0–2 or three or more to differentiate between significant bone disease or low level/absent bone disease. B7-H3 staining on plasma cells was not associated with a significant difference in lytic lesions (Figure 2E) but stromal staining was associated with higher lytic lesions (Figure 2F, p<0.001). The number of lytic lesions was compared in the double negative (negative plasma cell and stromal staining) to patients with positive B7-H3 (staining in either plasma cells or stroma). More B7-H3 double negative patients had no bone lesions at diagnosis (Figure 2G, p=0.004).
There were 48 (41.4%). patients with pathologic fracture within +/− 30 days of diagnosis (Table 1). Both plasma cell and stromal staining for B7-H3 were significantly associated with pathologic fracture at diagnosis (p<0.05, Figure 2H–I). Patients who were double negative for B7-H3 staining were compared to patients with either positive B7-H3 plasma cells or stroma, or with B7-H3 positivity in both plasma and stroma. The groups were compared by Fischer’s exact test (p=0.003) and there were significantly more pathologic fractures in double positive patients (55.9%) compared to single positive (30.8%) or double negative (16.6%) patients (Figure 2J). Bone marrow biopsies from two patients with bone disease showed membranous staining of OCs for B7-H3 (Figure 1E–F).
Bone lesions can serve as sites of clonal evolution, so we hypothesized that B7-H3 expression would impact progression free survival (PFS) in myeloma patients. We used log-rank test to compare higher B7-H3 staining intensity (2–3) versus 0 or 1 in a univariable model. We did not find B7-H3 staining intensity on plasma cells to be associated with PFS (Figure 3A) but higher B7-H3 staining intensity (2–3) on the stroma was associated with significantly worse PFS (Figure 3B, p<0.001).
Figure 3. Progression free survival by B7-H3 staining intensity.
A. PFS by B7-H3 immunohistochemical (IHC) staining intensity on plasma cells in bone marrow. B. PFS by B7-H3 IHC staining intensity on stroma in bone marrow.
B7-H3 expression on myeloma cells increases at relapse.
Since new therapeutics are typically tested in the relapsed setting, we used a cohort of 118 patients from Karolinska Institute with bone marrow aspirates at diagnosis, first relapse and triple class refractory relapse to determine how treatment affects B7-H3 expression. Of these patients, 40 had samples at all three time points (Supplemental Table 1). Using flow cytometry we found that the median expression of B7-H3 on CD138+ cell was 45 % at diagnosis but rose to 51% at first relapse (p=0.002) and 63% at triple class refractory relapse (p<0.001). (Figure 4A) although the mean fluorescence intensity ratio did not change significantly (Supplemental Figure 1A). None of the patients who expressed B7-H3 at diagnosis stopped expressing B7-H3 at subsequent relapses but 17% of patients who were negative at diagnosis (and had a subsequent biopsy) went on to express B7-H3 on plasma cells at subsequent relapses (Figure 4B).
Figure 4. B7-H3 expression at diagnosis and relapse by flow cytometry.
A. Percentage of CD138+ plasma cells expressing B7-H3 at diagnosis, relapse, and triple class refractory relapse determined by flow cytometry of marrow aspirates. Quarter lines in black and median in red. B. Percentage of CD138+ plasma cells expressing B7-H3 only showing patients negative at diagnosis (n=43).
An NK cell engager against B7-H3 enhances killing of myeloma cells.
After determining the suitability of B7-H3 as a target, we tested the B7-H3 TriKE against a variety of MM cell lines with varying sensitivity to NK cell-mediated cytotoxicity. The most NK sensitive target, H929 had the lowest expression of B7-H3 (Supplemental Figure 2A–B). RPMI-8226 (the most resistant to NK cell-mediated cytotoxicity), U266, and MM1S had similar intensities of B7-H3 expression by flow cytometry. Live cell imaging was performed over 48 hours with healthy donor NK cells at effector:target (E:T) ratios of 2:1 and 4:1 with or without 3nM B7-H3 TriKE (Figure 5A–B). At E:T of 2:1, the addition of B7-H3 TriKE improved NK cell mediated killing of H929 (p=0.0005), MM1S (p=0.0005), and U266 (p=0.0009). At E:T of 2:1 there was poor killing of RPMI-8226 even with TriKE but at E:T of 4:1 we saw improved killing with addition of TriKE (p=0.002). Assays with each cell line were performed in triplicate.
Figure 5. B7-H3 TriKE enhances killing of multiple myeloma cell lines by healthy donor and patient-derived NK cells.
A. Representative Incucyte Zoom Assay over 48 hours showing remaining live H929 cells after incubation with B7-H3 TriKE only, healthy donor-derived NK cells at 2:1, or NK 2:1 and B7-H3 TriKE. B. Summary graph of total myeloma cell killing after 48 hours of co-incubation with healthy donor-derived NK cells with or without B7-H3 TriKE. All E:T ratios shown were 2:1 except RPMI-8226 which were performed at 4:1. C. Representative Incucyte Zoom Assay over 48 hours showing remaining live RPMI-8226 cells after incubation with NK at 2:1 alone, with 3nM null TriKE, or with 3 nM B7-H3 TriKE. D. Representative Incucyte Zoom Assay over 48 hours showing remaining live B7-H3− RPMI-8226 cells after incubation with NK at 2:1 alone, with 3nM null TriKE, or with 3 nM B7-H3 TriKE. E. Single cells multi-plex ELISA was performed with healthy donor-derived NK cells incubated with IL-15 (negative control); or IL-15, IL-18, and IL-12 (positive control); or 3 nM B7-H3 TriKE for 48 hours. Polyfunctional strength index (PSI) was determined by the percentage of healthy donor-derived NK cells in a sample that secretes two or more proteins multiplied by the average signal intensity of secreted proteins from functional groups from each cell (left). Functional cytokine groups are color coded (right). F. Representative Incucyte Zoom Assay over 64 hours showing remaining live RPMI-8226 cells after incubation with patient-derived NK cells at 2:1 with or without B7-H3 TriKE. G. Summary graph of total myeloma cell killing after 64 hours of co-incubation with patient-derived NK cells with or without B7-H3 TriKE. H. Summary graph of CD107a staining of patient derived NK cells after incubation with matched marrow. N=4 *=p<0.05, **=p<0.01, ***=p<0.001
To determine the relative effect of the IL-15 moiety on NK cell killing we used a null (non-binding TriKE) with the same IL-15 linker and CD16 binding domain that does not bind MM cells. Although there was a trend towards increased killing of RPMI-8226 cells with the null TriKE, it was not statistically significant (Figure 5C). B7-H3 was then knocked out by the CRISPR/Cas9 gene editing system in RPMI-8226 and flow cytometry sorting was used to select B7-H3− cells. Killing assays were repeated with B7-H3− RPMI-8226 incubated with NK cells with either null or B7-H3 TriKE. No significant differences were seen in killing between the three groups (Figure 5D).
Single cell multiplex ELISA of healthy blood-derived NK cells was performed to test functionality after incubation with cytokines or B7-H3 TriKE. Polyfunctional strength index (PSI), a measure of cytokine and chemokine secretion, has been associated with improved clinical outcomes to chimeric antigen receptor (CAR) T cells [28, 29]. PSI was determined by the percentage of single cells in a sample that secreted two or more proteins multiplied by the average signal intensity of secreted proteins from functional groups from each cell (Figure 5E). PSI was higher with B7-H3 TriKE compared with IL-15 alone or IL-12, IL-15, and IL-18. Heatmap analysis of combinations of cytokines from the different treatment groups showed more combinations expressing chemoattractive (CCL-11, IP-10, MIP-1b, and RANTES) and inflammatory (IL-17A, IL-17F, IL-1b, IL-6, MCP-1, and MCP-4) profiles than either IL-15 alone or IL-12, IL-15, and IL-18 (Supplemental Figure 3A). NK incubated with B7-H3 TriKE secreted more cytolytic perforin and granzyme as well as increased TNF-a and GM-CSF (Supplemental Figure 3B).
Since MM patients have significant NK cell deficits, we tested NK cells isolated from the blood of five MM patients that were in remission and on maintenance therapy (3 on lenalidomide, 1 on daratumumab and lenalidomide, and 1 on carfilzomib and lenalidomide). We found that NK cells alone exhibited minimal killing (<20%) against myeloma cell lines. With the addition of B7-H3 TriKE, killing was significantly enhanced but required a longer period of incubation with TriKE compared to healthy donor NK cells (Figure 5F–G). At E:T of 2:1 we saw a significant difference in killing of H929 (p=0.019), MM1S (p=0.0002), RPMI-8226 (p= 0.009), and U266 (p=0.003) at 64 hours of incubation.
We performed co-incubation of autologous NK cells with bone marrow aspirate (collected on the same day from 2 relapsed and 2 newly diagnosed MM patients) and compared degranulation with or without B7-H3 TriKE (Figure 5H). The percentage of plasma cells per patient ranged from 18 to 47% by flow cytometry. Degranulation increased significantly in all four patients with B7-H3 TriKE compared to NK cells alone (p<0.001).
B7-H3 TriKE enhances killing of MDSC and OCs.
We next used bone marrow biopsies from MM patients to determine if MDSC routinely expressed B7-H3 and if this population could differentiate into OCs under inflammatory conditions. We collected six patients (n=1 at diagnosis, 1 at relapse, and 4 in remission after autologous stem transplant). CD14+ cells were isolated by magnetic beads and flow cytometry confirmed CD11b and B7-H3 expression. In five of six patients, MDSC expressed >75% B7-H3 expression (Figure 6A). We also collected peripheral blood and marrow from one patient (in remission after transplant) and attempted to differentiate OCs on glass slides by 14-day incubation with RANKL and M-CSF. Like the results reported in a breast cancer model [30], MDSC from the bone marrow differentiated into TRAP+, multi-nucleated OCs while MDSC from peripheral blood did not (Supplemental Figure 4B–C).
Figure 6. CD14+ Myeloid Derived Suppressor Cells express B7-H3 and are targeted by B7-H3 TriKE.
A. Summary graph of B7-H3 expression on MDSC derived from PBMC with either cytokines (IL-6 and G-CSF) or MM cell lines as well as bone marrow aspirates from newly diagnosed or relapsed MM patients. B. Representative Incucyte showing killing of MDSC derived from peripheral blood by peripheral bloodderived NK cells with or without 3 nM B7-H3 TriKE at E:T of 1:1. C. Summary graph of total myeloma cell killing after 48 hours of co-incubation with patient-derived NK cells with or without B7-H3 TriKE. E-F Healthy donor NK cells were incubated alone, with 3 nM B7-H3 TriKE, with MDSC at E:T of 2:1, or with 3 nM B7-H3 TriKE and MDSC at E:T of 2:1 for four hours. D. Summary graph of CD107a staining. N=4 E. Summary graph of intracellular IFN-© staining on NK cells after four hour co-culture with MDSC at E:T of 2:1. N=4. F. OC were grown on bone chips and stained for TRAP. On the left are bone chips with OC incubated with B7-H3 TriKE and on the right are NK cells and OC (2:1) incubated with B7-H3 TriKE. G. Summary data of the average number of TRAP+ OCs per visual field on each slide. N=3. H. Summary data on percent of bone chip erosion with OC alone, OC with 3n B7-H3 TriKE, NK cells at 2:1 or NK cells at 2:1 with 3 nM B7-H3 TriKE after 72 hours. N=3. *=p<0.05, **=p<0.01
To generate sufficient MDSC for in vitro killing assays, we isolated CD33+ cells from healthy donor blood with inflammatory cytokines or coincubation with MM cell lines. B7-H3 expression was assayed on MDSC generated by cytokines, MM cells, or patient samples (Figure 6 A, Supplemental Figure 4A).). We confirmed the immunosuppressive function of in vitro-derived MDSC by testing their effect on NK cell proliferation as described in the methods. (Supplemental Figure 4D).
Next, we incubated in vitro-derived MDSC with healthy donor derived-NK cells alone or with B7-H3 TriKE at E:T of 1:1 for 48 hours and found significantly enhanced killing with the B7-H3 TriKE (p=0.0024, Figure 6 B–C). A flow cytometry based killing assay showed increased degranulation against MDSC (p=0.02, Figure 6 D) and enhanced secretion of IFN-γ against MDSC with TriKE compared to NK cells alone (p=0.008, Figure 6 E Supplemental Figure 4E.). OCs were grown on bone chips and incubated with B7-H3 TriKE alone of NK cells with B7-H3 TriKE. Viable OCs were counted per visual field after TRAP staining (Figure 6 F). A significant reduction of OCs occurred after incubation with NK cells and TriKE compared to TriKE alone (p<0.001) (Figure 6 G). Bone erosion was measured as a percentage of the bone chip after 72 hours. Incubation with NK cells alone or NK cells with B7-H3 TriKE significantly decreased bone erosion compared to OCs alone. Incubation with B7-H3 TriKE alone did not significantly change bone erosion. While not statistically significant, co-incubation with NK cells and B7-H3 TriKE further suppressed bone erosion compared to NK cells alone (mean=12.5±6% versus 35.0±14%) (Figure 6H).
B7-H3 TriKE improves killing of fibroblasts.
The fibroblast cell line Hs5 has been used to model interactions between MM and the stroma [17, 31, 32]. We saw high B7-H3 expression on Hs5 fibroblasts by flow cytometry (Supplemental Figure 5A). When we incubated Hs5 with healthy donor derived-NK cells alone or with B7-H3 TriKE at E:T of 2:1 for 48 hours, we saw significantly enhanced killing against fibroblasts with the B7-H3 TriKE (p=0.0005, Figure 7 A–B). Live cell killing assays were performed with HS5 and MM cells line at 1:1 ratio and normalized to stained MM targets alone. Peripheral blood NK cells were incubated at an effector:target ratio of 2:1 with or without 3 nM B7-H3 TriKE. Although HS5 increased resistance of MM cells to NK cell mediated killing, the addition of the B7-H3 TriKE improved killing of co-incubated MM cells compared to NK cells alone (Figure 7C–D).
Figure 7. B7-H3 TriKE enhances killing of stromal cells.
A. Representative Incucyte showing killing of HS 5 fibroblasts by peripheral blood-derived NK cells with or without 3 nM B7-H3 TriKE at E:T of 2:1. B. Summary graph of HS 5 fibroblast killing after 48 hours of co-incubation with NK cells with or without B7-H3 TriKE. C. Representative killing assays with with HS5 and MM cells line at 1:1 ratio and normalized to stained MM targets alone. Peripheral blood NK cells were incubated at effector:target of 2:1 with or without 3 nM B7-H3 TriKE. D. Summary graph of MM killing with or without HS5 co-incubation after 48 hours of co-incubation with NK cells with or without B7-H3 TriKE.
Discussion
In a retrospective cohort with 243 patients, we showed that B7-H3 is expressed on MM and immunosuppressive cells in the marrow. Flow cytometry of aspirates and IHC staining of decalcified marrow showed similar rates of plasma cell positivity (64% versus 53%) but IHC staining showed that even patients with negative staining of plasma cells often were positive for B7-H3 on the stroma. Bone marrow aspirates collected at diagnosis, first relapse, and at triple class refractory relapse showed increasing B7-H3 expression. B7-H3 expression is regulated post-transcriptionally in several solid tumors [33, 34] but less is known in MM. In this cohort, we did not find any specific cytogenetic abnormalities or cytogenetic risk to be associated with B7-H3 expression, suggesting that most MM patients can benefit from B7-H3 targeting.
Unlike other therapeutically targeted antigens in MM that are predominantly immune cell specific, B7-H3 is unique due to its association with bone lesions. Here we show that both plasma cell and stromal expression were associated with pathologic fractures at diagnosis, but only stromal staining was associated with more lytic lesions and worse PFS. Bone disease remains a significant driver of morbidity in MM patients with 60% experiencing a fracture at some point [35] and most patients experiencing dysregulated bone turnover at diagnosis [36]. We showed here that CD14+ cells from marrow can differentiate into OCs and that the B7-H3 TriKE promotes killing of these MDSC by NK cells. Targeting this population may stop the formation of new lytic lesions and have positive impacts on bone health.
In addition to worse PFS, B7-H3 expression on stroma was also associated with a higher ISS and this was driven by high B2M. B2M is a surrogate marker for myeloma disease burden [37], suggesting that increased B7-H3 is associated with more proliferative disease and/or a higher number of MM cells in total. This may explain why B7-H3 expression increases in the relapsed setting as the disease becomes increasingly aggressive. Interestingly, all patients with pathologically confirmed AL-amyloidosis had either B7-H3 positive stroma, plasma cells, or both. Whether or not B7-H3 is a good target for patients with AL-amyloidosis will require further investigation in a larger cohort.
NK cells from MM patients have significant killing defects [38, 39], so it is notable that B7-H3 TriKE enhanced killing of MM cell lines with patient-derived NK cells and increased degranulation of autologous NK cells against MM cells from marrow. We hypothesize that restoration of NK cell function is due to the proliferative effects of the B7-H3 TriKE IL-15 moiety. The cytokine secretion profile of NK cells was also improved after incubation with the B7-H3 TriKE compared to IL-15 or IL-12, IL-15, and IL-18 in combination although the PSI has not been widely studied with NK cell therapeutics. Since NK cells orchestrate dendritic and T cell activation [40], treatment with B7-H3 TriKE might enhance downstream immune responses against myeloma. Future work should explore the synergistic effects of NK cell engagers with T cell redirecting therapies.
In addition to direct killing of MM cells and enhanced secretory function of NK cells, the B7-H3 TriKE also enhanced killing of fibroblasts, OCs, and MDSC. Fibroblasts promote resistance of MM to CAR-T therapy [41] and MDSC suppress T cell function [42] suggesting that antecedent therapy with B7-H3 TriKE might enhance CAR-T cell function.
In summary, these data from over 200 myeloma patients show that B7-H3 expression is an important predictor of worse prognosis in MM. The B7-H3 TriKE restores killing by patient-derived NK cells and may uniquely alter the tumor microenvironment. Given its unique targeting profile it may enhance the therapeutic potential of other available MM therapies and is planned for clinical testing.
Supplementary Material
Key Points.
High B7-H3 is associated with bone lesions, fractures, and worse PFS in >200 MM patients
An NK cell engager induced NK cells to kill MM, OCs, fibroblasts, and MDSC
Acknowledgements
Research reported in this publication was also supported by NIH grant 5K08 CA255426–03 and made possible by the Randy Shaver Cancer Research and Community Fund. We utilized the Biostatistics and Bioinformatics Core shared resource of the Masonic Cancer Center, University of Minnesota.
Footnotes
Conflict of Interest Disclosure
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Data sharing:
For original data please contact the corresponding author (merin008@umn.edu)
References
- 1.Korst C, et al. , The bone marrow NK-cell profile predicts MRD negativity in patients with multiple myeloma treated with daratumumab-based therapy. Blood, 2025. 145(25): p. 3007–3014. [DOI] [PubMed] [Google Scholar]
- 2.Mizuhara K, et al. , Tumour-derived exosomes promote the induction of monocytic myeloid-derived suppressor cells from peripheral blood mononuclear cells by delivering miR-106a-5p and miR-146a-5p in multiple myeloma. Br J Haematol, 2023. 203(3): p. 426–438. [DOI] [PubMed] [Google Scholar]
- 3.De Veirman K, et al. , Myeloid-derived suppressor cells induce multiple myeloma cell survival by activating the AMPK pathway. Cancer Lett, 2019. 442: p. 233–241. [DOI] [PubMed] [Google Scholar]
- 4.Romano A, et al. , Mechanisms of Action of the New Antibodies in Use in Multiple Myeloma. Front Oncol, 2021. 11: p. 684561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Casneuf T, et al. , Effects of daratumumab on natural killer cells and impact on clinical outcomes in relapsed or refractory multiple myeloma. Blood Adv, 2017. 1(23): p. 2105–2114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Janakiram M, et al. , The third group of the B7-CD28 immune checkpoint family: HHLA2, TMIGD2, B7x, and B7-H3. Immunol Rev, 2017. 276(1): p. 26–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lin L, et al. , B7-H3 promotes multiple myeloma cell survival and proliferation by ROS-dependent activation of Src/STAT3 and c-Cbl-mediated degradation of SOCS3. Leukemia, 2019. 33(6): p. 1475–1486. [DOI] [PubMed] [Google Scholar]
- 8.Vallera DA, et al. , NK-Cell-Mediated Targeting of Various Solid Tumors Using a B7-H3 Tri-Specific Killer Engager In Vitro and In Vivo. Cancers (Basel), 2020. 12(9). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Castriconi R, et al. , Identification of 4Ig-B7-H3 as a neuroblastoma-associated molecule that exerts a protective role from an NK cell-mediated lysis. Proc Natl Acad Sci U S A, 2004. 101(34): p. 12640–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Vigdorovich V, et al. , Structure and T cell inhibition properties of B7 family member, B7-H3. Structure, 2013. 21(5): p. 707–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lim S, et al. , Immunoregulatory Protein B7-H3 Reprograms Glucose Metabolism in Cancer Cells by ROS-Mediated Stabilization of HIF1alpha. Cancer Res, 2016. 76(8): p. 2231–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Luo D, et al. , B7-H3 regulates lipid metabolism of lung cancer through SREBP1-mediated expression of FASN. Biochem Biophys Res Commun, 2017. 482(4): p. 1246–1251. [DOI] [PubMed] [Google Scholar]
- 13.Steinberger P, et al. , Molecular characterization of human 4Ig-B7-H3, a member of the B7 family with four Ig-like domains. J Immunol, 2004. 172(4): p. 2352–9. [DOI] [PubMed] [Google Scholar]
- 14.Van der Vreken A, et al. , B7-H3 nanobody-based CAR T cells control multiple myeloma growth, while dual BCMA/B7-H3 CAR T cells overcome antigen escape. J Hematol Oncol, 2025. 18(1): p. 103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wang R, et al. , B7-H3 promotes colorectal cancer angiogenesis through activating the NF-kappaB pathway to induce VEGFA expression. Cell Death Dis, 2020. 11(1): p. 55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zhan S, et al. , Overexpression of B7-H3 in alpha- SMA-Positive Fibroblasts Is Associated With Cancer Progression and Survival in Gastric Adenocarcinomas. Front Oncol, 2019. 9: p. 1466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bar-Natan M, et al. , Bone marrow stroma protects myeloma cells from cytotoxic damage via induction of the oncoprotein MUC1. Br J Haematol, 2017. 176(6): p. 929–938. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yim J, et al. , Effects of B7-H3 expression on tumour-infiltrating immune cells and clinicopathological characteristics in non-small-cell lung cancer. Eur J Cancer, 2020. 133: p. 74–85. [DOI] [PubMed] [Google Scholar]
- 19.Sawant A and Ponnazhagan S, Myeloid-derived suppressor cells as OC progenitors: a novel target for controlling osteolytic bone metastasis. Cancer Res, 2013. 73(15): p. 4606–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Li H, et al. , Cancer-expanded myeloid-derived suppressor cells induce anergy of NK cells through membrane-bound TGF-beta 1. J Immunol, 2009. 182(1): p. 240–9. [DOI] [PubMed] [Google Scholar]
- 21.Oh Y, et al. , B7-H3 regulates OC differentiation via type I interferon-dependent IDO induction. Cell Death Dis, 2021. 12(11): p. 971. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Khaw MJ, et al. , Novel trispecific killer engager targeting B7-H3 enhances natural killer cell antitumor activity against head and neck cancer. J Immunother Cancer, 2025. 13(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mikhael JR, et al. , Management of newly diagnosed symptomatic multiple myeloma: updated Mayo Stratification of Myeloma and Risk-Adapted Therapy (mSMART) consensus guidelines 2013. Mayo Clin Proc, 2013. 88(4): p. 360–76. [DOI] [PubMed] [Google Scholar]
- 24.Greipp PR, et al. , International staging system for multiple myeloma. J Clin Oncol, 2005. 23(15): p. 3412–20. [DOI] [PubMed] [Google Scholar]
- 25.Phillips P and Pain R, Correcting the calcium. Br Med J, 1977. 1(6074): p. 1473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Lechner MG, Liebertz DJ, and Epstein AL, Characterization of cytokine-induced myeloid-derived suppressor cells from normal human peripheral blood mononuclear cells. J Immunol, 2010. 185(4): p. 2273–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Danilin S, et al. , Myeloid-derived suppressor cells expand during breast cancer progression and promote tumor-induced bone destruction. Oncoimmunology, 2012. 1(9): p. 1484–1494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Rossi J, et al. , Preinfusion polyfunctional anti-CD19 chimeric antigen receptor T cells are associated with clinical outcomes in NHL. Blood, 2018. 132(8): p. 804–814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Dong Z, et al. , Analysis of polyfunctionality for enhanced BAFF-R CAR T-cell therapy for hematologic malignancies. Blood Adv, 2024. 8(15): p. 4066–4076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sawant A, et al. , Myeloid-derived suppressor cells function as novel OC progenitors enhancing bone loss in breast cancer. Cancer Res, 2013. 73(2): p. 672–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Perez LE, et al. , Bone marrow stroma confers resistance to Apo2 ligand/TRAIL in multiple myeloma in part by regulating c-FLIP. J Immunol, 2008. 180(3): p. 1545–55. [DOI] [PubMed] [Google Scholar]
- 32.Binder M, et al. , Bone marrow stromal cells induce chromatin remodeling in multiple myeloma cells leading to transcriptional changes. Nat Commun, 2024. 15(1): p. 4139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Nygren MK, et al. , Identifying microRNAs regulating B7-H3 in breast cancer: the clinical impact of microRNA-29c. Br J Cancer, 2014. 110(8): p. 2072–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Shao L, et al. , MicroRNA-326 attenuates immune escape and prevents metastasis in lung adenocarcinoma by targeting PD-L1 and B7-H3. Cell Death Discov, 2021. 7(1): p. 145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Terpos E, et al. , Prognostic variables for survival and skeletal complications in patients with multiple myeloma osteolytic bone disease. Leukemia, 2010. 24(5): p. 1043–9. [DOI] [PubMed] [Google Scholar]
- 36.Gerov V, et al. , Circulating sRANKL, Periostin, and Osteopontin as Biomarkers for the Assessment of Activated OCogenesis in Myeloma Related Bone Disease. Cancers (Basel), 2023. 15(23). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Bataille R, Grenier J, and Sany J, Beta-2-microglobulin in myeloma: optimal use for staging, prognosis, and treatment--a prospective study of 160 patients. Blood, 1984. 63(2): p. 468–76. [PubMed] [Google Scholar]
- 38.Blanquart E, et al. , NK cells with adhesion defects and reduced cytotoxic functions are associated with a poor prognosis in multiple myeloma. Blood, 2024. 144(12): p. 1271–1283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Konjevic G, et al. , Decreased CD161 activating and increased CD158a inhibitory receptor expression on NK cells underlies impaired NK cell cytotoxicity in patients with multiple myeloma. J Clin Pathol, 2016. [DOI] [PubMed] [Google Scholar]
- 40.Deauvieau F, et al. , Human natural killer cells promote cross-presentation of tumor cell-derived antigens by dendritic cells. Int J Cancer, 2015. 136(5): p. 1085–94. [DOI] [PubMed] [Google Scholar]
- 41.Sakemura R, et al. , Targeting cancer-associated fibroblasts in the bone marrow prevents resistance to CART-cell therapy in multiple myeloma. Blood, 2022. 139(26): p. 3708–3721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Gorgun GT, et al. , Tumor-promoting immune-suppressive myeloid-derived suppressor cells in the multiple myeloma microenvironment in humans. Blood, 2013. 121(15): p. 2975–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
For original data please contact the corresponding author (merin008@umn.edu)