Human leukemia (K562), breast (MDA-MB231), ovarian (OVCAR8), lung (NCI-H46), colon (HCT116), bladder (RT112), and multiple myeloma (RPMI 8226, NCI-H929, MM1.S and U266) cells were stored at Genentech’s internal cell line repository, gCell. K562, MDA-MB231, NCI-H460, HCT-116, RPMI 8226, NCI H929 and U266 cells were obtained from ATCC. OVCAR8 cells were obtained from NCI-Frederick Cancer DCTD. RT112 were obtained from DSMZ and MM1.S cells were obtained from Northwestern University. Tumor cells were cultured in culture medium (RPMI 1640, 10% FBS, 1% penicillin/streptomycin, 1% glutamine). For functional assays, adherent tumor cells were isolated into single cells with 0.5% trypsin/EDTA and following centrifugation, were subsequently resuspended in culture medium.

Whole blood samples from healthy donors were supplied through the Samples for Science program, an IRB approved research program operated out of the Genentech Campus Health Center. NK cells were purified via negative selection using the RosetteSep Human NK Isolation Kit (StemCell Technologies) according to the manufacturer’s protocol. Briefly, 500 μL of reagent were added per 10 ml of whole blood for at least 30 min. Subsequently, whole blood samples were diluted 1:1 with 1X sterile PBS and overlaid on Ficoll-Paque (cat #45-001-749) and centrifuged for 20 min at 1,200 g (no brakes). The interface containing enriched NK cells was isolated and NK cells were enumerated and purity of NK cells (Live-Dead-CD45+CD3-CD14/19-CD56+) as determined by flow cytometry was >90% for all participants evaluated ( Supplementary Data 1 ).

Enriched blood NK cells were cultured in culture medium in the presence or absence of 10 µg/ml XmAb24306 for 18–20 hrs. Primed viable NK cells (>95% viability) ( Supplementary Data 1 ) were then washed, enumerated and co-cultured with target cells at various E:T ratios where specified. The same number of NK cells for XmAb24306 and untreated conditions were added to target cells. The viability of untreated or XmAb24306 treated NK cells remained at 90-95% during various time points of the K562 cytotoxicity assay.

NK cell cytotoxicity was performed using a flow cytometry cytotoxicity assay based on the measurement of the uptake of Annexin V and 7AAD dyes in target tumor cells. The target cancer cells were labeled with CellTrace Violet (0.5µM; ThermoFisher Scientific) according to the manufacturer’s instructions and then suspended in culture medium and plated at a density of 5×10⁴ cells per well in 96-well plates. Enriched NK cells (effector) that were either pre-activated with XmAb24306 (10µg/ml) overnight or cultured in control media were then added at various E:T ratios as indicated. The plates were centrifuged at 200 g for 2 min and incubated for the indicated times at 37°C, 5% CO₂. Control wells were set up without NK cells to monitor background tumor cell death in the absence of effector NK cells. At the end of the cytotoxicity assay, the cell mixtures were stained for surface antibodies and subsequently resuspended in Annexin V binding buffer (1X) containing 1:100 dilution of Annexin V-PE and 7AAD as per the manufacturer’s protocol. Dead cells were reported as the sum of AnnexinV+7AAD double positive cells and 7AAD+AnnexinV- single positive cells. Apoptotic cells were reported as 7AAD-Annexin V+. Tumor cell and NK cell death were monitored using either a BD Fortessa or Cytek Aurora flow cytometers and analyzed using either FlowJo software (Version 10.6.2, Treestar) or cloud based OMIQ. Specific NK induced cell lysis was calculated using the following formula: % NK cell induced death = (% tumor death in tumor + NK cells) - (% tumor death in tumor - NK cells). K562 and solid tumor cell line killing assays were tested with at least 6 individual donors.

To evaluate the ADCC capacity of XmAb24306 primed NK cells, we measured NK cell mediated cytotoxicity against MM cells in the presence of the anti-CD38 monoclonal antibody dara. Briefly, target MM cell lines including RPMI 8226, NCI-H929, U266, MM1.S were labeled with CellTrace Violet according to the manufacturer’s protocol and 5×10⁴ cells were incubated with 10 μg/ml of dara for at least 1 hour at 37°C, 5% CO₂ before addition of effector NK cells that had been preactivated with XmAb24306 (10µg/ml). Various E:T ratios (0:1, 0.2:1, 0.5:1, 1:1, 2:1, 5:1) were used for these assays and co-cultures were allowed to proceed for a total of 5 h. The viability of MM cells prior to the initiation of the ADCC assay was >90%. Cells were then harvested and stained with extracellular surface antibodies followed by Annexin V and 7AAD staining to label dead target cells. Amount of NK cell ADCC was calculated using the formula used to calculate cytotoxicity above. % NK cell fratricide= (% NK cell death in presence of dara with or without XmAb24306) – (% NK cell death in the absence of dara with or without XmAb24306). For evaluating the polyfunctional NK cell cytokine responses, brefeldin A and monensin (to inhibit intracellular protein transport) were added during the 5 h incubation and subsequently, extracellular and intracellular staining were completed as described below. Experiments measuring ADCC were conducted with n=3-4 individual donors as specified and are pooled data from 2-3 independent experiments.

Cell counts for prepared single-cell suspensions of NK cells were determined using a Vi-CELL XR (Beckman Coulter). Briefly, 1×10⁵ enriched blood NK cells from single-cell suspensions were incubated with a mixture of fluorescence-conjugated anti-human antibodies for 30 min at 4°C. Stained cells were then washed twice, and following extracellular surface staining, where appropriate, the Foxp3 Fixation/Permeabilization Concentrate and Diluent Kit (Thermo Fisher Scientific) was used for intracellular staining. Samples were subsequently acquired with an LSR Fortessa flow cytometer (BD) using FACSDiva software (BD) or Cytek Aurora flow cytometer (Cytek Biosciences) using SpectroFlo Software v2.2.0.2. To evaluate NK cell degranulation and cytokine production, NK cells were cocultured with the appropriate target cells for 5 h in the presence of anti-CD107a and 1X protein transport inhibitor cocktail (ThermoFisher Scientific). This was followed by surface and intracellular staining for cytokine production as described above. Analysis was completed using FlowJo software version 10.2 (TreeStar) or cloud-based OMIQ (Dotmatics). A comprehensive list of all antibodies used can be found in Supplementary Table 1 .

To generate spheroids using HCT-116 colorectal tumor cells, 5×10³ CellTrace Far Red (2.5 µM) (ThermoFisher Scientific) labeled cells were plated and centrifuged (290 g for 3 min) on Day 0 in 96 well Nunclon Sphera plates (ThermoFisher Scientific). Nunclon Sphera plates were subsequently centrifuged (100 g, 3 min) and cultured at 37°C, 5% CO₂. Spheroids were monitored for development over 5 days. On Day 3, 1x10⁵ Carboxyfluorescein succinimidyl ester (CFSE) (2.5 μM) (ThermoFisher Scientific) labeled NK cells that had been stimulated overnight with or without XmAb24306 (10µg/ml) were added to spheroids and cocultures were then subjected to time lapse confocal imaging as described below. Three technical replicates were established for each donor evaluated.

Spheroids were imaged every 4 hours for 5 days at 37°C and 5% CO₂. Z-stacks were acquired in phase, 488 and 647 channels at 10 μm steps (140 μm total) with a 10× Plan Fluor objective (NA: 0.3, Nikon). Imaging was performed on a Nikon Ti-E perfect focus inverted microscope equipped with a spinning disk confocal CSU-X1 (Andor, Oxford Instruments) motorized X, Y stage (Nikon), environmental chamber (OkoLab) and Prime 95B sCMOS camera (Teledyne Photometrics), all controlled by NIS-Elements software (Nikon). Z-stacks were processed into maximum intensity projections, spheroids were traced in the phase image and area was measured at every time point, then normalized to the initial measurement at t=0. Total integrated intensity of the CellTrace Far Red fluorescence was measured within spheroid ROIs in the 647 channel and normalized to t=0 measurements. 6 donors with 3 wells each were assayed with or without XmAb24306 treated NK cells, or spheroids alone.

The results are presented as mean +/- standard deviation (s.d.) or with box plots, as appropriate. Data sets with sample size greater than or equal to 10 were evaluated for normality using Kolmogorov-Smirnov, Shapiro-Wilk and D’Agostino & Pearson normality tests. If there was insufficient evidence to reject normality, a one-way ANOVA with Tukey multiple comparisons test was applied. In the case of failed normality tests or small sample sizes, Wilcoxon-matched paired rank tests were used to evaluate differences between two paired groups, and the Friedman test with Dunn’s multiple comparisons to evaluate multiple matched group comparisons. For statistical evaluation, P<0.05 was considered significant. Not significant (NS)= P>0.05, *P<0.05, **P<0.005, ***P<0.0005 and ****P<0.0001. Corrections for multiple hypothesis testing were not performed for this exploratory analysis. Correlation values were calculated with Pearson tests. The number of donors used were N=6-11 per group and is indicated in each respective figure legend. Python was used for graphical visualization of data. All statistical analyses were performed using GraphPad Prism 9.4 (Dotmatics).

To determine whether XmAb24306 may enhance the killing efficiency of NK cells directly ex vivo, we enriched primary NK cells from the blood of healthy participants ( Supplementary Figure 1 ) and incubated NK cells with or without XmAb24306 (10µg/ml) for 18–20 hrs. After incubation, live NK cells from untreated and XmAb24306 treated conditions were counted and equal numbers were then co-cultured with cell trace violet (CTV) labeled chronic myelogenous leukemia cell line, K562, at various effector to target (E:T) ratios. The efficiency with which NK cells killed K562 targets was measured at defined time points ( Figures 1A, B ). Treatment of NK cells with XmAb24306 led to more efficient target cell killing ( Figure 1A ) and more rapid induction of early apoptosis ( Figure 1B ) compared to the activity of the same NK cells that were not primed with XmAb24306. XmAb24306 significantly increased the efficiency of killing at all E:T ratios examined, including lower E:T ratios (0.2:1 and 0.5:1) that more likely model the in vivo situation. In these lower E:T ratio conditions, efficiency of NK cell killing was increased by about two-fold in the 5 h cytotoxicity assay ( Figure 1A , right).

Granule exocytosis is the predominant mechanism used by NK cells to eliminate their targets (20). The cell surface expression of CD107a, a lysosomal protein, is considered a sensitive marker of NK cell degranulation and release of cytotoxic granules. To determine the degranulation capacity of XmAb24306 treated NK cells against K562 cells, we co-cultured NK cells that were treated with or without XmAb24306 with K562 target cells for 5 h and measured the expression of CD107a as a surrogate marker for NK cell degranulation by flow cytometry. Our data shows that XmAb24306 increases CD107a expression, suggesting enhanced degranulation capacity of NK cells within 5 h against K562 target cells ( Figure 1C ).

K562 tumor cells lack HLA-I expression, which predisposes them to be highly sensitive to NK cell mediated killing (43, 44). In contrast, NK cell killing of HLA-I expressing tumor cells is less efficient due to the presence of inhibitory signaling. To determine if XmAb24306 could also increase NK cell cytotoxicity against HLA-I expressing cell lines, we used various solid tumor cell lines derived from lung, colon, ovarian, bladder and breast cancers. We observed that while NK cell cytotoxicity against these targets was less than that observed for K562, XmAb24306 consistently and significantly enhanced the cytotoxic capacity of NK cells against these various tumor cell lines (E:T=2:1) ( Figure 1D ) (45, 46). These results suggest that XmAb24306 can enhance NK cell killing against a broad range of tumor cell lines in vitro.

In addition to enhancing killing of tumor cell lines, we wanted to know whether XmAb24306 would enhance killing of three-dimensional spheroids as well. We evaluated the effects of NK cells (CFSE-labeled) on the integrity of colorectal (HCT116) tumor spheroids fluorescently labeled with CellTrace Far Red dye, using bright-field and fluorescence confocal time lapse microscopy ( Figure 2A ; Movie S1 ). HCT116 are known to have high levels of HLA-I and NKG2DL (47). Within 24 h after NK cell addition, XmAb24306-treated NK cells began to infiltrate spheroids and intersperse with the HCT116 cells. The integrity and delineation of the tight tumor spheroid border significantly decreased after coculture with XmAb24306 treated NK cells ( Figure 2A ; Movie S3 ) in contrast to the untreated NK cells ( Figure 2A ; Movie S2 ). This disruption of the spheroid border coincided with a decrease in the relative spheroid area ( Figure 2B ) and CellTrace (HCT116 -label) intensity ( Figure 2C ) over the time course, indicating increased cytolytic activity of the XmAb24306-treated NK cells. Taken together, these data suggest that XmAb24306 enhances the capacity of NK cells to disintegrate and kill 3D colorectal tumor spheroids in vitro.

To understand how the ADCC function of NK cells would be influenced by XmAb24306, we measured dara-mediated cytotoxicity against RPMI 8226, a MM cell line with high cell surface CD38 expression (48). Target RPMI 8226 were labeled with CTV and co-cultured with either untreated NK cells or those pretreated with XmAb24306 (E:T= 2:1). XmAb24306 and dara alone can each enhance NK cell killing of RPMI 8226 cells, and the combination of dara and XmAb24306 shows an improvement in cell killing over the single agents ( Figure 3A ). These data show that XmAb24306 can enhance the direct killing capacity of NK cells against the RPMI 8226 MM cell line even in the absence of dara, and that the combination treatment of XmAb24306 and dara in vitro show an additive effect leading to a significantly enhanced ADCC effect compared to dara alone ( Figure 3A ).

In addition to their cytotoxic functionality, NK cells also produce effector cytokines upon activation (21–23). In this regard, we compared cytokine production (IFN-γ, TNF-α and MIP1-β) and the cytotoxic degranulation activity (CD107a) of NK cells in response to treatment with XmAb24306, dara and the combination of XmAb24306 and dara during cocultures with RPMI 8226 ( Figure 3B ). The fraction of polyfunctional NK cells producing any combination of two effector molecules was significantly higher in the presence of XmAb24306 and dara ( Figure 3B ).

As RPMI 8226 may not represent the heterogeneity of target expression and inhibitory molecules in MM, we asked whether XmAb24306 would enhance NK cell ADCC in response to other MM cell lines. We evaluated NK cell ADCC in response to NCI-H929 (CD38^hi), MM1.S (CD38^hi), U266 (CD38^lo) and L363 (CD38^lo) (E:T=2:1) which also express varying levels of NKG2DL, HLA-I and HLA-E ( Supplementary Data 2 ) (49–54). XmAb24306 alone significantly enhanced the direct killing capacity and ADCC against all MM cell lines. However, the combination with dara did not significantly increase ADCC above the dara alone condition for the other CD38^hi expressing cell lines ( Figure 3C , NCI-H929, MM1.S). Of note, dara led to a reduction in NK cell mediated killing of CD38^lo U266 and L363 cells, an expected result given the low expression of CD38 on these targets and potentially enhanced NK cell death due to dara-mediated fratricide of the CD38+ NK cells. Despite the lower CD38 expression on U266 and L363 cells ( Supplementary Data 2 ) (49–54), XmAb24306 alone enhanced NK cell killing of these target cell lines in the absence of dara ( Figure 3C ). Taken together, this data shows that the combination of XmAb24306 with dara increased ADCC on CD38^hi MM cell lines and that XmAb24306 alone was able to enhance NK cell killing against CD38^lo MM cell lines.

A subset of NK cells express CD38 (37). Dara has been shown to mediate NK cell fratricide by binding to CD38 expressed on NK cells (37). We therefore asked whether the combination of XmAb24306 and dara would augment NK cell fratricide. Target RPMI 8226 were labeled with CTV and co-cultured with either untreated NK cells or those pretreated with XmAb24306 (E:T=2:1) in the presence or absence of dara and NK cell death was evaluated within 5 h. Our results show that in the absence of dara, XmAb24306 had minimal effects on NK cell death ( Figure 4A ). However, when combined with dara, XmAb24306 significantly increased the degree of NK cell death (fratricide) compared to dara alone ( Figure 4A ).

To determine whether the enhanced fratricide in the presence of XmAb24306 and dara decreased ADCC mediated killing by NK cells, we performed correlative analyses between % cytotoxicity mediated by ADCC and % NK cell fratricide. We did not observe any negative correlation between higher rates of NK cell fratricide and NK cell ADCC either in the presence of dara alone or the combination of XmAb24306 and dara ( Figure 4B ). Additionally, because NK cell fratricide may reduce the effective E:T ratio, we tested whether XmAb24306 treatment can still induce robust tumor killing even at lower E:T ratios in vitro. To this end, we co cultured XmAb24306 treated and untreated NK cells with CTV labeled RPMI 8226 at various E:T ratios for 5 h and measured NK cell mediated cytotoxicity in the presence or absence of dara. At the lowest E:T ratio of 0.2:1, addition of dara drives minimal MM cell killing (5–10%). However, the combination of XmAb24306 and dara leads to 5-fold increased ADCC compared with dara alone at this E:T ratio. Equally important, the combination treatment at 0.2:1 E:T ratio induces similar levels of ADCC compared with single agent dara when tested at 5 times higher E:T ratio (1:1) ( Figure 4C ) (purple dots Vs. red diamonds). Similarly, the combination of XmAb24306 and dara at E:T ratio of 2:1 also yielded similar ADCC when compared with dara alone at more than 2 times the number of effectors (E:T ratio of 5:1) (purple dots Vs. blue triangles). Taken together, these data show that XmAb24306 can enhance the ability of NK cells to efficiently and robustly eliminate MM cell lines even under conditions of low E:T ratio which may happen due to fratricide.