NEO-201 was generated and provided by Precision Biologics, Inc., Bethesda, MD, USA (19, 20). Recombinant human IL-15 used in this study was provided by National Institutes of Health (NIH)-Cancer Therapy Evaluation Program (CTEP) as a sterile filtered solution endotoxin-free in aqueous buffer.

The following human cancer cell lines were obtained from the American Type Culture Collection (ATCC): CFPAC-1 (pancreatic); OV-90 (ovarian); HCT116 (colon); KLE CRL-1622 (endometrial).

The following human cancer cell lines were obtained from the National Cancer Institute (NCI): ACI52, ACI80, ACI98, ACI126, ACI158, ARK-1, ARK-2, EC1, EC2, EFE184 (endometrial); TC4601 (non-transformed endometrial epithelium); OVCAR-8 (ovarian).

All human carcinoma cell lines were grown and maintained in a culture medium (Corning Life Science, Manassas, VA, USA) designated by the provider supplemented with 10% USA-sourced and heat-inactivated HyClone Fetal Bovine Serum (FBS; GE Healthcare Life Sciences, Issaquah, WA, USA), 1% penicillin/streptomycin (Corning Life Science, Manassas, VA, USA) and maintained at 37 °C in incubator under 5% CO₂.

Human cancer cells used for tumor induction in mice were tested by Molecular Testing of Biological Materials (MTBM) as required by the NCI ACUC Committee and confirmed to contain no mouse viruses.

Expression of tumor antigen recognized by NEO-201 on human cancer cell lines was analyzed by flow cytometry. Human cancer cell lines (1.0 × 10⁶) were harvested and first incubated with 1 μL per test of LIVE/DEAD Fixable Aqua (Thermo Fisher Scientific, Waltham, MA, USA) in 1× phosphate-buffered saline (PBS) for 30 min at 4°C to accomplish live vs. dead cell discrimination. Cells were then centrifuged, washed twice with cold PBS, and then stained in 1× PBS + 1% BSA (Teknova, Hollister, CA, USA) for 30 min at 4°C with Pacific Blue-conjugated NEO-201 antibody (BioLegend, San Diego, CA, USA).

After staining, cells were washed twice with cold PBS and examined using BD FACSCanto II flow cytometer (BD Biosciences, San Jose, CA, USA). Analysis of cellular fluorescence was performed using BD FACSDiva software (BD Biosciences, San Jose, CA, USA). Percentage (%) of NEO-201 positive cells was calculated with the following formula: % NEO-201 stained cells (cell reactive with NEO-201 in the stained sample) - % unstained cells (% of cells with autofluorescence in the unstained sample). Cells with staining values >10% were considered positive for NEO-201 reactivity. Relative median fluorescence intensity (MFI) was calculated with the following formula: MFI NEO-201 stained cells/MFI unstained cells.

Formalin-fixed, paraffin-embedded (FFPE) sections of human tumor tissues and human endometrial cancer cell lines were analyzed for NEO-201 target antigen expression using an immunohistochemistry protocol previously described (22).

Slides were digitalized at 20× objective (0.5 × 0.5µm per pixel) using Aperio AT2 scanner (Leica Biosystems) and analyzed using HALO (Indica Labs, v3.6). The appropriateness of staining and regions of interest annotation were completed by a board-certified pathologist to include viable tumor and exclude tumor necrosis, non-tumor tissues, and histology artifact. Quantification of tumor positivity and immune cell infiltrates were performed using image analysis, which included identifying stain vectors, optimizing cell detection algorithms, and thresholding chromogenic IHC staining based on positive and negative controls.

Tissues and cancer cell lines were scored for the expression of the antigen recognized by NEO-201 and percentage of positive tumor cells.

For quantification of tumor expression of NEO-201 target antigen, tissues and cancer cell lines were binned based on staining intensity as negative, 1+ (mild staining; tumor tissues or cancer cell lines with a complete staining of the membrane in <10% of the sample analyzed), 2+ (moderate staining), 3+ (strong staining). Tissues and cancer cell lines with 2+ and 3+ staining intensity were defined as tumor tissues or cancer cell lines with a complete staining of the membrane in more than 10% of the sample analyzed.

The % of positive cells was determined for each bin and the tumor H-score was calculated as follow: H-score = [1*(%1+) + 2*(%2+) + 3*(%3+)] (28).

Tissue microarrays (TMA) of 185 endometrial tumor samples were obtained from Dr. Faceuglia P. at University of Southern California.

TMA were immunolabeled with NEO-201, CD4, and FoxP3 monoclonal antibodies using multiplex immunofluorescence. DAPI was used for fluorescent nuclear staining and Gold antifade mounting media was used with coverslips. Whole slide fluorescent imaging was performed with a Akoya Phenoimager HT scanner with a 20x objective (0.5 × 0.5µm per pixel) and analyzed using QuPath (v 0.5.1) (29). Digital image analysis included identifying stain vectors, de-arraying tissue microarrays, defining regions of interest, detecting cells, excluding artifact and normal adjacent tissues, and quantifying immunofluorescent signal. Following automated TMA grid detection, manual pathologist review was performed to exclude spots or regions of spots that were non-assessable, non-tumor containing, or obscured by artifact. Nuclei segmentation and cell detection was performed using a pretrained convolutional neural network model employing a star-convex polygon approach. Staining was quantified for detected cells and thresholds were determined for each assay to bin cells based on fluorescent intensity as follows: 0 for no staining, 1 for mild staining, 2 for moderate, and 3 for strong staining. Tumor H-scores were calculated as follows: H-score = [1*(%1+) + 2*(%2+) + 3*(%3+)] (28).

Exome samples were pooled and sequenced on HiSeq using Agilent SureSelect Human All Exon V7 and paired-end sequencing. The samples had 150M to 194M pass filter reads, with Q30 above 93%. The samples were mapped, and variants were called using DRAGEN (Dynamic Read Analysis for GENomics) pipeline provided by the CCR sequencing facility at NCI (https://crtp.ccr.cancer.gov/sf/). Percent total mapping against reference genome Human - hg38 is about 99% and Uniquely mapped reads were above 70%. Library complexity (i.e. percentage of non-duplicate reads) was determined by measuring the percentage of unique fragments in the mapped reads using MarkDuplicate utility. Percent duplicate reads are between 23% to 29%. There are 85% to 87% of reads mapped on target. Coverage statistics were also measured using DRAGEN. Raw sequencing depth coverage over the target region was between 457x and 589x and mapped sequencing depth coverage over target (after alignment and marking duplicates) was between 215x to 276x. The mean insert size for these samples was between 180 and 200 bases. More than 95% of the target region have the coverage above 20x. DRAGEN was run for germline joint genotyping. There are about 40K to 58K germline joint genotyping variants per sample, including 37K to 50K SNPs and 3K to 9K INDELs. The Ti/Tv ratio was between 2.5 and 2.7.

The ADCC activity of NEO-201 against human carcinoma cell lines was evaluated using a radioactive ADCC assay. The following human cancer cell lines were used as target cells: CFPAC-1 (pancreatic cancer cell line; positive control), OVCAR-8 (ovarian cancer cell line; negative control), OV-90 (ovarian cancer cell line), ACI158 (endometrial cancer cell line).

Human peripheral blood mononuclear cells (PBMCs) and isolated NK cells from the same healthy donor were used as effector cells.

PBMCs were collected from anonymous healthy donors under protocol 99-CC-0168, approved by the Institutional Review Board of the National Cancer Institute.

NK cells were isolated from PBMCs from healthy donors using the EasySep Human NK Cell Isolation Kit (StemCell Technologies, Vancouver, BC, Canada) according to the manufacturer’s protocol.

Before the ADCC assay, both PBMCs and NK cells were treated with recombinant human IL-15 at different concentrations (2, 1, 0.5 and 0.25 ng/mL) or vehicle control medium (RPMI-1640 medium supplemented with l-glutamine, 10% FBS, and 1% penicillin/streptomycin) for 48 hours.

On the day of the assay, target cells (2x10⁶) were pulsed with 60µCi 111In-oxyquinoline (GE Healthcare) at 37°C for 20 minutes and then 3,000 cells per well were plated in 96-well round-bottom culture plates in vehicle control medium.

Then PBMCs were added at effector cell-to-target cell (E:T) ratios of 50:1 and NK cells at E:T ratios of 12:1 and 6:1 in presence of 10 µg/mL of NEO-201 or human IgG1 isotype control antibody (Thermo Fisher Scientific, Waltham, MA, USA).

For blocking studies, on the day of the ADCC assay, purified NK cells untreated or treated with IL-15 (0.25 ng/mL) were incubated at 37°C for 2h with 15 µg/mL of anti-human CD16-neutralizing mAb (eBioscience, San Diego, CA) prior to being used as effectors at 6:1 E:T ratio.

Spontaneous release was determined by incubating target cells with vehicle control medium alone, and complete lysis by incubation with 0.05% Triton X-100.

After 4h incubation at 37°C under 5% CO₂, cells were harvested and 111In-oxyquinoline release was detected using a specific gamma counter.

Specific ADCC lysis was determined using the following equation:

In vivo experiments were performed on 6–8 weeks old female athymic nude mice. Mice were maintained on a 12h light/dark cycle, with food and water provided ad libitum.

To assess the IL-15 ability to enhance NEO-201 efficacy on survival, 1 × 10⁶ OV-90 cells were injected into the peritoneal cavity of each mouse. Tumors were allowed to grow for 2 weeks before 9 mice per group were randomized into the following groups:

NEO-201 was administered IP on days 1, 4, and 8 of treatment, while NK cells were administered IP on days 2, 5, and 9 together with IL-15. Before injection, isolated NK cells were cultured overnight in vehicle control medium. After the first treatment, mice were followed for survival for 120 days. Mice were evaluated biweekly for signs of drug-related toxicity and disease progression based on distress, physical exam changes, and cachexia. Animal care was provided in accordance with the procedures in the Guide for the Care and Use of Laboratory Animals. Experiments were carried out according to a protocol approved by the National Cancer Institute Animal Care and Use Committee.

Differences between multiple treatments in the in vitro ADCC assay were evaluated by 2way ANOVA followed by Turkey’s multiple comparison test, using GraphPad Prism 10.0.2 software (GraphPad Software, La Jolla, CA, USA). Differences were considered significant when the p value was < 0.05.

Survival curves were analyzed using the Kaplan–Meier method and compared using the log-rank (Mantel-Cox) test using GraphPad Prism 10.0.2 software. Differences in survival curves were considered significant when the p value was < 0.05.

Flow cytometry analysis was used to profile a panel of human endometrial and ovarian cancer cell lines for NEO-201 binding. The staining profile is summarized in Table 1.

The human pancreatic cancer cell line CFPAC-1 was used as positive control for NEO-201 binding, since previous studies showed that this cell line is highly positive for NEO-201 binding in flow cytometry (20, 27).

Assessment of the binding activity of NEO-201 revealed that only 2/9 (22%) of human endometrial cancer cell lines (ACI80 and ACI158) were positive for NEO-201 binding. The cell line with the strongest reactivity for NEO-201 was ACI158 (71.6% of cells positive) (Table 1).

Between the two ovarian cancer cell lines tested, OV-90 was found to be reactive with NEO-201 (30.4%), while the cell line OVCAR-8 was negative for NEO-201 binding (Table 1). Representative flow cytometry plot for each cell line is shown in Supplementary Figure 1.

NEO-201 binding profile on ovarian cancer was already evaluated by TMA in a previous manuscript, where we reported the reactivity of NEO-201 in 627 tissues from more than 10 different ovarian cancer subtypes including serous, germ cell, clear cell, endometroid, mucinous, sarcoma, transitional, and adenocarcinoma (22). In this manuscript, to confirm the data obtained in flow cytometry, NEO-201 binding was evaluated by IHC in several human endometrial cancer cell lines. A colon cancer tissue was used as positive control (Figure 1A). IHC confirmed results from flow cytometry.

Among all human endometrial cancer cell lines analyzed, only the cell line ACI158 (endometrial serous adenocarcinoma) reacted with NEO-201 in IHC (18.30% cells positive for NEO-201 staining, with H-score of 20.95). The cell line ACI80, that was NEO-201 positive in flow cytometry, did not react with NEO-201 in IHC. All other human endometrial cancer cell lines analyzed by IHC had less than 2% of cells positive for NEO-201 staining (Supplementary Table 1).

The binding of NEO-201 was also evaluated in tumor samples, using a TMA containing 185 endometrial tumor samples derived from subjects with uterine cancer. Types of uterine cancers included in the TMA were endometrioid G1 (118) and G2 (67) (Supplementary Table 2).

Many of the tissues analyzed derived from patients with FIGO (International Federation of Gynecology and Obstetrics) stage IA and IB (Supplementary Table 2).

Analysis of TMA showed that a total of 16 tissues had more than 10% of NEO-201 positive cells. In particular, 10 tissues had 10-24% of cells NEO-201 positive, 2 tissues had 25-49% of NEO-201 positive cells (Figure 1B; D4 and M4), 2 tissues had 50-74% of NEO-201 positive cells (Figure 1B; B13 and M3), and 2 tissues had 75-100% of NEO-201 positive cells (Figure 1B; E1 and E2).

A strong membranous staining by NEO-201 at the level of apical surfaces of glandular structures of endometrial tumor tissues was observed in most of these 16 tissues.

Occasionally, limited staining was observed at the level of apical surfaces of columnar epithelial cells. A positive staining was also observed within glandular lumina. This may indicate the presence of inflammatory cells infiltrating the lumina or potentially sloughed epithelial cells or membrane secretions from the endometrial glandular epithelium (Figures 1B, C).

A deeper analysis of endometrial cancer tissue represented in Figure 1B (E2) revealed also the presence of cells consistent with infiltrating immune cells in the lumina. Staining of this specific endometrial tumor tissue infiltrated with immune cells showed that NEO-201 stained only the membrane and cytoplasm of tumor cells (79.5% tumor cells positive for NEO-201), while the immune cells were primarily CD4⁺ (Figure 1C). A very small fraction of CD4⁺ cells were also FoxP3⁺ (0.02%), indicating a limited infiltration of tumor tissue by regulatory T cells (Tregs). However, CD4⁺ and CD4⁺/FoxP3⁺ cells were negative for NEO-201.

To evaluate the correlation between genetic mutations carried by gynecological tumors and their sensitivity or resistance to NEO-201, we performed whole exome and RNA sequencing on different human endometrial cancer cell lines, including cell lines positive or negative for NEO-201 binding in IHC.

We searched for mutations that were commonly present in endometrial cancer cell lines reactive with NEO-201 but not in endometrial cancer cell lines not reactive with NEO-201, or for mutations specifically expressed in a particular cancer cell line.

The mutation burden varied widely across cell lines, with the highest number of mutations in uterine serous cell line ACI126 and lowest in endometrial adenocarcinoma line EC1 (Figure 2A). The human colon cancer cell line HCT116 was used as positive control for the mutational burden since these cells exhibit microsatellite instability (MSI) and an increased rate of mutations. HCT116 showed higher mutational burden than control untransformed human endometrial epithelial cells TC4601 (low mutational count) (Figure 2A). Alterations were predominantly made up of missense mutations or frame-shift deletions. Histological subtype of origin (serous or non-serous) did not correlate with mutational burden (Figure 2B).

Focusing down on specific mutations commonly found in this dataset, the mucin-related genes MUC5AC and MUC3A universally carried missense mutations in all the cell lines. Of note, the genes ORC4C45 (nonsense mutation) and FOLR3 (frame shift deletion) appeared to be commonly inactivated in these cell lines, suggesting the possibility of tumor-suppressor functions (Figure 2B).

Several genes were more closely examined based on their known association with colon and endometrial cancers and for possible combination strategies (Figure 2C). We evaluated whether any of these were found in the ACI158 cell line that was highly reactive with NEO-201. Interestingly, we observed that the ACI158 cell line (only cell line reactive with NEO-201 in IHC) and EC1 cell line (not reactive with NEO-201 in IHC) do not carry any missense mutation, gene amplification, frame shift deletion or insertion in this set of genes considered as variants of interest. On the contrary, in all the other endometrial cancer cell lines that do not react with NEO-201 in IHC we observed the following mutations in variants of interest: amplification of HER2 gene in ARK-1 and ARK-2 cells; mutations of Mismatch Repair (MMR) genes (MLH1, MSH2, MSH3, MSH6, PMS2) in EC2, ACI80, ACI98, ACI126 cell lines; mutation of KRAS and AKT1 gene in ARK-2 cell line; mutation of mitogen-activated protein kinase (MAPK) genes (MAP2K2, MAP2K5) and CTNNB1 gene in the ACI98 cell line (Figure 2C).

The ability of NEO-201 to mediate ADCC against human cancer cell lines expressing its target antigen has been demonstrated in previous studies (20, 21, 27).

In this study we evaluated the ability of NEO-201 to mediate ADCC in gynecological cancers using the following cell lines tested for the expression of NEO-201 target antigen in flow cytometry: CFPAC-1 (97.9% of cells expressing NEO-201 target antigen): positive control; OVCAR-8 (8.2% of cells expressing NEO-201 target antigen): negative control; OV-90 (30.4% cells expressing NEO-201 target antigen) and ACI158 (71.6% cells expressing NEO-201 target antigen).

Before the ADCC assay, both PBMCs and NK cells, isolated from PBMCs from same healthy donors, were treated with recombinant human IL-15 at different concentrations (2, 1, 0.5 and 0.25 ng/mL) for 48 hours. These experiments show that IL-15 enhanced statistically NEO-201-mediated ADCC against CFPAC-1 at 1ng/mL (p<0.01), OV-90 at 2 ng/mL (p<0.01), and ACI158 at 2 and 1ng/mL (p<0.01) with PBMCs as effectors (Supplementary Figure 2).

When NK cells were used as effectors, the best enhancement of NEO-201-mediated ADCC was achieved with IL-15 at 0.25 ng/mL in CFPAC-1 and ACI158 cell lines with NK cells at12:1 E:T ratio (Supplementary Figure 2). For this reason, we chose the concentration of 0.25ng/mL for future experiments.

To evaluate the ability of IL-15 to enhance ADCC mediated by NEO-201 with both PBMCs and NK cells as effectors in multiple healthy donors, we treated PBMCs and NK cells from three different healthy donors without and with IL-15 at the concentration of 0.25ng/mL for 48 hours before using them in the ADCC assay.

Results from experiments in Figure 3 (mean of 3 experiments performed with three different healthy donors) show that NEO-201 mediates the killing of CFPAC-1, OV-90 and ACI158 cancer cell lines via ADCC with both PBMCs and NK cells as effectors. Results also confirmed that OVCAR-8 is an optimal negative control to test the specificity of the ADCC mediated by NEO-201. No difference in tumor lysis between NEO-201 and IgG isotype control was observed with both PBMCs and NK cells as effector cells with or without treatment with IL-15. This shows that NEO-201 does not mediate ADCC against cancer cells that do not express its target antigen (Figure 3B).

In cancer cell lines CFPAC-1 and ACI158 the combination of IL-15 + NEO-201 resulted in a statistically significant higher lysis compared to IL-15 alone with both PBMCs and NK cells as effectors (Figures 3A, C).

In the cell line ACI158, the mean of percentage of specific lysis from three different healthy donors showed that IL-15 enhanced ADCC mediated by NEO-201 in a statistically significant manner only when NK cells were used as effector cells at 12:1 E:T ratio (Figure 3C).

On the other hand, although in both CPFAC-1 and OV-90 cell lines we observed that IL-15 enhanced ADCC mediated by NEO-201 with NK at 12:1 ratio in a statistically significant manner in one healthy donor, this enhancement was not statistically significant when we calculated the mean of results from 3 different healthy donors (Figures 3A, D). This phenomenon is due to the variability in percentage of lysis between different donors.

Altogether, these data suggest that IL-15 can enhance ADCC mediated by NEO-201 against gynecological cancers expressing NEO-201 target antigen. Furthermore, since the percentage of lysis via ADCC mediated by NK cells is similar to that mediated by PBMCs in CFPAC-1, OV-90 and ACI-158 cancer cell lines, we decided to use NK cells as effector cells in the in vivo experiments.

This observation confirms results of a previous study, which demonstrated that NEO-201 utilizes NK cells as one of the main effectors to mediate ADCC (27).

In a previous study, we demonstrated that treating NK cells with an anti-CD16 antibody, prior to using them in the ADCC assay, significantly decreased ADCC activity mediated by NEO-201 against the CFPAC-1 cell line when NK cells alone or NK cells boosted with the IL-15 superagonist were used as effector cells (27).

In this study, to confirm that tumor cells lysis of endometrial (ACI158) and ovarian (OV-90) cancer cell lines is mediated by NEO-201 via ADCC and that IL-15 is able to enhance NEO-201-mediated ADCC, we used NK cells isolated from three healthy donors at 6:1 E:T ratio as effectors cells. NK cells were untreated or treated with IL-15 (0.25 ng/mL) for 48h prior to being used in the ADCC assays. The day of the ADCC assay, the anti-CD16 antibody was added to NK cells 2 hours prior to using them as effector cells. As shown in Figure 4, the anti-CD16 antibody significantly decreased ADCC mediated by NEO-201 in both untreated and IL-15-treated NK cells (p<0.001) in all cell lines, confirming that NEO-201 mediates ADCC to kill tumor cells expressing its target antigen and that IL-15 enhances specifically NEO-201-mediated ADCC.

To assess the ability of IL-15 to enhance NEO-201 efficacy on overall survival in mice, we used a murine model previously tested for survival in vivo studies (22). Briefly, cancer cells from the OV-90 cancer cell line were inoculated into the peritoneal cavity of 6–8 weeks old female athymic nude mice to reproduce the condition of disseminated ovarian cancer and peritoneal carcinomatosis. Tumor-bearing mice were treated with NK cells + NEO-201 with or without IL-15.To have the best efficacy as effector for the ADCC mediated by NEO-201 in vivo, we used isolated NK cells from a healthy donor with a good ADCC in vitro against OV-90 cells (Supplementary Figure 3).

After treatments described in materials and methods, tumor-bearing mice were followed for survival for 120 days. Mice treated with NK cells + NEO-201 + IL-15 experienced the longest median survival compared to all other groups (Figure 5, Supplementary Table 3). Although the risk of dying from cancer in mice treated with NK cells + NEO-201 + IL-15 was 2.92 times lower compared to mice treated with NK cells + NEO-201, the difference in median survival between these two groups was not statistically significant (Figure 5, Supplementary Table 3).

Instead, mice treated with NK cells + NEO-201 + IL-15 showed a statistically significant prolonged median survival compared to mice treated with NK cells + IL-15 (p=0.0122; 8.34 times lower risk of developing tumors) (Figure 5, Supplementary Table 3).

It is also important to note that mice treated with NK cells + NEO-201 did not show a statistically significant prolonged median survival compared to mice treated with NK cells + IL-15 (Figure 5, Supplementary Table 3).