Preventative Cancer Vaccine-Elicited Human Anti-MUC1 Antibodies Have Multiple Effector Functions

Abstract

Background/objectives: Mucin-1 (MUC1) is a transmembrane glycoprotein that is overexpressed and hypoglycosylated in premalignant and malignant epithelial cells compared to normal cells, creating a target antigen for humoral and cellular immunity. Healthy individuals with a history of advanced colonic adenomas and at high risk for colon cancer were enrolled in a clinical trial to evaluate the feasibility of using a MUC1 peptide vaccine to prevent colon cancer. Anti-MUC1 antibodies elicited by this vaccine were cloned using peripheral blood B cells and sera collected two weeks after a one-year booster. Twelve of these fully human monoclonal antibodies (mAb) were tested for binding to MUC1+ target cells, and three with the highest binding were further evaluated for various effector functions important for tumor rejection.

Methods: Immune cells were incubated together with target cells expressing variations in the number, distance, and membrane anchoring properties of the MUC1 epitope in the presence of each mAb.

Results: All three mAbs mediated antibody-dependent cytokine release (ADCR), antibody-dependent cellular cytotoxicity (ADCC), and antibody-dependent cellular phagocytosis (ADCP). Two also mediated antibody-dependent trogocytosis/trogoptosis (ADCT). None were capable of complement-dependent cytotoxicity (CDC).

Conclusions: ADCP and ADCT functions were more efficient when antibodies bound epitopes proximal to and anchored to the membrane, providing insight for future therapeutic antibody validation strategies.

Keywords: NK cell; O-glycosylation; epitope properties; monocyte; mucin-1; neutrophil; phagocytosis; trogocytosis; tumor; vaccine.

Figure 1

Binding of human anti-MUC1 mAbs and control antibodies on target cells. Surface staining with indicated antibodies and goat anti-human IgG secondaries on Jurkat MUC1 22TR (A) or Raji MUC1 22TR (B). As cytometer voltages were set independently for each experiment, the geometric mean fluorescence intensity (MFI) of each primary-antibody containing a sample was normalized to a secondary-only negative control. Each dot represents an individual staining (n = 4–25, Jurkat MUC1 22TR utilizing either PGK or EF1α promoters to drive MUC1 gene expression; n = 5–13 Raji MUC1 22TR); bars depict mean ± SEM. Comparisons made by Kruskal-Wallis test with Dunn’s correction for multiple testing. * p < 0.05, ** p < 0.01, **** p < 0.0001.

Figure 2

Human anti-MUC1 mAbs mediate antibody-dependent cytokine release (ADCR). Healthy donor peripheral blood mononuclear cells (PBMC) were incubated for four hours with the Jurkat MUC1 22TR (A) and Raji MUC1 22TR (B) cells at an effector to target ratio (E:T) of 25:1 and individual mAbs (10 µg/mL). Thirteen cytokines in the LEGENDplex Human Inflammation panel 1 were assayed from the supernatants of the co-incubations in two independent experiments. Cytokines measured above the limit of detection are shown. The gMFI of each cytokine was converted to absolute picogram quantities, and each was normalized to the isotype control Herceptin condition. All eight cytokines were simultaneously compared by antibody condition in a Kruskal-Wallis rank-sum test with Dunn’s correction for multiple comparisons; * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 3

Human anti-MUC1 mAbs mediate antibody-dependent cellular cytotoxicity (ADCC). Jurkat MUC1 22TR (A) and Raji MUC1 22TR (B) targets were incubated with and without human NK cells from a healthy donor stimulated overnight with IL-2 at E:T ratios from 8:1 to 20:1 and the indicated human IgG antibodies. Each percentage of mAb-dependent cell death was calculated by taking the percentage of non-viable cells and subtracting non-antibody-mediated cell death that occurred in control wells without antibodies. Each dot plotted is the average of duplicates from n = 3–12 Jurkat MUC1 22TR or n = 2–3 Raji MUC1 22TR independent experiments. Bars depict the mean ± SEM. Data were compared by two-way ANOVA with Šídák’s multiple comparisons test; * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Figure 4

The ability of human anti-MUC1 mAbs to mediate antibody-dependent cellular phagocytosis (ADCP) is target-dependent. CellTrace Yellow-labeled Jurkat MUC1 22TR (A) and Raji MUC1 22TR (B) target cells were coincubated 1:1 with CellTrace Violet-labeled human THP-1 monocytes. %ADCP was calculated by taking the percentage of THP-1 cells that were also CellTrace Yellow+ and subtracting the background of double positive events that occurred with no primary mAbs. Plotted for each dot is the average of duplicates from n = 4–7 Jurkat MUC1 22TR or n = 11–14 Raji MUC1 22TR independent experiments. Bars depict the mean ± SEM. Statistical comparisons were made through one-way ANOVA with Dunnett’s test to correct for multiple comparisons, **** p < 0.0001.

Figure 5

Human anti-MUC1 mAbs mediate antibody-dependent cellular trogocytosis. Jurkat MUC1 22TR^hi (A) and Raji MUC1 22TR (B) target cells labeled with calcein-red and DiO and coincubated with CellTrace Violet-labeled healthy donor neutrophils stimulated overnight with IFNγ and G-CSF, at an E:T ratio of 5:1 for 4 h with and without antibodies. Antibody-dependent cellular trogocytosis was measured by taking the %DiO+Calcein−CellTrace+ neutrophils and subtracting the same percentage in wells with no primary antibodies, representing the background trogocytosis. Each dot shows the average of duplicates from n = 3 independent experiments. Bars depict the mean ± SEM. Statistical comparisons were made through one-way ANOVA with Dunnett’s test to correct for multiple comparisons; *** p < 0.001.

Figure 6

Human anti-MUC1 mAbs do not mediate complement-dependent cytotoxicity (CDC). Jurkat MUC1 22TR (A) and Raji MUC1 22TR (B) target cells were incubated for 30–120 min with human mAbs (10 µg/mL) and 15% normal human serum. Jurkat CD20 and parental Raji cells were used as additional controls (A). %CDC was calculated by taking the percentage of non-viable cells and subtracting the background cell death that occurred with no primary mAbs. Each dot shows a single value or average of duplicates from n = 2–4 independent experiments. Bars depict the mean ± SEM. Statistical comparisons were made through two-way (A) and one-way (B) ANOVA with Dunnett’s test to correct for multiple comparisons; **** p < 0.0001.

Figure 7

MUC1 constructs that vary in epitope number, proximity to membrane, and retention on cell surface. (A). 22TR: MUC1 with 22 tandem repeats of 20 amino acids from the VNTR region. It can be cleaved in the SEA domain into alpha and beta subunits, which remain non-covalently associated. 2TR: MUC1 construct with 2 tandem repeats of 20 amino acids from the VNTR region. m1: MUC1 with 2 tandem repeats, no cleavable domain or intracellular cytoplasmic domain, and an mCherry reporter and an N-terminal myc tag. m2: MUC1 with two tandem repeats, no cleavable domain, the MUC1 cytoplasmic domain, and an mCherry reporter and an N-terminal myc tag. N = amino-terminus, C = Carboxy-terminus, IR = imperfect repeats, VNTR = variable number tandem repeats, SEA = sperm protein, enterokinase, agrin domain TM = transmembrane. (B). The level of mCherry expression representing overall MUC1 variant expression of m1 and m2 in lentivirally transduced Jurkat cells as measured by flow cytometry. Numbers represent the geometric MFI for each population of cells. Parental Jurkats are untransduced and shown as a negative control. (C,D). Geometric mean fluorescent intensity (gMFI) of anti-CD227 (C) or anti-myc antibody staining (D) on the surface of intact cells (Surface) or combined surface and intracellular staining (Surface + IC). Relative MFI is the gMFI normalized to the gMFI background on unstained cells. Each dot represents one of two independent experiments. Each bar is the mean, and error bars depict ± SEM. The relative antibody binding of CD227 and myc-tag between m1 and m2 cells was compared by student’s t-tests with Holm-Sidak’s correction method for multiple comparisons, * p < 0.05, ns = not significant.

Figure 8

Amount of antibody bound does not correlate with efficiency of ADCC or ADCP but does with ADCT. Cells were stained with 10 µg/mL of each primary antibody for 15 min prior to (A) 4 h incubation with human NK cells that had been stimulated overnight with IL-2 at effector to target ratios ranging from 10:1 to 20:1, (B) 1 h incubation with human THP-1 cells at a 1:1 ratio, or (C) 4 h incubation with human neutrophils stimulated overnight with hIFNγ and G-CSF. Jurkat cells transduced with human CD20 were used with rituximab as positive controls. %ADCC, ADCP, and trogocytosis were calculated as in Figure 3, Figure 4 and Figure 5. Gray error bars represent mean ± SEM for average relative MFI and % antibody-mediated function. There is no significant correlation between the amount of antibody bound (average relative MFI) and average %ADCC (r² = 0.07, ns 10:1; r² = 0.05, ns 20:1) or %ADCP (r² = 0.06, ns), but there is for trogocytosis (r² = 0.81, p < 0.0001).