Peritoneal natural killer cells from epithelial ovarian cancer patients show an altered phenotype and bind to the tumour marker MUC16 (CA125)

Abstract

The ovarian tumour marker MUC16 (CA125) inhibits the cytotoxic responses of human natural killer (NK) cells and down-regulates CD16. Here we show that approximately 10% of the peripheral blood NK cells (PBNK) from the epithelial ovarian cancer (EOC) patients are CD16(-) CD56(br) whereas 40% of the peritoneal fluid NK (PFNK) carry this phenotype, which is usually associated with NK cells from the lymph nodes or human decidua. PBNK from healthy donors exposed to PF show a significant increase in the CD16(-) CD56(br) population. This shift in phenotype is not caused by increased apoptosis of the CD16(+) CD56(dim) cells or selective proliferation of the CD16(-) CD56(br) NK cells. Thus, the terminal differentiation of the CD16(-) CD56(br) NK cells to CD16(+) CD56(dim) subset that occurs during normal NK cell development may actually be a reversible step. A majority of the NK cell receptors (NKp46, NKp44, NKG2D, CD244, CD226, CD158a, CD158b, and CD158e) studied were down-regulated in the PFNK. MUC16 binds selectively to 30-40% of CD16(+) CD56(dim) NK cells in EOC patients indicating that phenotypic alterations in these cells are mediated by tumour-derived soluble factors. Similar to EOC, MUC16 in early pregnancy also binds to NK cells suggesting shared mechanisms of NK cell suppression in feto-maternal tolerance and immune evasion by ovarian cancers.

Figure 1

Distribution of CD16⁺ CD56^dim and CD16^– CD56^br NK subsets in the PB and PF of EOC patients. (a) The gating of the CD16^– CD56^br (bold box) and CD16⁺ CD56^dim (thin box) for PT25 are shown. Similar gates were used for all samples in this study. (b) The numbers of CD16⁺ CD56^dim and CD16^– CD56^br as a percentage of total NK cells are shown for all eight HD PBNK and nine EOC PB and PFNK samples. The data point for each HD or EOC sample represents a mean of eight independent measurements. (c) Relative expression of CD16 on the HD and EOC NK cells is shown. Statistical significance (P < 0·05, by Wilcoxon Sign Ranked test) is represented by matched symbols.

Figure 2

Relative expression of NK cell receptors on the PB and PFNK cells of EOC patients. Expression of NKp46 (a), NKp44 (b), NKG2D (c), CD94/NKG2A (d), CD244 (e), CD226 (f), CD158a (g), CD158b (h), and CD158e (i) on the CD16⁺ CD56^dim (open bars) and CD16^– CD56^br NK (dark bars) cell subsets are indicated. Each bar shown is an average reading obtained from eight HD or PB and PF samples from nine EOC patients. Complete statistical analysis is provided in Tables S3–+S11.

Figure 3

Phenotypic changes in HD derived PBNK following treatment with EOC PF. (a) PBMC were treated with PF from EOC patients (test) or incubated in media only (control) for 72 h in triplicate. Effect of PF from PT45 was tested on PBMC from HD4, HD5, and HD30. PF from PT15, PT17, and PT30 was tested on PBMC from HD4, HD24, and HD31. Flow cytometry was performed and the numbers of CD16^– CD56^br cells as a percentage of total NK cells were determined. Each bar represents the average of the triplicate data obtained for the three HD PBMC samples (in effect, n = 9). (b) Relative expression of CD16 on the CD16⁺ CD56^dim HD PBNK cells that were treated or untreated with PF is indicated. Each bar is a mean of three independent measurements. (c) Average expression of NKp46 on the CD16⁺ CD56^dim and CD16^– CD56^br HD PBNK subsets following treatment with PF or media alone is show. The data represents the mean of triplicate measurements for each PF sample in three independent assays. APC, antigen-presenting cell.

Figure 4

Treatment with PF does not result in increased proliferation of NK cells. (a) Proliferation of entire population of HD4 derived PBMC incubated in PF of EOC PT45 for 72 hr is shown. Comparable results were obtained in all HD PBMC samples tested in this experiment. (b) PBMC from HD4 were treated for 24, 48, and 72 hr with media only (control) or PF (test) from EOC patient PT45. Proliferation of the CD16⁺ CD56^dim (shaded histogram) and the CD16^– CD56^br (clear histogram) NK cell subsets was determined at the designated time intervals. (c) Combined proliferation at the 24, 48, and 72 hr time points from duplicate experiments of PBNK from HD4 and HD31 in EOC PT45 PF. Similar results were obtained when PBNK from HD4 and HD31 were incubated with PF from EOC PT30.

Figure 5

Treatment with PF does not result in increased apoptosis of the NK cells. PBMC from HD4, HD29, and HD31 were incubated with media only (control) or PF (test) from EOC patients PT30 and PT45 for 72 hr. The average of the percentage of viable cells for all three HD incubated with media or PF samples is shown.

Figure 6

MUC16 is present on PBL and PFL of EOC patients. (a) MUC16 on the PBL (solid line) and PFL (dotted line) of three EOC patients is shown. All nine patients studied showed similar flow cytometry patterns. The anti-MUC16 VK-8 antibody and a FITC-labelled goat anti-mouse secondary were used for detection. (b) MUC16 on the PBL was detected by using both the VK-8 and the OC125 antibodies as indicated. (c) The VK-8 (bold line) and OC125 (thin line) did not detect any appreciable amounts of MUC16 on PBL from HD. Data from HD4 are shown, although tests on PBL samples from an additional three donors gave identical results. The shaded histogram in all panels shows binding of isotype-matched control for both VK-8 and OC125.

Figure 7

MUC16 is not endogenously expressed on PBMC but binds to the cells. (a) MUC16 expression in PBMC from HD4, HD22, and HD25, Lanes 2, 3, and 4, respectively, was determined by RT-PCR using primer set 1. RT–PCR product from OVCAR-3 mRNA was used as control, lane 1. (b) MUC16 expression on PBMC from PT6 and PT8, lanes 2 and 3, respectively, detected by using primer set 1. MUC16 expression in OVCAR-3 cells is in lane 1. (c) Primer set 2 was used to determine MUC16 expression in OVCAR-3 cells and in PB from PT22, and PT24, lanes 1–3, respectively. (d) PBMC from HD1, HD3, and HD4 treated with PF from PT36 for 72 hr were analysed by flow cytometry for MUC16. Control PBMC from the HD were incubated in media only. Average of the data from all three HD is shown. This data is representative of five separate experiments, each involving incubation of PBMC from three HD with PF from different EOC patients.

Figure 8

Restricted binding of MUC16 to NK cell subsets. (a) MUC16 is present on a significant population of the PB and PF NK cells. Data shown is the average from seven HD and nine EOC patients. (b) Distribution of MUC16 on the CD16⁺ CD56^dim and the CD16^– CD56^br NK cell subsets from the PB and PF of nine EOC patients is shown. Statistical significance (P < 0·05, by Wilcoxon Sign Ranked test) is represented by matched symbols.

Figure 9

MUC16 is present on PBNK cells of pregnant women. PBMC samples from three HD were analysed for MUC16 before they conceived and then subsequently when they were 9 weeks into their pregnancy. The number of MUC16⁺ NK cells as a percent of total PBNK is shown.