Systemic dysregulation of CEACAM1 in melanoma patients

Abstract

It was previously shown that CEACAM1 on melanoma cells strongly predicts poor outcome. Here, we show a statistically significant increase of serum CEACAM1 in 64 active melanoma patients, as compared to 48 patients with no evidence of disease and 37 healthy donors. Among active patients, higher serum CEACAM1 correlated with LDH values and with decreased survival. Multivariate analysis with neutralization of LDH showed that increased serum CEACAM1 carries a hazard ratio of 2.40. In vitro, soluble CEACAM1 was derived from CEACAM1(+), but neither from CEACAM1(-) melanoma cells nor from CEACAM1(+) lymphocytes, and directly correlated with the number of CEACAM1(+) melanoma cells. Production of soluble CEACAM1 depended on intact de novo protein synthesis and secretion machineries, but not on metalloproteinase function. An unusually high percentage of CEACAM1(+) circulating NK and T lymphocytes was demonstrated in melanoma patients. CEACAM1 inhibited killing activity in functional assays. CEACAM1 expression could not be induced on lymphocytes by serum from patients with high CEACAM1 expression. Further, expression of other NK receptors was impaired, which collectively indicate on a general abnormality. In conclusion, the systemic dysregulation of CEACAM1 in melanoma patients further denotes the role of CEACAM1 in melanoma and may provide a basis for new tumor monitoring and prognostic platforms.

Fig. 1

Establishment of soluble CEACAM1 quantification method. Quantification of increasing concentrations of CEACAM1 (open diamonds) or CEACAM5 (filled squares) standard proteins using NC8 mAb (a) or Kat4c mAb (b). Y-axis denotes the optical density at wavelength of 405 nm. This figure shows a representative experiment out of five performed. c Detection of recombinant N domain of CEACAM1 compared to the full protein. Y-axis denotes the optical density at wavelength of 405 nm. A representative experiment is shown

Fig. 2

Serum CEACAM1 in melanoma patients. a Serum samples obtained either from healthy donors (N = 37), patients with no evidence of disease (NED) (N = 48), or with evidence of disease (WED) (N = 64) were analyzed for soluble CEACAM1 concentration. This figure shows the determined serum CEACAM1 concentration in each individual (black dot). Each serum sample has been independently analyzed twice in triplicates. Group comparison (lines) and statistical significance using non-parametric two-tailed t test (asterisks) are indicated in the figure. *P value <0.05 and ***P value <0.001. b The serum CEACAM1 of patients with evidence of disease categorized into two subgroups: patients who died of disease (DOD) during follow-up and patients who remained alive with disease (AWD). Horizontal lines indicate the median value of the group. P value in non-parametric two sided t test is indicated in each plot. This figure shows the mean results of two independent quantifications in triplicates of the same samples. c Serum CEACAM1 concentration is plotted against time-to-death in months for each patient who died of disease. Correlation was calculated with Spearman’s correlation (r index is indicated in the figure). d Kaplan–Meier plot demonstrating the effect of CEACAM1 on survival of WED patients after neutralization of LDH. P value of Cox regression was 0.06

Fig. 3

No difference in serum levels of four immunomodulating factors between healthy donors and melanoma patients. This figure shows scatter distribution of all samples. There are three main groups of samples, healthy donors (filled squares), no evidence of disease (NED) patients (filled upright triangles) and patients with evidence of disease (ED) (filled inverse triangles). Each individual shape represents a single sample from the same group. Each of the four panels represents the results for a given immunomodulating factor, as indicated in the figure. Y-axis denotes quantified concentration. Horizontal lines indicate the median value of the group. P values in Kruskal–Wallis test are indicated in each plot. This figure shows the mean results of two independent quantifications in triplicates of the same samples

Fig. 4

Soluble CEACAM1 originates in vitro from CEACAM1-positive melanoma cells. a One hundred thousand cells from each indicated type were seeded per well in six-well plates for 3 days. Soluble CEACAM1 was quantified in the conditioned medium, cells were harvested, counted and stained for CEACAM1 expression. Black bars represent amount of CEACAM1 normalized per 10⁵ melanoma cells (primary Y-axis). Open diamonds represent MFI of CEACAM1 staining (secondary Y-axis). b Plotting the number of seeded melanoma cells or lymphocytes (X-axis) against quantified soluble CEACAM1 in the conditioned medium (Y-axis). This figure shows a representative experiment out of three performed. c CEACAM1 surface expression by lymphocyte cultures. Y-axis denotes the median fluorescence intensity (MFI). This figure shows representative staining values

Fig. 5

Production of soluble CEACAM1 by melanoma cells. a Melanoma cells were incubated for 3 days in the presence of 5 μM BB94 (metalloproteinase inhibitor with broad specificity) or DMSO in equivalent dilution (control). Serum CEACAM1 was quantified in the supernatant. This figure shows the average of three performed experiments. b Melanoma cells were incubated for 1 day in the presence of 1 μg/ml or the absence of cycloheximide. Soluble CEACAM1 was quantified in the supernatant and corrected for the ratio of harvested cells between the different treatment groups. This figure shows the average of three performed experiments. c Melanoma cells were incubated for overnight period in the presence or absence of Monensin, an intracellular vesicular transport inhibitor. Soluble CEACAM1 was quantified in the supernatant (c) and cells were concomitantly stained intracellularly for CEACAM1 (d). An average of three performed experiments is shown in c and a representative histogram is shown in d

Fig. 6

Unusually high expression of CEACAM1 on circulating lymphocytes among melanoma patients. This figure shows scatter distribution of CEACAM1 expression profile on circulating NK cells (a) or T cells (b). There are three main groups of samples, healthy donors (filled squares), patients with no evidence of disease (NED, filled upright triangles) and patients with evidence of disease (WED, filled inverse triangles). Each individual shape represents a single sample from the same group. Y-axis denotes the percent of CEACAM1-positive circulating lymphocytes. c, d WED patients were further categorized into patients who died of disease (DOD) during follow-up and patients who remained alive with disease (AWD). This figure shows CEACAM1 scatter distribution on circulating NK cells (c) or T cells (d) in these subgroups. Horizontal lines indicate the median value of the group. Non-parametric two sided t test was used to compare between different groups, as indicated in each plot. *P value <0.05, **P value <0.01 and ***P value <0.001

Fig. 7

Enhanced CEACAM1 expression is functional and inhibits NK killing activity. a Plots show the CEACAM1 expression profile on gated peripheral blood NK cells. Samples were derived either from healthy donors or melanoma patients, as indicated in the figure. b Peripheral blood lymphocytes were tested for natural killing activity against NK-sensitive 221 cells. Target cells were either mock transfected (221/Mock, black bars) or stable transfected with CEACAM1 cDNA (221/CEACAM1, gray bars). Effector-to-target ratio was 50-to-1. Y-axis denotes the percent of specific lysis of target cells. This figure shows a representative experiment out of three performed. *P value <0.05. c Peripheral blood lymphocytes derived from a healthy donor were cultured either in culture medium or in serum. Serum was derived either from an allogeneic healthy donor or from melanoma patients with either low or high percentage of CEACAM1-positive lymphocytes, as indicated in the figure. Peripheral blood lymphocytes from four different donors were tested, each in three different sera samples from each category. This figure shows the staining results of gated lymphocytes of a representative experiment

Fig. 8

Dysregulated expression of NK activating receptors by circulating NK cells. This figure shows scatter distribution of various NK activating receptors (indicated in the figure) expression profile on circulating NK cells. There are three main groups of samples, healthy donors (filled squares), patients with no evidence of disease (NED, filled upright triangles) and patients with evidence of disease (WED, filled inverse triangles). Each individual shape represents a single sample from the same group. Y-axis denotes the percent of receptor-positive circulating NK cells. Horizontal lines indicate the median value of the group, which is indicated numerically below. Non-parametric two sided t test was used to compare between different groups, as indicated in each plot. *P value <0.05, **P value <0.01