Serum selenium, selenoprotein P and glutathione peroxidase 3 as predictors of mortality and recurrence following breast cancer diagnosis: A multicentre cohort study

Abstract

The trace element selenium is of essential importance for the synthesis of a set of redox active proteins. We investigated three complementary serum selenium status biomarkers in relation to overall survival and recurrence following diagnosis of primary invasive breast cancer in a large prospective cohort study. The Sweden Cancerome Analysis Network - Breast Initiative (SCAN-B) is a prospective population-based study including multiple participating hospitals. Main analyses included 1996 patients with a new diagnosis of primary invasive breast cancer, with blood sampling at the time of diagnosis. In sera of the patients, total serum selenium, selenoprotein P (SELENOP), and glutathione peroxidase 3 (GPx3) activity was analysed. All three biomarkers showed a positive correlation (p < 0.001), supporting the high quality of samples and analytical techniques. During a total of 13,306 person years of follow-up, 310 deaths and 167 recurrent breast cancer events occurred. In fully adjusted Cox models, all three biomarkers correlated inversely with mortality (p trend <0.001) and compared with the lowest quintile, hazard ratios (95% confidence interval) for overall survival in the highest quintile of selenium, SELENOP and GPx3 were 0.42 (0.28-0.63), 0.51 (0.36-0.73) and 0.52 (0.36-0.75), respectively. Low GPx3 activity was associated with more recurrences (Q5 vs Q1: fully adjusted HR (95%CI); 0.57 (0.35-0.92), (p trend = 0.005). Patients with low selenium status according to all three biomarkers (triple deficient) had the highest mortality risk with an overall survival probability of ∼50% after 8 years, in particular as compared to those having at least one marker in the highest quintile; fully adjusted HR (95%CI); 0.30 (0.21-0.43). Prediction of mortality based on all three biomarkers outperformed established tumour characteristics like histologic grade, number of involved lymph nodes or tumour size. An assessment of Se status at breast cancer diagnosis identifies patients at exceptionally high risk for a poor prognosis.

Keywords: Biomarkers; Micronutrient; Mortality; Prognostic factors; Prospective study.

Conflict of interest statement

LS holds shares of selenOmed GmbH, a company involved in Se status assessment and supplementation; no other relationships or activities that could appear to have influenced the submitted work.

Copyright © 2021 The Authors. Published by Elsevier B.V. All rights reserved.

Figures

Graphical abstract

Fig. 1

Flow chart explaining inclusion and exclusion criteria.

Fig. 2

Spearman's correlation analysis in patients with primary invasive breast cancer. In the total cohort, all three biomarkers of Se show significant correlations, p<0.001. (A) Total serum Se and SELENOP display a tight correlation, R = 0.604. (B) Total serum Se correlates with the GPx3 activity in sera of the patients, R = 0.294. (C) Serum SELENOP and GPx3 show a similarly strong correlation as total Se and GPx3, with R = 0.279. N (A,B) = 1993, 3 data points missing in the figures, n(C) = 1996, Spearman's R, two-tailed.

Fig. 3

Spearman's correlation of biomarkers in patients who display all biomarkers in the first or last quintile. The Venn diagram and the plots in the top section of figure (A,B,C) highlight the significant correlations between the biomarkers in samples of the 121 patients who are assigned to the first quintile regarding each of the three biomarkers. The correlations are non-significant for patients in the highest quintile regarding all three biomarkers, as shown in the lower Venn diagram and the plots in the lower panel of figure (D,E,F). This comparison underlines the saturation of selenoprotein expression under high Se status.

Fig. 4

Kaplan Meier curves for overall survival (A,B,C), and for recurrence free survival (D,E,F), by quintiles of the three biomarkers. Log-Rank-Test was used to evaluate differences. Y-axis-limits were set between 0.4 and 1.0 for better visualization, no points are missing. P-value was calculated with log-rank test.

Fig. 5

Overall survival (A) and recurrence free survival (B) assessed for the triple deficient group. Triple deficient stands for patients who are in the first quintile regarding all three biomarkers; ∼ (Se < 57 μg/l and SELENOP <3 mg/l and GPx3 <167 U/l). Purple curve stands for patients who are in the fifth quintile regarding at least one biomarker ∼ (Se > 84 μg/l or SELENOP >5 mg/l or GPx3 >248 U/l). Q = Quintile, HR=Hazard ratio, CI=Confidence interval, Fully adj. = Fully adjusted for confounders of breast cancer mortality as listed in Table 3, Table 4 (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

Predictive value of Se status for mortality. The first panel (A) compares the predictors individually, (B) compares the predictors in combined models. AUCt (y-axis) was computed at each time of death, marked with the symbol ○. An AUC of 0.5, depicted by the dotted line, represents a random predictor without any value, while an AUC of 1.0 is a prediction model with 100% specificity and 100% sensitivity. iAUC = Integrated area under the curve, Comp. Se = Composite Se status.