Selenium is an essential trace element for most organisms, including humans, and plays a key role in multiple biological processes by selenium-containing proteins (Se-proteins) [[1], [2], [3]], which are encoded by a total of 25 genes [4]. Glutathione peroxidase, selenoprotein P (SeP) and selenoalbumin (SeAlb) account for most of selenium species in human plasma [5,6]. Glutathione peroxidase isoform 3 is the only isoform that has been reported in plasma so far [[7], [8], [9]]. Since chromatographic methods cannot differentiate isoforms, we call GPx serum determinations in this study just serum GPx for simplicity. Nonetheless, it is likely, that the reported serum GPx levels are mostly reflecting GPx3. GPx and SeP, which incorporate selenium as selenocysteinyl residues (SeCys) during polypeptide chain biosynthesis, account for about 15–20 % and more than 50 % of total serum selenium, respectively [10,11]. The GPx family of proteins plays an important role in the redox homeostasis, but also participates in tumor suppression, ferroptosis balance and other pathways [12]. SeP main function is the transport of selenium from the liver to peripheral tissues [10], but might also have antioxidant properties and has been related to type 2 diabetes [13] and Alzheimer [14]. Selenoalbumin (SeAlb) contains non-specifically bound free inorganic selenium and makes up < 1–35 % of total serum selenium [15]. Despite there is no known direct biological function, SeAlb allows circulating inorganic selenium to be stored [16] and transported to the liver for the synthesis of SeP. Free selenomethionine (SeMet), a methionine-like amino acid containing selenium in place of sulfur, has no established biological function. Other minor fraction of circulating selenometabolites (Se-metabolites), including not only inorganic selenium and seleno-aminoacids, but also small organic selenium-containing compounds (<1500 Da), may reflect selenium excess potentially leading to a non-specific binding of selenium to other circulating proteins and additional synthesis of bioactive Se-metabolites with pathophysiological effects [17,18]. Serum Se-metabolites may partly reflect the body ability to remove unbound circulating selenium.

Selenium-related health effects include type 2 diabetes [[19], [20], [21], [22]], Keshan's disease and other cardiomyopathies [23,24], thyroid diseases [25], several types of cancer [26] including lung cancer and chronic obstructive pulmonary disease [27], COVID-19 [28], cognitive decline and neurological disorders [[29], [30], [31], [32]] and mortality [32,33], among others. In some reports, both selenium deficiency and excess have been associated with adverse health outcomes, showing a characteristic U-shaped relationship [[33], [34], [35]]. Identifying potential determinants of serum selenium and individual species concentrations can help to understand controversial selenium-related health effects. Population-based samples with comprehensive lifestyle, genetic and selenium speciation data are, however, scarce. Additionally, while there are limited studies evaluating the association of selenium with non-genetic [[36], [37], [38]] and genetic [[39], [40], [41], [42]] factors, they explore them independently and only for specific selenium markers. In this study, we evaluated the contribution of non-genetic and genetic factors to the concentrations of an extended panel of serum selenium markers, specifically total serum selenium and selenium species (i.e. GPx, SeP, SeAlb and Se-metabolites) in the Aragon Workers Health Study (AWHS), a cohort study primarily made up of male factory workers from Spain. Since blood selenium is more stable to inflammatory processes compared to serum [43], in order to check consistency of findings comparing to serum selenium and, to potentially gain insight into selenium excretion [44,45], we performed additional secondary analysis of total blood and urine selenium.

In particular, we first evaluated the cross-sectional association of sociodemographic and lifestyle factors with selenium markers. Subsequently we evaluated genetic influences by using a targeted approach, based on previous knowledge of selenium biology and, also, an exploratory approach, for novel variants discovery. In addition, we studied the cumulative effect of multiple genetic variants (i.e., polygenic score) on each selenium marker. Finally, we assessed the relative contribution of non-genetic (sociodemographic and lifestyle) and genetic factors to variation of each selenium biomarker concentrations.