Selenium (Se) is a human micronutrient and is critical for many biological functions. Selenium is ingested from solid food sources and absorbed as selenomethionine and selenocysteine (Ha et al., 2019). The US Food and Nutrition Board at the Institute of Medicine of the National Academies recommends a daily dietary reference intake of 55 mcg in both men and women aged 14 years and older, with a slightly higher intake recommendation for those who are pregnant or lactating (Institute of Medicine, 2000). Once absorbed in the small intestine, selenium is incorporated into one of two Se-containing proteins in vivo: (a) selenomethionine-containing proteins, which are nearly identical to their methionine-containing counterparts (Schrauzer, 2000); and (b) selenocysteine (Sec)-containing proteins, which are specialized proteins essential to human function (Figure 1). Sec-containing proteins are most biologically relevant and are commonly known as selenoproteins (Minich, 2022).

Approximately 80% of selenium is found in association with selenoproteins (Minich, 2022). To date, there are 25 known selenoproteins (Kryukov et al., 2003). There are three major classes of selenoproteins (Pitts et al., 2014). First, glutathione peroxidases function to protect from oxidative damage by reducing hydrogen peroxide and peroxide radicals. There are eight known glutathione peroxidases, but only five (GPX1, GPX2, GPX3, GPX4, and GPX6) have been associated with selenocysteine in their active sites. Thioredoxin reductases (Trx1, Trx2, and Trx3) reduce disulfides to oxidize thiol-dependent peroxidases, thereby exhibiting antioxidant activity as well (Lu and Holmgren, 2014). Thioredoxin reductases also play a role in cell growth and p53 activity (Mustacich and Powis, 2000). Deiodinases (DIO1, DIO2, and DIO3) primarily assist in regulating plasma thyroid hormone (T₃) levels, and conversion of T₃ to T₄ (Luongo et al., 2019). The remainder of selenoproteins are named with the prefix SELENO-and conduct a variety of functions, such as reduction of lipid hydroperoxides and regulation of protein folding and endoplastic reticulum (ER) stress (Zhang and Song, 2021). With the exception of deiodinases, all selenoproteins have some method of antioxidant activity. A summary of selenoproteins and their functions are included in Table 1.

Selenoprotein activity is essential to many biological processes, such as fetal development (Hogan and Perkins, 2022), hormone regulation, and reproduction (Kurokawa and Berry, 2013). Selenium deficiency during pregnancy leads to reduced placental weight and reduced fetal blood glucose levels in mice (Hofstee et al., 2019), likely due to deficiencies in SELENOP and SELENON. In addition, methylation at Se-dependent CpG sites during pregnancy is associated with poorer muscle tone in newborns (Tian et al., 2020). Selenium is highly present in the testes as well, and selenium deficiency adversely impacts testes development and spermatozoa production (Behne et al., 1996). Several selenoproteins (GPX1, GPX2, GPX3, GPX4, SELENOM, SELENOP, SELENOS, and SELENBP1) are expressed in the intestine, and they may play a role in mitigating excessive immune responses that may lead to inflammatory bowel diseases (Speckmann and Steinbrenner, 2014).

Most known selenoproteins are expressed in the neurons of adult mouse brains (Zhang et al., 2008) and in human brains (Zhang and Song, 2021). Within the brain, there is the most selenoprotein expression in the olfactory bulb, cerebral cortex, hippocampus, and cerebellar cortex (Zhang et al., 2008), and selenium is transported to these areas by SELENOP (Hill et al., 2003; Schomburg et al., 2003). When selenoprotein formation is malfunctioning, studies have shown cerebellar hypoplasia and atrophy and cerebral atrophy (Agamy et al., 2010; Wirth et al., 2014), indicating the necessity of selenoproteins during brain development. Therefore, deficiencies in selenium and/or selenoproteins poses some threat for neurological conditions (Schweizer et al., 2021).

Autism spectrum disorder (ASD) affects 1 in 44 children in the United States (Maenner et al., 2021), and is characterized by repetitive behaviors and issues with social interaction (Mughal et al., 2022). A recent study from our group demonstrated lower serum and nail selenium levels in boys aged 24 to 47 months with ASD compared to sex-matched controls (Mehta et al., 2021). In addition, there was a negative correlation between Autism Diagnostic Observation Schedule (ADOS) scores and selenium levels. The purpose of this review is to examine the role of selenium and selenoproteins in the human brain, and how this may be implicated in ASD etiology. To our knowledge, there has not been a quantitative study of Selenoproteins in people with ASD.

As discussed, most selenoproteins function as protection from oxidative stress, particularly within the hippocampus and cerebellum of the brain. ASD etiology is established as linked with oxidative stress in the brain (Raymond et al., 2014; Bjorklund et al., 2018; Hu et al., 2020; Manivasagam et al., 2020; Pangrazzi et al., 2020). Sources of oxidative stress during neurodevelopment may include maternal immune activation and environmental toxins, both of which have been linked to ASD. However, the compensation of partially functioning proteins such as SHANK3, FMRP, or CHD8 during neural development would require increased metabolic activity and, subsequently, more oxidative stress.

Studies examining biomarker levels in children with ASD found a significantly higher level of oxidative stress markers and plasma glutamate in comparison to neurotypical controls (Adams et al., 2011). The leading explanation for the high levels of oxidative stress biomarkers is an imbalance in selenoprotein antioxidant defense (Raymond et al., 2014), such as glutathione. One meta-analysis found decreased blood levels of glutathione and glutathione peroxidases in ASD patients relative to controls (Frustaci et al., 2012), along with three independent studies published after the meta-analysis (Chauhan et al., 2012; Rose et al., 2012; Gu et al., 2013). This glutathione expression and activity imbalance seem to be within the cerebellum and temporal cortex. Aside from glutathione, other antioxidants such as ceruloplasmin and transferrin have been seen at lower levels in serum of ASD patients (Chauhan et al., 2004); interestingly, while these are not selenoproteins, ceruloplasmin and transferrin are associated with transport of copper and iron, respectively. Of note, some discuss high levels of oxidative stress markers as an independent factor of ASD etiology, rather than a result of antioxidant imbalance (Chauhan and Chauhan, 2006).

The increased oxidative stress in the brain may be due to impairments in the blood–brain barrier (BBB): one study found that ASD patients have lower levels of endothelial adhesion molecules, which suggest blockage of peripheral leukocytes into the central nervous system (Onore et al., 2012). Another group identified alterations in BBB integrity in post-mortem brain tissues of ASD patients, indicated by varied expression levels of tight junction and similar BBB molecules (Fiorentino et al., 2016).

Prior to our study described above, it has been suggested that dyshomeostasis in selenium may be associated with ASD incidence (Raymond et al., 2014). Since our study, others have also reported an association between ASD and low plasma Se levels (Wu et al., 2022; Zhang et al., 2022). However, as a whole, the literature on this subject has been equivocal. In a literature review of 10 studies comparing hair trace element levels in ASD and controls, four of them found a significant difference in Selenium levels. Two found a significant increase in Selenium levels in children with ASD, and two found a significant decrease (Tinkov et al., 2019). Another meta-analysis found no significant differences in mean hair or erythrocyte selenium concentrations among 12 studies (Saghazadeh et al., 2017). A retrospective study of prenatal levels of trace elements in Norwegian mothers found no association between Se levels and risk for Autism in pregnancy (Skogheim et al., 2021) while a prospective study in the United States found an associated between elevated maternal Se levels and increased ASD risk (Lee et al., 2021). There are several possible factors that account for the variability of these studies. The first is variability of measurement. Older studies used a variety of techniques to measure Se concentration, rather than the inductively coupled plasma mass spectrometry (ICP-MS) that has become standard over the past few years. Second, the source material in these studies varied significantly. In both ASD and control populations, the variability of Se in hair and nails was greater than in serum (Mehta et al., 2021). Finally, there is evidence that normal ranges of Se change depending on age and gender (e). Many previous studies recruited children with ASD from ages 2–18, which may have resulted in non-significant findings.

We propose several hypotheses for the etiology of ASD related to selenoproteins (Figure 2). Increased oxidative stress during neurodevelopment leads to a lessened ability for selenoproteins to function efficiently. Thus, oxidative stress may lead to increased neuronal apoptosis, as seen in many studies investigating glutathione peroxidases 1 and 4 (Crack et al., 2001; Flentjar et al., 2002; Ufer et al., 2008). Second, ferroptosis may be decreased in oxidative stress conditions. It has been shown that TRX1 and GPX4 work together to modulate ferroptosis (Bai et al., 2021). Finally, oxidative stress may lead to incorrect formation of neuronal connections and/or firing patterns during neurodevelopment. Raymond et al. hypothesized that the antioxidant function of glutathione and other selenoproteins affect the epigenetic regulation of gene expression (Raymond et al., 2014), thus altered oxidative stress levels in the brain may lead to abnormal neuronal connections. Taken together, there are several plausible hypotheses for the etiology of ASD through oxidative stress mechanisms and selenoproteins function.

We attempted to establish a pattern of selenium formulation used as a dietary supplement in clinical trials using information available at the www.clinicaltrials.gov. The keyword “Selenium” was searched in www.clinicaltrials.gov website, which resulted in 381 hits. Of those, only 222 studies were completed. To which, only 69 studies were conducted on children under age 17 and only one in ASD. The alone study was conducted in Indonesia to study the “Effect of High Selenium Functional Food and Selenium Supplement” (NCT05218577). Interestingly, this study used beef liver as a source of selenium supplement. However, when we searched with condition keyword as “Autism” and others as “selenium” we found total two studies focusing on Autism including the first one conducted in Indonesia (NCT05218577). The second study was performed at the Arkansas Children’s Hospital Research Institute (NCT00572741), where investigators studied treatment of oxidative stress and the metabolic pathology of autism with multi-vitamins and multi minerals containing 50 mcg of Selenium.

We thought to perform additional search for “Selenium supplementation” alone irrespective of its application in different disease pathology. We got 95 hits. Of those only 63 were completed. When we attempted to establish a selenium formulation pattern in these studies, we found variation in type of selenium used in these clinical trials. The major forms of Selenium used in these studies were, Selenium-yeast, sodium selenite (selenase) and selenomethionine. However, in many studies with limited information available on the clinicaltrials.gov website, it was not clear which chemical form of the selenium was used, especially when selenium was part of multi nutrients containing various vitamins, minerals including selenium, but no specific chemical form of selenium was mentioned making it even harder to establish preferred chemical form of Selenium. In our opinion, every chemical form of selenium has its own advantages and disadvantages in terms of bioavailability, toxicity, suitability of absorption, ease of synthesis and formulation.

Based on the chemical composition and formulation, selenium supplementation can be categorized in following four categories, (i) Selenium enriched yeast, (ii) Selenium containing amino acids (iii) Selenium salts (iv) Selenium nanoparticles (Figure 3).

Selenium enriched yeast: Selenium enriched yeast is simply known as a selenium yeast and produced by culturing Saccharomyces cerevisiae in a selenium rich media. Saccharomyces cerevisiae commonly known as a brewer’s yeast, which produces selenomethionine in presence of selenium along with normal methionine due to the similar chemical characteristics of selenium as of sulfur. Irrespective of amount of selenium content in a culture media, complete replacement of methionine by selenomethionine has not been reported. The highest quantity of selenomethionine produced by a single yeast cell is 3,000 ppm (Schrauzer, 2006). Various safety concerns have been raised to selenium yeast as a dietary supplement due to variable quantities of selenomethionine present in selenium yeast and there was no rigorous quality control check for the presence of other contaminations. However, later some quality control parameters were placed to ensure the safety of selenomethionine produced from the Saccharomyces cerevisiae as a selenium yeast including limits on moisture content in the yeast, limits on allowed lead and arsenic presence, limits on allowed live bacterial and mold counts (Schrauzer, 2006). Moreover, selenium yeast is used as feed additives to supplement selenium.

Selenium and selenoproteins are involved in healthy brain function and development, primarily through their antioxidant functions. Oxidative stress is associated with ASD; studies show a correlation between increased oxidative stress biomarkers and presence of ASD. Many have also shown a decrease in selenoproteins such as glutathiones in ASD patients, suggesting that selenoprotein dyshomeostasis may in fact play a role in ASD etiology. However, current literature studying selenium dyshomeostasis and ASD are variable in nature and thus inconclusive. Further, selenium supplementation trials (whether with ASD cohorts or others) lack clear definitions of the type of selenium used for supplementation. Larger and better-defined trials are needed to better understand the role of selenium and selenoproteins in the etiology of ASD, as well as the ideal type of selenium needed for supplementation trials.

SB, SM, and MP contributed to conception and design of the review. SB wrote the first draft of the manuscript. SM and MP wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.