Keywords

Aluminum in the Environment and Human Exposure

Al stands as the most abundant metallic element and ranks third in abundance among the Earth’s crust constituents. Although natural processes and acidic rain redistributes it in the nature, thus contributing to the natural occurrence of the metal in food and water, growing industrialization has been responsible for increasing the presence of Al in the environment.

To date, no physiological functions for Al have been described in mammals and, therefore, it has sometimes been regarded as not presenting a significant health hazard. In addition to the insoluble nature of the metal and its low absorption, an efficient renal elimination prevents Al accumulation in the body, thereby reducing the risk of acute human toxicity under non-pathological conditions. Despite human natural barriers (i.e., skin, gastrointestinal barrier, lungs, etc.) protect general population from environmental exposures, patients suffering from chronic renal failure may be at risk of Al toxicity (Fenwick et al. 2005).

Al has been extensively used in the industry, and it is currently added to a vast number of products available to everyone, including drinking water, many processed foods, infant formulae, cosmetics, toothpaste, antiperspirants, and various medical preparations and medicines (for a review, see Bondy 2016). These widespread uses make human exposure to Al practically unavoidable. Therefore, once presumed Al is ubiquitous in the environment, it is not so unreasonable to expect a wide range of sources of exposure.

Considering the general population, food and water represent the most common form of human exposure to the metal (Bondy 2016; Crisponi et al. 2013). The concentration of Al in food is extremely variable, due both to the original content and to food interaction with the material it contacts when stored or cooked. For example, when food or beverages are stored in Al-derived packaging formats, Al content is five to seven times higher compared to the same type of food from other containers (Duggan et al. 1992). Even though the Al content in most plant food is low (i.e., less than 25 μg/g of dry food weight), relatively high levels of Al have been reported in some spices, such as marjoram and thyme, soy-based milk formulas, tea leaves, and coffee beans (Burrell and Exley 2010; Crisponi et al. 2013; Malik et al. 2008). As for animal-derived food, Al levels in some dairy products (i.e., milk, cheese, etc.) have been found to be beyond the permissible limits (AI-Ashmawy 2011). The increasing contamination of rivers and seas has also prompted the accumulation and storage of Al in such crustaceans as crayfishes (Woodburn et al. 2011).

On the other hand, some data have endorsed that both the inhalation of Al particles and dermal absorption upon contact with the skin may also account, although to a lesser extent, for the body burden of Al (Darbre 2005; Pauluhn 2009).

Although Al total intake considerably varies upon country, place of residence, and diet composition, several authors have estimated Al typical adult intake to be ranged from 3 to 12 mg/day (Domingo et al. 2011; Krewski et al. 2007). Al absorption, though, is generally low, being almost 95% of the total Al ingested directly excreted through feces. In point of fact, the total Al absorption may vary from 0.01 to 1% of the total metal intake (Moore et al. 2000; Wilhelm et al. 1990). The presence of certain compounds in the diet such as citrate, lactate, ascorbate, gluconate, succinate, tartrate, malate, and oxalate can increase the rate of absorption of Al (Krewski et al. 2007). Likewise, low plasma levels of magnesium (Mg) and Fe (Cannata et al. 1991), as well as enhanced vitamin D status, may increase Al absorption (Krewski et al. 2007; Schwalfenberg and Genuis 2015). Therefore, Al bioavailability is highly dependent on individual differences, fact that merits to be taken into account to control confounding variables in both experimental and epidemiological studies.

Once absorbed, Al has a half-life of several hours in the blood. Indeed, Al is primarily bound to plasma transferrin (Tf) (i.e., 90%) and, to a lesser extent, to low molecular weight molecules, such as citrate (i.e., 10%) (Ohman and Martin 1994). Even though the mechanisms through which Al enter the brain are not yet fully understood, this process seems to be governed by two different mechanisms. Firstly, Al can enter the brain from blood. As a matter of fact, it is well known for more than 25 years that transferrin can mediate Al transport across the blood-brain barrier (BBB) by transferrin receptor (TfR)-mediated endocytosis of Al transferrin (Bondy 2016; Yokel and McNamara 2001). On the other hand, though, there is evidence to suggest that a second mechanism transporting Al citrate across the BBB into the brain independently of transferrin may exist. Indeed, transferrin concentration is very low in cerebrospinal fluid (CSF), and presumably brain extracellular fluid, whereas the citrate concentration is higher in CSF than in plasma, favoring Al citrate as the predominant Al species in brain extracellular fluid (Martin and Bruce 1997; Yokel and McNamara 2001). Although most brain Al is quickly removed, some experimental research evidenced that its half-life may be about 150 days in rats (Yokel et al. 2001).

At physiological pH, Al exhibits the trivalent oxidation state (i.e., Al3+), which is crucial in determining the physicochemical properties and biological interactions inherent to the metal.

The main mechanism of Al toxicity involves the disruption of the homeostasis of metals, such as Mg, calcium (Ca), and Fe (Harris et al. 1996; Yokel and McNamara 2001). Indeed, the physical and chemical properties of Al (i.e., the small radius of Al3+, its affinity to oxygen atoms, carboxylate, deprotonated hydroxyl and phosphate groups, etc.) allow it to effectively mimic these metals in their respective biological functions and trigger biochemical abnormalities, thereby defining Al individual’s toxicokinetics (Yokel and McNamara 2001).

Aluminum in the Brain: Molecular and Functional Interactions

Al is unequally distributed in brain areas and neural cells. Indeed, Al accumulates in glia largely than in neurons (Oshiro et al. 2000). Experimental studies in rats and mice showed that Al accumulates in the brain cortex, hippocampus, and cerebellum (Bellés et al. 1998; Esparza et al. 2003; García et al. 2009) after either parenteral or oral exposures. Accordingly, several authors have measured brain levels of Al in AD patients and non-demented controls, and both the hippocampus and the amygdala stand as the most relevant areas containing Al. Despite data are not always consistent (Akatsu et al. 2012), statistical treatments and the control of confounding variables have been found to influence the statistical significance of the result (Rusina et al. 2011).

It is well known that such metal ions as Al are able to interact with different proteins to induce conformational changes that eventually result in misfolding, aggregation, or oligomerization, thus leading to an altered turnover and removal of the protein. Protein misfolding and aggregation is a key pathophysiological mechanism on AD. Hence, an increased interest on the possible contribution of Al on the amyloid (Aβ) cascade hypothesis for AD has generated a deal of research. Unfortunately, results have not always been consistent. Indeed, many studies reported that Al promotes the expression of the precursor (APP) of the Aβ protein, increases the levels of β-40 and β-42 fragments in the brain, and boosts the aggregation of Aβ protein (Zatta et al. 2009; Bolognin et al. 2011; Praticò et al. 2002; Banks et al. 2006). Other in vivo approaches, though, did not replicate the results previously reported for the Aβ pathway (Ribes et al. 2010; Akiyama et al. 2012). Further, it has been shown that Al promotes both the phosphorylation and the aggregation of highly phosphorylated proteins such as tau protein (Leterrier et al. 1992; Nübling et al. 2012). According to this, Al has been detected in neuritic plaques and tangle-bearing neurons, pointing at the involvement of this metal in the pathogenesis of AD (Miu et al. 2003; Yumoto et al. 2009). Moreover, Al has also been related to altered synaptic plasticity in the hippocampus of rats when chronically and orally administered at high doses (Colomina et al. 2002).

Aluminum and Glucose Metabolism

Despite the high prevalence of Al in the environment, there is a gap of knowledge on its interaction with physiological systems. An escalating body of experimental research has demonstrated so far that Al inhibits a vast number of ATP-dependent reactions, thereby interfering with energy-dependent processes (Caspers et al. 1994; Joshi et al. 1994; Kaizer et al. 2007; Silva et al. 2005; Singla and Dhawan 2012). Nonetheless, the exact mechanisms remain to be fully unraveled. Thus, Al3+ binds to ATP 107 times more tightly than does Mg2+, but upon in vivo testing not every ATP-dependent reactions are inhibited (Joshi 1990).

Many experimental approaches have also endorsed that Al exposure may impair glucose utilization, upon altering activities of glucose-metabolizing enzymes, such as glucose-6-phosphate dehydrogenase (G6PD), hexokinase, or glutamate dehydrogenase (Dua et al. 2010; Joshi et al. 1994). Strikingly, G6PD enzyme has been shown to reduce its activity in the presence of Al in such brain regions that are also affected in AD (Joshi et al. 1994; Jovanović et al. 2014). In point of fact, accumulated evidence indicates that AD is a metabolic neurodegenerative disease. Thus, impaired cerebral glucose metabolism represents an invariant pathophysiological feature in AD, and its occurrence mostly precedes cognitive dysfunction and pathological alterations. Therefore, delving into the consequences associated with abnormal cerebral glucose metabolism will provide valuable clues for treatment strategies as well as ideal diagnostic approaches in AD.

Aluminum and Iron Interactions

Over the last years, a considerable amount of literature has grown up shedding light on the disruption of Fe homeostasis by Al. Fe, an essential trace metal, displays a great deal of biological functions, including oxygen transport and exchange, metabolic protein synthesis, and enzyme cofactor (Aisen et al. 2001). Because of its biological importance and high redox potential, Fe is strictly regulated by Tf, transferrin receptor (TfR), and ferritin. Thus, increases in TfR allow more Fe to enter the cell, while a decrease in ferritin levels enables more free iron to reach the respiratory chain and other Fe-requiring systems. Under Fe-replete conditions, TfR decreases and levels of ferritin increase, thereby allowing the metal to be stored in a complex with ferritin, which prevents iron-mediated oxidative stress (Aisen et al. 2001). Several in vitro studies have reported that Al exhibits the same effect on TfR and ferritin as Fe does in a deprived status. Thus, Al increases the number of TfRs, which leads to an increase in Fe absorption, and also decreases ferritin, which might result in higher levels of free Fe. These effects of Al on Fe homeostasis might explain the increases in oxidative stress and inflammation in both in vitro and in vivo upon exposure to Al (Kim et al. 2007).

The total body burden of Fe has been found to increase with age in a sex-dependent manner (Joshi et al. 1994). While males progressively increase Fe levels from 300 to 1800 mg between 20–25 and 80–90 years, Fe stores in women remain at 300 mg from 20 to 25 years until the premenopausal state. Then, Fe stores begin to increase until reaching 1300 mg at the age of 80–90 years (Joshi et al. 1994). Furthermore, women increase Fe storage parameters from premenopause to postmenopause. Strikingly, such Fe increases correlate with the increase in HOMA-R index, which indicated insulin resistance (Van den Bosch et al. 2001).

In view of the influence of Fe status in Al absorption, we speculated that this sex-related pattern of storage can influence the onset and course of neurodegeneration. The sharp increase in Fe stores from middle age to elderly shows some parallelisms in temporal patterns observed for AD prevalence in women. According to Fe status, young women, which display low levels of Fe storage, would be more able to store Al but protected from Al-Fe interactions, and therefore from oxidative stress. By contrast, postmenopausal women would have higher Fe stores, but Al deposits would remain high because of a long-life exposure and efficient storage favored by moderated Fe levels in serum.

Aluminum and Oxidative Stress

As previously stated, no biological role for Al is known yet. However, it is well accepted that it can induce severe toxic manifestations in mammals. Given the nondividing nature of neurons, the brain has sometimes been considered to be the most vulnerable tissue to the toxic effects of Al (Kumar and Gill 2014). Indeed, a constellation of experimental research has highlighted neuropathological, neurobehavioral, neurophysical, and neurochemical changes upon Al administration (Akiyama et al. 2012; Colomina et al. 2002; Verstraeten et al. 1997; Ribes et al. 2008, 2010). Further, the brain is particularly sensitive to oxidative stress due to an increased level of free radicals and decreased level of antioxidants following toxic insult (Kumar and Gill 2014). Several authors have suggested that Al exert a strong prooxidant activity despite its non-redox status (Exley 2004; Kumar and Gill 2014; Yuan et al. 2012).

To date, there are described many potential mechanisms underlying Al-related prooxidant toxicity, of which the effect of Al on Fe homeostasis is of special interest (Ward et al. 2001; Wu et al. 2012). As a matter of fact, the interaction between both agents generates labile Fe from Fe-containing enzymes and proteins, thereby increasing the intracellular pool of free Fe, which in turn leads to the formation of reactive oxygen species (ROS).

On the other hand, Al oxidative toxicity has also been related to increased lipid peroxidation, decreased membrane fluidity, and oxidized high-density lipoprotein (Johnson et al. 2005; Kaneko et al. 2007; Silva et al. 2002; Zatta et al. 2003). For example, Al has been shown to potentiate the free radical damage initiated by Fe3+ in lipid peroxidation, probably by facilitating the action of OH− radicals in the membrane of phospholipids (Zatta et al. 2003). Other mechanisms, such as the formation of superoxide Al3+ semi-reduced radical, have been suggested to explain Al prooxidant effects (Ruipérez et al. 2012; Exley 2004). In general terms, the interaction with lipid substrates as well as with other prooxidant metals or elements such as O2− are subjects of study in this regard (for review, see Exley 2004). Additionally, Al3+ decreases the activity of some antioxidant enzymes such as catalase, superoxide dismutase, and glutathione peroxidase (Fattoretti et al. 2003; Sánchez-Iglesias et al. 2009), thus aggravating neuronal damage induced by oxidative stress.

An Al-dependent oxidative environment is also characterized by a sharp decrease in mitochondrial activity (Han et al. 2013). Specifically, Al3+ disrupts mitochondrial bioenergetics and decreases the respiratory efficiency and the capacity of the mitochondria to modulate and control the energy production through the phosphorylation system (Iglesias-González et al. 2016).

Aluminum and Neurotransmission

Several studies have indicated Al is able to disrupt the cholinergic system, which is in turn implicated in AD pathogenesis. Both in vivo and in vitro studies have consistently shown changes in acetylcholinesterase (AChE) activity, as well as in ACh-evoked neurotransmission (Yokel et al. 1994; Bielarczyk et al. 1998; Szutowicz et al. 2000; Yellamma et al. 2010). Accordingly, the group of Petronijević found activity changes in AChE and lipid peroxides in a series of different studies with high Al doses administered to Gerbils (Mićić et al. 2003; Vučetić-Arsić et al. 2013). Despite the possible relevance of this pathway, few research has evaluated cholinergic function together with other parameters of interest. Strikingly, estradiol administration has shown to ameliorate alterations in cholinergic parameters and oxidative stress induced by Al intoxication (Mohamd et al. 2011).

Moreover, the neurotransmitters serotonin (5-HT) and dopamine (DA) (Abu-Taweel et al. 2012), as well as glutamate and aspartate (Liu et al. 2010), have been reported to decrease upon Al exposure. It is well known that neurotransmitter systems are modulated by sex hormones. In this sense, differences between sexes have been reported for the septo-hippocampal cholinergic system (Mitsushima 2011), monoamines 5-HT, and DA, as well as for glutamate and GABA (Barth et al. 2015). Therefore, the effects of Al in neurotransmitter systems could be different depending on sex, but no data exist so far in this regard.

Worldwide Advices and Al Regulations

Needless to say, to date, the detrimental effects of Al to human health are well established, and increasing eagerness to regulate its uses has become noticeable. Thus, many relevant regulatory agencies, including the Agency for Toxic Substances and Disease Registry (ATSDR), the US Environmental Protection Agency (EPA), the World Health Organization (WHO), the Food and Agriculture Organization (FAO), and the European Food Safety Authority (EFSA), have published a great body of reports on Al toxicity. However, there is still widespread mistrust about its potential deleterious effects upon silent chronic exposures. By way of example, there is still no convincing evidence to associate the Al found in food and drinking water with increased risk of AD. Neither is there clear evidence to suggest increased risk of AD nor some types of cancer upon using Al-containing antiperspirant or cosmetics. At most, the US FDA has warned that increasing Al concentration in an antiperspirant product may result in skin irritation. Further, even if adverse effects to vaccines with Al adjuvants have occurred, recent controlled trials found that the immunologic response to certain vaccines containing Al was no greater, and in some cases less than, that after identical vaccination without Al adjuvants (for a review, see Willhite et al. 2014).

The FAO expert committee on food additives and food contaminants had originally recommended a tolerable weekly intake (TWI) of Al of 7 mg/kg, which was lately reduced to 1 mg/kg upon reconsidering the reproductive and neurological detrimental effects of the metal (FAO 2006). In Europe, the EFSA stated in 2011 a TWI equivalent to 280 μg/kg per day (Anon 2011). Nonetheless, some authors have questioned these values since the EFSA assumed back then that gastrointestinal uptake of all ingested Al materials was equivalent to that measured for Al citrate (Willhite et al. 2014). On the other hand, the WHO recommended a maximum drinking water concentration of 0.2 mg Al/L (WHO 2004; WHO 2010).

Given that it has not yet been established which levels of Al are safe upon chronic exposures in human populations, an effort is needed to demand more regulations for the use of this metal in drinking water, dairy products, pharmaceuticals, and occupational exposures.

Alzheimer’s Disease and Environmental Al Exposures

Alzheimer’s disease (AD) is one of the most devastating neurodegenerative diseases, accounting for more than 80% of dementia cases in the elderly. It is a complex neurodegenerative disorder characterized by a neurological progressive impairment affecting several cognitive domains, behavior, and personality. Clinical phenotype is accompanied by three main neuropathological hallmarks: diffuse loss of neurons, neuronal cytoskeletal alterations or neurofibrillary tangles (NFT) produced by hyperphosphorylated tau protein aggregations, and extracellular Aβ protein deposits or senile plaques (Torreilles and Touchon 2002).

Two forms have been described for AD: the familial form, which is less frequent (1–5%) and mainly genetic, and late-onset AD (LOAD), which is most prevalent and heterogeneous in both onset and progression (Ridge et al. 2013).

The familial forms of the disease are mostly associated with mutations exhibiting an autosomal dominant pattern of inheritance. Thus, three mutated human genes encoding for (1) APP and the enzymes related to APP processing, (2) presenilin 1 (PSEN1), (3) and presenilin 2 (PSEN2) are crucial to the establishment of the disease (Levy-Lahad et al. 1998; Schellenberg et al. 1992). All of these genetic mutations lead to abnormal processing of APP and give rise to the Aβ cascade hypothesis. Although crucial, this hypothesis fails to explain by itself all the molecular, cellular, and clinical events observed in the different forms of AD. Before the identification of familiar forms, anatomical-pathological and biochemical studies of postmortem human brains from AD patients described deficits in the cholinergic system. In addition, considerable pieces of experimental and human studies have supported that a dysfunctional cholinergic system is sufficient to produce learning and memory deficits (Muir 1997). Thenceforth, the earliest cholinergic hypothesis of AD emerged. Degeneration of cholinergic neurons in the basal telencephalon (i.e., Meynert nucleus and medial septum nucleus) innervating the hippocampus, amygdala, and frontal cortex has been associated with severe cognitive deficits implicated in AD (Muir 1997). Moreover, pharmacological interventions with cholinergic agonists have endorsed the contribution of this system to cognitive decline (Giacobini 2003).

On the other hand, environmental risk factors accumulating over years are constantly challenging the integrity of the brain and thereby contributing together with risk genetic factors to the onset and progression of LOAD. Accordingly, APOE4 genotype is the largest genetic risk for AD accounting for approximately 60% cases (Higgins et al. 1997). Indeed, being a carrier of one ε4 allele increases the risk for AD in 2–3-folds, whereas the risk rises about 12-fold when carrying two ε4 alleles (Roses 1996). Interestingly, emerging lines of evidence supported an APOE4-sex interaction in humans. Women carrying ε4 have been shown to display more pronounced AD-like changes in neuroimaging, neuropathological, and neuropsychological measures than men (Beydoun et al. 2013; Ungar et al. 2014). As for environmental factors, it is worth asking for these agents and to which extent they are contributing to the onset and progression of the disease. In this sense, the hypothesis of Al and AD has become the subject of intense debate over the last decades. The putative link between dietary Al and neurodegenerative disorders has been addressed in a large volume of clinical (Wills and Savory 1989; Yumoto et al. 2009), occupational (Riihimäki et al. 2000), and epidemiological surveys (Flaten 1990; Rondeau et al. 2008). Moreover, some anatomopathological findings in the brain of AD patients (Walton and Wang 2009; Yumoto et al. 2009) and some experimental studies (Praticò et al. 2002; Ribes et al. 2008; Ribes et al. 2010; Walton and Wang 2009) have also provided links between Al and AD. A collection of different studies performed until 2014 are reviewed in Willhite et al. (2014).

While some experimental results have not been widely replicable, epidemiological studies showed considerable consistent associations. In a recent meta-analysis of epidemiological studies, Wang et al. (2016) assessed the relation between Al exposure and AD. They included 8 studies and a total population of 10,567 individuals, the source of Al exposure they evaluated was drinking water and occupational exposure, and the follow-up duration from the cohort studies ranged between 8 and 48 years. The primary result of this meta-analysis was a significant association between Al exposure and the risk for AD (OR = 1.71, 95% CI, 1.35–2.18). Further, the authors also reported differences between groups exposed at 100 μg/L or higher Al concentrations in drinking water and those exposed to levels below 100 μg/L (OR = 1.95, 95% CI, 1.47–2.59). They concluded a possible link between Al exposure and AD (Wang et al. 2016). Authors also highlighted the importance of obtaining data from long-term Al exposure from food consumption to establish a possible dose-dependent link between Al and AD. The results from this study are in line with existing literature, thus indicating the importance of time of exposure and the exposure level in chronic Al exposure.

Notwithstanding the numerous scientific efforts and our actual knowledge of mechanisms involved in Al neurotoxicity, there is still no consensus on the real implication of Al and AD. Probably controlling for confounding factors in both epidemiological and experimental approaches would help to disentangle this complex relation. Remarkably, no information exists on sex possible differences in Al neurotoxicity.

Have Sex Differences Been Overlooked in AD and Al Toxicity?

Needless to say, experimental investigations involving male individuals are to date much more abundant than those using females. The female’s estrous cycle is often singled out as the driving reason researchers prefer to use male subjects, but this selective discrimination is to blame for the lack of empirical data regarding the differences between both sexes. Nowadays, it is well recognized that they differ in such several behavioral processes as emotion (Girbovan and Plamondon 2013), impulsivity (Bayless et al. 2012; Weafer and de Wit 2014), basal activity (Simpson and Kelly 2012), learning and memory (Jonasson 2005; Li and Singh 2014), or attention (Bayless et al. 2012).

AD prevalence varies by age, sex, ethnicity, and geographic region, suggesting environmental and genetic factors may play an important modulating role (Mazure and Swendsen 2016). Indeed, as it has been suggested on many occasions, sex differences are evident when analyzing the prevalence and severity of AD. In fact, clinical and preclinical studies have shown that women not only are more prone to develop AD than men but also show significantly age-related faster decline and greater deterioration of cognition than they actually do (Cornutiu 2015; Li and Singh 2014). Some investigations have also described sex-genetic interactions. As previously stated, APOE4 confers greater AD risk associated with tau pathology in women (Altmann et al. 2014). Similarly, the development of Aβ pathology in several transgenic mouse models of AD is greater in females (Maynard et al. 2006). These sex differences are also evident for metal brain levels in Cu and Mn, suggesting natural sex differences may contribute to the increased propensity of females to develop AD (Maynard et al. 2006).

Despite early epidemiological studies have clearly related differences among sexes as for AD onset (Gao et al. 1998), current clinical AD research sometimes considers males and females having equal risk toward developing the disease (Altmann et al. 2014).

Conclusions

Upon taken together Al implication in oxidative stress, mitochondrial dysfunction, Fe and Ca dyshomeostasis, neuroinflammation, microtubule alterations, cholinergic disruption, as well as compromised axonal transport and Aβ aggregation, Al implication in cognitive deficits and neurodegeneration is undeniable. Therefore, the contribution of Al to AD is plausible. However, it has not yet been established which levels of Al are needed, which factors are essential, or how long the exposure to it must be to induce functional brain deficits. It is not unrealistic to hypothesize that some populations may be protected against Al exposure or show some kind of resistance to it, a point that it is important to take into account as a possible source of bias in epidemiological studies. The major challenge for future researchers is identifying which variables are needed to be controlled in epidemiological studies and further designing more focused and translational experimental studies. The exposure pattern including time of exposure, dose-response effects, and the time elapsed between exposure and cognitive evaluations are of special importance. The identification of factors contributing to either resilience or exacerbated vulnerability to Al neurotoxicity must be taken into account in epidemiological and experimental studies. Clearly, research is needed to establish sex and age Al-related interactions, as no data exist so far in this regard.