Abstract
Epidemiological evidence has linked the consumption of milk and dairy products to an increased risk of developing Parkinson’s disease (PD). However, the potential pathogenetic mechanisms remain to date yet to be ascertained.
Galactose (D-Gal) is a hexose sugar which, when given orally and by other routes of administration, is picked up by the brain after a few hours. For doses above 100 mg/kg, it appears that galactose can cause biochemical and neuropathological alterations in neurons and astrocytes, similar to those observed in PD. These quantities can be reached and surpassed with the simple daily consumption of two glasses of milk, the main dietary source of D-Gal in humans.
It is likely that, relative to other neurons, dopaminergic neurons are more vulnerable to D-Gal induced damage, because of their greater vulnerability to oxidative stress.
Mitochondrial dysfunction plays a critical role in both PD and D-Gal toxicity, and mutations of the genes commonly involved in PD and in mitochondrial homeostasis could enhance this mechanism.
If this hypothesis were to be confirmed, dietary interventions, such as reducing the sources of galactose in the diet, and/or increasing the intake of protective molecules, could help reduce the occurrence of this disease in the aging population.
Glossary of terms
advanced glycation end products
(receptor) α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (receptor)
antioxidant response element
acetylsalicylic acid
apoptosis regulator (Bcl2 family)
blood brain barrier
B-cell lymphoma 2
B-cell lymphoma-extralarge (Bcl2 family)
C-C Motif Chemokine Ligand 2
catalase
transcription factor, also known as proto-oncogene c-Maf or V-maf musculoaponeurotic fibrosarcoma oncogene homolog
regulator gene that codes for a transcription factor
Clueless protein, human orthologue of “hClu”
3,4-dihydroxyphenethylamine (dopamine)
D-galactose
protein deglycase (PARK 7)
damage-regulated autophagy modulator
excitatory amino acid transporter (glutamate transporter)
F-Box Protein 7 (PD associated gene)
gamma-aminobutyric acid
glutamate dehydrogenase
galactokinase
glucose
glutaminase 2
glutamate
glutamate neurotoxicity
glutathione peroxidase 1
glutathione reductase
glutamine synthetase
glutathione
heme oxygenase-1
interleukin 1
interleukin 6
leucine-rich repeat kinase 2 (PD associated gene)
malondialdehyde
mouse double minute 2
mitochondrial outer membrane (permeabilization)
manganese-dependent superoxide dismutase
mitochondrial permeability transition pore
(1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine)
mammalian target of rapamycin
nuclear factor kappa-light-chain-enhancer of activated B cells
N-methyl-D-aspartate (receptor)
Notch homolog 1, translocation-associated
apoptosis regulator (Bcl2 family)
Nuclear factor (erythroid-derived 2)-like 2
Nuclear factor (erythroid-derived 2)-like 2
oxidative phosphorylation
Parkinson disease
PPAR/peroxisome proliferator activated receptor 1-α
PTEN-induced putative kinase 1
phosphatase and tensin homolog
apoptosis regulator (Bcl2 family)
tumor protein 53
receptor for advanced glycation end products
reactive oxygen species
cytochrome C oxidase assembly protein
superoxide dismutase
substantia nigra pars compacta
synaptic vesicle glycoprotein 2A
(cycle) tricarboxylic acid (cycle)
transcription factor A, mitochondrial
TP53-inducible glycolysis and apoptosis regulator
toll like receptor
tumor necrosis factor
dehydrothromboxane B2
cysteine/glutamate transporter
Introduction
Parkinson’s disease is one of the most frequent causes of disability in adult and elderly people. Its prevalence has risen dramatically in China since the 80 s when PD affected 1 in 1000 subjects, while today 1 in 200 will develop the disease, as is the case in Europe and the United States. This epidemiological trend suggests that environmental variables such as lifestyle habits or diet may play a role in the pathogenesis of PD [1]. Idiopathic PD is histologically characterized by neurodegeneration of the dopaminergic substantia nigra pars compacta (SNpc). Although the reason is far from being ascertained, some possible pathogenetic factors acting through enhanced inflammation and oxidative stress, which may be involved in the destruction of the dopaminergic neurons, have been identified [2–6]. A positive association between milk and other dairy product consumption and the risk of developing PD has also been reported [7–11], while other studies have found a strong association only with milk consumption, and not with yoghurt or other dairy products [12].
Drinking more than 2 glasses of milk per day was associated with a reduced neuronal density of the substantia nigra in human brain tissue examined post-mortem. A pesticide found in both the examined brains and in milk was suggested to be the pathogenetic factor [13]. Some authors state “a clear association between dairy products and PD exists, but a rational explanation why such foods may represent a risk for PD is lacking” [14]. Other authors reported an increased risk of developing neurodegenerative disorders with higher consumption of milk and dairy [15], emphasizing that the effects of galactose (D-Gal) on cognitive decline have not been fully evaluated. Nevertheless, even if a diet rich in saturated and trans fatty acids, of which dairy products are our richest dietary source, has been associated with a higher risk of dementia [16], this association fails to explain the complexity of the observations in PD. Our hypothesis is that D-Gal may represent the missing link between milk and PD: in this review, we will describe the potential mechanisms through which D-Gal may exert this effect.
Methods
Following the observation that D-Gal is experimentally used to induce aging, oxidative stress, oxidative imbalance, mitochondrial dysfunction, neuroinflammation, neurogenesis inhibition and overexpression of p53 in animal and human tissue, we formulated the hypothesis that it could effectively explain the association between milk and Parkinson’s disease. We searched PubMed and Google Scholar with search terms related to each of the above-listed topics and selected 336 studies, relevant to our research. The studies were reviewed in order to test our hypothesis. In doing so, we tried to identify the biochemical and molecular mechanisms involved in the neurotoxicity induced by D-Gal, and compare them with the biological features of the familiar and sporadic forms of PD.
Results
Galactose-induced cell alterations
D-Gal is a reducing sugar. It is widely used to induce aging in liver, lungs and brain tissue both in human and animal models [17, 18]. D-Gal induces aging through several mechanisms, which are shown in Table 1, including: increase of IL-6, TNF and NF-kB, and lipid peroxidation; ROS production; enhanced excretion of 11-dehydroTXB [2]; increase in metyl-malonil aldehyde (MDA); reduction in superoxide dismutase (SOD) and glutathione (GSH); increase in serum receptors for advanced glycation end products (RAGEs); reduction in catalase activity; increased expression of cerebral beta amyloid and presenilin 1 precursors; increase in protein carbonyl in liver, brain and kidney; mitochondrial enzyme damage (complex I, II and III); decrease in pre- and postsynaptic proteins in the hippocampal area and cerebral cortex; increase in apoptosis [19–45]; increased expression in BAX, PUMA, Bxl-xle and caspase 3 [46–56]; increased levels of the oncogenic protein TP53 (Tumor protein 53 or p53) [57–61]. Other mechanisms for D-Gal-induced aging, which have been described in animal models, are shown in Table 2. They include: reduction of life span, cognitive impairment, increased oxidative stress and decreased total antioxidant activity, decreased immune response, increased production of AGEs, increased mitochondrial DNA mutations and mitochondrial dysfunction [31–35, 62–67], NF-kB mediated inflammation [66, 68–72]. FOXO3a and Nrf2 suppression, along with the inhibition of their antioxidant targets (SOD, CAT, GR, GSH and HO-1) was reported in one study where aging was induced by subcutaneous administration of D-Gal (150 mg/kg) [73]. Furthermore, D-Gal promotes oxidative phosphorylation of glutamine and glutamate (GLU) instead of pyruvate in the Krebs cycle [74]. D-Gal grown cells contain more mitochondria, consume more oxygen and are more susceptible to damage induced by mitochondrial toxins when compared to glucose (Glc) grown cells [75–78]. Finally, some authors hypothesized a possible role of the renin-angiotensin axis in the D-Gal and aluminum trichloride induced cerebral damage [79, 80]. Excess of D-Gal is converted into galactitol, which causes osmotic stress and ROS production [34], and a possible increase in oxidative stress caused by an unbalance in the redox system [81–83]. D-Gal amounts causing such effects are relatively low. Data show that 100 mg/Kg of D-Gal can accelerate aging in mice. Such a dose is easily met and exceeded in the typical Western diet, as it corresponds to the amount of D-Gal contained in 1–2 glasses of milk [84]. Michaëlsson K, et al. [85], observed increased mortality rates at moderate levels of milk consumption (greater than 1–2 glasses of milk per day for women, and 3 glasses per day for men). Data were obtained from 2 large population-based cohorts. The authors previously [84] reported increased oxidative stress and higher inflammatory markers in human serum and urine (IL-6, 8-iso-PGF2α) following high consumption of milk, and hypothesized that milk could induce oxidative stress through D-Gal. In experimental models, the age of the model and the duration of exposure can be key factors for D-Gal toxicity. Actually, some studies did not confirm the role of D-Gal in brain aging [86, 87], but mice used in these studies were very young (4–8 weeks), and exposure time was short (6–8 weeks). Both of these variables could have affected the results, as demonstrated by Hao L, et al. [23]. Galactosemic patients could provide a model for studying the effects of D-Gal overload, but, in this respect, very few studies are currently available. It is not possible, though, to translate these results obtained in animal and in vitro models, to humans. Further investigation of the effect of D-Gal on human tissues is needed.
D-Gal toxicity features
D-Gal toxicity features
D Gal potential contribution to PD
As anticipated, research suggests that AGEs, ROSs, mitochondrial dysfunction, neuroinflammation and GLU excitotoxicity can be all involved in D-Gal neurotoxicity.
AGEs and ROSs
Recent evidence associated ROSs, AGEs, and mitochondrial dysfunction to aging, cell death and to neurodegenerative disorders, like Alzheimer Disease (AD) and PD. AGEs are found in Lewy’s bodies in PD brains and alpha-synuclein (α-Syn), the main component of Lewy bodies, is susceptible to glycation, at least under experimental conditions. Glycation may represent an interesting therapeutic target in these diseases [88–94]. D-Gal is a reducing sugar and shows approximately ten times the glycation activity of glucose [20, 95–99]. ROSs are typically considered neurotoxic molecules: they oxidize essential macromolecules, like enzymes and cytoskeleton proteins, damaging them. More specifically, ROSs have been associated with cognitive impairment in animal models. Many studies confirm the association between ROSs and neurodegeneration [88, 100–104], and ROSs and PD [101, 105–110]. Nevertheless, also human studies, as well, show a possible role for oxidative stress in some cognitive impairment and neurodegenerative diseases, like AD or vascular dementia, thus supporting the hypothesis that oxidative stress can play an important role in the pathogenesis of neurodegeneration [110–118]. ROSs are involved in synaptic plasticity, memory deposition and in cognitive functions, at physiological concentrations. Unfortunately, the line separating these two opposite effects remains blurry. According to Massaad CA, Klann E [108], the balance between ROSs production and clearance is the critical factor in determining whether the beneficial effects of ROSs are outweighed by their ability to cause oxidative stress. The brain is highly enriched with fatty acids (mainly polyunsaturated fatty acids) and it is lower in antioxidant activity in comparison with other tissues; for example, it has only about 10% of the antioxidant capacity of the liver. Because of its elevated energy requirements, the brain produces very high levels of ROSs, compared to other organs and is, at the same time, extremely susceptible to excessive oxidative stress. Furthermore, ROSs are particularly active in the brain and neuronal tissues, since the metabolism of many excitatory amino acids and neurotransmitters generates ROSs. Glial cells and neurons, which are post-mitotic cells, are particularly sensitive to free radicals [88]. The increased ROSs production has been associated with some features of neurodegenerative diseases, such as accumulation of protein aggregates, increase in intracellular free Ca2 +, release of excitatory amino acids, autophagy and apoptosis. These impairments can lead to neuronal death, mediated by glutamate neurotoxicity (GNT), oxidative damage and apoptosis [119–123]. The binding of AGEs to their receptors (RAGEs) produces a cascade of events that leads to the activation of NF-kB, ROSs generation and a pro-inflammatory response [34, 124–126]. Dopaminergic neurons are generally considered extremely vulnerable to oxidative damage, caused not only during dopamine (DA) metabolism [6], irrelevant for other authors [127], but also by other features, related to the enormous axonal fields and the huge number of synapses in DA neurons. Unlike most neurons, relying exclusively on monovalent cation channels to drive pacemaking, DA neurons also engage type L-ion channels that allow Ca2+ to enter the cytoplasm, increasing intracellular Ca2+ concentrations. Intracellular Ca2+ accumulation has a metabolic cost, since Ca2+ must be rapidly sequestered or pumped back across the steep plasma membrane concentration gradient, and this process requires energy stored in ATP. This work implies an extraordinary enhancement of mitochondrial oxidative phosphorylation, which leads to an increased ROS production [127–129]. In at least two studies, oxidative damage was exacerbated by PARK7/DJ-1 deletion [129, 130].
Mitochondrial dysfunction and the role of genes
Research data supports the fundamental role of mitochondrial dysfunction in the pathogenesis of PD. Although the pathogenetic mechanisms are not fully clarified, evidence suggests that DA neurons are selectively vulnerable to mitochondrial dysfunction [131] and to mitochondrial DNA (mtDNA) deletions. It has been proposed that drugs and genetic approaches able to modulate the mitochondrial dynamics might have potential applications in the future of PD therapy [132–135]. Mitochondrial dynamics (fission, fusion, migration), as well as an appropriate mitochondrial turnover, are fundamental elements in synaptic homeostasis, neuronal survival and neurotransmission. Human and molecular studies on animal models of PD have shown that mitochondrial functional impairments occur in the early stages of PD, both in sporadic and familiar forms of the disease [127, 137]. Some of the PD associated genes, like dardarin (LRRK2), PINK1 (PARK6), FBXO7 (PARK15), Parkin (PARK2), DJ1 (PARK7) and possibly also CLU (clueless protein, human orthologue of “hClu”, which acts between PINK1 and Parkin) can contribute to mitochondrial dysfunction, with the subsequent increase of ROSs and DA neuronal damage. These events can lead to DA neuronal apoptosis and neuronal loss and then to the clinical manifestation of the disease. Mutations of DJ-1 or PINK1 and Parkin make cells more vulnerable to oxidative stress and mitochondrial toxins, like the ones involved in the sporadic forms of PD. These observations suggest a gene-environmental interaction in the onset and progression of the disease [136, 138–143]. Mitochondrial quality control is a fundamental step in cell homeostasis and neuroprotection [144]. PINK and Parkin, and probably also FBXO7, are central to mitochondrial quality control: they promote damaged mitochondria’s clearance via the autophagy pathway (mitophagy) [144–147]. Damaged mitochondria accumulate PINK1, which then recruits Parkin, resulting in ubiquitination of mitochondrial proteins [148]. DJ-1 e FBXO7 also appear to participate in the process, in different ways, and the p53/Bax pathway may be involved [146, 149–152]. An accumulation of mtDNA deletions was also observed in the substantia nigra of PD patients. This evidence suggests that mtDNA mutations, in some cases, could predispose DA neurons to death by the impairment of their respiratory chain. Additionally, a small portion of alpha-synuclein migrates in mitochondria and its accumulation, in the brain of PD patients, can impair the complex I activity [132, 153]. Moreover, the presence of mitochondrial complex I deficiency was demonstrated in the substantia nigra and platelets of PD patients [120, 155]. Mitochondrial dysfunction not only affects energy production in the cell, it also enhances ROSs production and oxidative stress, and dysfunctional mitochondria exacerbate apoptosis in case of cellular stress [6, 157]. Mitochondrial homeostasis processes act by the removal of damaged mitochondria and the generation of new mitochondria [158], so an imbalance in mitochondria turnover can have serious consequences on the fragile DA neurons. As previously described, one of the features of the D-Gal-induced cellular senescence (probably mediated by p53) is the decrease of Nrf2 expression. This transcription factor is considered an interesting therapeutic target in PD, as it may be able to slow or stop DA neuronal loss. Nrf2 up-regulation can be induced with different molecules, either synthetic or plant-derived, with physical exercise and genetic therapies. Nrf2 up-regulation was shown to protect DA neurons from experimentally-induced oxidative stress, for instance with Paraquat or MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine) [126, 159–168]. Nrf2, encoded by the gene NFE2L2, is a target of PGC-1α (PPAR/peroxisome proliferator activated receptor 1-α), a key regulator of mitochondrial biogenesis and function. Nrf-1 and Nrf-2 are important contributors to the increase in transcription of key mitochondrial enzymes, and they have been shown to interact with Tfam (transcription factor A, mitochondrial), which drives mitochondrial DNA replication and repair [158, 169]. Following their observations on neuroblastoma cell lines, Ivankovic D, et al. [148] proposed that these events can be triggered by PINK/Parkin-induced mitophagy and that they can contribute to the maintenance of the mitochondrial turnover, removing the aberrant organelles and replacing them with new ones. Since the nuclear translocation of Nrf2 appears to be inhibited in AD and enhanced in PD [170], it would be important to further investigate the variants of the NFE2L2-Nfr2 gene. In fact, some variants of this gene have been associated to a decreased efficiency of Nrf2 in conferring protection against oxidative stress to the cell [171–174]. Nevertheless, these theories require further confirmation [175], as the data could possibly be due to other co-participating factors, such as ethnicity, environment, diet or lifestyle. D-Gal is widely used to induce cell senescence and in this role it is considered responsible for the associated mitochondrial dysfunction, damage to the mitochondrial complex I, II and III, and mitochondrial damage on murine models of brain aging [32–35, 176–179]. D-Gal promotes oxidative phosphorylation, increasing the use of glutamine and GLU instead of pyruvate, to fuel the Krebs cycle [74]. Cells adapted to grow in D-Gal media produce more mitochondria, use more oxygen and appear to be more susceptible to mitochondrial toxicants compared to cells growth in glucose media. In fact, this sugar is widely used to detect and study mitochondrial dysfunction and to increase mitochondrial toxicity of certain substances [75–78, 181].
Neuroinflammation and microgliosis: The role of p53
Although the precise mechanisms remain to be clarified, neuroinflammation appears to be heavily involved in many neurodegenerative disorders, including PD. Neuroinflammation, with the associated increased levels of inflammatory cytokines, like IL-1β, IL-6 e TNF [182–184] and microglial activation [185, 186], is actually considered one of the main pathogenetic mechanisms in PD [182–184, 187–195]. The hallmark of neuroinflammation is the activation of microglia. The substantia nigra has one of the highest density of microglia in the brain [196, 197], so the nigrostriatal dopaminergic neurons are particularly vulnerable to inflammatory attack [198]. The key elements involved in neuroinflammation include: oxidative stress and mitochondrial dysfunction [188], p53 mediated TLR up-regulation [201–204], p53 mediated cMAF down-regulation [190], p53-NF-kB interplay, AGE-RAGE binding, periodontal disease (on the basis of systemic inflammation) [199], autoimmunity [189], genetic mutations involving immune system cells like LRRK2 [191], environmental pollutants [200], neurotoxins that can induce microgliosis and astrogliosis [194].
In vivo and in vitro studies show that D-Gal consistently increases p53 and p53 downstream gene expression (p21, Bax, Noxa, PUMA) [43, 214]. P-53 mediates overexpression of toll-like receptors (TLR) in human cells, increasing the inflammatory response [201–204]. These receptors, which play a critical role in pathogen recognition, once engaged, activate the NF-kB pathway, which, in turn, regulates the induction of pro-inflammatory cytokines like TNF, IL-1, and IL-6 [205]. When microglia and astrocytes’ TLR receptors are dysregulated, they can exacerbate CNS destruction [206]. Furthermore, p53 mediates microglial activation and pro-inflammatory phenotypes, by a downstream gene, miR 155, which leads a p53-driven microglia evoked neurotoxicity. These mechanisms have been demonstrated in vitro and in vivo, on murine models [207–212]. D-Gal induced neurotoxicity, in many studies, was associated with increased levels of inflammatory cytokines (IL-1β, IL-6- TNF), microgliosis and astrogliosis [66, 68–72], and D-Gal can also upregulate p53 [48, 213].
Glutamate (GLU)
According to evidence, the glutamatergic system and GLU toxicity are also involved in PD pathogenesis [187, 215–218]. GLU is the main excitatory neurotransmitter in the nervous system and is known to be neurotoxic when excessively released or incompletely recycled. There is some evidence that the loss of dopaminergic innervation in the striatum, due to the loss of nigral dopaminergic neurons, is associated with glutamatergic hyperactivity. This denervation could cause a cascade of functional modification in the basal nuclei activity [218, 219]. During inflammation, microglia, astrocytes and neurons can release GLU. This process could be implicated in neuronal death by glutamate neurotoxicity (GNT). TNFα has been found to induce the expression and release of glutaminase from neurons. Activated microglia can kill neurons by several mechanisms. Two of them are the release of GLU and glutaminase and the consequent excitotoxicity caused by the higher levels of extracellular GLU. This event can be prevented by NMDA blockers [187, 220].
Some studies also show that p53 can impair GLU uptake by astrocytes, contributing to GNT [221]. Interestingly, Bcl2, Bax and c-Myc proteins, and caspase-3 have shown to be involved in the GLU-induced cell death in cultured cortical neurons [224].
GNT, mitochondrial dysfunction and cell death
In GNT, mitochondrial dysfunction is considered a major pathogenic feature. GLU activity on NMDA receptors increases ROSs production and has been associated with mitochondrial hyperpolarization and fragmentation [225–228]. GLU inhibits cystine uptake via the cystine/GLU antiporter, leading to glutathione (GSH) depletion. GSH does not cross the Blood Brain Barrier, for which it must be synthesized inside the nervous system [5, 230]. Low GSH levels in the substantia nigra are a common feature in PD patients and have also been reported in preclinical stages of the disease. GSH depletion leads to ROSs accumulation and oxidative stress. GLU can, moreover, interact with different classes of postsynaptic receptors. Acute astrocyte neuronal swelling follows GLU binding to its NMDA receptor. This edema is due to the uptake of extracellular Na+ and Cl+, which causes membrane depolarization and the subsequent opening of the calcium channels. After exposure to GLU, mitochondria show impaired capacity to regulate Ca2+ homeostasis, mitochondrial dysfunction, mitochondrial fragmentation and the opening of the mitochondrial permeability transition pore (mPTP), this latter a primary mediator of cell death [231]. Oxidative stress, itself caused by ROSs accumulation, also perturbs the mitochondrial mechanisms that regulate Ca2+ participating in the opening of the mPTP and causing mitochondrial dysfunction and fragmentation [227, 233]. Inhibition of p53 activation by pifithrin-alpha, a p53 inhibitor, or cultured p53 null cells, are protected from mPT induction, NF-kB activation, and cell swelling [222, 223].
Cell death induced by GNT is partly mediated by overstimulation of the postsynaptic GLU receptor system, but also by a non-receptor mediated oxidative toxicity [228, 232]. Oxidative stress and mitochondrial dysfunction are considered primary events in GNT, even if the exact mechanisms need to be fully elucidated, and could involve different pathways [229, 232–234]. Inhibitors of some molecules, like PI3Kα, proved to be potent protectors against GLU toxicity, preventing ROSs generation, mitochondrial hyperpolarization, and lipid peroxidation in neurons [234]. GLU was shown to upregulate GSH-peroxidase-1 (GPx-1), Ca ion influx, NO production, cytochrome c release, Bax/Bcl-2 ratio, caspase-3 and 9 activity, ROSs, H2O2, mitochondrial fission markers and malondialdehyde; and to downregulate GSH, GLU reductase, SOD and catalase, resulting in oxidative stress and enhanced cell apoptosis [228, 235].
GLU and p53
Activation of striatal GLU receptors produced a significant elevation in p53 levels in striatal neurons [223]. Acetylsalicylic acid (ASA) and p53 antisense oligonucleotides experimentally abolished GLU-induced apoptosis [236]. These findings suggest that p53 induction may be linked to apoptosis caused by GNT.
DA release modulation by GLU; D-Gal and GLU
Dopamine (DA) release modulation by GLU is a controversial issue, but could contribute to the comprehension of the potential effects of D-Gal on the CNS. Some in vivo studies on animal models showed that GLU exposure increased DA release, but results are not consistent [237–243]. GLU effect on DA release was also evaluated by in vitro studies, using receptor-specific antagonists and implementing electric stimulus to evoke DA release. The AMPA receptor activation generates ROSs, like H2O2, which opens potassium channels inducing hyperpolarization and inhibiting the release of DA in the synaptic cleft. Accordingly, GLU could inhibit DA release by the elevation of H2O2, as happens in GNT [244]. GNT has been proved to be amplified under chronic lead exposure. PD has been associated with chronic exposure to heavy metals, among which, apart from lead [245–248], also mercury: this latter heavy metal could interfere with GLU uptake, increasing its release and lowering the threshold for cellular neurotoxicity [249–252]. Aluminum [253–256], cadmium [257] and other heavy metals might have similar effects. There are at least two mechanisms involved in protecting neurons from GNT: GLU uptake by EEAT (excitatory amino acid transporters), and GLU uptake and metabolism by astrocytes; details will follow. When the extracellular levels of this neurotransmitter are elevated (0.5–1.0 mM), GLU metabolism can be shifted towards oxidative deamination (by the enzyme glutamate dehydrogenase, or GDH) and the Kreb cycle. Dopaminergic neurons seem to express higher concentration of GDH and EAAT3 membrane proteins, and this could indicate a higher vulnerability of these cells to GLU toxicity [259]. Once taken up by the brain, D-Gal and Glc are partly converted into amino acids, mainly GLU, glutamine and GABA, by the transamination of alfa-ketoglutarate and oxaloacetate generated in the tricarboxylic acids cycle (TCA or Krebs cycle). The rate of this transformation is much higher in the brain than in the liver, where its rate is around 5% that of the brain. In experimental models, D-Gal, administered by various routes, is initially picked up by the liver, but, after a few hours, it is the brain to uptake it at a rate 20% higher than the liver, and 50% higher than skeletal muscle. This leads to a longer persistence in the brain of D-Gal metabolites (especially, GLU), as compared to Glc,. In fact, if 4 hours after Glc administration GLU family amino acids, are no longer detectable in the brain, their level is still detectable at the same time after the same dose of D-Gal [259]. D-Gal could increase GLU availability by upregulating p53 protein [57–61]. In fact, glutaminase 2 (GLS2) is upregulated by p53 [260–263]. It is also worth noting that in astrocytes exposed to D-Gal a decrease in the activity of astrocytic glutamine synthetase (GS) has been reported [264].
Astrocytes, PD and Gal
Astrocyte dysfunction has been associated with neurodegenerative diseases, such as PD [265–268]. Astrocytes play a critical role in brain homeostasis and neuron integrity. Astrocytes have more antioxidants than neurons do. In fact, the latter often rely on astrocytes for protection against oxidative stress. For example, astrocyte Nfr2-dependent ARE gene expression is higher than in neurons. Without close interaction with astrocytes, neurons would be much more vulnerable to oxidative stress and could not survive [269–271]. One of the major tasks of astrocytes is to protect neurons against excitotoxicity by taking up excess ammonia and GLU, and converting them into glutamine via the enzyme GS [272]. This reaction implies a considerable metabolic cost, with a drop of pH that extends to the mitochondria. This event knocks out the cytosol-to-mitochondrial matrix pH gradient, which slows down the mitochondrial activity [273]. Astrocytes uptake the extracellular GLU, control its release and the glutamine-glutamate cycle, removing excess ammonia thanks to the enzyme GS, an enzyme located mainly in astrocytes that catalyzes the condensation of GLU and ammonia to form glutamine [274]. In certain pathological situations, astrocytes and microglia can release considerable quantity of GLU and expose neurons to excitotoxic insult [275]. A decrease in GS activity results in an accumulation of GLU. A selective reduction, with no systemic dysregulation, of GS activity, was reported in PD patients [276]. Astrocytes are also involved in DA metabolism, cytokine release, pH regulation and many other metabolic processes [274]. In certain situations, astrocytes can participate in the inflammatory process, through the release of cytokines such as IL-6 and CCL2. Their true role, as well as the mechanisms involved, still need to be clarified [277–279]. In experimental models, D-Gal caused astrocyte degeneration [36], senescence and even apoptosis. D-Gal also caused a decrease in GS activity. Senescent astrocytes were more vulnerable to GNT [264]. Cui Q, et al. [265] using animal models of PD, showed that astrocytes inhibit the motor suppression striatal pathway. They showed that in mice, external globus pallidus astrocytes, which are regulated by DA, critically control the environmental level of GLU, which in turn gates striatopallidal transmission via the activation of presynaptic metabotropic GLU receptors. With a chronic loss of DA, the astrocytes lose the ability to regulate GLU levels. They found that in the models of Parkinson’s disease, astrocytes are dysfunctional. According to their research, dysfunctional astrocytes could contribute to the hypokinetic symptoms of PD by the overactivity of the striatal pathway. They suggest astrocytes could be used as a therapeutic target to relieve the motor symptoms associated with PD. Some other studies show microglia activation, and subsequent aggression to special regions of the brain, e.g. substantia nigra, by α-syn. This evidence suggests that astrocytes could play a co-causal role in PD pathogenesis. Astrogliosis and microgliosis are key features, yet non-specific, in PD neuronal degeneration. Astrogliosis and microgliosis can be induced, in animal models, by D-Gal exposure [36, 281].
Gal, SV2A, xCT and oxidative stress
Selective transporters for D-Gal, like SV2A, have been recently identified at a synaptic level. SV2A is the target for a second-generation antiepileptic drug (levetiracetam-LEV). SV2A acts as a proton-coupled symporter and could play a key role in the synaptic regulation [282, 283]. The latest studies reported that LEV influences xCT (cystine/GLU exchange transporter) expression in vivo. xCT is the light subunit of the two system xc-cystine/GLU antiporter subunits (fully functional also in the absence of its corresponding heavy subunit), or Sxc [230, 284]. xCT was previously mentioned in the GNT paragraph, since it plays a key role in cysteine availability for the synthesis of GSH. Miyazaki I, et al. [285] demonstrated the neuroprotective effect of LEV on cultured striatal astrocytes from hemi-parkinsonian mice through a substantial increase of the xCT expression and GSH availability. Therefore, it is possible to speculate on the role that D-Gal could play in the observed damage to astrocytes, as reported in previously mentioned murine models, via the SV2A receptor.
Galactose affects neurogenesis
Some studies showed impairment in hippocampal neurogenesis in patients affected by neurodegenerative disorders, including PD [286–289]. Chronic exposure to D-Gal in animal models decreases neurogenesis in the hippocampus of the adult brain [21, 290–293] and is experimentally used for this purpose [291]. The D-Gal inhibitory effect on neurogenesis could occur through the Notch1/p53 pathway. Notch-1 is a human gene encoding for a transmembrane receptor. The Notch-1 promoter is known to be involved in the development, asymmetric cell division, differentiation and proliferation of neurons [294–297]. Notch1 promoter contains p53 response elements and is normally expressed in this area of the brain. P53, or the tumor suppressor gene (TP53), plays a key role in cell growth and apoptosis, and it is the most extensively studied gene involved in human cancers. The mechanisms that lead to defective neurogenesis by accumulation of α-synuclein are still not fully clarified. Apparently α-synuclein up-regulates p53 expression, which in turn causes down-regulation of Notch1, and consequent inhibition of neurogenesis in the hippocampal gyrus dentatus. Studies show that p53 overexpression negatively affects neurogenesis in many areas of the brain during brain-development and that its suppression/deletion enhances neurogenesis in the same areas [300–303]. According to some researchers, this effect is mediated by ROSs production [304]. Inhibition of hippocampal neurogenesis was also reported in a murine model by some antidepressant drugs, by down-regulating the expression of p21, a p53 target gene [305]. As repeatedly mentioned before, D-Gal is associated with p53 overexpression [57–60]. The overexpression of p53, induced by D-Gal, could, possibly, explain the neurogenesis inhibition observed in the previously mentioned murine models.
D-Gal in food
As shown in Table 3, D-Gal is found in many food products, either animal or plant food. The main sources of this sugar, in the Western diet, are milk and some milk products, such as yoghurts and fresh cheeses. The amount of available galactose in food products other than milk and dairy, is very modest. D-Gal content in dairy products is undesirable [306] and depends mainly on the microbial species used in their production and on the ripening and aging times. There’s a lack of data on D-Gal content, free or bound to other monosaccharaides, in food products. Almost unknown is the bioavailability of the complexed D-Gal in food products other than dairy food. Some of them are “prebiotics”, like the bifidogenic galacto-oligosaccharides (GOS), which are D-Gal polymers [307]. D-Gal, complexed with several other molecules, is ubiquitous. Unfortunately, the bioavailability of alpha-1,6 and beta-1,3 linked D-Gal in food products is unknown, yet we know that beta-1,4 linked D-Gal can be released either in animal or in plant tissues. Fermented foods may contain free Gal. The content of the free Gal in food products other than milk and dairy products, for galactosemic patients appears to be negligible, apart from chickpeas which, according to some authors, have a moderate free Gal content [308]. Gal concentrations, in the same kind of dairy product, can differ depending on aging time and methods of production [309]. According to the literature, intake of fruit and vegetables does not seem to change Gal-1-phosphate measurements. This is probably due to the endogenous production of Gal, estimated to be 1,000–2,100 mg/day. Another possible explanation could be Gal bound with other indigestible components, in plant food, that prevents its systemic absorption [310].
The Free Galactose Content of Foods mg/100 g
The Free Galactose Content of Foods mg/100 g
Adapted from (Kim HO, 2007; Weese SJ, et al., 2003; Portnoi PA, MacDonald A, 2009; Portnoi PA, MacDonald A, 2015; Gross KC, Acosta PB, 1991; Abbot Nutrition; 2010; Portnoi PA, MacDonald A-bis., 2015; National Food Composition Database in Finland (Finali); ASIEM, 2010; Van Calcar SC et al., 2014.) [310–315].
Data shows surprising similarities between PD and D-Gal toxicity in rodents. Galactose is widely used to induce aging in in vitro and in vivo animal models, and to test mitochondrial toxins. How Gal causes organismal aging is poorly understood [18]. Adding galactose to in vitro medium leads cells to a metabolic shift from glycolysis to mitochondrial oxidation and the underlying mechanisms are only partly understood. Gal is metabolized by D-galactokinase and galactose-1-phosphate uridyltransferase, while at levels greater than normal, it can be oxidized into aldehydes and hydrogen peroxide by galactose oxidase. High amount of D-galactose may also lead to the production of galactitol, by aldose reductase [19]. Galactitol accumulates in cells, resulting in oxidative stress and cellular damage [34].
Gal also leads to structural changes in cells, especially concerning mitochondria (number, condensation, cristae) [326]. Nevertheless, its biological effects are not constant but related to some variables, particularly to the cell type [77]. However, one of the main effects of Gal is to increase mitochondrial oxidation and, according to available data, Gal can cause oxidative stress, oxidative imbalance, inflammation, decreased neurogenesis and cell death. Even if the bulk of data derives from animal models, there are a few documents, listed in Table 5, showing similar effects in cell lines derived from human tissues, which support the hypothesis that Gal could cause/enhance the above-mentioned pathogenetic mechanisms in humans.
D-Gal toxicity in human cell lines
D-Gal toxicity in human cell lines
Possible role of D-Gal in PD pathogenesis. In this figure, we represent the main pathogenetic events, currently studied and more detailed in Table 2, in PD and the potential role of D-Gal (represented as small yellow squares). According to the evidence, the main pathogenetic mechanisms in PD are: oxidative stress, neuroinflammation, loss of neurons due to increased DA neuronal death and decreased neurogenesis. In fact, both the familial and the sporadic forms of PD, and the D-Gal neurotoxicity, share many of the represented features: oxidative stress (mitochondrial dysfunction, decreased antioxidant proteins, increased ROSs production), neuroinflammation (increased IL-6, TNF, NF-kB, microgliosis), astrocyte dysfunction, loss of neurons (enhanced neuronal death, inhibited neurogenesis), AGEs, glutamate toxicity (GNT). Several of the genes associated with the familial form of PD are involved in mitochondrial homeostasis, particularly in quality control, a fundamental step in mitochondrial turnover. P53 seems to be involved in all pathogenetic pathways, and represents an interesting unifying theory.
Although the many features, markers and investigated mechanisms of D-Gal toxicity are complex and difficult to interconnect and explain, they do however display a common pattern: p53 up-regulation. The same observation applies to PD: p53 is up-regulated either in the familiar or the sporadic forms of the disease, and p53 up-regulation represent an intriguing unifying theory.
P53 up-regulation can cause oxidative stress with oxidative imbalance, inhibit antioxidant proteins and enhance ROSs production and mitochondrial dysfunction by altering mitochondrial membrane permeability via the activation of some pro-apoptotic downstream genes like Bax, Noxa and Puma. It can enhance the expression of microglial and astrocytic inflammatory response genes TLR, influencing, then, microglia and astrocytes polarization towards a more pro-inflammatory-phenotype via miR-155-cMAF [327], which leads to increased production of inflammatory cytokines (IL-1, IL-6, TNF). P53 can inhibit adult neurogenesis and also enhance α-syn synthesis via direct interaction with its promoter [328].
Gal could cause p53 up-regulation either indirectly -via oxidative stress- or possibly also via another unknown mechanism. Given the protective effects of antioxidants revealed by in vivo and in vitro data, it is likely that Gal induces p53 up-regulation via oxidative stress.
Furthermore: in Michaëlsson’s cohort [84, 85] and in Petruski-Ivleva’s study [15], Gal has been proposed as the possible explanation of the observed “milk adverse effect” on human health. In both cohorts, galactose’s toxicity would appear above 75–100 mg/kg (2–3 glasses of milk), which corresponds to experimental data. Interestingly, similar doses -more than 2 glasses of milk- also induced lower neuron density in substantia nigra in the Abbott cohort. In the latter study, this effect – attributed to heptachlor epoxide, a pesticide found in milk- was not observed in smokers. It is worth to note that nicotine was found to down-regulate p53, possibly via up-regulating survivin [329, 330].
Even if PD follows a defined clinical pattern, its pathogenesis remains essentially unknown. The neuronal loss in this disease appears to be caused by a multifactorial cascade of pathological events where environmental factors can play a significant role. Among the environmental factors, diet provides a wide range of elements able to protect from oxidative stress and inflammation, or enhance them, either directly (biochemical reactions) or indirectly (influencing genes expression).
In this review we described various mechanisms involved in the pathogenesis of Parkinson Disease. We can summarize them as follows:
increased advanced glycation end-products (AGEs) and their receptors (RAGEs) [91, 319]; mitochondrial dysfunction [5, 157]. Furthermore, some genetic mutations, commonly associated with PD (eg: PARK-1, PARK-6, LRRK2, Parkin, DJ1) predispose to mitochondrial dysfunction. Carriers of these mutations are more vulnerable to oxidants and more susceptible to agents that affect mitochondria. oxidative imbalance: overexpression of p53, Bax, Noxa, Puma, and decreased expression of antioxidant proteins [151, 321–323]; neuroinflammation, increased inflammatory cytokines (IL-6, TNF) [182–184, 188–195]; As illustrated in Table 4, D-Gal exposure induces similar changes and could contribute, directly or indirectly, to any of the listed mechanisms, namely: oxidative stress/mitochondrial dysfunction (particularly in the presence of the aforementioned PD associated genetic mutations) [75, 258]. oxidative imbalance: decreased GSH and antioxidant proteins [21, 36]; neurogenesis inhibition in the adult brain (mostly in hippocampal gyrus dentatus) [21, 293]; oxidative imbalance: overexpression of p53, Bax, Noxa, Puma [39–43, 213]; neuroinflammation, increased inflammatory cytokines (IL-6, TNF) [21, 68–72];
Comparison between the changes observed in PD vs induced/observed in D-Gal toxicity
Therefore, it is reasonable to suppose that D-Gal may interact with other genetic, dietary and environmental factors and enhance oxidative and mitochondrial damage, accelerating neuronal senescence, neuronal loss and the PD pathogenesis. These other factors include mitochondrial poisons, some medications (acetaminophen, statins, valproic acid, anthracyclines, etc.) [331–336], oxidant agents, heavy metals, GLU or agents capable of increasing their concentrations in the brain. They include also other molecules capable of impairing the redox system - e.g., influencing the expression of the antioxidant proteins – as well as chronic inflammation and suboptimal or inadequate intake of dietary antioxidants, stress, and so on. This interplay might have a considerable impact on individuals with a genetic vulnerability to PD and/or to oxidative and inflammatory stress. This hypothesis would support the claim that milk and dairy intake are associated with PD. Since the bulk of the available literature derives from animal studies, it is not possible to confirm the validity of these conclusions with respect to humans. Nevertheless, given the importance of this topic, further studies are recommended, to evaluate the impact of dietary milk and dairy on the development of human PD and on human organic senescence. In fact, if this hypothesis were confirmed, it would support a role for dietary interventions: reducing the intake of Gal and increasing the intake of protective foods and supplements, and could therefore help reduce the incidence and/or the severity of this disease.
Conflicts of interest
The authors declare no conflicts of interest.