Redox Homeostasis in Neural Plasticity and the Aged Brain

Abstract

Currently, humans can easily live for 60 years and more. This increase in life expectancy produces myriad changes in our bodies that diminish the individual’s physical and mental capacities and affect as well the functional capacity of individuals to interact appropriately with their social and physical environments. The oxidative theory of aging predicts an accumulation of oxidative damage to proteins, lipids, and DNA with age; as a consequence, the aged brain gradually suffers loss in neuronal functions, increasing the risk of developing neurodegenerative diseases and cognitive impairment. To date, there are no effective treatments to prevent age-related cognitive decline, making it urgent to identify the neural mechanisms that are altered during aging. In this chapter, we discuss the mechanisms that underlie synaptic plasticity, emphasizing the relationship between redox balance and neuronal function, and we also address current evidence supporting oxidative stress as an important contributing factor in brain aging.

1. Introduction

Currently, humans can easily live for 60 years and more. This increase in life expectancy produces myriad changes in our bodies that diminish the individual’s physical and mental capacities and that affect as well the functional capacity of the individual to interact appropriately with the environment. To date, there are no effective treatments to deter the age-related cognitive decline. Thus, in order to identify potential therapeutic targets and thus improve the quality of life of aging individuals it becomes urgent to decipher the neural mechanisms that are altered during aging. This knowledge will allow the design of molecules for the prevention, delay or even the reversal of age-related cognitive malfunction.

During aging, the brain undergoes a progressive accumulation of oxidative damage to macromolecules, such as synaptic proteins, lipids, and DNA, which gradually alters neuronal functions and increases the risk of developing neurodegenerative diseases and cognitive impairments.

As you would expect, various nutritional interventions that somehow prevent or counteract this oxidative accumulation, they have proved effective in mitigating the effects of brain aging. Although there is no consensus, the evidence suggests that the consumption of dietary antioxidants may have a beneficial effect on the mental health of aging individuals by a mechanism that involves improvement of synaptic plasticity (SP) [1], as well as an increase in the flow of blood and neurogenesis [2]. Caloric restriction, a significant decrease in caloric intake, is another nutritional intervention in animal models with positive impact in synaptic plasticity, which produces attenuation in the effects related to aging in the CA3 region of the hippocampus [3, 4].

In particular, aging entails synaptic failure, so to achieve a better understanding of the changes associated with the aging process and how to prevent, it is necessary to understand how it affects the mechanisms underlying synaptic plasticity.

This chapter discusses the following topics:

1. Cellular mechanisms of synaptic plasticity.

2. Redox regulation mechanisms acting in synaptic plasticity.

3. Disturbances of neuronal redox homeostasis and accumulation of oxidative damage during aging.

2. Cellular mechanisms of synaptic plasticity

Synaptic plasticity, defined as functional and structural changes occurring in the synapse in response to specific neuronal activity, is critical for the processing of information by the brain. It is widely accepted that changes in synaptic connections, which represent the cellular basis of memories, are encoded and stored in the central nervous system. At the cellular level, these changes occur when a postsynaptic neuron responds to a given presynaptic stimulation caused by either depolarization or neurotransmitter release. This postsynaptic response initiates a series of metabolic changes that promote, in turn, the stimulation of gene expression necessary to enable and consolidate long-lasting structural and functional changes. In its most general form, the hypothesis linking SP and memory states that neuronal activity-dependent plasticity is induced in certain synapses during memory formation. This fact is necessary and sufficient to store information, outlining the type of memory involved in the area of the brain in which plasticity is observed [5–7].

At the electrophysiological level, SP entails an increase or a decrease in synaptic efficiency; these changes are known as long-term potentiation (LTP) (Figure 1) or long-term depression (LTD), respectively [8]. If changes occur on the same synapse that was stimulated, they give rise to homosynaptic plasticity. Alternatively, if changes occur in synapses other than those stimulated, they originate heterosynaptic plasticity [9, 10]. The cellular and molecular mechanisms of LTP and LTD have been extensively studied in the hippocampus [11], an area involved in the formation of spatial memory in rodents and humans. In the hippocampus, many forms of plasticity, including LTP and LTD, require an increase in calcium concentration in the postsynaptic neuron [7, 12, 13].

Figure 1.

Cartoon depicting the signaling pathways involved in synaptic plasticity. In response to glutamate, NMDA (N-methyl-D-aspartate) receptors (NMDARs) and L-type voltage-dependent calcium channels (VDCCs) open allowing the influx of calcium into the cell. The increase in cytosolic calcium concentration activates the release of calcium from the endoplasmic reticulum (ER) by ryanodine receptors (RyRs) through a calcium-induced calcium-released mechanism (CICR) amplifying calcium signaling. In young synapses, this calcium elevation leads to the activation of calcium-dependent kinases, which are relevant to synaptic plasticity and learning and memory. NMDAR activation also induces the generation of ROS by the activation of NADPH oxidase. Mitochondrial-generated ROS can also contribute to the increase of calcium concentration through the modulation of the activity of NMDARs and RyRs. In aged synapses, the generation of ROS by mitochondria is exacerbated leading to oxidative stress. Ion imbalance, lipoperoxidation, DNA damage, and metabolic impairment are signatures of aged conditions. Increased density of VDCC and calcium-dependent potassium channels produce an increase of the slow component after hyperpolarization (sAHP), contributing to the aging process.

In the hippocampal CA1 area, activity-dependent postsynaptic calcium increments are initiated by calcium influx from the extracellular space through glutamate receptors of the N-methyl-D-aspartate (NMDA) type or through voltage-dependent calcium channels (VDCCs) (Figure 1). The resulting calcium signals are amplified and propagated through the calcium-induced calcium release (CICR) mechanism, which engages calcium release from intracellular stores [14]. Much of the experimental work concerning the possible role of LTP in learning has been focused on NMDA receptor (NMDAR)-dependent LTP [15–19]. These studies have shown that pharmacological NMDAR blockage impairs both learning and SP [16, 18–20]. Similarly, spatial memory is impaired in mice with a mutation in the NMDAR R1 subunit; this impairment correlates with alterations in LTP or LTD [21]. Hippocampal NMDAR-dependent SP includes classical associative long-term plasticity between the perforant/dentate gyrus pathway and between neurons in the CA3 and CA1 circuits; the coupling between the excitatory postsynaptic potentials (EPSPs) is associated with action potentials [22]. In addition, post-tetanic potentiation (PTP) and paired pulses facilitation (PPF) are two forms of short-term plasticity, which critically require NMDAR activation [23–25]. A number of NMDAR-independent SP also occur in the hippocampus, including LTP between the mossy fibers of the dentate gyrus and CA3 neurons; this plasticity is induced by the brief application of certain growth factors or neurotransmitters [26, 27].

At glutamatergic synapses, glutamate released from the presynaptic vesicles diffuses across the synaptic cleft and then binds to ionotropic receptors for α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) and to NMDAR present in the postsynaptic membrane. The activation of AMPA receptors (AMPARs) allows sodium influx and potassium efflux through these channels, causing excitatory postsynaptic currents and transient postsynaptic depolarization, which can be recorded as EPSP. This depolarization allows the removal of NMDAR blockage by magnesium ions (Mg²⁺), which occurs under resting conditions and prevents ion conduction through the channel. However, the strong depolarization is typically achieved during a train of high-frequency presynaptic activity or direct current injection to the postsynaptic neuron, release Mg²⁺ from the channel pore, leading to NMDAR opening and allowing calcium/sodium influx into the postsynaptic terminal [28].

Calcium ion (Ca²⁺) is a versatile cellular messenger engaged in numerous functions, including muscle contraction, apoptosis, cell growth, cell proliferation, synaptic plasticity, and gene expression regulation [29]. The extracellular Ca²⁺ concentration ranges from 1.5 to 2 mM, whereas the intracellular resting concentration is four orders of magnitude lower and ranges from 70 to 100 nM [30–32]. Increments in neuronal intracellular calcium concentration are of great physiological importance because they play an important role in neuronal functions such as neurotransmitter release, excitability, neuronal plasticity, and gene expression, all functions that are associated with SP and memory formation [30, 31, 33, 34]. In particular, postsynaptic calcium influx triggers amplification and signal propagation by CICR; this process involves the activation of intracellular calcium channels present in endoplasmic reticulum (ER) [14, 30, 34, 35], a continuous membranous network distributed throughout the neuronal cell, including dendrites and dendritic spines, the soma, and the surrounding the nucleus and extending into the axon to reach the presynaptic terminals, where it is closely associated with mitochondria [36] (Figure 1).

Two different types of intracellular calcium channels are involved in calcium release from the ER: the inositol 1, 4, 5-trisphosphate (IP3) receptor (IP3R) and the ryanodine receptor (RyR) channels [14, 30, 34, 35]. In areas such as the hippocampus, calcium release is mediated primarily through RyR [37] (Figure 1), providing the largest share of the increase of calcium in the postsynaptic spines. The RyR channel has been cloned, purified, and sequenced from several species. Three isoforms of this receptor have been identified: RyR1, RyR2, and RyR3, each one encoded by a different gene. All three RyR isoforms have about 5000 amino acid residues and present 65% of identity among them [38, 39]. RyR1 is the isoform primarily expressed in skeletal muscle, RyR2 in cardiac muscle, and RyR3 in diaphragm muscle [34, 40, 41]. All isoforms are expressed in the brain, RyR2 being the most abundant isoform [41–43]. At the anatomical level, RyR channels have been found in the frontal cortex, olfactory bulb, thalamus, amygdala, cingulate cortex, piriform cortex, entorhinal cortex, occipital cortex, and hippocampus [40–42, 44].

The ryanodine receptor is a homotetramer with a molecular mass of >2 MDa in which each subunit has a weight of ∽560 kDa [38, 39]. It has a large N-terminal segment located in the cytoplasm that represents about 90% of the protein, while its C-terminal is small and also facing the cytoplasm. The presence of a large N-terminal site allows its scaffold function for various proteins that play central roles in signal transduction pathways mediated by calcium [45]. The RyR channel can be activated by calcium, caffeine, 4-chloro-m-cresol (4-CMC), and ryanodine, the alkaloid from which its name comes. However, ryanodine has a dual effect on the receptor because at low concentrations (<1 μM) it activates the channel, whereas at high concentrations (>10 μM) it abolishes channel activity [37, 46]. RyR channels are also susceptible to inhibition by dantrolene and are modulated by adenosine triphosphate (ATP), H+, Mg²⁺, kinases, phosphatases, and reactive oxygen species (ROS) [38].

Different authors have shown that RyR-mediated Ca²⁺ signals have a role in SP, learning, and memory. In primary hippocampal cultures, RyR agonists such as ryanodine and caffeine generate calcium signals that trigger neuritic growth [47] as well as dendritic remodelling [43]. At the electrophysiological level, RyR inhibition by inhibitory concentrations of ryanodine suppresses sustained LTP, while ryanodine concentrations that activate RyR promote hippocampal LTP induction [48]. Moreover, RyR2/RyR3 expression in young rats increases after a hippocampal-dependent behavioral task [43, 44].

Using a genetic approach, it has been demonstrated that the absence of the RyR2 or the RyR3 isoforms, but not of RyR1, impairs spatial memory, while ryanodine concentrations that activate RyR channels promote memory consolidation in rats [39], mice [49], and chickens [50]. By contrast, the RyR inhibitor dantrolene alters spatial memory [35].

3. Reactive oxygen species in synaptic plasticity

As mentioned in the previous section, NMDAR activation is required for various forms of hippocampal plasticity, which produce a number of second messengers, including cAMP, nitric oxide, arachidonic acid, and, certainly, calcium. Recently, ROS have been added to the list of molecules generated upon activation of NMDAR [51, 52].

The role of ROS as synaptic transmission modulators is complex and depends on the concentration and duration of the oxidant stimulus [53, 54]. However, after LTP induction, neuronal ROS generation has been detected [54, 55]. This observation not only suggests that ROS may be necessary for LTP induction but also suggests that ROS generation is important for normal neuronal activity, as discussed subsequently.

Pharmacological activation of NMDAR induces the production of superoxide (O₂^.-) [51], which raises the question of whether ROS generation meets some physiological function. Ascribing to O₂^.- only a neurotoxic effect is not correct, because inhibition of O₂^.- production results in significant reduction of LTP induction [54, 55]. The NADPH oxidase complex is a recently characterized superoxide-generating source during NMDAR activation in neurons [52, 56]. This oxidase has been extensively characterized in immune cells [57, 58], but compelling evidence also suggests that superoxide anion generation mediated by this enzyme plays an important role in LTP induction. Incubation of brain slices with superoxide dismutase (SOD), an antioxidant enzyme that catalyzes the removal of superoxide, or with permeable and impermeable superoxide scavengers results in LTP attenuation [55, 59, 60]. Consistent with the above reports, hippocampal LTP is also affected in transgenic mice overexpressing Cu/Zn-SOD isoform [59, 61]. Thus, superoxide anion may be produced as a result of, or in conjunction with, other molecules necessary for LTP. Although superoxide has been implicated to SP-related kinases, such as protein kinase C (PKC) [62] and extracellular signal-regulated kinase (ERK) [63], its mechanism of action is not yet defined.

In addition to the modulating effect of superoxide on SP and memory, hydrogen peroxide (H₂O₂) also affects signaling pathways involved in SP [64]. For instance, H₂O₂ regulates the activity of several kinases, such as PKC and the mitogen-activated protein kinase (MAPK) family [65–67], as well as phosphatases such as calcineurin [66]. Accordingly, incubation of hippocampal slices with catalase, an enzyme that degrades H₂O₂ to water, exhibits altered hippocampal LTP, further supporting a role for H₂O₂ in SP [53, 68, 69].

Remarkably, another electrophysiological study reported conflicting results that are difficult to interpret [53], illustrating the complexity of redox signaling. Using hippocampal slices, it was found that a high concentration of H₂O₂ reduces synaptic responses, while exposure to an intermediate concentration has no apparent effect on the expression of pre-established LTP, but prevents induction of a new LTP. Surprisingly, a low concentration of H₂O₂ increased LTP expression compared to control (absence of peroxide). Another important effect of a low concentration of H₂O₂ is the suppression of LTD [53]. It seems that high concentrations of H₂O₂ can give rise to secondary oxidative reactions unrelated to a physiological response; by contrast, low concentrations of H₂O₂ can be part of the normal mechanism for LTP induction.

Within the context of ROS as cellular messengers, both Fenton and Haber-Weiss reactions are catalyzed by iron and involve superoxide anion and hydrogen peroxide. This feature implies that the mentioned reactions may have biological relevance [70]. While there are few studies that make a direct relation between iron-generated ROS and SP [71, 72], there are clinical data showing that iron deficiency during childhood has a strong impact on some cognitive functions related to SP and that this effect persists in the adult individual despite restoration of normal iron levels [73–75].

4. Redox imbalance in the aged brain

One of the most plausible theories of aging, which involves free radicals, was proposed by Harman in the mid-1950s [76]. This theory suggests that free radicals generated during aerobic metabolism produce cumulative damage in various cellular components, resulting in a loss of function characteristic in the physiology of organisms during aging. Subsequently, this hypothesis was refined including mitochondria as the main source of free radical production in physiological processes [77], as the inner membrane of mitochondria consumes nearly 90% of the oxygen generated by cellular respiration. According to this theory, ROS generated as by-products produce oxidative stress that damages the mitochondria itself, which results in dysfunctional mitochondria with the passing of years.

Particularly, the brain is sensitive to oxidative damage due to several factors. Among them, we can mention its mitochondrial high metabolic rate, the presence of high concentrations of polyunsaturated fatty acids, the presence of transition metals such as iron that are involved in the generation of the hydroxyl radical and the consequent lipid peroxidation, and, finally, a lower antioxidant capacity compared with other organs [78, 79].

This theory predicts that by controlling oxidative stress it is possible to delay the effects of aging. However, studies in Drosophila melanogaster, Caenorhabditis elegans, and mice, in which the expression of antioxidant genes was experimentally increased, did not achieve the expected impact on lifespan [80–82]. Furthermore, a recent study using post-mortem brains showed no significant differences in glutathione levels in aged individuals compared to younger ones, although a number of increments were found in some brain regions as the frontal and occipital cortex, caudate nucleus, and cerebellum [83]. The fact that this theory cannot fully explain the changes associated with aging shows how complex and multifactorial the aging process is.

Evidence supporting this theory in patients or in Parkinson’s and Alzheimer’s disease animal models suggests a decrease of glutathione and glutathione transferase activity with age in selected areas of the brain and ventricular cerebrospinal fluid [84]. In this regard, it has been described that ROS and oxidative damage may contribute to cerebral aging and also to a higher prevalence of neurodegenerative diseases. Thus, whereas diverse evidences correlate neuronal changes in redox status with progressive aging in the brain [85–90], others show that the cumulative effect of oxidative stress in neurodegenerative diseases may depend on the organism [91]. Multiple lines of evidence suggest that in the brain changes in the redox environment induce cognitive decline with age, by alterations in synaptic function or intracellular calcium regulation [92, 93]. Aging is associated with impairment in the ability to store, retain, and retrieve information, affecting declarative memory in humans, primates, dogs, and rodents [94–98].

Initially, the deficit in cognitive function associated with aging was partly attributed to a decrease in the number of neurons in the hippocampus [31, 36, 99], a critical region for spatial memory [100–102], or by diminished activity in the prefrontal cortex, the brain area involved in working memory, attention, and planning [95, 103]. However, subsequent studies reported that the loss of neurons does not contribute significantly to the cognitive decline associated with aging; rather, a decrease in synaptic connectivity has been linked to cognitive impairment [99, 104]. On the other hand, numerous evidences suggest that cognitive impairment associated with aging may be due to changes in gene expression [40, 102, 105–109], alteration of calcium homeostasis [32, 93, 110, 111], or changes in the redox state of the cell [14, 34, 86]. All these changes will trigger alterations in the LTP and LTD, two models of SP considered the cellular mechanism of learning and memory [112].

Intracellular calcium deregulation affects a number of synaptic components and calcium-dependent processes [31, 113]. The activity of NMDAR depends on neuronal redox state, so that variations of oxidative stress strongly impact both synaptic responses and NMDAR-dependent plasticity [88, 90]. This property is of great interest since the spatial memory impairment occurred in middle-aged animals is also dependent on NMDAR function [114]. Interestingly, manipulating ROS levels by genetic overexpression of antioxidant enzymes has revealed a direct relationship between the cognitive impairment in aged animals and the regulation of NMDAR [89].

Another synaptic process affected by aging is the increase in the slow component of the Ca²⁺ activated K⁺- after hyperpolarization (sAHP) in neurons of the hippocampal CA1 region in aged rodents [115–117]. While it has been described that the sAHP can be modified by an increase in L-type voltage-dependent calcium channels [102, 105] or an increase in calcium-dependent potassium channel density [118], enhanced calcium release from intracellular compartments could also make a significant contribution [85, 111, 119]. In fact, among these factors, the latter becomes crucial because oxidative stress increases during aging and RyR channels are highly sensitive to cellular redox state, increasing their activity in response to oxidation [34, 120, 121].

Certainly, further research is needed in the search to understand and combat the deleterious effects of aging; knowledge gained from these studies should help the development of effective treatments for brain pathologies associated with aging.

Acknowledgments

I like to thank Dr. Mario Párraga for his valuable comments and review of the manuscript.