|Home | About | Journals | Submit | Contact Us | Français|
It is now well established that brain plasticity is an inherent property not only of the developing, but also of the adult brain. Numerous beneficial effects of exercise, including improved memory, cognitive function and neuroprotection, have been shown to involve an important neuroplastic component. However, whether major adaptive cardiovascular adjustments during exercise, needed to ensure proper blood perfusion of peripheral tissues, also require brain neuroplasticity, is presently unknown. This review will critically evaluate current knowledge on proposed mechanisms that likely underlie the continuous resetting of baroreflex control of heart rate during/after exercise and following exercise training. Accumulating evidence indicates that not only somatosensory afferents (conveyed by skeletal muscle receptors, baroreceptors and/or cardiopulmonary receptors), but also projections arising from central command neurons (in particular peptidergic hypothalamic preautonomic neurons) converge into the nucleus tractus solitarii (NTS) in the dorsal brainstem, to coordinate complex cardiovascular adaptations during dynamic exercise. This review focuses in particular on a reciprocally interconnected network between the NTS and the hypothalamic paraventricular nucleus (PVN), which is proposed to act as a pivotal anatomical and functional substrate underlying integrative feed-forward and feed-back cardiovascular adjustments during exercise. Recent findings supporting neuroplastic adaptive changes within the NTS-PVN reciprocal network (e.g., remodeling of afferent inputs, structural and functional neuronal plasticity, and changes in neurotransmitter content), will be discussed within the context of their role as important underlying cellular mechanisms supporting the tonic activation and improved efficacy of these central pathways in response to circulatory demand at rest and during exercise, both in sedentary and trained individuals. We hope this review will stimulate more comprehensive studies aimed at understanding cellular and molecular mechanisms within CNS neuronal networks contributing to exercise-induced neuroplasticity and cardiovascular adjustments.
It has long been thought that brain plasticity, defined here as functional or structural changes that occur in response to perturbations in the external environment or internal milieu, was a phenomenon limited to critical periods during development. During the past decades, however, it has been widely recognized that activity-dependent brain remodelling also occurs in the adult, mature brain, giving rise to the concept that plasticity is in fact an inherent property not only of the developing, but also of the adult brain (Pascual-Leone et al., 2005). Typical situations under which brain plasticity has been observed in the mature brain include novel experiences and exposure to rich environments. A characteristic example is the functional and structural synaptic plasticity associated with long-term potentiation, a critical cellular mechanism for memory storage in the mammalian brain (Engert & Bonhoeffer, 1999; Luscher et al., 2000).
An increasing body of evidence supports that physical activity and exercise training can also induce remarkable functional and neuroanatomical plasticity in the mature brain, including neurogenesis, angiogenesis, synaptic plasticity and dendritic morphological remodelling. Exercise-induced neuroplastic changes are thought to play critical roles in mediating important beneficial effects associated with physical activity, including improved memory, cognitive function and neuroprotection, among others (see (Dishman et al., 2006; Cotman et al., 2007; Mueller, 2007; Draganski & May, 2008; van Praag, 2008) for recent reviews on this topic). In addition, exercise training has been shown to induce important adaptive and beneficial autonomic and cardiovascular adjustments, in order to ensure proper blood perfusion of peripheral tissues according to metabolic demands (Ludbrook & Graham, 1985; Mitchell, 1990; Rowell & O’Leary, 1990; Michelini & Morris, 1999; Michelini, 2001). Whether exercise-induced cardiovascular and autonomic adjustments also involve neuroplasticity as an underlying mechanism, is at present not fully elucidated. In this review, we will provide a summary of current knowledge on neural control of the circulation during exercise, and discuss recent findings from our laboratories and others in the field, supporting central nervous system neuroplasticity as a likely mechanism underlying autonomic and cardiovascular adjustments during exercise.
Rapid adjustments of cardiovascular function, essential for the maintenance of body homeostasis, are mediated by the autonomic nervous system, and depend on neural reflexes primarily integrated within the brainstem (Dampney, 1994). Receptors in the cardiovascular system (intrinsic as well as extrinsic) detect changes in pressure, volume, flow, blood gases, pH, temperature, movement, etc, and send this peripheral sensory information to the brain for neural processing. Among the intrinsic mechanisms controlling cardiovascular function, the arterial mechanoreceptors (also known as baroreceptors) are important controllers implicated in maintaining short-term stability of arterial pressure. Baroreceptors detect vessel stretch caused by pressure increases, codifying this information into bursts of action potentials (proportional to pressure changes), which are then sent to the nucleus tractus solitarii (NTS) in the dorsal brainstem, the primary termination site for the cardiovascular afferents. Second order NTS glutamatergic neurons (as well as higher order NTS neurons) excite both parasympathetic preganglionic cell bodies (in the dorsal motor nucleus of the vagus and nucleus ambiguous) and the sympathetic cell groups in the caudal ventrolateral medulla (CVLM, GABAergic inhibitory neurons) that project to, and inhibit rostral ventrolateral medulla (RVLM) glutamatergic neurons, thus decreasing sympathetic preganglionic neuronal outflow. During unloading of baroreceptors, increased sympathetic and decreased parasympathetic outflows (regulated by this simple reflex loop integrated within the brainstem), cause a marked bradycardic response, diminished cardiac output, increased venous capacitance and decreased total peripheral resistance (Palkovits, 1988; Dampney, 1994; Sved & Gordon, 1994). Opposite autonomic effects, i.e. reduced parasympathetic and increased sympathetic drive, are observed during baroreceptors’ loading.
Reflex control of the circulation is also critical during exercise activity. In order to maintain proper blood pressure, and to efficiently adjust blood supply to different vascular beds according to regional metabolic demands, adaptive responses to acute exercise mainly involve a moderate increase in blood pressure accompanied by a marked tachycardia (Ludbrook & Graham, 1985; Michelini & Morris, 1999). Current knowledge on circulatory control during exercise underscores the coexistence of two main neural mechanisms (Mitchell, 1990; Rowell & O’Leary, 1990; Waldrop et al., 1996): (a) a feed-forward mechanism, so-called “central command”, which involves higher brain centers such as the motor cortex, hypothalamic and mesencephalic locomotor regions that activates parallel circuits controlling locomotor, cardiovascular and ventilatory functions, and (b) a feed-back control mechanism, driven by receptors from cardiovascular areas and active muscles. For example, we and others have shown that dynamic interactions between these feed-forward and feed-back circuits are associated with beneficial adjustments in both parasympathetic and sympathetic tone to the heart and in the sympathetic outflow to vessels, enabling the co-existence of exercise tachycardia in the presence of moderately increased pressure levels (Ludbrook & Graham, 1985; Mitchell, 1990; Rowell & O’Leary, 1990; Michelini & Morris, 1999; Michelini, 2001) and the appearance of resting bradycardia in trained individuals (Clausen, 1977; Scheuer & Tipton, 1977; Negrao et al., 1992a; Negrao et al., 1992b; DiCarlo & Bishop, 2001).
Accumulating anatomical and functional studies support a reciprocal interconnectivity between the NTS and the hypothalamic paraventricular nucleus (PVN) as a pivotal neuronal circuit underlying integrative autonomic adjustments of cardiovascular control during exercise (Michelini & Bonagamba, 1988; Michelini, 1994; Dufloth et al., 1997; Michelini & Morris, 1999; Braga et al., 2000; Michelini, 2001; Jackson et al., 2005; Martins et al., 2005; Higa-Taniguchi et al., 2007; Michelini, 2007a, b). More recently, work from our laboratories, as well as others, have started to investigate whether neuroplastic mechanisms within this circuitry contribute to cardiovascular adaptive changes in trained individuals. This work is summarized and discussed in this review. It is important to acknowledge that while our work specifically focuses on the hypothalamic PVN, other hypothalamic areas, in particular the dorsomedial hypothalamus (DMH) and the perifornical area (areas also known as the “hypothalamic defense area”), have been shown to play important roles in the control of baroreflex function, particularly during conditions of arousal and alerting responses (Hilton, 1966; Coote et al., 1979; Dampney et al., 2005).
It is well established that a rich interconnectivity between the brainstem and the hypothalamus constitutes a critical neuronal substrate for the central integration of viscerosceptive and somatosensory information, which in turn trigger the generation of complex patterns of autonomic and cardiovascular homeostatic responses (Palkovits, 1999) In this sense, peripheral information conveyed by the baroreceptors to the NTS (Palkovits, 1988; van Giersbergen et al., 1992; Dampney, 1994; Sved & Gordon, 1994) is transmitted directly and/or indirectly via NEergic projections to higher brain centers, including the amygdala, cortex, and the hypothalamus, in particular, the PVN (Sawchenko & Swanson, 1981; Kalia et al., 1985; Palkovits, 1988). These sensory hypothalamic NEergic afferent inputs originate primarily from the A1 and A2 catecholaminergic cell groups, located in the NTS and caudal ventrolateral medulla, respectively (Sawchenko & Swanson, 1982b). Within the PVN, the ascending NEergic afferents innervate and modulate the activity of magnocellular and parvocellular neurosecretory neurons (Swanson & Kuypers, 1980; Swanson & Sawchenko, 1980), as well as preautonomic neurons (Sawchenko & Swanson, 1982b; Liposits et al., 1986; Daftary et al., 2000; Higa-Taniguchi et al., 2007; Yang et al., 2008). Preautonomic PVN neurons have been shown to play an important role in the generation and coordination of homeostatic autonomic responses (Swanson & Sawchenko, 1980; Palkovits, 1999). This neuronal population is neurochemically heterogeneous, being oxytocin (OT) and vasopressin (VP) the most predominant peptides found (Swanson & Kuypers, 1980; Sofroniew & Schrell, 1981; Sawchenko & Swanson, 1982a). In addition to directly innervating preganglionic neurons in the spinal cord, preautonomic PVN neurons project back to the brainstem, to influence preganglionic neurons (e.g., in the dorsal motor nucleus of the vagus (DMV) and nucleus ambiguous) either directly, or indirectly, in the latter case through innervation of intermediate catecholaminergic neurons (e.g., A5 and C1 cell groups) (Luiten et al., 1985; Palkovits, 1999). Importantly, these peptidergic, medullary-projecting PVN neurons also directly innervate the sensory NEergic neurons in the A1 and A2 cell groups that provide viscero-somatosensory inputs to the PVN (Swanson & Kuypers, 1980; Luiten et al., 1985; Toth et al., 1999), indicating that PVN neurons are reciprocally connected with catecholaminergic neurons in the dorsal (A2) and ventral (A1) brainstem.
Recent studies from our laboratories have shown that baroreceptor reflex control of heart rate, as well as the adjustment of tachycardic response during exercise, could be reset and modulated, respectively, by this reciprocal interconnectivity between the NTS ® PVN (NEergic), and PVN ® NTS (OTergic, VPergic), acting thus as an anatomical substrate for a prompt feedback control loop for cardiovascular adjustments during exercise. In the following sections, we will summarize recent work supporting the importance of the NTS-PVN-NTS suprabulbar circuit in the reflex autonomic control of the cardiovascular system, as well as studies addressing whether and how functional and structural remodeling within this interconnected circuit contributes to adjustment of baroreceptor function during dynamic exercise and exercise training.
The effect of exercise on the NTS-PVN-NTS pathway, and its influence on cardiovascular adjustments during exercise, is currently being investigated in our laboratories using a comprehensive, multidisciplinary approach. In a first set of studies, we aimed to determine whether acute exercise and/or exercise training affected OTergic and VPergic drive to the dorsal brainstem (DBS, comprising the NTS and DMV). To this end, we combined RT-PCR, in situ hybridization, immunohistochemistry and radioimmunoassay for measurements of VP and OT expression / content in brains from sedentary and trained rats obtained at rest, or immediately after an acute bout of exercise on a treadmill (Martins et al., 2005). When compared to sedentary controls, trained rats exhibited a robust increase in hypothalamic OT mRNA expression (Figure 1B), without any changes in OT receptor expression within the NTS/DMV area. OT immunoreactivity within the NTS/DMV area was also markedly increased after training (Figure 1A). Although we could not detect parallel changes in VPergic terminals, we observed a marked increase in the sensitivity of V1 receptors to the endogenous ligand in the NTS following exercise training (quantitative autoradiography, (De Souza et al., 2001). In addition, trained rats performing an acute bout of exercise showed a significant increase in both OT and VP peptide content within the DBS (Dufloth et al., 1997; Michelini & Morris, 1999; Braga et al., 2000; Michelini, 2001). In sedentary rats, only a tendency for increased VP content in the DBS was observed after an acute bout of exercise (Michelini, 2007a). Given that the sole source of OT and VP in the DBS originates from descending PVN projections (Buijs et al., 1978; Sofroniew & Schrell, 1981; Sawchenko & Swanson, 1982a), our results supports specific activation of these preautonomic PVN-DBS projecting neurons during during an acute bout of exercise and after exercise training.
To shed light into the functional significance of exercise-induced increases on OTergic and VPergic input to the DBS, we compared exercise-induced tachycardia and pressure responses before and after microinjection of OT and VP receptors antagonists/agonists within the NTS/DMV, in conscious sedentary and trained rats running on a treadmill (Dufloth et al., 1997; Braga et al., 2000). NTS OT receptor blockade caused a marked increase on the tachycardic response at all exercise intensities only in the trained group without changes in sedentary controls (Braga et al., 2000) (Figure 2A,B). On the other hand, NTS V1 receptor blockade reduced the magnitude of exercise tachycardia in both groups with larger effect on trained rats (Figure 2C,D) (Dufloth et al., 1997). These functional studies are in agreement with our measurements of DBS peptide content, and further support the activation of PVN-NTS descending peptidergic pathways during an acute bout of exercise (Dufloth et al., 1997; Michelini & Morris, 1999; Michelini, 2007a).
The effects of OTergic and VPergic drives on baroreceptor reflex control of heart rate were also investigated. Administration of sub-pressor doses of OT within the NTS in conscious sedentary rats at rest (mimicking activation of OTergic pathways to this area), improved the gain of reflex bradycardia, and displaced the inferior plateau towards a lower heart rate value, thus increasing the operational range of the reflex (Higa et al., 2002). Conversely, blockade of OT receptors into this area displaced the lower plateau towards high heart rate values, and significantly decreased both the gain and the operating range of the reflex (Higa et al., 2002). Further supporting a role for OT within the NTS, Peters et al (2008) in a recent study provided evidence that OT released from PVN axons acts on a subset of second-order neurons within the medial NTS to enhance visceral afferent transmission. OT actions appeared to be mediated by both presynaptic (increased release probability of glutamate), and postsynaptic (OT-sensitive inward current) mechanisms (Peters et al., 2008).
Differently from OT, sub-pressor doses of VP given into the NTS of sedentary rats at rest did not change the sensitivity of reflex bradycardia, but displaced the set-point of the reflex towards higher heart rate values (Michelini & Bonagamba, 1988). Importantly, similar changes in baroreflex function were observed in sedentary animals doing dynamic exercise (Rowell & O’Leary, 1990; DiCarlo & Bishop, 1992, 2001), in response to stimulation of mesencephalic locomotor region or skeletal muscle receptor afferents (McIlveen et al., 2001; Degtyarenko & Kaufman, 2005; Potts, 2006), or after exercise training (Negrao et al., 1992b; Krieger et al., 2001). We also found that VP within the NTS opposes the sympathetic withdrawn during transient pressure increases (Michelini, 1994, 2007a). Based on these observations, we proposed that PVN-NTS VPergic projections activated during exercise act to restrain afferent information conveyed by the baroreceptors, resulting in a reduced bradycardic response for a given pressure increase (Michelini, 1994; Michelini & Morris, 1999).This notion is also supported by an elegant in vitro study by Bailey et al. (Bailey et al., 2006), showing that solitary tract-evoked excitatory post-synaptic currents (EPSCs) recorded on second-order NTS neurons was diminished by VP administration, supporting a VP-mediated inhibition of baroreceptor afferent transmission within the NTS.
Recently, we also found exercise training to improve the operational range and the gain of the baroreflex (Ceroni et al., 2009), displacing the set-point towards higher values during and acute bout of exercise. While in these studies we did not test directly the role of OT and VP in mediating such changes, the results are in line with the studies summarized above in sedentary rats. Thus, it is reasonable to speculate that OT and VP effects on baroreflex function are similar between sedentary and trained animals (Ceroni et al., 2009). In summary, these studies suggest that PVN-NTS VP and OT projections are tonically active, influencing in opposite, though complementary fashion, baroreceptor signaling, reflex control of the heart, and necessary autonomic adjustments during exercise-induced pressure increase. Thus, while VPergic inputs resets reflex control toward high pressure and heart rate values, facilitating the appearance of exercise tachycardia in sedentary and trained individuals (Michelini & Bonagamba, 1988; Michelini, 1994; Dufloth et al., 1997; Michelini, 2007a), OT inputs facilitate the slowdown of the heart for a given pressure increase, specifically in trained individuals (Braga et al., 2000; Michelini, 2001; Martins et al., 2005; Michelini, 2007a) (see Table 1). This complementary and balanced increased activity of the excitatory (VP) and inhibitory (OT) stimuli is thought to be necessary to improve the efficacy of the cardiovascular system in trained individuals, to better adjust to physiological demands during exercise.
As summarized above, numerous studies from our laboratories, as well as others in the field, support an important role for brainstem and hypothalamic neuronal circuits in cardiovascular adaptation during exercise training. However, the precise underlying mechanisms contributing to these adaptive responses remain largely unknown. Physical activity-induced brain neuroplasticity has been shown to contribute to major beneficial effects associated with exercise, including improved memory, cognitive function and neuroprotection ((Dishman et al., 2006; Cotman et al., 2007; Mueller, 2007; Draganski & May, 2008; van Praag, 2008)). However, whether similar mechanisms also underlie optimization of the NTS-PVN-NTS reciprocal network during exercise is incompletely understood.
One of the strongest evidence for structural-functional reorganisation in the adult brain originates from studies in the PVN and SON nuclei, in particular with respect to the magnocellular neurosecretory system. Under conditions of high hormonal demand, such as dehydration and lactation, the magnocellular neurosecretory system undergoes a robust, dynamic and reversible reorganization. This activity-dependent plasticity, which includes rearrangements of neuronal-glial microenvironments, re-wiring of afferent synaptic inputs, changes in intrinsic membrane properties, and reorganization dendritic geometries, among others, serves to optimize network activity and hormone release during these challenging physiological conditions. Interestingly, the ability of these hypothalamic centers to display such a dramatic plasticity in the developed brain, seems to be dependent on their ability to retain immature molecular features, including a polysialylated form of neural cell adhesion molecule, which has been associated with similar neuroplastic events during development and regeneration. This topic has been the subject of excellent reviews, to which the readers are referred for more detailed information (Armstrong & Stern, 1998; Hatton, 2004; Theodosis et al., 2004).
In addition to contribute to adaptive responses during physiological conditions, the involvement of hypothalamic neuroplastic mechanisms in maladaptive responses during pathological conditions has also been demonstrated. An exacerbated activity within the PVN for example, has been shown to contribute to increased neurohumoral drive in various disease conditions, including heart failure, hypertension and diabetes (Li & Patel, 2003; Zucker et al., 2004; Dampney et al., 2005; Zheng et al., 2006). While the precise underlying mechanisms are yet to be elucidated, emerging evidence supports an important contribution of neuroplastic mechanisms, including afferent input remodelling (Biancardi et al., 2009), changes in ion channel activity (Sonner et al., 2008), and expression and function of neurotransmitter receptors (Li & Pan, 2006; Li et al., 2008). These changes are thought to enhance, in a maladaptive and imbalanced manner, PVN neurotransmitter efficacy and neuronal excitability, ultimately resulting in an abnormally elevated neurohumoral outflow during these conditions. Importantly, manipulations that improve neurohumoral drive during hypertension or heart failure, in particular exercise (Zheng et al., 2005; Kleiber et al., 2008), modify the efficacy of these neurotransmitter systems within the PVN, alleviating in turn the elevated neurohumoral drive found in these conditions. These studies suggest then that exercise training beneficial effect may be dependent on its ability to reverse maladaptive neuroplastic changes observed in these disorders. This raises the interesting question as to whether physical activity per se could also induce neuroplastic changes within the NTS-PVN-NTS network, contributing in turn to physiological cardiovascular adaptive responses after exercise training.
A potential mechanism underlying enhanced activity of the PVN-NTS pathway after exercise training involves an increased membrane excitability of the output neurons within this circuit. To address whether this was the case, we obtained electrophysiological patch-clamp recordings from retrogradely-labeled PVN-NTS neurons in brain slices obtained from sedentary and exercise-trained rats (Figure 3A, (Jackson et al., 2005). Our results support an increased intrinsic excitability of PVN-NTS neurons in exercise-trained rats. Briefly, action potential recorded from PVN-NTS neurons in exercise-trained rats were larger in amplitude and displayed faster rise times than those recorded from sedentary rats (Jackson et al., 2005). (Figure 3B). More importantly, we found the input-output function of PVN-NTS neurons (which reflects their ability to discharge action potentials in response to varying degrees of neuronal inputs) to be significantly enhanced in exercise-trained rats. PVN-NTS neurons were able to fire a higher number of action potentials per stimulus, as compared to sedentary rats, showing little or no adaptation as a function of intensity of the stimulus (Jackson et al., 2005) (Figure 3C). Despite changes in neuronal surface area observed in PVN-NTS neurons during exercise (see 5.3 below), we found no changes in neuronal input resistance. Thus, the increased ability of PVN-NTS neurons to fire repetitively in response to depolarizing pulses during exercise was not a direct consequence of a change in input resistance, but rather a diminished expression of after-hyperpolarizing potential (AHPs) during this condition (Jackson et al., 2005). The AHP results from the progressive accumulation of intracellular Ca2+ and subsequent activation of Ca2+-dependent SK potassium channels (Greffrath et al., 1998), and is a key membrane property influencing spike frequency adaptation in PVN neurons during repetitive firing (Bains & Ferguson, 1998; Stern, 2001). Thus, exercise-induced changes in PVN-NTS repetitive firing properties could be brought about by multiple combined mechanisms, including a reduction in intracellular Ca2+ accumulation during spiking, a reduction in the Ca2+ sensitivity of the Ca2+-dependent SK channels, and/or intrinsic changes in the SK channels per se. These are important mechanistic questions that deserved to be further explored. Interestingly, changes in the activity/density of SK channels have been recently shown to contribute to physiological optimization of neurosecretory oxytocin neurons during lactation (Teruyama & Armstrong, 2005), as well as to abnormally enhanced preautonomic PVN neuronal excitability during hypertension (Chen et al., 2009). Thus, these studies indicate that changes in the expression/function of an individual ion channel within hypothalamic networks could be part of both adaptive and maladaptive neuroplastic mechanisms.
Exercise-induced increments in PVN membrane excitability appeared to be cell-specific and related to preautonomic PVN neurons, since these changes were absent or reduced in magnocellular neurosecretory neurons (Jackson et al., 2005) (Figure 3B and C). These results are in accordance with our previous observations that low intensity exercise training did not change circulating levels of VP and OT, causing only a small reduction in plasma levels of OT after an acute bout of exercise in the trained rats (Dufloth et al., 1997; Braga et al., 2000). The enhanced input-output function observed in PVN-NTS neurons during exercise training will likely be translated into an improved stimulus-secretion coupling at the brainstem terminals (Poulain & Wakerley, 1982; Bicknell et al., 1988), resulting in turn in an enhanced peptide release during activation of the descending pathways. In summary, these studies support an increased excitability of PVN-NTS neurons during exercise training, and suggest that this could be a mechanism contributing to enhanced activation and peptide secretion from VPergic and OTergic terminals during exercise training.
Activity-dependent neuronal structural plasticity, including changes in dendritic size and branching patterns, constitutes another mechanism by which neuronal networks regulate their integrative and output properties (Rall, 1959; Mainen & Sejnowski, 1996). We have recently completed a set of studies that support that dendritic remodeling also occurs in the PVN-DVC pathway during exercise training. Retrogradely-labeled PVN-NTS neurons were intracellularly filled with biocytin, and reconstructed in three-dimensions. Morphometric data was collected from neurons obtained from both sedentary and trained rats. As summarized in Figure 4, we found neuronal surface area of PVN-NTS neurons to be increased in exercise-trained rats. Moreover, the total number of dendritic trees, as well as the number of dendritic branches was enhanced after exercise training, when compared to sedentary rats. Changes in the size and branching patterns of the dendritic trees could effectively alter the electrotonic, and thus, the integrative properties of PVN-NTS neurons during exercise training. For example, an expanded dendritic tree would increase the possibility for synaptic integration, e.g., synchronous remote inputs have a better opportunity for linear summation at the soma. Moreover, an enlarged dendritic area and branching could facilitate the formation of new synaptic inputs (see below). Thus, it is possible that as a consequence of exercise training, PVN-NTS neurons undergo both functional and structural plasticity, contributing in turn to the increased activity and efficacy of this descending pathway during exercise. It is important to mention that exercise training-induced dendritic plasticity has been also reported in other brain regions, including the hippocampus, cerebellum and cardiorespiratory brain centers (Nelson & Iwamoto, 2006; Dietrich et al., 2008).
Similarly to other CNS neurons, neuronal activity in PVN-NTS neurons results from the combined actions of intrinsic and extrinsic factors (Stern, 2001). Thus, we also evaluated whether in addition to the changes in intrinsic neuronal mechanisms summarized above, structural and/or functional remodelling of PVN afferent pathways also contributed to PVN-NTS activation after exercise training.
Noradrenergic fibers originating mostly from brainstem cells groups (NTS and caudal ventrolateral medulla, A2/C2 and A1/C1 cell groups, respectively), as well as from the locus coeruleus (A6), constitute a major afferent visceroceptive inputs to the PVN. These inputs convey critical peripheral cardiovascular information necessary for the proper generation of autonomic and neuroendocrine homeostatic responses by the PVN (Cunningham et al., 1990). Previous observations that exercise increased hypothalamic noradrenergic input (Dishman et al., 2000), and that blockade of PVN adrenoceptors prevented exercise-mediated changes in plasma norepinephrine (Scheurink et al., 1990), prompted us to investigate whether plastic remodeling in this major PVN afferent input occurred as a consequence of exercise training. Using a combination of retrograde tract tracing along with immunofluorescence for dopamine β-hydroxylase (DBH, a marker for NEergic terminals), we compared the density of DBH immunoreactive terminals within specific PVN subnuclei, as well as on identified OTergic PVN-NTS projecting neurons (Higa-Taniguchi et al., 2007). Exercise training markedly increased overall PVN DBH immunoreactivity, most predominantly in caudal, preautonomic regions of the PVN (Higa-Taniguchi et al., 2007). Moreover, 3D reconstructions of identified PVN-NTS neurons along with their apposing DBHir boutons, indicated a significant increase in the density of somato-dendritic overlapping NEergic boutons in these neurons (Higa-Taniguchi et al., 2007) (Figure 5). In summary, these studies support the view that PVN NEergic afferent inputs undergo plastic remodeling during exercise training in a region- and cell-specific dependent manner. Our findings are in general agreement with previous studies indicating changes in central catecholaminergic function during ET (Chaouloff et al., 1989; Lambert & Jonsdottir, 1998; Dishman et al., 2000).
Whereas the precise functional consequences of NEergic afferent remodeling during exercise training are yet to be determined, a denser PVN DBH immunoreactivity in OT-PVN-NTS neurons would be consistent with an enhanced synaptic action of NEergic transmitters in this neuronal population during exercise training. In addition to affecting neuronal excitability (Boudaba et al., 2003; Yang et al., 2008), it has been shown that noradrenaline can modulate OT mRNA expression within the PVN (Vacher et al., 2002). Thus, the enhanced NEergic action during exercise could also contribute to the increased PVN/NTS OT mRNA levels found during exercise training (Ludbrook & Graham, 1985).
Accumulating evidence over the past few decades supports highly specific cardiovascular adjustments during dynamic exercise, aiming at ensuring proper redistribution of blood flow and perfusion of peripheral tissues, including active muscles. A characteristic, though unique adaptive modification observed during exercise includes a moderate increase in pressure, which is not accompanied by reflex bradycardia, but rather by a marked tachycardic response (Mitchell, 1990; Rowell & O’Leary, 1990). Up to date, all the mechanisms proposed to explain the coexistence of tachycardia and increased pressure during exercise are centred on the NTS as a key integrative brain area (McIlveen et al., 2001; Degtyarenko & Kaufman, 2005; Mueller & Hasser, 2006; Potts, 2006; Michelini, 2007b). As summarized throughout this review, recent results further expand on this concept, supporting a model in which a reciprocal interconnectivity between the NTS and the hypothalamic PVN acts as a pivotal anatomical and functional substrate underlying integrative feed-forward and feed-back adjustments of circulatory control during exercise. According to this model (see Figure 6), exercise-related peripheral information, conveyed in part by the baroreceptors to the NTS (Palkovits, 1988; van Giersbergen et al., 1992; Dampney, 1994; Sved & Gordon, 1994), is transmitted directly and/or indirectly via CAergic projections to the PVN. This results in turn in the simultaneous activation of OTergic and VPergic preautonomic neurons projecting back to the NTS, which drive the necessary autonomic adjustments during exercise-induced pressure increase.
The sustained activation and increased efficacy of this NTS-PVN-NTS pathway during exercise appears to be mediated and/or sustained by neuroplastic adaptive mechanisms occurring at different loci within this interconnected network. These include remodeling of visceroceptive afferent inputs, structural and functional plasticity of PVN-NTS projecting neurons, as well as changes in neurotransmitter content within the NTS (see working model in Figure 6).
While the summarized data support an important role for PVN-NTS reciprocal connections in cardiovascular adaptations during exercise, it is evident that numerous important venues remain open for future investigation. For example, a critical challenge for future research in this field would involve the identification of precise molecular signals and pathways underlying exercise-induced neuroplasticity. In this sense, neurotrophic factors, in particular the brain-derived neurotrophic factor stands as a likely molecular candidate driving neuronal/synaptic plasticity during exercise training (Hartmann et al., 2001; Adlard et al., 2004). Another relevant aspect of exercise-induced neuroplasticity is its potential implication in pathological conditions. In this sense, it is well established that exercise training has multiple beneficial cardiovascular effects in disease conditions, particularly in hypertensive disorders. For example, exercise training reduces cardiovascular risk in hypertension, by diminishing both sympathetic activity and pressure in hypertensive individuals (Amaral et al., 2000; Amaral et al., 2001; Melo et al., 2003).Thus, it would be important in future studies to determine whether exercise training-induced neuroplastic changes in the PVN-NTS bidirectional loop could also promote favorable autonomic adjustments, and/or reverse detrimental changes caused by hypertension.
This work was supported by Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP: 06/50548-9 and 02/11937-9; LCM); Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPq: 302734/02-3 and 304178/05-5, LCM), The National Institutes of Health (R01 HL68725, JES), and the American Heart Association (AHA 0640092N; JES).