Ventral tegmental area dopamine revisited: effects of acute and repeated stress

Heterogeneity in structure and function of VTA dopamine neurons

Prior to the advancement of current labeling techniques, the VTA was not considered a separate structure from the cell bodies of the substantia nigra (SN). The first anatomical description of the VTA was made with Golgi and Nissl preparations by Tsai (1925a, b), who concluded that the cell-free space overlying the sulcus, along with smaller cell size and close relationship to the tracti mammillo- and olfacto-tegmentalis, warranted a separation from the SN. Later anatomical investigations validated Tsai's initial hypothesis that this area contains a discrete population of dopaminergic cells serving a distinct function from SN dopaminergic neurons, leading other researchers to initially term the region the ventral tegmental area of Tsai.

Dopamine cells have been isolated in many animals, including fish (Lefranc et al. 1969), birds (Fuxe and Ljunggren 1965), rats (Carlsson et al. 1965), and other mammals (Fuxe and Owman 1965), but the VTA as a structure appears to be evolutionarily conserved only in higher-order vertebrates. Lower vertebrates do not show a defined VTA, with the “peripeduncular area” containing both dopamine and serotonin cells (Dube and Parent 1982), and broader development of the VTA observed in only a few teleosts and reptiles (Oades and Halliday 1987). However, there is a high degree of similarity between the VTA of mammals, including opossum (Crutcher and Humbertson 1978), rat (Lindvall and Bjorklund 1974; Phillipson 1979a, b, c), rabbit (Blessing et al. 1978), dog (Shimada et al. 1976), cat (Pin et al. 1968; Poitras and Parent 1978; Taber 1961), non-human primate (Felten et al. 1974; Garver and Sladek 1975; Hubbard and Di Carlo 1974; Jacobowitz and MacLean 1978; Tanaka et al. 1982), and human (Bogerts 1981; Bogerts et al. 1983; Nobin and Bjorklund 1973; Olson et al. 1973). Further, the number of VTA dopamine neurons increases with phylogenetic order, such that Balb/C mice have an estimated 25,000 dopamine neurons, albino rats 40,000, and a 33-year-old man 450,000 (German et al. 1983).

Regardless of homology between higher-order species, researchers have struggled to clearly define the boundaries and function of the VTA. As reviewed below, the VTA is a heterogeneous structure in regards to cytoarchitecture, neuro-chemical, and electrophysiological profiles, and afferent/efferent connections, so it is not surprising that there is evidence that VTA dopamine neurons may serve multiple functions, such as reward and aversion.

Heterogeneity in dopaminergic cytoarchitecture

The VTA is characterized by considerable heterogeneity in dopaminergic cytoarchitecture. In mammalian species, the VTA is comprised of four major zones or subnuclei (Fig. 1). The rostrally located parafasciculus retroflexus area (PFR) and caudally located ventral tegmental tail (VTT) contain few dopaminergic cell bodies, while the paranigral nucleus (PN) and parabrachial pigmented area (PBP) are rich in dopaminergic neurons. Additionally, the adjacent midline nuclei—the caudal linear nucleus (CLi), interfasicular nucleus (IF), and rostral linear nucleus of the raphe (RLi)—are often considered VTA subregions (Oades and Halliday 1987; Swanson 1982). However, even within these subregions, dopaminergic cell body characteristics are still not homogeneous (for thorough review, see Ikemoto 2007).

As described in detail by Ikemoto (2007), the PFR, restricted to the anterior portion of the VTA, contains a low density of small- to medium-sized dopaminergic cell bodies, which show light to moderate immunoreactivity for tyrosine hydroxylase (TH, the rate-limiting enzyme in dopamine biosynthesis, used as a marker of dopaminergic cells) and are continuous with dopaminergic cell bodies in the posterior hypothalamic area. The VTT also has a low density of dopaminergic cell bodies, which are small and moderately stained for TH. The densest TH-positive staining is found in the middle two thirds of the VTA, divided into the PN and PBP (some have characterized an additional subregion separating the PN and PBP, the paraintrafasicular nucleus, PIF). The PBP begins to emerge in the anterior VTA but spans the majority of the posterior VTA. The PBP is heterogeneous in terms of cytoarchitecture, leading to inconsistently defined borders in the rat and mouse. The PBP contains both large and medium cell bodies, with no unified orientation. Within the anterior VTA, the PBP contains large, intensely stained cell bodies, which are continuous with the anterior SN pars compacta. In the posterior VTA, the PBP is located dorsolateral to the PN and contains cell bodies and fibers that form a net-like structure. The PN is restricted to the posterior VTA and contains TH-positive cell bodies oriented mediolaterally, tilting toward the IF, that are relatively medium in size and medium to darkly stained. The midline nuclei, which are often considered part of the VTA, are also rich in TH-positive cell bodies. Most notably, the IF contains the densest population of dopaminergic cell bodies in the ventral midbrain. The CLi also contains a dense population of relatively homogenous dopaminergic cell bodies, which are medium in size and medium-dark in TH staining.

It is clear from this and other existing immunohistochemical data that a great deal of cytoarchitectonic heterogeneity exists not only within the VTA but also within specific subregions of the VTA. Not only has this led to a difficulty in establishing clear boundaries of the VTA and its subregions, but this heterogeneity in structure points to further heterogeneity in neurochemical and electrophysiological profiles, as well as overall function.

Heterogeneity in neurochemical profile

VTA neurons also differ in their neurotransmitter profile. VTA neurons have typically been classified as principal (primarily dopaminergic), secondary (GABAergic), or tertiary (other) on the basis of immunohistochemistry for TH, as well as electrophysiological and pharmacological properties (Cameron et al. 1997; Grace and Onn 1989; Johnson and North 1992). Tertiary neurons are hyperpolarized by opioids and serotonin, and while one third of these have been identified as atypical-dopaminergic, the neurochemical profile of the remaining two thirds has yet to be clearly characterized (Cameron et al. 1997; Lammel et al. 2014). Altogether, the VTA is comprised of approximately 65 % dopaminergic neurons, 35 % GABAergic neurons, and less than 3 % glutamatergic neurons (Nair-Roberts et al. 2008; Sesack and Grace 2010). However, it should be noted that VTA dopamine neurons projecting to the NAc can also co-release glutamate (Hnasko et al. 2012; Stuber et al. 2010), further highlighting the neurochemical heterogeneity within the region.

Heterogeneity in electrophysiological profile

Despite this immense anatomical and neurochemical heterogeneity in the VTA, it has until recently been common practice of most in vivo electrophysiological studies to consider VTA dopamine neurons as a single homogenous population (reviewed in Lammel et al. 2014; Ungless and Grace 2012). As in vivo electrophysiological measurements do not allow for direct confirmation of the neurochemical identity of the neurons being recorded, neurons are putatively characterized based on standard classification criteria: broad action potentials, low-frequency pacemaker activity, D2-agonist-induced hyperpolarization, and/or the presence of large Ih currents generated by hyperpolarization-activated cyclic nucleotide–regulated cation channels, or HCN channels (Kitai et al. 1999; Ungless and Grace 2012).

However, it has been established that these conventional criteria are not necessarily reliable (as reviewed extensively in Ungless and Grace 2012). Briefly, the presence of large Ih currents within the VTA can be observed in non-dopaminergic neurons (Margolis et al. 2006; Margolis et al. 2008; Zhang et al. 2010). Furthermore, some verified VTA dopamine neurons are not responsive to dopamine bath application (Bannon and Roth 1983; Lammel et al. 2008) and others have very small or negligible Ih currents (Brischoux et al. 2009; Ford et al. 2006; Hnasko et al. 2012; Jones and Kauer 1999; Lammel et al. 2008; Lammel et al. 2011; Margolis et al. 2006; Zhang et al. 2010). As such, it appears that VTA dopamine neuron heterogeneity extends to electrophysiological profiles as well. Unfortunately, this electrophysiological heterogeneity has resulted in some populations of dopamine neurons going unstudied in many prior reports, confounding previous conclusions drawn about VTA dopamine neuron function.

Not surprisingly, it appears that these electrophysiologically distinct dopamine neurons are located within discrete anatomical subregions of the VTA. Most in vivo electrophysiological studies have used these conventional classification methods described above to identify putative dopamine neurons, and as such have primarily focused on dorsal portions of the VTA, specifically within a region medial to the medial terminal nucleus of the accessory optical tract (Lammel et al. 2014; Ungless et al. 2010; Zhang et al. 2010), where putative dopamine neurons fit these conventional criteria. Thus, the studies focusing on the function of these specific, “conventional” dopamine neurons in this small portion of the VTA may not be applicable to “non-conventional” dopamine neurons in other subregions of the VTA. These other regions of the VTA, such as the ventromedial posterior VTA consisting of the PN and PBP, have been largely ignored as many of the dopaminergic neurons in this area do not conform to established conventional criteria such as large Ih (Lammel et al. 2008). Therefore, it has been proposed that while the correlation between Ih and dopamine phenotype may be high in the commonly targeted dorsolateral region of the VTA (specifically the anterior PBP), other subregions such as the PN and posterior PBP, which have been largely ignored, contain dopamine neurons with a distinct electrophysiological profile. Electrophysiological characterization of dopaminergic neurons within these other anatomical subregions of the VTA should be elucidated in future work, as these areas have been implicated in vastly different behavioral functions (discussed in section Heterogeneity in VTA dopamine neuron function).

Heterogeneity in efferent connections

VTA connectivity is also critical because recent anatomical studies demonstrate localized projection targets of VTA sub-regions, which in turn have important behavioral and functional implications. VTA dopamine neurons project throughout the brain in a non-overlapping mediolateral topography at an approximate 45° angle to the midline (Albanese and Minciacchi 1983; Fallon 1981; Ikemoto 2007). While few have closely examined the heterogeneity in efferent and afferent connections within VTA subregions, the most intensively mapped connections have been between the VTA subregions and the striatum. Ikemoto (2007) demonstrated that dopamine-rich cell bodies in the ventromedially located PN and dorsoposteromedial portions of the PBP selectively project to the medial nucleus accumbens (NAc) shell, medial prefrontal cortex (mPFC), and medial olfactory tubercle (OT, Fig. 2) regions heavily implicated in reward and reward processing (although these regions certainly mediate other functions as well). Along the mediolateral projection topography, lateral PBP dopamine neurons send dense projections to the ventrolateral striatum, which has been less heavily implicated in reward-related functions (Ikemoto 2007). The PFR and VTT do not contain dense dopamine cell bodies, but the sparse dopamine cell bodies of the PFR selectively project to the diagonal band. Projection targets of the midline nuclei are also distinct, with dopamine neurons in the IF projecting selectively to the dorsomedial NAc shell, and RLi to the diagonal band and pallidal zone of the OT (Ikemoto 2007). Future work needs to evaluate VTA heterogeneity in other dopamine projection sites, such as the mPFC and BLA.

Heterogeneity in afferent connections

There may also be subregional differences in afferent connectivity to the VTA, although this has not yet been thoroughly investigated. VTA dopamine neurons receive innervation from widespread regions throughout the brain. The direct monosynaptic inputs to midbrain dopamine neurons have been thoroughly mapped recently (Watabe-Uchida et al. 2012). Notably, VTA dopaminergic neurons receive the most innervation from the ventral striatum, particularly the NAc, while the densest innervation originates from the dorsal raphe nucleus (DRN). NAc cells projecting to VTA dopamine neurons form extremely dense patches within the NAc. Moreover, these projection cells are morphologically distinct from NAc GABAergic medium spiny neurons, indicating distinct heterogeneity in the ventral striatum. Future circuit tracing experiments should investigate whether this afferent heterogeneity extends to the VTA.

Heterogeneity in VTA dopamine neuron function

The aforementioned heterogeneity in anatomy, neurochemistry, electrophysiological profile, and connectivity points to diversity in the overall behavioral functions VTA dopamine neurons mediate. As many of the dopamine projection targets of the VTA have been heavily implicated in reward, there has been considerable attention paid to VTA dopamine neurons in these processes. Both natural rewards (Berridge 1996) and drugs of abuse (Di Chiara and Imperato 1988) stimulate release of dopamine from VTA neurons projecting to the NAc, leading to a well-accepted hypothesis that this connection at least partially drives reward-related functions.

However, recent evidence has also shown heterogeneity within these projection targets in terms of reward-related function. Cocaine infused directly into the medial NAcSh produces significantly greater changes in locomotion compared to cocaine infused directly into the lateral NAcSh (Ikemoto 2002, 2007). Rats will selectively self-administer cocaine and amphetamine into the medial but not lateral NAcSh (Ikemoto 2002; Ikemoto and Donahue 2005), with similar differences observed between the medial and lateral OT (Ikemoto 2002; Ikemoto and Donahue 2005). Therefore, it is not surprising that the posteromedial VTA, centered around the PN and posteromedial PBP and projecting to the medial NAcSh and OT, has been shown to also play a stronger role in reward processes than the anterior and lateral portions of the VTA (Rodd-Henricks et al. 2002; Sellings and Clarke 2003; Sellings et al. 2006). Specifically, cocaine, nicotine, opiates, ethanol, and cannabinoids are all selectively self-administered into the posterior but not anterior VTA (Ikemoto et al. 2006; Ikemoto and Wise 2002; Rodd et al. 2005; Rodd-Henricks et al. 2000; Zangen et al. 2006).

This mesocorticolimbic dopamine circuitry stemming from the VTA and projecting to the ventral striatum and medial prefrontal cortex thus plays a fundamental role in reward. This is paralleled by subsequent reports that rewards, as well as their predictive cues, can elicit strong phasic firing within the dopamine cell bodies of the VTA (Schultz 1997, 1998). These pioneering findings have led to a prominent hypothesis that this system primarily serves to mediate reward, hedonia, and related energizing processes (but see Salamone and Correa 2012). Given this overwhelming evidence, it was initially surprising and controversial to many researchers that VTA dopamine could be involved in stress and other aversive events (Thierry et al. 1976), which will be the focus of the remainder of this review.

Effects of acute stress on VTA dopamine neuron activity

Electrophysiological studies have shown that aversive stimuli inhibit putative VTA dopamine neuron firing (e.g., Mantz et al. 1989; Mirenowicz and Schultz 1996; Schultz and Romo 1987; Ungless et al. 2004). However, microdialysis studies examining extracellular dopamine and its metabolites collected over minutes and hours have found a robust dopaminergic increase during stress in VTA projection targets. Various stressors such as restraint, footshock, tail pinch/shock, social threat, and others potently increase extracellular dopamine in the NAc and mPFC (section Microdialysis evidence; see Tables 1, 2, 3, 4, 5, 6, 7, 8, and 9 and Fig. 3). Across these studies, the nature and degree of the dopaminergic increase vary according to stressor and intensity. Recent electrophysiological studies have also found a discrete subset of VTA dopamine neurons that increase firing in response to aversive stimulation, corroborating the observed microdialysis results (section Electrophysiological evidence). Additional studies have found long-lasting neuroadaptive changes on VTA dopamine neurons after a single stress exposure, highlighting that acute stress can alter VTA dopamine neuron responsivity to future stimulation, whether by additional stressors or rewards (section Evidence for neuroadaptations on VTA dopamine neurons).

Microdialysis evidence

Early postmortem studies found altered dopamine and dopamine metabolite concentrations in brains of rodents following stress (Deutch et al. 1991; Deutch et al. 1985; Dunn and File 1983; Fadda et al. 1978; Kramarcy et al. 1984). With advances in microdialysis techniques to monitor in vivo extracellular monamine levels in awake, freely moving animals in the late 1980s, researchers began to more directly assess dopamine in response to a variety of stressors.

Imperato and colleagues (1989) were among the first to use microdialysis to demonstrate a significant increase in extracellular dopamine in response to restraint stress. She and others have found that restraint stress reliably increases extracellular dopamine in the NAc (Table 1) and mPFC (Table 2) to roughly equivalent degrees (average maximal percent change from baseline from Tables 1 and 2 142.5 and 155 % for NAc and mPFC, respectively).

Studies examining extracellular dopamine in the mPFC in response to immobilization have found a slightly greater response (average maximal percent change from baseline 200 %, Table 2), indicating that there may be a difference in severity between these two similar stressors. This may be explained by a confound within the methods of the experiments utilizing restraint stress. All but two studies (Garrido et al. 2013; Mokler et al. 2007) were conducted in the light phase of the light–dark cycle. Restraint stress during the dark (active) phase results in significantly reduced body weight gain and development of stomach ulcers, whereas no such effects are produced by restraint during the light (inactive) phase (Koolhaas et al. 2011; Pare and Glavin 1986; Rybkin et al. 1997). Wild Norway rats spend the light phase hiding in narrow burrow systems (Koolhaas et al. 2011), so restraint may be a less potent stressor during this phase. Rather than the physical compression used in restraint stress, immobilization involves restricting paw movement in a less constrained manner and as such may be a more powerful stressor during the light phase.

While some studies have demonstrated that dopamine levels in both the NAc and mPFC remain elevated for the duration of restraint (Cuadra et al. 2001; Cuadra et al. 1999; Garrido et al. 2013; Jackson and Moghaddam 2004; Mokler et al. 2007), when restraint is prolonged (>60 min), dopamine levels return to baseline within 70–120 min (Imperato et al. 1992; Imperato et al. 1993; Imperato et al. 1991; Imperato et al. 1989; Imperato et al. 1990; Puglisi-Allegra et al. 1991). Thus, there appears to be a habituation of the dopamine response upon extended stressor presentation. However, as restraint is the only stressor examined in microdialysis studies to date that is amenable to such prolonged presentation, it is not clear if this habituation would extend to other types of stress.

Mild footshock has also been shown to potently increase dopamine in both the NAc (Table 3) and mPFC (Table 4) to comparable degrees (average maximal percent change from baseline 169.22 and 194 % for NAc and mPFC, respectively). Notably, all reports of microdialysis during footshock stress have used less than 0.55-mA intensity, generally considered to be mild. Future work could examine the relationship between footshock intensity and extracellular dopamine in the NAc and mPFC. Mild to moderate tailshock (1.0 mA) also produces significant increases in extracellular dopamine in the mPFC (Table 7, average maximal percent change from baseline 169 %). Like footshock, the effects of varying intensities of tailshock on extracellular mesocorticolimbic dopamine efflux have not been examined and could be the focus of future work.

Microdialysis has also been used to examine dopamine responses to acute tail pinch (Tables 5 and 6). Unlike restraint and footshock, tail pinch stress may differentially increase extracellular dopamine in the NAc and mPFC, with greater dopamine efflux observed in the mPFC (average maximal percent change from baseline 124 and 184 % for NAc and mPFC, respectively). As none of the studies examined the dopamine response in both the NAc and mPFC, it is possible that there are differences in intensity of tail pinch pressure between labs. However, it may also be the case that very mild tail pinch stress is insufficient to activate VTA dopamine neurons, and the mPFC dopamine response is due to another function, such as novelty.

Likewise, handling, often considered a very mild stressor, has differential effects on extracellular dopamine in the NAc and mPFC (average maximal percent change from baseline 126 and 197 % for NAc and mPFC, respectively, Tables 8 and 9). Duration of handling stress does not appear to reliably affect extracellular dopamine concentrations, but there may be strain differences in reactivity to handling, as the greatest changes in mPFC dopamine were observed in Wistar as opposed to Sprague–Dawley rats.

However, not all stressors examined have produced increases in extracellular NAc and mPFC DA. Acute forced swim stress, often thought to be a much milder stressor than footshock and restraint stress (Jordan et al. 1994), does not alter extracellular dopamine in the NAc or the mPFC (Azzi et al. 1998; Jordan et al. 1994). Likewise, the similarly mild stressor of airpuff to the face or low-dose cytokine (IL-8) injection does not alter extracellular dopamine in either brain region, although these stressors work synergistically to increase dopamine when administered concurrently (Merali et al. 1997).

One potentially important distinction is that all the above-mentioned stressors involve direct physical tactile contact/stimulation of the animal. However, stressors that do not involve direct contact with the animals’ body can also elicit strong increases in extracellular dopamine in both the NAc and mPFC. The “psychological” stress of observing and smelling nine other rats receiving severe (3.0 mA) footshocks elicits a significant increase in extracellular dopamine in the NAc shell, but not core (Wu et al. 1999), one of the only studies to examine the difference in responsivity to stress between these subregions of the NAc. Additionally, presentation of a predator (fox) odor produces a gradual increase to 205 % baseline levels in extracellular dopamine in the mPFC (Wu et al. 2003).

It is possible that these findings showing stress-induced elevations in extracellular NAc and mPFC dopamine are in line with a hypothesis that VTA dopamine neurons primarily subserve reward-related functions as opposed to stress-related functions. The removal of a stressor or aversive stimulus is negative reinforcement and can strengthen subsequent associated behaviors (Thorndike et al. 1932). Considerable behavioral evidence has demonstrated that the termination of a stressor or aversive stimulus can serve as a reward (e.g., Navratilova et al. 2012; Tanimoto et al. 2004). Thus, it could be expected that rather than stress activating these dopaminergic neurons, it is actually the offset of stress that excites VTA dopamine neurons, resulting in the observed extracellular dopamine increases in VTA projection sites. Indeed, approximately half of the VTA dopamine neurons inhibited by footshock also show excitation at the termination of the aversive stimulation (Brischoux et al. 2009).

Some microdialysis studies are in support of this explanation. Although restraint stress produces a sustained elevation in extracellular dopamine in both the NAc and mPFC (Tables 1 and 2), when restraint or immobilization is prolonged until dopamine levels return to baseline, most have shown that there is a strong, rapid increase in extracellular dopamine levels again upon release (Imperato et al. 1992; Imperato et al. 1991; Jackson and Moghaddam 2004; Lillrank et al. 1999; Pozzi et al. 2002; Puglisi-Allegra et al. 1991; Swanson et al. 2004). However, it is difficult to evaluate a dopamine response to the termination of a stressor, as most other types of stress studied (e.g., footshock and tail pinch) are much shorter in duration, rarely spanning greater than two microdialysis samples, and do not show a return to baseline prior to the termination of the stressor. Therefore, it cannot be concluded that any significant increases after termination of the stressor are due to negative reinforcement as opposed to carryover from the aversive experience. Regardless, this hypothesis that the dopamine increase is due to negative reinforcement as opposed to stress itself cannot explain the sustained dopaminergic increases observed in microdialysis studies where the stressors or aversive stimuli outlast the sampling time.

Thus, while the temporal resolution and correlational nature of these microdialysis experiments could not conclusively prove that VTA dopamine neurons are excited by stress as opposed to the removal of a stressor, the magnitude and duration of dopaminergic increases in these target areas indicate a likely effect on VTA firing in response to stress. Overall, as summarized in Fig. 3, most stressors elicit an increase in extracellular dopamine in VTA projection targets, with the most potent stressors eliciting the greatest changes from baseline. However, with milder stressors, there seems to be a greater increase in mPFC dopamine compared with NAc dopamine, indicating a possible lower threshold of stimulation or alternative function for mPFC projecting dopamine neurons. Alternatively, mPFC projecting dopamine neurons may be less sensitive to different types of stressors.

Electrophysiological evidence

Initially, extracellular increases in dopamine concentration in VTA projection targets were difficult to reconcile, as most studies had shown a suppression of VTA dopamine neuronal firing during stress or aversive stimulus presentation (Guarraci and Kapp 1999; Mantz et al. 1989; Mirenowicz and Schultz 1996; Schultz and Romo 1987; Ungless et al. 2004). However, recent evidence demonstrates that there is in fact a subset of dopamine neurons within the VTA that are rapidly and potently excited by stressful, aversive stimuli. Single unit recordings in awake rats showed that both firing rate and burst firing are increased in putative VTA dopamine neurons during restraint stress (Anstrom and Woodward 2005), while multi-unit recording showed similar increases in burst firing, but not firing rate, during social defeat stress (Anstrom et al. 2009). Burst firing is thought to play an important functional role in dopamine release, as frequencies may overwhelm the dopamine transporter, causing supralinear increases in extracellular dopamine concentration (Gonon 1988).

As described in section Heterogeneity in electrophysiological profile, most in vivo electrophysiological studies have focused on the dorsolateral VTA dopamine neurons using classic criteria, particularly large Ih current. When examining ventromedial VTA dopamine neurons, which were characterized by smaller Ih currents, Brischoux and colleagues (2009) found that there was a subset of neurons rapidly and strongly excited by stress (Fig. 4). Similarly, others have found increased activity in these “non-conventional” VTA dopamine neurons in response to aversive stimulus presentation (Cohen et al. 2012; Zweifel et al. 2011). It is highly likely that a subpopulation of dopamine neurons responsive to aversive or stressful stimuli have been overlooked in prior work due to sampling bias and mischaracterization of VTA dopamine neurons (Brischoux et al. 2009; Ungless et al. 2010; Ungless and Grace 2012). In light of this growing evidence, it has been proposed that there are at least two subpopulations of VTA dopamine neurons: one group encoding reward-prediction error that is suppressed by aversive stimulation, and a second group, with atypical Ih and high baseline burst firing, that is phasically stimulated by aversive stimuli (Ungless et al. 2010).

Evidence for neuroadaptations on VTA dopamine neurons

Exposure to a single acute stressor can also promote long-lasting neuroplastic changes in VTA dopamine neurons in a manner similar to exposure to drugs of abuse (Dong et al. 2004; Graziane et al. 2013; Niehaus et al. 2010; Saal et al. 2003). Acute stress induces long-term potentiation (LTP) at glutamatergic synapses onto VTA dopamine neurons, while concurrently blocking the formation of LTP at GABAergic synapses (Graziane et al. 2013; Niehaus et al. 2010). During induction of LTP at glutamatergic synapses, new AMPA receptors are inserted, increasing the AMPA/NMDA ratio and increasing later excitability of the postsynaptic neuron (Malinow and Malenka 2002). This alteration in the AMPA/NMDA ratio enhances calcium permeability and changes calcium dynamics in the synapse, such that subthreshold stimulation can induce robust LTP (Polter and Kauer 2014). Acute exposure to stress increases this ratio of AMPA to NMDA receptors within excitatory synapses on VTA dopamine neurons (Dong et al. 2004; Graziane et al. 2013; Saal et al. 2003). However, consistent with the theme of importance of VTA heterogeneity, distinct regional differences in AMPA/NMDA ratio alterations within the VTA have been observed after acute stress exposure. Injection of formalin into the hindpaw, an intense noxious stimulus, results in a significant increase in the AMPA/NMDA ratio in medial VTA dopamine neurons projecting to the mPFC, whereas VTA dopamine neurons projecting to the NAc shell do not exhibit such alterations (Lammel et al. 2011). These increases in AMPA/NMDA ratio are present within 2 h of stress and have been observed for at least 24 h (Daftary et al. 2009). Furthermore, intra-VTA antagonism of both AMPA and NMDA receptors prevents tail pinch–induced dopamine efflux in the mPFC, although the NAc has not been examined (Butts and Phillips 2013).

Recent evidence also demonstrates that acute exposure to stress can block the induction of LTP at GABA_A synapses on VTA dopamine neurons (Graziane et al. 2013; Niehaus et al. 2010). VTA dopamine neurons are relatively depolarized at baseline, and thus typically at or very close to action potential threshold (Graziane et al. 2013; Johnson and North 1992). This loss of LTP at inhibitory synapses on VTA dopamine neurons may represent the removal of a brake on the system, which combined with the induction of LTP at excitatory synapses can lead to increased responsivity of VTA dopamine neurons to future stimulation, whether by additional stressors or rewards such as drugs of abuse.

Effects of repeated and chronic stress on VTA dopamine neuron activity

Activation of VTA dopamine neurons during acute stress exposure and subsequent neuroadaptations may result in altered VTA dopamine response to later stimulation. The effects of repeated or chronic stress on VTA dopamine neurons have remained largely unstudied by electrophysiological measures, as electrophysiological evidence for a subset of VTA dopamine neuron activation by stress has only recently emerged. However, a few in vivo microdialysis studies indicate that repeated exposure to stress might indeed alter dopaminergic release in VTA projection targets, particularly the NAc and mPFC. Importantly, repeated stress exposure can affect both tonic (basal levels) and phasic (release in response to stimulation) dopamine in the NAc and mPFC.

The schedule, intensity, and nature of stressors or aversive stimuli again have differential effects on extracellular NAc and mPFC dopamine, related to altered VTA dopamine neuron activity. For example, repeated intermittent exposure to social defeat stress increases dopaminergic tone in the NAc (Miczek et al. 2011), while chronic social defeat reduces overall dopaminergic tone in both males and females (Miczek et al. 2011; Shimamoto et al. 2011). Chronic, inescapable restraint stress, a relatively severe stressor, also decreases dopamine tone in the NAc (Mangiavacchi et al. 2001), while other animal models used to study symptoms of depression such as chronic cold and chronic mild stress have no effect on basal dopaminergic tone in the NAc, striatum, or mPFC (Di Chiara et al. 1999; Gresch et al. 1994). Both chronic restraint and repeated social defeat stress increase both spontaneous and burst firing of VTA dopamine neurons (Anstrom et al. 2009; Anstrom and Woodward 2005; Cao et al. 2010; Krishnan et al. 2007). Notably, studies investigating individual differences in responses to chronic social defeat have found that these effects on VTA firing are only observed in susceptible mice that exhibit behavioral signs of anhedonia (Cao et al. 2010; Feder et al. 2009; Krishnan et al. 2007). Furthermore, these effects are long-lasting, still observed 3 weeks after stress termination (Razzoli et al. 2011).

In addition to altered tonic dopamine in VTA projection targets, the phasic dopamine response in the NAc and mPFC to subsequent stressors is also altered. While daily restraint stress for 6 consecutive days results in a habituation of the extracellular dopamine response in the NAc across time, when restraint is again repeated after 72 h, the extracellular dopamine phasic response in the NAc is equivalent to the response on the first day (Imperato et al. 1992; Imperato et al. 1993). Repeated footshock stress (Young 2004) and intermittent social defeat stress (Holly et al. 2015), on the other hand, do not show such habituation in the phasic extracellular NAc dopamine response, while a sensitized response is observed after the much milder stresses of repeated tail pinch (Naef et al. 2013) or forced swim (Jordan et al. 1994; Petty et al. 1997).

In addition to altered responses to repeated presentations of the same stressor, a history of repeated stress can also alter the subsequent phasic extracellular dopamine response to a different stressor. Prior history of chronic variable stress results in a significantly greater phasic extracellular dopamine response in the mPFC in response to restraint stress (Cuadra et al. 2001; Cuadra et al. 1999), as well as both the NAc and mPFC during tail pinch stress (Di Chiara et al. 1999) compared to previously non-stressed controls. Continuous chronic cold exposure, another model of repeated stress shown to elicit anhedonic-like responses in rodents, also produces greater mPFC phasic extracellular dopamine responses to tail pinch (Finlay et al. 1995) and tail shock (Gresch et al. 1994; Murphy et al. 2003). A similar sensitized effect of extracellular NAc dopamine is observed in animals with a history of isolation rearing in response to footshock stress (Fulford and Marsden 1998), and history of prior social defeat stress results in greater NAc and mPFC response to social threat compared to previously non-stressed controls (Tidey and Miczek 1996, 1997; Watt et al. 2014). When repeated footshock stress is paired with a conditioned stimulus (CS), an augmented response is observed in both the NAc and mPFC, with the CS alone significantly elevating extracellular dopamine above baseline (Feenstra et al. 2001; Young 2004; Young et al. 1993).

Overall, while the effects of repeated stress on VTA dopamine neuron activity and related tonic and phasic dopamine levels in the NAc and mPFC have received only limited attention, current evidence points to a clear effect of repeated stress on subsequent tonic dopamine activity as well as subsequent response to both identical and different stressors. As with the effects of acute stress, the nature, intensity, and schedule of repeated stress are critical, such that mild or intermittent stressors may potentiate basal VTA dopamine neuron activity and more severe or chronic uncontrollable stressors may reduce basal VTA dopamine activity, but the response to later stressors of a different nature is generally cross-sensitized.