Desperately Driven and No Brakes

The relationship between stress and drug use

Stress has been postulated to play a significant role in the initiation and maintenance of drug abuse and has been identified as a key factor leading to relapse to drug use in humans (Kreek and Koob, 1998). In controlled experiments psychological stress was found to induce strong cravings for cocaine in addicts (Sinha et al., 1999; Sinha et al., 2000). A few studies have examined the effects of childhood abuse on stress response and HPA axis regulation. Girls (aged 7-15 years) with a history of CSA (n=13) showed significantly lower basal and net ovine CRH stimulated ACTH levels relative to controls (n=13; (De Bellis et al., 1994a)). Basal urinary and plasma cortisol levels and ovine CRH-stimulated levels, however, were similar in victims of CSA to those in controls. These results may reflect a form of HPA dysregulation with pituitary hypo-responsiveness to ovine CRH and exaggerated adrenal response to reduced levels of ACTH in children. In contrast, Heim et al., (Heim et al., 2001) reported opposite dysregulatory consequences of CSA in adults. Abused women without major depressive disorder exhibited greater than usual ACTH responses to ovine CRF administration, whereas abused women with major depressive disorder and depressed women without early abuse had blunted ACTH responses relative to controls. Abused women without major depressive disorder exhibited lower baseline and ACTH stimulated plasma cortisol concentrations. Similarly, men with childhood trauma histories exhibited increased ACTH and cortisol response to dexamethasone/CRF. Increased response was associated with the severity, duration, and earlier onset of the abuse (Heim et al., 2008). These findings indicate that sensitization of the anterior pituitary and counter-regulatory adaptation of the adrenal cortex occurs in abused individuals without major depressive disorder or post-traumatic stress disorder.

Animal model studies have shown that stress influences response to drugs of abuse in several ways. First repeated exposure to stressful situations increases individual reactivity to addictive drugs. A variety of stressors including repeated tail-pinch stress (Piazza et al., 1990), restraint stress (Deroche et al., 1992a), social stress (Deroche et al., 1994), and food-deprivation stress (Deroche et al., 1992a), augments the locomotor response to systemic amphetamine or morphine. Stress-induced sensitization to amphetamine and morphine depends on an intact HPA axis, and does not occur in animals in which stress-induced corticosterone secretion is suppressed (Deroche et al., 1992a; Deroche et al., 1994; Marinelli et al., 1996). Repeated corticosterone administration alone (without exposure to stress) is sufficient to sensitize locomotor response to amphetamine, as was prenatal stress that occurred as a consequence of maternal restraint during the last week of pregnancy (Deroche et al., 1992b). Glucocorticoids themselves have direct euphoric effects in some individuals and reinforcing properties in laboratory animals, as evidenced by the development of intravenous self-administration of corticosterone, which occurs at plasma corticosterone levels comparable to those induced by stress (Piazza et al., 1993). However, there are substantial individual differences in susceptibility to corticosterone self-administration (Piazza et al., 1993).

Perhaps most importantly, exposure to a variety of stressors has been found to reinstate previously extinguished seeking of heroin (Shaham et al., 2000; Shaham and Stewart, 1995), cocaine (Ahmed and Koob, 1997; Erb et al., 1996), alcohol (Le et al., 2000; Le et al., 1998), and nicotine (Buczek et al., 1999). In some studies stress exerted an even more powerful reinstating effect than re-exposure to the drug (Shaham et al., 1996; Shaham and Stewart, 1996). Stress reinstates previously extinguished drug-induced place preferences (Wang et al., 2000). Metyrapone (which blocks stress-induced corticosterone secretion but does not modify basal corticosterone levels) attenuates stress-induced relapse of cocaine self-administration, without inducing a nonspecific disruption of motor or food-directed behaviors (Deroche et al., 1994).

In conclusion, preclinical studies have shown that stress is a prominent factor in the initiation, maintenance and reinstatement of substance use (Deroche et al., 1992a; Erb et al., 2001; Goeders, 1997; Kabbaj et al., 2001; Piazza et al., 1990; Shaham et al., 2000; Shalev et al., 2002; Shalev et al., 2001; Stewart, 2000). Enhanced or dysregulated stress responses due to childhood abuse may, at least in part, mediate the increased vulnerability of abuse survivors to drug addiction (McEwen, 2000a; Rodriguez de Fonseca and Navarro, 1998; Sinha, 2001; Stewart et al., 1997; Triffleman et al., 1995).

Increased stress reactivity alters brain development

The HPA axis changes increase stress reactivity (tenet #1), but these changes have unique effects on brain development. The dramatic and profound effects of CRH (Brunson et al., 2001) and stress hormones, work in concert with monoamine neuromodulators and excitatory amino acids (McEwen, 2000b), to modify basic neural processes. Administration of glucocorticoids during early life in laboratory animals permanently reduces brain weight and DNA content (Ardeleanu and Strerescu, 1978), suppresses postnatal mitosis of granule cells in cerebellum and dentate gyrus (Bohn, 1980), interferes with glial cell division (Lauder, 1983), and reduces the number of dendritic spines in various brain regions (Schapiro, 1971). More recent work clearly shows that early exposure to stress-induced molecular signals impacts myelination (Leussis and Andersen, 2008; Meyer, 1983; Tsuneishi et al., 1991), neuronal arborization(McEwen, 2000b), neurogenesis(Gould and Tanapat, 1999; Mirescu and Gould, 2006), and synaptogenesis (Andersen and Teicher, 2004; Garcia, 2002).

The effects of stress on the brain, however, are not universal. Specific brain regions differ in their sensitivity to stress-induced effects. Susceptibility may be influenced by genetics (Caspi et al., 2002; Caspi et al., 2003; Koenen et al., 2005), gender (Barna et al., 2003; De Bellis and Keshavan, 2003; Teicher et al., 2004), timing (Andersen, 2003; Andersen et al., 2008; Leussis and Andersen, 2008; Perlman et al., 2007), density of glucocorticoid receptors (Benesova and Pavlik, 1985; Haynes et al., 2001; McEwen et al., 1992; Pryce, 2007), and capacity of local neurons to release CRH in response to stress (Chen et al., 2004). In this review, we focus on how the timing of stress exposure may facilitate the expression of precocial drug use through regional brain changes. Brain regions that appear to show morphometric vulnerability to the effects of early stress or childhood abuse include corpus callosum, hippocampus, cerebellum and neocortex (Andersen et al., 2008). To begin, we will discuss the delayed impact of early life stress on the hippocampus, which manifests between adolescence and adulthood (Andersen and Teicher, 2004). Changes in the nucleus accumbens and the prefrontal cortex will also be discussed.

Early adversity affects hippocampus development

Combined clinical and preclinical research suggests that exposure to stress during early life produces delayed effects on hippocampal development. Changes in hippocampal volume seem to appear during adulthood. Clinical studies evaluating hippocampal morphometry in survivors of childhood abuse have observed significant volume loss (Andersen et al., 2008; Bremner et al., 1997; Driessen et al., 2000; Stein, 1997; Vermetten et al., 2006; Vythilingam et al., 2002). In contrast, studies in abused children with PTSD failed to find any evidence of hippocampal volume loss (Carrion et al., 2001; De Bellis et al., 1999b; De Bellis et al., 2002), and in fact, a significant increase in white matter volume was reported (Tupler and De Bellis, 2006). Thus, the effects of childhood abuse are associated with hippocampal volume loss in adulthood, but not during childhood or early adolescence (Andersen et al., 2008; Teicher et al., 2003). We recently found reductions in bilateral hippocampal volume following CSA in a young adult sample are maximally if the abuse occurred between 3-5 years of age and between 11-13 years of age (Andersen et al., 2008). These periods correspond to overproduction phases of human hippocampal gray matter (Gogtay et al., 2006). Preclinical observations seem to reconcile these seemingly disparate results between studies examining children or adults. Early isolation stress in developing rats prevents the normal peripubertal overproduction of synapses in CA1 and CA3 region of the hippocampus of rats; however, early stress does not prevent pruning, which leads to an enduring deficit in synaptic density by late adolescence/early adulthood (60 days of age in the rat) (Andersen and Teicher, 2004). Thus, it is likely that exposure to early stress alters the trajectory of hippocampal development, with the adverse impact of early stress becoming manifest during the transition from adolescence to early adulthood.

The role of the hippocampus is to provide contextual gating of information coming from the prefrontal cortex at the level of the nucleus accumbens (Grace et al., 2007), and therefore is involved in drug sensitization processes. Loss of hippocampal synaptic density, or gray matter volume, that appears to occur in early stress exposed individuals as they pass through adolescence may interfere with or alter this gating function. Currently, little is known about the effects of early stress-induced hippocampal changes on vulnerability to substance abuse. However, excitotoxic lesions of the ventromedial hippocampus during the first week of life in rats increases the amount and frequency of drug-taking behavior, but not the break point in a progressive ratio schedule of self-administration of methamphetamine (Brady et al., 2008). These data are consistent with reduced gating of cortical inputs into the accumbens, and would render a stress-exposed pre-teen more susceptible to drug abuse at a younger age than his peers. An alternative explanation may be that stress-induced hippocampal changes reduce the negative feedback control of HPA axis (tenet #1; Goursaud et al., 2006). The findings of greater vulnerability to use drugs following hippocampal manipulations in preclinical studies are consistent with the clinical findings discussed above.

Developmental stress exposure programs the accumbens dopamine system for reduced reward and elevated anhedonia

It is well established that exposure to early life stress increases feelings of dysphoria, anhedonia, and anxiety (Ruedi-Bettschen et al., 2006). A detailed, comprehensive review presented by Matthews and Robbins in 2003 (Matthews and Robbins, 2003) suggests that early life stress dampens the reward system. Repeated maternal separation is associated with enduring reductions in behavioral responses to appetitive stimuli in the adult rat compared with handled controls (Matthews et al., 1996). As observed across multiple experimental procedures, including self-administration, intercranial self-stimulation, and sucrose preferences, the data consistently suggest that appetitive stimuli elicit less vigorous responses in separated rats (Matthews and Robbins, 2003). For example, isolated rats also show blunted positive and negative contrast effects with sucrose as the comparator solution, suggesting reduced reward processing to this natural stimulus (Matthews and Robbins, 2003). The use of drugs helps to overcome this feeling of anhedonia, and it follows that the stressed subject will be more sensitive to the effects of drugs in an attempt to normalize this basal state.

Stress exposure works, in part, by altering the dopamine (DA) system within the nucleus accumbens. Based on adult outcomes of maternal deprivation, behavioral evidence, such as increases in locomotor activity in response to novelty (Brake et al., 2004), to cocaine (Brake et al., 2004), and to tail-pinch stress support the hypothesis that maternal deprivation enhances baseline sensitivity of DA within the nucleus accumbens. We found that pre-weaning maternal deprivation increased DA content and decreases serotonin turnover (5-HT/5-HIAA ratio) in the nucleus accumbens and amygdala during adulthood (Andersen et al., 1999). Direct measurement show that isolates have elevated levels of extracellular DA in this region in response to cocaine (Kosten et al., 2003), an effect that may be mediated by lower levels of the DA transporter (Brake et al., 2004; Meaney et al., 2002). Stress effects on the nucleus accumbens, however, are not only restricted to early life events. Hall et al (Hall et al., 1998) found that post-weaning stress from isolation rearing produced an enduring decrease in 5-HIAA levels and an increase in DA levels and augmentation of stimulant-induced DA release in the nucleus accumbens. Hence, chronic early stress may increase susceptibility to substance abuse by altering the development of the mesolimbic DA system, but the effects of stress may emerge at any age.

One possible mechanism through which sustained increases in accumbens DA produces feelings of dysphoria and anhedonia is through its actions on the transcriptional regulating factor, CREB (Nestler and Carlezon, 2006). Notably, however, increased accumbens CREB levels were not observed in maternally separated animals (Lippmann et al., 2007) and suggest that the anhedonia may be driven by another mechanism or changes elsewhere in the brain. Corticosterone and/or CRH effects on the mesolimbic DA system may modulate stress-induced modulation of behavioral sensitivity to drugs of abuse (Barrot et al., 1999; Deroche et al., 1995; Koob, 1999; 2000; Marinelli and Piazza, 2002; Piazza et al., 1996; Rouge-Pont et al., 1998; Self, 1998). Chronic stress produces neural adaptations in the nucleus accumbens - ventral tegmental area that are very similar to the effects of drug exposure (Fitzgerald et al., 1996; Ortiz et al., 1996).

Exposure to childhood abuse may also increase risk for substance abuse by affecting the DA system and sensitivity to stimulants. Childhood abuse has been associated with increased peripheral levels of DA or homovanillic acid (De Bellis et al., 1999a; De Bellis et al., 1994b), and reduced plasma levels of dopamine beta-hydroxylase (Galvin et al., 1995) the enzyme responsible for the conversion of DA to norepinephrine.

Adolescent stress affects the prefrontal cortex

In contrast to the delayed effects of stress on hippocampal morphometry, stress exerts its maximal effects on the prefrontal cortex during adolescence (Andersen et al., 2008; Hall, 1998; Leussis and Andersen, 2008). Moreover, these effects are observable without a delay as observed for the hippocampus (Andersen and Teicher, 2004). The protracted development of the prefrontal cortex (Crews et al., 2007; Spear, 2000) may render it increasingly vulnerable to the effects of stress during adolescence. In addition, high levels of glucocorticoid receptors are expressed in the cortex during this stage, which may further increase the effects of stress (Pryce, 2007). Stress exposure during adolescence, including CSA in clinical studies (Andersen et al., 2008) or social isolation in preclinical studies (Hall, 1998; Leussis and Andersen, 2008), is associated with a decrease in prefrontal gray matter and synaptic density, with little to no change in other brain regions. Based on pharmacological studies, this synaptic loss reflects stress-induced increases in glutamatergic activity (Leussis et al., 2008). Enhanced glutamatergic activity in the prefrontal cortex is consistent with increased conditioning aspects of drugs of abuse (see tenet #3).

Indeed, studies that have examined the relationship between adolescent stress and drug abuse show increased vulnerability. Ibotenic lesions of the medial prefrontal cortex of rats, which presumably attenuates innervation of the accumbens, resulted in a greater behavioral response to stress and increased stress-induced dopamine release into the accumbens (Brake et al., 2000). Isolation reared, but not maternal deprived rats, showed locomotor sensitization to intermittent 1.5 mg/kg amphetamine (Weiss et al., 2001). In this regard, drugs that produce addictions with a high cue-associated drug-seeking (i.e., heroin and crack (Franken et al., 2003)) may be more likely to be affected by later stressors. As discussed in tenet #3, the prefrontal cortex is involved in the behavioral expression of drug-seeking once the drug-cue association is formed.

Maturation of the adolescent brain is associated with vulnerability to drug-associated cues

Epidemiological research (Figure 2) shows that most addictions do not emerge until adolescence. The maturation of the prefrontal cortex and its connectivity to other regions may be an important factor in the timing of this process (Ernst et al., 2006). Developmental differences in reward processing have been observed in BOLD fMRI studies in children. Immature frontal cortical responsiveness to reward is more diffuse and attenuated compared with adults (Durston, 2003). In contrast, children show greater activation in the accumbens (Ernst et al., 2005). Maturation leads to a more spatially restricted (less diffuse) cortical pattern of activiation (Rubia et al., 2000), which likely reflects pruning of synaptic connections.

Building on the work of Kalivas et al, Brenhouse and colleagues {Brenhouse, 2008 #7113} recently demonstrated that D1 dopamine receptors on fibers projecting from the prefrontal cortex to the nucleus accumbens are normally over-expressed during adolescence, but D1 receptors in this region and located on these terminals are lower in younger and older animals. This observation is consistent with previous preclinical studies in rats that suggested that stimulants have diminished effects in frontal regions of the brain relative to subcortical actions before adolescence (Andersen et al., 2001; Leslie et al., 2004). Increased D1 receptors in the prefrontal cortex enhance the sensitivity to cocaine-associated environments/cues in adolescents, who require lower doses of cocaine than younger or older animals to form significant place preferences (Badanic et al., 2006; Brenhouse et al., 2008a). Once formed, these adolescent drug-context associations are more resistant to extinction than adult associations (Brenhouse and Andersen, 2008b). In summary, tenet #3 suggests that addictive-like behaviors come on-line during adolescence both in stressed and non-stressed individuals, due partly to the maturation of the prefrontal cortex.