Targeting Stress Neuroadaptations for Addiction Treatment: A Commentary on Kaye et al. (2017)

Kaye et al. (2017—this issue) identify the importance of targeting stress for addiction treatment and how unpredictable stress plays a pivotal role in addiction etiology. They present an elegant model examining startle potentiation during unpredictable (vs. predictable) stressors in cued threat tasks to further elucidate the role of brain stress system neuroadaptations in addiction. Of note, modeling stress reactivity in addiction and, more broadly, examining neurobiological targets for stress-induced drug-motivated behaviors lend support to the seamless translation between preclinical and human research. Ultimately, the feasibility of both forward and back translation based on this program of research will likely lead to the development of novel treatments that are better able to target specific stress neurocircuitry to attenuate drug craving and self-administration and increase options to providers for the treatment of stress-precipitated substance use.

The review by Kaye et al. (2017) should be particularly encouraging to the scientific community seeking to expand on the startle potentiation model and to examine or develop other models of addiction that tap into stress pathophysiology. Kaye and colleagues focus on the corticotrophin-releasing factor (CRF) and norepinephrine (NE) systems and the central extended amygdala (CeA). Indeed, there is an increasing amount of support for the role of the CRF and NE systems in addiction and stress-precipitated substance use (Back et al., 2010; Fox et al., 2012, 2014; McKee et al., 2015), particularly in rodent models of stress-induced reinstatement of drug seeking (Lê et al., 2011; Mantsch et al., 2016; Zislis et al., 2007). Unfortunately, there has been little translation of these findings to humans, and too few human laboratory studies have examined CRF or noradrenergic mechanisms underlying substance use.

The pathophysiology of stress and neuroadaptations within brain stress systems is complex and expands beyond the CRF and NE systems and the CeA. Further investigation of other stress systems and stress-related neurocircuitry is also relevant and important in the etiology and treatment of addiction. Research remains limited on the role of urocortin, dynorphin, neuropeptide Y, hypocretins/orexin, and brain regions such as the striatum and prefrontal cortex (PFC) in addiction etiology and treatment. For example, preclinical and human work suggests that exposure to stress impairs PFC structure and function (Arnsten, 2009). Specifically, stress floods the PFC with dopamine and norepinephrine, which leads to dendritic spine loss (over time) and reduces functional connectivity within the PFC. These anatomical changes result in executive function decrements generally and more so during stress. Guanfacine, an α₂-adrenergic agonist, increases neuronal firing within the PFC, improves connectivity, and increases the ability to regulate behavior (Arnsten, 2010). Work such as this identifies that guanfacine may improve stress-related decrements in PFC function and subsequent modulation of behavior in response to stress. In humans, guanfacine improves executive function and attenuates stress-related effects on substance use (McKee et al., 2015). Thus, it remains crucial to consider bi-directional translation of findings between preclinical and clinical research to further identify, understand, and characterize potentially novel mechanisms and treatment targets underlying stress neuroadaptations in addiction.

One avenue that will further our understanding of stress neuroadaptations in addiction is positron emission tomography (PET) neuroimaging. New and better tracers are constantly being developed to better probe for treatment targets associated with addiction (e.g., neuroinflammation, glucocorticoids) and could be used to elucidate neurochemical mechanisms underlying stress-precipitated substance use in the living human brain. For example, new PET technology has allowed for the identification of highly localized dopamine transients in PET data and can be used to examine dynamic changes in dopamine release during in vivo drug self-administration (Cosgrove et al., 2014). This technological advancement is a direct translation of microdialysis studies examining dopamine release in animals (Wickham et al., 2013). The inclusion of stress into such translational investigations would increase our understanding of how stress increases the reinforcement value of substances (e.g., McKee et al., 2011) as one example.

Other considerations for the scientific community would be to expand on unpredictable stressors to also consider uncontrollable or chronic stressors. Kaye and colleagues (2017) nicely convey the importance of adequately examining the role of varying stressor types on stress neuroadaptations in addiction. They correctly highlight that research is needed to disentangle unpredictability from uncontrollability. Additional factors to consider include chronicity and timing to further complicate matters. For example, childhood adversity is a stressor that can be considered unpredictable, uncontrollable, and potentially chronic in nature. Childhood adversity also requires “exposure” before age 16 or 18 (depending on varying definitions). Childhood adversity has been identified to increase risk of substance use disorders, result in anatomical changes (e.g., reductions in white matter), and alter stress reactivity (Huang et al., 2012; McLaughlin et al., 2010; Myers et al., 2014). Preclinical work indicates that chronic stress during early development can increase drug self-administration in adulthood (Lopez et al., 2011). Bi-directional translational work carefully parsing apart unpredictability, uncontrollability, chronicity, and timing would greatly assist in our understanding of how stress facilitates substance use and point to how we may intervene.

We should also consider the impact of individual difference factors that may attenuate or facilitate the impact of stress on neuroadaptations. For example, much work identifies that women may be more susceptible to using substances for stress and negative-affect regulation compared with men (Verplaetse et al., 2015). Other key individual difference factors that may influence stress neuroadaptations and stress-related substance use may include but are certainly not limited to genetic predisposition, environment, age at first drug use, years of daily drug use, early and/or chronic exposure to stress, early negative emotionality, general stress reactivity, and, as mentioned by Kaye and colleagues, general startle reactivity. A crucial future direction for the field is to identify individual differences that affect the development of stress neuroadaptations, particularly in response to chronic substance use.

Targeting stress neuroadaptations for addiction treatment is an important avenue of research for both pharmacotherapeutic and behavioral treatment development. Recent findings highlight that pharmacotherapy targeting stress pathophysiology can improve clinical outcomes in alcohol-dependent patients (Ryan et al., 2017). Stress and negative affect reduction are key components of efficacious behavioral treatments (e.g., cognitive–behavioral therapy and mindfulness) for substance users. Further, as highlighted by Kaye et al. (2017), alterations in environment could have significant treatment implications on substance use such that individuals may encounter fewer unpredictable stressors and decrease subsequent drug use if environmental factors improve. The scientific community should be encouraged to pursue programmatic research on the role of stress neuroadaptations and stress pathophysiology in substance use and use bi-directional translation between preclinical and human research as a means to further elucidate the complex mechanisms underlying stress-precipitated substance use. Kaye et al. should be commended for providing an exemplary model for such programmatic work, which will ultimately lead to the development of novel and better treatments for substance use disorders.

Acknowledgment

Work on this commentary was supported by National Institutes of Health Grants T32DA007238, P50DA033945 (Office of Research on Women’s Health & National Institute on Drug Abuse), and R01AA022285.