Neural and molecular correlates of psychological pain during major depression, and its link with suicidal ideas

Highlights

• Mean psychological pain during depression was measured with a visual analog scale.

• Mean psychological pain was associated with the at-rest activity of 11 brain clusters.

• Medial prefrontal cortex activity predicted change in mental pain during treatment.

• The peripheral expression of 41 selected genes was not correlated with mental pain.

• But 5HTR2B, 5HTR3A, TPH1, and OPRL1 were correlated with several brain clusters.

Abstract

Objectives

Psychological pain increases the risk of suicidal ideas and acts, and represents a potential therapeutic target. However, the mechanisms of mental pain remain unclear. Here, we assessed the peripheral transcriptomic and central neural correlates of mental pain during a depressive episode.

Methods

172 adult un-medicated depressed patients were recruited. Leucocytes were extracted for RNA quantification at baseline (T0) and after 8 weeks (T8) of an antidepressant treatment. Ninety-nine genes of the cortisol, immune, opioid, serotonergic, and kynurenine systems were a priori selected, and 41 were sufficiently expressed to be analyzed. At both T0 and T8, mean level of mental pain over the last 15 days was measured with a visual analog scale. A subset of 38 patients was additionally scanned with Magnetic Resonance Imaging at T0. Resting-state sequences of 4 networks (default-mode, basal ganglia, central executive, salience) were examined.

Results

Mean psychological pain scores significantly decreased between T0 and T8. At conservative p-corrected levels, T0 mental pain was significantly correlated with 11 brain clusters encompassing the prefrontal, parietal, and temporal cortices, the striatum, and the cerebellum. There was no direct association between peripheral gene expression and mean mental pain at any time points or in terms of temporal changes. However, expressions of 5HTR2B at p-corrected levels, and 5HTR3A, TPH1, and OPRL1 were correlated with the activity of several identified brain clusters at T0. Finally, while suicidal ideas and mental pain were correlated, the neural and molecular correlates of suicidal ideas were not the same.

Conclusion

Our study suggests that the serotonergic and nociceptin systems are associated with the activity of a cortico-subcortical brain network underlying the perception of mental pain during depression. Mental pain may be a necessary but insufficient condition for the emergence of suicidal ideation during depression.

Introduction

Psychological pain (also named psychache or mental pain) is an important evolutionary-selected mental function (Eisenberger, 2015b). People experience psychological pain many times in their lifetime at varying degree, most notably in social situations such as a relationship breakup, job dismissal, or loss of a loved one. Psychological pain accompanies complex feelings such as worry, sadness, humiliation, shame, guilt, loneliness, jealousy, and others. Psychological pain has been conceptualized as an alarm system of social disconnection (Eisenberger, 2012), a condition with profound risks for health and survival in humans (Holt-Lunstad et al., 2010). It therefore represents a core mechanism for social and emotional homeostasis maintenance. Over the last years, research has been conducted in healthy volunteers to uncover the mechanisms of psychological pain. Studies have used different paradigms, notably social exclusion games or recall of sadness/grief (Eisenberger, 2012; Meerwijk et al., 2013). At the brain network level, psychological pain has been associated with the anterior cingulate and medial prefrontal cortex, orbitofrontal cortex, anterior insula, posterior cingulate cortex, and occipital cortex in addition to subcortical structures (caudate, putamen, nucleus accumbens, thalamus), and cerebellum (Cacioppo et al., 2013; Meerwijk et al., 2013; Eisenberger, 2015a; Hsu et al., 2015)

While usually within normal variation range, mental pain – just like physical pain – can become pathological and detrimental when extremely intense and/or prolonged. It is found in several psychiatric conditions, particularly major depressive episodes and borderline personality disorder. It has also been associated with a higher risk of suicidal ideation and behavior independently from depression (Ducasse et al., 2017). Recent studies using the opioid agonist buprenorphine (Yovell et al., 2016) or the anesthetic ketamine (Grunebaum et al., 2018) suggest that targeting psychological pain could be a relevant way to reduce short-term suicidal risk. Improving our understanding of the mechanisms of pathological psychological pain could shed light on potential specific therapeutic means.

Yet, psychological pain in the context of mental disorders has given rise to limited investigations. Using autobiographical scripts following a recent suicide attempt in ten women, Reisch et al. (2010) reported increased activation in bilateral medial prefrontal/anterior cingulate cortex, and hippocampus during recall of mental pain vs. suicide action. In 39 depressed individuals, van Heeringen et al. (2010) showed increased cerebral blood flow perfusion at rest in right inferior frontal gyrus, dorsolateral prefontal cortex, occipital cortex, and left temporal cortex in those with high vs. low levels of mental pain. Olié et al. (2017) found differential activation of the left posterior insula and supra-marginal gyrus during a social exclusion task in medicated depressed women with a past history of suicidal acts in comparison to patient and healthy controls. Jollant et al. (2017) reported a negative correlation between psychological pain and N-acetylaspartate levels, a marker of neuronal integrity, in dorsomedial prefrontal cortex. At the molecular level, we are not aware of any study specifically investigating psychological pain in patients. Hsu et al. (2015) showed reduced opioid release in depressed patients during social rejection. Antidepressants are known to reduce painful states associated with depression suggesting a potential role for the serotonergic system, in addition to the opioid system as mentioned above.

In the present study, we aimed at jointly exploring the central mechanisms and peripheral surrogates of psychological pain. To this aim, we combined for the first time brain imaging, peripheral transcriptomics, and a psychological pain measure in un-medicated depressed patients. First, the subjective perception of psychological pain was measured with a simple visual analog scale used in several studies of suicidal behavior (Olie et al., 2010; Cáceda et al., 2017; Jollant et al., 2017, Jollant et al., 2019). Second, as psychological pain is the expression of stressful conditions and impaired homeostasis, we focused on genes related to systems implicated in alarm, defense, stress, pain, and homeostatic functions. This includes the immune/inflammation, cortisol, serotonergic, kynurenine, and opioid systems, biological systems also associated with suicidal behavior (Brundin et al., 2017; Lutz et al., 2018; Zhou et al., 2018). Third, recent studies have reported correlations between peripheral transcriptomic markers and brain activity (Ibrahim et al., 2017; Lopez et al., 2017). We therefore used here peripheral expression levels (based on RNA-sequencing) as biomarkers for these diverse systems. Finally, four complementary resting-state brain networks were examined: the Basal Ganglia (BGN), Salience (SN), Central-Executive (CEN), and Default-Mode (DMN) networks. DMN uniquely shows deactivation during cognitively-demanding tasks (Raichle et al., 2001), and has been associated with self-monitoring, autobiographical processing and social cognition, in relation to ventromedial prefrontal cortex and posterior cingulate cortex/precuneus (Bressler and Menon, 2010). SN is a network anchored in the anterior insula, anterior cingulate cortex, and subcortical areas and it has been involved in the orientation of attention to the most salient stimuli (Seeley et al., 2007; Menon and Uddin, 2010). CEN links dorsolateral prefrontal frontal and parietal cortices and is engaged in higher-order cognitive control (Seeley et al., 2007). Finally, we investigated the BGN as subcortical regions have been involved in mental pain (Meerwijk et al., 2013).

We hypothesized that:

• several regions of the above-mentioned networks would be correlated with psychological pain (H1);

• the peripheral expression of several genes of the above-mentioned systems would be directly correlated with psychological pain (H2);

• brain activity would mediate the relationship between gene expression and psychological pain perception (H3).

As a secondary objective, we aimed at confirming the association between psychological pain and suicidal ideas.