Serotonin transporter gene expression predicts the worsening of suicidal ideation and suicide attempts along a long-term follow-up of a Major Depressive Episode

Abstract

The quest for biomarkers in suicidal behaviors has been elusive so far, despite their potential utility in clinical practice. One of the most robust biological findings in suicidal behaviors is the alteration of the serotonin transporter function in suicidal individuals. Our main objective was to investigate the predictive value of the serotonin transporter gene expression (SLC6A4) for suicidal ideation and as secondary, for suicide attempts in individuals with a major depressive episode (MDE). A 30-week prospective study was conducted on 148 patients with a MDE and 100 healthy controls including 4 evaluation times (0, 2, 8 and 30 weeks). Blood samples and clinical data were collected and SLC6A4 mRNA levels were measured from peripheral blood mononuclear cells using RT-qPCR. We first demonstrated the stability and reproducibility of SLC6A4 mRNA expression measures over time in healthy controls (F=0.658; p=0.579; η²=0.008; ICC=0.91, 95% CI [0.87–0.94]). Baseline SLC6A4 expression level (OR=0.563 [0.340–0.932], p=0.026) as well as early changes in SLC6A4 expression between baseline and the 2^nd week (β=0.200, p=0.042) predicted the worsening of suicidal ideation (WSI) in the following 8 weeks. Moreover, changes in SLC6A4 expression between the 2^nd and 8^th weeks predicted the occurrence of a suicide attempt within 30 weeks (OR=10.976 [1.438–83.768], p=0.021). Altogether, the baseline level and the changes in SLC6A4 mRNA expression during a MDE might predict the WSI and the occurrence of suicidal attempts and could be a useful biomarker in clinical practice.

Introduction

Suicide is a major cause of premature death worldwide and a serious public health issue. Suicidal behaviors are the most serious complications of mood disorders. Almost 90% of suicide completers suffer from a psychiatric disorder, mainly from a major depressive episode (MDE) (Arsenault-Lapierre et al., 2004). Although efficient treatment of mood disorders is an effective prevention strategy (Zalsman et al., 2017), a worsening of suicidal ideation (WSI) occurs in up to 30% of patients with a MDE during antidepressant treatment (Perroud et al., 2009, Zisook et al., 2009).

Many researches, over the past decades, focused on identifying subgroups of patients with a higher risk of suicidal ideation or behaviors, to predict and to prevent these complications in the context of clinical care. Suicidal behaviors have been associated with several risk factors such as a previous psychiatric disorder, single marital status, unemployment, childhood maltreatments, family history of suicidal behaviors or previous suicide attempts, impulsivity and alcohol use disorder (Fawcett et al., 1990, Fergusson et al., 2000, Mann, 1998, McCauley et al., 1997, Tidemalm et al., 2011). Moreover, depression severity, previous suicide attempts, retirement, weight loss, vascular or neurological diseases and being widowed have also been identified as risk factors for WSI during antidepressant treatment (Brent, 2016; Coughlin et al., 2016, Perroud et al., 2009). We notably reported that emergence or WSI in the first weeks after initiating an antidepressant treatment was mainly related to the worsening of depressive symptoms (with a attributable risk of 67.5%) (Courtet et al., 2014). Pharmacogenomics studies of antidepressant treatment-emergent suicidal events have been performed in large cohorts of depressed patients (reviewed in (Brent et al., 2010)). Associations with polymorphisms in genes involved in different systems have been reported but rarely replicated. Despite all of these findings, to date, no specific clinical or genetic markers or any combination has allowed to effectively predict the WSI in the context of clinical care.

In such a context, biomarkers could help clinicians to identify specific subpopulations at high risk for suicidality. Based on the assumption that suicidality is a phenomenon resulting from a complex interaction of genetic background, early environmental stress and proximal biological variation (Mann and Currier, 2010, Turecki, 2014), changes in gene expression offer a good opportunity to capture biological dysregulation associated with suicide risk. Several studies examined gene expression differences in peripheral blood in an attempt to describe biomarkers of MDE and antidepressant response (Iga et al., 2008) but few of them regarding suicide risk in depressed individuals (Chang et al., 2016). Two types of methodology are used in this research area, the hypothesis-free and candidate gene approaches. Interestingly, three studies of the same team described changes in blood expression levels that could predict suicidal ideations states and future hospitalizations due to suicidal behaviors using a hypothesis-free approach and different cohorts of males and females patients independently with heterogeneous psychiatric disorders (Le-Niculescu et al., 2013, Levey et al., 2016, Niculescu et al., 2015b). Of note, Mullins et al. did not replicate the results of one of these previous studies (Mullins et al., 2014). Moreover, a cross-sectional study that used a candidate gene-based approach including mRNA expression of BDNF, FKBP5 and NR3C1 genes in peripheral blood, did not find any differences between depressed patients with and without suicidal ideation (Roy et al., 2017).

To the best of our knowledge, no previous study focused on the serotonin transporter gene expression variation. This gene has been robustly associated to suicidal behavior (Antypa et al., 2013, Purselle and Nemeroff, 2003). The serotonin transporter (5-HTT) is a major biological regulator of the serotonergic system (Lesch et al., 1994, Rudnick, 2006), a dysfunction of which is one of the most studied potential biological distal risk factor for suicide (Turecki, 2014). 5-HTT is a 7-transmembrane protein localized in presynaptic neuron that regulates the serotonin action by reuptaking the monoamine after its release in the synapse. The 5-HTT is encoded by the SLC6A4 gene located on chromosome 17. Post mortem studies revealed a decreased concentration of the 5-HTT protein in the prefrontal cortex, the hypothalamus and the brainstem of depressed patients died by suicide compared to matched controls (Purselle and Nemeroff, 2003, Stanley et al., 1982). In this context, some authors have also compared the expression levels of SLC6A4 mRNA in post mortem brains between suicide completer and healthy-matched control samples, without showing any difference (Arango et al., 2001, Little et al., 1997, Perroud et al., 2010). By quantifying regional brain 5-HTT binding in vivo using PET, Miller et al. (2013) reported that depressed suicide attempters had lower 5-HTT binding in midbrain compared with depressed non-attempters and healthy controls. They suggested that low 5-HTT binding was more related to suicidal behavior than to depression while another study did not replicate this finding (Oquendo et al., 2016).

Although brain blood correlation is still a matter of debate in biomarker development in psychiatry (Belzeaux et al., 2017a), several lines of evidence demonstrate the interest of studying SLC6A4 gene expression in blood samples. Indeed, various investigators have reported differences of SLC6A4 mRNA levels expression between depressed patients and controls (Belzeaux et al., 2012, Belzeaux et al., 2014b, Iga et al., 2005, Lima et al., 2005, Tsao et al., 2006). We previously demonstrated that SLC6A4 variation could be due to the occurrence and treatment of MDE (Belzeaux et al., 2014a).

As a consequence, we aimed to investigate if SLC6A4 gene expression in peripheral mononuclear cells (PBMCs) could predict the WSI during an 8-week follow-up among patients with a severe MDE. First, we aimed to examine the stability and reproducibility of SLC6A4 mRNA expression over time in healthy control. A secondary aim of the study, using a confirmatory approach, was to explore the link between the variation in SLC6A4 gene expression and the occurrence of suicide attempts in a 30-week follow-up among the MDE patients.