Major depressive disorder (MDD), defined as at least one discrete depressive episode that lasts for more than two weeks, is a debilitating mental disorder that is estimated to affect about 6% of the adult population worldwide every year (1). When measured as “years lived with disability,” MDD constitutes the second leading burden of disease (2). Importantly, MDD is one of the main risk factors for death by suicide. According to the World Health Organization, every year over 800,000 people die due to self-inflicted injuries (3). This accounts for 1.5% of total death and makes suicide one of the leading causes of death in people 15–34 years of age (4). In this regard, it is also noteworthy that the number of non-fatal suicide attempts significantly exceeds the number of suicides, i.e., for each death by suicide, there are an estimated 25 suicide attempts (5, 6). In the context of MDD, results from the ESEMED study, a European cross-sectional household survey, indicate that attributable risk for lifetime suicide attempt was 28% for MDD and additionally, major depressive episode together with alcohol dependency determined most strongly the progression from suicidal ideation to suicide attempt (7).

Currently, clinicians determine suicide risk in psychiatric patients primarily based on information regarding prior suicidal behavior and a family history of suicide as provided by patients. However, reliability of self-reported information concerning suicidal ideation is limited (8) and information regarding the family history of suicidal behavior might be inaccurate due to the associated stigma.

Given the significant personal as well as economic impact that suicide imposes especially in patients with MDD, reliable markers to objectively identify suicide risk are required to improve diagnosis, classifications and outcomes.

The etiology of MDD and suicidality is currently considered to involve a complex interplay of genetic and environmental factors (9). Accordingly, research targeting the identification of potential biomarkers for suicide risk in patients with MDD has focused on structural and functional brain alterations assessed by brain imaging as well as on biochemical and genomic markers associated with systems implicated in depression and suicidal behavior (10). However, based on the well-established association of early as well as of acute adverse life events with MDD and suicide risk (11–16), epigenetic markers have recently come into focus (17). While epigenetic mechanisms are classically defined as modulations of DNA organization in form of DNA methylation and histone modifications (18–20), more recently, non-coding RNAs have been identified as post-transcriptional regulators that mediate stress-induced changes in the context of psychiatric disorders (20). Within the group of non-coding RNAs, microRNAs (miRNAs) were first described as single-stranded RNA molecules with a length of 19–25 base-pairs that bind to 3’-untranslated regions (3’UTRs) of their respective target mRNAs resulting in repression of translation or alterations in mRNA stability (21–23). Since then, various studies have investigated the involvement of miRNAs in the etiopathology of psychiatric disorders including MDD and suicidal ideation (24). Importantly, alterations in miRNA levels are not limited to the brain but distinct changes were also detected in blood samples of depressed patients highlighting their potential as putative biomarkers for MDD and suicidal behavior (24, 25).

We previously described alterations in miR-30a and miR-30e as well as in miR-200a levels in post-mortem brain samples of individuals with MDD that died by suicide compared to mentally healthy individuals that died of unrelated causes (26). Given the high prevalence of suicidality in MDD patients and the need for reliable biomarkers to determine suicide risk, we aimed to assess whether these observed dysregulations in miRNA expression could be replicated in a pilot study with depressed patients with and without suicidality and suicide attempt.

In the present study, we report significantly higher relative expression levels of three miRNAs (i.e., miR-30a, miR-30e, and miR-200a) in blood serum samples from suicidal MDD patients compared to an MDD group that did not display current suicide risk. It is noteworthy that our previous study demonstrated that expression of these miRNAs was dysregulated in post-mortem brain samples of MDD patients that died by suicide (26). Additionally, we propose transcription factor EGR1 as potential regulator of miRNA expression.

Studies dedicated to the identification of suitable biomarkers for MDD and therapeutic responses to psychopharmacologic treatment are numerous and subsequently potential molecular markers have been identified on the genomic, epigenomic, transcriptomic, and proteomic level (30, 31).

Contrarily, studies that aim to identify markers for suicidality are less frequent. In this regard, several studies assessed alterations in miRNA expression in different brain regions of suicide victims (32, 33). However, while these findings offer important insights in potential pathomechanisms underlying suicidal behavior, interpretation of results is complicated by an often unclear psychiatric diagnose and/or by the lack of an appropriate control group (32). Consequently, post-mortem studies often do not allow for a differentiation between miRNA regulation linked to suicidality and changes brought forward by the underlying psychiatric illness. Therefore, a replication of post-mortem results in clinical samples is of significant importance.

In the context of central nervous system (CNS)-related disorders, including psychiatric disorders and MDD, numerous studies have described alterations in miRNA levels in peripheral blood samples (25), and it has been shown that alterations in brain miRNA expression appear to be reflected in the periphery (34, 35). Albeit the need for reliable biomarkers that objectively predict suicide risk in MDD patients and the demonstrated potential of peripheral miRNAs as suitable biomarkers in the context of CNS-related diseases (34), a recent systemic review identified only two studies that assessed miRNAs as potential biomarkers for suicidality in peripheral blood samples (33) and neither study included patients with recent suicide attempt (36, 37).

This is of importance, as the observed higher levels of miR-30a, miR-30e, and miR-200a in suicidal MDD patients compared to those without current suicide risk in this study, could be mostly attributed to a subgroup of patients that had recently attempted suicide compared to those without a recent suicide attempt.

Our observation is in line with the understanding that suicidal behavior, defined as suicide attempt or suicide completion, is characterized by overlapping, psychobiological abnormalities that are discernable from suicidal ideation (38). Accordingly, it is generally thought that vulnerability to suicidal behavior is mediated partly by an underlying genetic predisposition, which interacts with environmental and probable epigenetic factors (39), and might be reflected by changes in miRNA expression. Indeed, several studies demonstrated changed miRNA expression profiles in post-mortem as well as peripheral blood samples from MDD patients (40, 41). In particular, our previous study identified altered expression levels of miR-30a, miR-30e, and miR-200a in prefrontal cortex (PFC) samples of suicide victims diagnosed with depression compared to mentally healthy controls that died of unrelated causes (26). Interestingly, an association analyses in MDD patients and control subjects identified polymorphisms within the miR-30e precursor to be positively correlated with depression and its symptomatic onset (42). Moreover, results from a social defeat stress mouse model indicated a downregulation of miR-30a and miR-30e in the context of stress and depression-like behavior and in hippocampal neurogenesis (43). Intriguingly, we have recently found that the higher prevalence for MDD and panic disorders in women with peripartum cardiomyopathy was associated with elevated levels of miR-30e (44). Finally, alterations in miR-200a expression have been reported in a rodent model of unpredictable chronic mild stress as well as in plasma samples of patients with Alzheimer’s disease (45–47).

Regarding potential upstream regulators of miR-30a, miR-30e, and miR-200a transcription, predictions based on two independent database searches identified EGR1 as a potential transcriptional regulator of the assessed miRNAs. EGR1, also known as NGFI-A, constitutes a nuclear phosphoprotein attributed with an activation potential as well as with a negative regulatory function (48–51). Among others, EGR1 was found to be critically involved in neuroplasticity (52). Decreased EGR1 expression was found in post-mortem PFC samples of MDD patients in comparison to healthy individuals (53) and its downregulation was also detected in animal models of depression and anxiety in specific brain areas including the hippocampus (53, 54). Importantly, a study based on transcriptome sequencing data derived from post-mortem PFC samples found a gradual downregulation of a gene cluster that included EGR1 in non-suicidal MDD patients compared to control with a further decrease being detected in suicidal MDD patients (55). Given the opposite regulation of EGR1 (i.e., decreased in the context of MDD and suicidality) and of the miRNAs assessed in this study (i.e., increased in the context of SI and SA), one might speculate that EGR1 could function as a negative regulator of these miRNAs. In this regard, repressor function of EGR1 has been previously shown in neuronal cells (56–58) and positive as well as negative regulatory effects of EGR1 on miRNA expression have been described in various tissues and disease states (59–63) and EGR1 binding sites within the predicted promoter region of miR-30a and miR-30e were identified (64). A potential transcriptional regulation of the relevant miRNAs by EGR1 and its direction should be concern of future studies.

Concerning potential targets of miRNA-30a, miR-30e, and miR-200a, in silico analysis identified 20 mRNAs that were predicted to be regulated by all three miRNAs via their respective 3’ UTRs. To our knowledge, out of these predicted targets, only three [i.e., muscleblind like splicing regulator 1 (MBNL1), receptor accessory protein 3 (REEP3), retinoic acid receptor beta (RARB)] have previously been reported in the context of depression or depression-like phenotypes in the literature and none were described to be associated with suicidality (65–68). In this regard, only RARB has been reported in the context of MDD in humans (68) and rodents (66), while MBNL1 and REEP3 were reported only in rodent models (65, 67). The respective studies utilized screening approaches and we are unaware of studies that have investigated the function of the respective genes in the context of depression and/or suicidality. While our in silico analysis gives a first indication of potential targets of the assessed miRNAs in the context of depression and suicidality, further studies are needed to confirm these targets in vitro and to assess their respective relevancy in the context of suicidality in patients with MDD.

In conclusion, we found that miRNAs previously described to be dysregulated in post-mortem brain samples of MDD patients that died by suicide, were elevated in serum samples of MDD patients that displayed current suicide risk in comparison to MDD patients that did not show acute suicidal ideation. Our results suggest that the observed miRNAs might function as a trait marker of suicidal behavior (i.e., suicide attempt and suicide completion) as the overall higher levels could be mainly attributed to a patient subgroup that attempted suicide in the week prior to study entry. Therefore, further studies might focus not only on the potential application of these miRNAs as biomarkers for suicidal behavior, but also on their respective involvement in the pathomechanisms that underlie suicidal behavior in general and in the context of MDD in particular.

The original contributions presented in this study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by Ethics Committee, Hanover Medical School. The patients/participants provided their written informed consent to participate in this study.

BS: formal analysis and writing—original draft. KX and NG: investigation, formal analysis, methodology, and visualization. KS: investigation and formal analysis. AP: methodology and writing—review and editing. JF: supervision, methodology, and writing—review and editing. IR and A-LV: resources. TT, KK, and EP: supervision, funding acquisition, project administration, conceptualization, formal analysis, and writing—review and editing. All authors contributed to the article and approved the submitted version.