Introduction

It is now widely accepted that the psychological framework of each individual is shaped by a complex interplay of environmental, biological and genetic factors; the latter constitute an intriguing field of research, as a better understanding of the genomic variants concurring to a specific personality archetype or psychological trait may serve as an invaluable tool to comprehend and predict the intricacies of personal and collective mechanisms of information processing. Genetics has played a prominent role in psychiatric research since the very beginning of the era of etiological investigation of mental diseases (Burmeister et al., 2008), especially, regarding bipolar disorder and schizophrenia (Escamilla and Zavala, 2008; Henriksen et al., 2017). This intuition was later applied also to psychological and personality traits (Jang et al., 1996). Psychiatry genetics also benefitted greatly from recent advances in sequencing technologies; large-scale investigations such as genome-wide association studies (GWASs), in which thousands or even hundreds of thousands of genomes are correlated to a specific trait with a hypothesis-free approach, rather than considering a single gene-trait interconnection (Collins and Sullivan, 2013; Visscher et al., 2017), are now feasible.

Starting from this promising foothold, research into the field has expanded exponentially; among the various psychopathological domains, personality traits were one of the most obvious subjects to investigate, due to their unambiguous characterization (Rao and Broadbear, 2019), which retains its validity even in presence of potentially confounding sociocultural factors (Ayinde and Gureje, 2021) (at least, if compared to other psychological entities) (Zimmerman et al., 2015).

Most of the research on these topics employs the NEO Personality Inventory (named after its main original components, neuroticism, extraversion and openness to experiences) the same paper is already cited shortly after (McCrae), a well-recognized tool that explores five personality dimensions while providing an unparalleled perspective on multiple psychological domains (McCrae et al., 2005).

Some of these traits share numerous features and, rather than being distinct entities, contribute to complex personality structures that in certain occasions also predispose to psychiatric morbidity. One such case is constituted by alexithymia, emotional dysregulation and neuroticism, which all fall into a category characterized by inadequate, exaggerated or inhibited emotional responses, with frequent mood shifts and increased incidence of psychiatric conditions (Bradley et al., 2011; Ormel et al., 2013; Hemming et al., 2019). These entities in turn serve as a manifestation of this spectrum of features, hence identifying their common genetic roots might serve as an adequate instrument to describe the higher rank system encompassing them. These traits have an estimated heritability factor between 30 and 60% (evaluated through twin concordance studies) (Picardi et al., 2011; Hawn et al., 2015; Boomsma et al., 2018).

Several association analyses have been carried out on these traits (see the Materials and methods section); however, an investigation on their causative factors while considering them as a complex, multifacet – but still individual – entity might provide a whole new perspective on the subject. As shedding light on these constructs might prompt the development of new clinical and procedural approaches, and better understanding of their genetic basis might serve as a solid starting point, a review concerning these specific personality traits and their gene associations might as well prove valuable. To achieve this, the first step was to clearly define these traits separately, to subsequently merge the inferred information in a comprehensive report.

Neuroticism

Neuroticism as a concept was first theorized explicitly by Eysenck around 1950 (Eysenck, 1952).

People high in neuroticism (both meant as test scores and theoretical conception) are more prone to negative emotions than average; they also tend to be more impulsive, less able to delay gratification and more likely to suffer from life stressors. Neuroticism can, thus, be defined as the tendency to experience negative emotions such as fear, anxiety, irritability, feelings of guilt and anger (Widiger and Oltmanns, 2007). These difficulties often heavily influence the individual’s ability to navigate effectively social contexts (Messina et al., 2010).

The concept of neuroticism is not only useful in the field of personality psychology but also when taking into account psychiatric clinical practice. In fact, it has been proven that people with high neuroticism scores have a higher chance of being diagnosed with a mental health disorder, especially regarding internalizing conditions such as anxiety and depressive disorders (Khan et al., 2005; Gale et al., 2016).

Currently, neuroticism is measured through a wide variety of scales. Perhaps the most well-known is the NEO Personality Inventory, a test made up of 240 items (McCrae et al., 2005). Other tests often used are Eysenck Personality Inventory (EPI) and Eysenck Personality Questionnaire (EPQ) (Knowles and Kreitman, 1965; Loo, 1979). Both these tests have a similar outlook on what ‘neuroticism’ means while differing in some aspects (Rocklin and Revelle, 1981).

Alexithymia

The alexithymia construct, first introduced by Nemiah and Sifneos in the early 70s (Sifneos, 1986), was the point of arrival of a decades-long research into the cognitive style of patients with psychosomatic diseases. The salient features initially identified encompassed the difficulty in recognizing and describing feelings in the context of an externally oriented cognitive style, as well as the perplexity in distinguishing between feelings and the bodily sensations of emotional arousal. This inability to correctly interpret bodily inputs can often lead to psychosomatic symptoms and reduced insight. Moreover, alexithymia has been linked to psychopathology in the spheres of borderline personality disorder, eating disorders and psychosis (Mannarini et al., 2016). Neuroscientific studies on alexithymia point toward the importance of amygdala functioning (Goerlich, 2018).

The main tools for the evaluation of alexithymia are the Toronto Alexithymia Scale (Taylor et al., 2003) and the Schalling-Sifneos Personality Scale (Sifneos, 1986). Alexithymia has been often associated with several psychopathological conditions, most prominently depression and eating disorders (Taylor, 1984). Upon the realization of its validity in the clinical and research fields, alexithymia was included as an item in various assessment and diagnostic questionnaires and furtherly investigated by dedicated inventories (Sifneos, 1973; Montagne et al., 2007; de Vroege et al., 2018).

Emotion dysregulation

Emotion dysregulation (ED) is a wide and multifacet psychological concept (Thompson, 2019). It encompasses several different psychological traits and is associated with Borderline personality disorder (Carpenter and Trull, 2013), autism spectrum disorders (Cai et al., 2018), attention deficit hyperactivity disorder (Shaw et al., 2014), post traumatic disorder (Powers et al., 2015) and bipolar disorder (Bayes et al., 2016). Moreover, it often leads to substance abuse (Garke et al., 2021), self-harm (Gratz and Roemer, 2008) or even suicidal behavior (Raudales et al., 2020). ED is thought to be connected to abuse and trauma, especially during childhood (Dvir et al., 2014). This personality feature is characterized by incapacity or difficulty in modulating emotions in order to fit them to the social context. People with ED often suffer from poorer attention, labile mood and overly intense emotions. This trait often produces abnormal behavior both in the externalizing and internalizing spectrum. ED has been shown to be linked to neuroticism (Paulus et al., 2016), and interestingly, there is even evidence of this connection from a neuroscientific standpoint (Yang et al., 2020; Silverman et al., 2019). According to Gratz and Roemer (2004), ED can be defined as a deficit in awareness and acceptance of emotions, with a lack of control of one’s impulsive behavior. As such, ED often results in difficulties in employing appropriate strategies in social contexts.

Due to its broad definition, ED is investigated through a variety of scales, such as Difficulties in Emotion Regulation Scale (Gratz and Roemer, 2004), Emotion Dysregulation Scale (Powers et al., 2015), Emotional Expressivity Scale (Burgin et al., 2012), Connor-Davidson Resilience Scale (Connor and Davidson, 2003), Emotion Regulation Checklist (Shields and Cicchetti, 1997), Personality Assessment Inventory (Venta et al., 2018) and Sleep and Emotional Reactivity in Alcohol Use Disorder (NCT04979507, 2021). The multitude of instruments employed to assess ED might strengthen the belief that it might be a somewhat ill-defined concept in current literature (Cole, 2014); this translates into the tendency to study it from different perspectives and angles. However, such a tendency might as well negatively affect the reproducibility of the samples, which were attributed this trait by using different scales.

Objective

The aim of this systematic review is to investigate the influence of genetics on these three traits and to categorize known data on the subject, both from a GWAS and genetic association study perspective.

Materials and methods

This review adheres to the 2020 PRISMA guidelines (Page et al., 2021).

Eligibility criteria

Included articles were observational studies, either cross-sectional or longitudinal. The inclusion criteria were as follows: original article, written in English, reporting results of genetic analysis on humans in combination with measurements of neuroticism, ED or alexithymia using validated tools (questionnaires or tasks). The main outcome measures were all associations of either neuroticism, ED or alexithymia with specific alleles or intergenic variants. Exclusion criteria were: the study being a systematic review, a meta-analysis, an opinion article and methodological or technical contributions with no analysis over clinical data.

Information sources and search strategy

The authors used the electronic database PubMed in order to select studies. The following string was used for the systematic search:

(emotion regulation[Title/Abstract] OR emotion dysregulation[Title/Abstract] OR alexithymia[Title/Abstract] OR neuroticism[Title/Abstract]) AND (genetics OR gene OR genome OR GWAS OR genome-wide association study) NOT review[pt] NOT systematic review[pt] NOT meta-analysis[pt] AND eng[la].

The last search was run on 14 July 2021.

Selection process

Three authors (O.B.B, G.P.M. and V.P.) independently assessed the abstracts of potentially eligible studies. Eligibility assessment was performed in an unblinded standardized manner. If there was doubt about whether the study was eligible for inclusion, the reviewers examined the full text of the articles. The published protocol required consensus in case the authors disagreed on the inclusion of a specific study. In case, the opinion was not unanimous, a majority vote would have been taken between all authors. The authors agreed on all the eligibility assessments of the studies, and no consensus vote needed to take place.

Data collection process and data items

Four authors (O.B.B., V.P., G.P.M. and B.B.) independently extracted the following categories of data from each included study: study design (GWAS or genetic association), population (number of subjects, ethnicity), genes studied, polymorphisms and their effect on the traits.

Risk of bias

Risk of bias for individual studies was assessed using the STrengtheningthe REportingof Genetic Association Studies variant of the STROBE checklist for genetic association studies (Little et al., 2009). GWAS were not assessed through risk of bias, as no proper tool is available for such aim. An online tool was used to produce the summary graph S1 (McGuinness and Higgins, 2021).

Results

A total of 1340 studies were found after running the search line through PubMed. Three hundred one studies were included for evaluation of the manuscript, 1039 were excluded on the basis of title and abstract and 99 were excluded after manuscript review and application of inclusion criteria. Two hundred two studies (Rinieris et al., 1980; Ball et al., 1997; Gelernter et al., 1998; Jorm et al., 1998, 2000; Deary et al., 1999; Flory et al., 1999; Gustavsson et al., 1999; Sirota et al., 1999; Du et al., 2000; Greenberg et al., 2000; Henderson et al., 2000; Lerman et al., 2000; Persson et al., 2000; Sher et al., 2000; Jang et al., 2001; Close Kirkwood et al., 2002; Stoltenberg et al., 2002; Brummett et al., 2003; Eley et al., 2003; Fullerton et al., 2003; Sen et al., 2003,2004; Umekage et al., 2003; Westberg et al., 2003,2009; Jacob et al., 2004; Samochowiec et al., 2004; Tsai et al., 2004; Ham et al., 2005; Kato et al., 2005; Kusumi et al., 2005; Lang et al., 2005; Neale et al., 2005; Olsson et al., 2005; Tochigi et al., 2005, 2006; Wacker et al., 2005; Willis-Owen et al., 2005; Beem et al., 2006; Dragan and Oniszczenko, 2006; Hettema et al., 2006,2008, 2009, 2013,2015; Hibino et al., 2006; Hoth et al., 2006; Koller et al., 2006; van Rijn et al., 2006; Fiocco et al., 2007,2009; Gatt et al., 2007; Gutknecht et al., 2007; Hünnerkopf et al., 2007; Middeldorp et al., 2007, 2010, 2010; Pascual et al., 2007; Urata et al., 2007; Waga and Iwahashi, 2007; Wasserman et al., 2007; Wray et al., 2007, 2008, 2008, 2009; Gillespie et al., 2008; Heck et al., 2008; Kazantseva et al., 2008,2011; Rietschel et al., 2008; Shifman et al., 2008; van den Oord et al., 2008; Wachleski et al., 2008; Harro et al., 2009; Joffe et al., 2009; Juhasz et al., 2009,2010; Kochanska et al., 2009; Murakami et al., 2009; Schmitz et al., 2009; Terracciano et al., 2009,2010; Unschuld et al., 2009; Antypa and Van der Does, 2010; Calboli et al., 2010; Luciano et al., 2010,2012,2018,2021; Vinberg et al., 2010; Zuo et al., 2010; DeYoung et al., 2011; Ellis et al., 2011; Swart et al., 2011; Verschoor and Markus, 2011; Walter et al., 2011; Amin et al., 2012; Amstadter et al., 2012; Boscarino et al., 2012; Dar-Nimrod et al., 2012; Jutras-Aswad et al., 2012; Kano et al., 2012; Kuepper et al., 2012; Kurrikoff et al., 2012; Markus and Capello, 2012; McIntosh et al., 2012; Montag et al., 2012,2013; Pełka-Wysiecka et al., 2012; Propper et al., 2012; Wahlstrom et al., 2012; Wingbermühle et al., 2012; Aragam et al., 2013; Barbato et al., 2013; Glocke et al., 2013; Grazioplene et al., 2013; Kim et al., 2013,2015; Kuhnen et al., 2013; Lehto et al., 2013,2015,2016; Mandelli et al., 2013; Markus, 2013; Strohmaier et al., 2013; Zayats et al., 2013; Chang et al., 2014,2017,2020; Criado et al., 2014; Gong et al., 2014; Haram et al., 2014; Laas et al., 2014; Lovallo et al., 2014; Plieger et al., 2014; Weiss et al., 2014; Jurczak et al., 2015; Koh et al., 2015,2016; Kotyuk et al., 2015; Kruschwitz et al., 2015; Mezzavilla et al., 2015; Narita et al., 2015,2018; Peciña et al., 2015; Roelofs et al., 2015; Voigt et al., 2015; Jung et al., 2016; Larsen et al., 2016; Madsen et al., 2016; Okbay et al., 2016; Petito et al., 2016; Powers et al., 2016; Schneider-Hassloff et al., 2016; Schneider-Matyka et al., 2016; Smith et al., 2016; Fan et al., 2017; Halldorsdottir et al., 2017; Iliadis et al., 2017; Lo et al., 2017; Viddal et al., 2017; Xu et al., 2017; Bîlc et al., 2018; Bizaoui et al., 2018; Chapman et al., 2018; Nagel et al., 2018,2020; Noroña et al., 2018; Panitz et al., 2018; Sacchinelli et al., 2018; Terock et al., 2018, 2021; Turley et al., 2018; Yen et al., 2018; An et al., 2019; Chong et al., 2019; Luo et al., 2019; Michałowska-Sawczyn et al., 2019; Rodríguez-Ramos et al., 2019; Zou et al., 2019; Grzywacz et al., 2020; Hill et al., 2020; Hou et al., 2020; Kataja et al., 2020; Li et al., 2020; Morris et al., 2020; Salinas et al., 2020; Suchanecka et al., 2020; Tabak et al., 2020; Vaht et al., 2020; Zhao and Liu, 2020; Belonogova et al., 2021; Byrd et al., 2021; Heilbronner et al., 2021; Nestor et al., 2021) were finally selected: 27 GWAS and 175 observational genetic association studies. Of the latter, 142 were on neuroticism, 20 on alexitimia and 13 on emotional dysregulation (Fig. 1). Full results are shown in Supplementary Material S2, Supplemental Digital Content 1, https://links.lww.com/PG/A295, due to the unwieldy size of the result table. Below are the results concerning the most represented genes among the genetic association studies on neuroticism: serotonin transporter (SLC6A4, Table 1), catechol-ortho-methyltransferase (COMT, Table 2), monoamine oxidase type A (MAO-A, Table 3) and brain-derived neurotrophic factor (BDNF, Table 4). The outcomes of the screening on alexithymia are shown in Table 5 and on ED in Table 6. Results concerning GWAS studies are shown in Table 7.

Table 1 - SLC6A4 review results (neuroticism) Authors Publication year Tool used to assess neuroticism Polymorphisms studied Genetic association (S = short allele; L = long allele) Sample population Antypa and Van der Does 2010 NEO-PI-R 5-HTTLPR biallelic and triallelic - 250 Dutch students (mostly females) Ball et al. 1997 NEO-FFI 5-HTTLPR biallelic, STIN2-VNTR - 2085 German twins Brummett et al. 2003 NEO-PI-R 5-HTTLPR biallelic L 103 depressed patients and 99 controls Chang et al. 2017 MPI 5-HTTLPR triallelic (rs25531G coded equivalent to S) S, both biallelic and triallelic (only in men) 1340 Taiwaneses Chang et al. 2020 MPI 5-HTTLPR triallelic S in males, L in females 2236 Han Chinese adults, including 736 patients with GAD and 1500 healthy participants Deary et al. 1999 NEO-FFI 5-HTTLPR biallelic, STIN2-VNTR - 809 men and 783 women from Scotland Dragan and Oniszczenko 2006 NEO-FFI 5-HTTLPR biallelic S 200 Polish (only females) Du et al. 2000 NEO-FFI 5-HTTLPR biallelic S (only in males) 186 Canadians Flory et al. 1999 NEO-FFI 5-HTTLPR biallelic - 271 Americans Gelernter et al. 1998 NEO-FFI 5-HTTLPR biallelic, STIN2-VNTR - 322 Americans Greenberg et al. 2000 NEO-PI-R 5-HTTLPR biallelic S 397 American sisters (thus females only) Gustavsson et al. 1999 KSP 5-HTTLPR biallelic, STIN2-VNTR - Two healthy samples from Sweden (127 and 178) Harro et al. 2009 EBBFI 5-HTTLPR biallelic S 1176 Estonian children Hettema et al. 2015 NEO-FFI, EPQ 5-HTTLPR biallelic, rs3813034, rs140701, rs6354, rs2020936 - 928 Americans Jacob et al. 2004 NEO-PI, TPQ 5-HTTLPR biallelic S (only within Cluster C patients) 320 patients with personality disorders (mainly cluster B and C) and 281 healthy controls, from Germany Jang et al. 2001 NEO-PI-R 5-HTTLPR biallelic S 388 American siblings Jorm et al. 1998 EPQ-R 5-HTTLPR biallelic - 759 Polish Juhasz et al. 2010 NEO-PI-R, BFI 5-HTTLPR biallelic - 1188 English (history of anxious or depressive psychopathology)