Keywords

1 Borderline Personality Disorder

Borderline personality disorder (BPD) is a serious mental disorder characterized by a pervasive pattern of instability in affect regulation, impulse control, interpersonal relationships, and self-image. BPD is discussed to be part of the bipolar (Agius et al. 2012) or of the atypical depressive spectrum (Gremaud-Heitz et al. 2014) but less of the schizophrenia spectrum (Dammann 2004).

Clinical signs of BPD include emotional dysregulation, impulsive aggression, repeated self-injury, and chronic suicidal tendencies, which make these patients frequent users of mental health resources (Lieb et al. 2004; Leichsenring et al. 2011). The nine criteria for borderline personality (Diagnostic and Statistical Manual of Mental Disorders: DSM-IV) (APA 1994) can be organized into four sectors of psychopathology: (a) affective disturbance (a range of intense dysphoric affects, sometimes experienced as aversive tension or tremendous mood reactivity), (b) disturbed cognition (overvalued ideas of being bad, experiences of dissociation in terms of depersonalization and derealization, and nondelusional suspiciousness and ideas of reference), (c) impulsivity (e.g., deliberately physically self-destructive and more general forms of impulsivity, disordered eating), and (d) intense unstable relationships (e.g., profound fear of abandonment or tumultuous quality to close relationships). The lifetime prevalence has been estimated to be about 5.9 % (Grant et al. 2008). The point prevalence rates are estimated to be about 0.7 % in European national community samples (Torgersen et al. 2001). Additionally, findings showed that BPD was more common in women than in men (about 70 % and 30 %, respectively) (Widiger and Weissman 1991). The disorder often co-occurs with mood, anxiety, and substance use disorders and is also associated with other personality disorders. The disorder is characterized by severe psychosocial impairment (Skodol et al. 2002a) and a high mortality rate due to suicide – up to 10 % of patients commit suicide, a rate almost 50 times higher than in the general population. Several measures (semi-structural interviews) are highly reliable in the care of these patients (Zimmerman 1994). Borderline personality seems to be less stable over time than expected for personality disorders (Zanarini et al. 2006). Causal factors are only partly known, but genetic factors (Siever et al. 2002; Skodol et al. 2002b; Torgersen et al. 2008) and adverse events during childhood, such as physical and sexual abuse (Soloff et al. 2002), contribute to the development of the disorder.

Etiologies such as trauma, abuse, childhood adversity, and exposure to war appear to influence posttraumatic stress disorder, BPD, emotional dysregulation, empathy, and impulsivity via epigenetic mechanisms (Ripoll et al. 2013). Accumulating evidence points to severe emotional and relationship dysfunction as the core epigenetic expressions of BPD (Steele and Siever 2010). In adulthood, BPD is typified by disorganization within and across interpersonal domains of functioning (Clarkin et al. 2004). When interacting with their infants, mothers with BPD show marked withdrawal and frightening or frightened behavior, leading to disorganized infant-mother attachments. Linked to both infant disorganization and BPD is a maternal state of mind typified by unresolved mourning regarding past loss or trauma. Early risk factors for BPD in adulthood include maternal withdrawal in infancy and separation of 1 month or more from mother in the first 5 years of life (Steele and Siever 2010). The evolutionary based complex gene-environment picture emerging confers risk or protection against BPD pathology in ways consistent with infants varying biological sensitivity to context (Dammann 2003). Maternal separation is influencing epigenetic regulation (Toda et al. 2014).

Epigenetic mechanisms are relevant for the description of the endogenous oxytocin system, playing a central role in social cognition, disturbed in BPD (Puglia et al. 2015). Various forms of epigenetic regulation at the levels of DNA methylation, histone modification, and noncoding RNAs (ncRNAs) can modulate transcriptional and translational events required for memory processes. By changing the cellular profile in the brain’s emotional, reward, and memory circuits, these epigenetic modifications have also been linked to perseverant, pathogenic memories typical for posttraumatic stress disorder (PTSD) and BPD (Sillivan et al. 2015). Because epigenetic marks (e.g., DNA methylation) can be modulated by environmental stimuli, DNA methylation may be important in the development of anxiety (Alisch et al. 2014), suicidality (Turecki 2014), and vulnerability for stress (Tsuji et al. 2014) common in BPD.

2 Aberrant DNA Methylation in Borderline Personality Disorder

Aberrant epigenetic modifications are hallmarks of altered gene expression and contribute to several diseases including cancer and psychiatric illness (Dammann et al. 2000; Esteller 2007; Jones and Baylin 2007; Tsankova et al. 2007; van Vliet et al. 2007; Abdolmaleky et al. 2008). Epigenetic regulation is mediated through altered chromatin modifications, including changes in DNA methylation levels and aberrant histone marks. Gene promoters are often located in G/C- and CpG-rich DNA sequences which are termed CpG island promoters (Bird 1986). Active CpG island promoters are associated with unmethylated DNA sequences and open chromatin structures. In contrast, inactive promoters are in a repressed chromatin state: CpG sites are hypermethylated and histones are deacetylated. More than 70 % of genes (e.g., housekeeping genes) are transcribed from unmethylated CpG island promoters (Saxonov et al. 2006; Weber et al. 2007). The other genes are transcribed from G/C- and CpG-poor and are rather methylated in the human genome.

For borderline personality disorder, the epigenetic regulation of some disease-associated genes has been investigated in detail (Table 9.1). In a first study, we have analyzed the DNA methylation pattern of 14 neuropsychiatric genes: COMT, DAT1, GABRA1, GNB3, GRIN2B, HTR1B, 5-hydroxytryptamine (serotonin) receptor 2A (HTR2A), 5-HTT, monoamine oxidase A (MAOA), monoamine oxidase B (MAOB), NOS1, glucocorticoid receptor (GR/NR3C1), TPH1, and TH (Dammann et al. 2011). DNA methylation was analyzed by bisulfite restriction analysis and pyrosequencing in whole blood samples of 26 patients diagnosed with BPD and 11 controls. Aberrant methylation was undetectable by bisulfite restriction analysis. By utilizing pyrosequencing, a significant hypermethylation of HTR2A (1.2-fold), GR/NR3C1 (1.2-fold), MAOB (1.1-fold), MAOB (1.1-fold), and soluble COMT (1.1-fold) was revealed for BPD patients (Table 9.1). Methylation of DAT1, GABRA1, GNB3, GRIN2B, HTR1B, 5-HTT, NOS1, TPH1, and TH was not yet analyzed by pyrosequencing in further details (Dammann et al. 2011).

Table 9.1 Summary of DNA methylation analysis in borderline personality disorder (BPD)
Full size table

3 Aberrant DNA Methylation of GR/NR3C1 and BDNF in BPD

For the neuron-specific glucocorticoid receptor (GR/NR3C1), decreased levels of mRNA as well as increased methylation of the promoter were observed in hippocampus obtained from suicide victims with a history of childhood abuse (McGowan et al. 2009). We and others have reported increased methylation of GR/NR3C1 in borderline personality disorder (Dammann et al. 2011; Perroud et al. 2011; Steiger et al. 2013; Martin-Blanco et al. 2014). In Fig. 9.1, we have indicated CpG sites that exhibited significant changes in methylation levels in these studies. Perroud et al. have analyzed the methylation of exon 1F of GR/NR3C1 in 101 BPD and 99 major depressive disorder (MDD) subjects (Perroud et al. 2011). They report that childhood maltreatment and its severity (e.g., the number of maltreatments) positively correlated with increased NR3C1 methylation (Perroud et al. 2011). For example, they revealed a significant correlation between sexual abuse and GR/NR3C1 methylation showing a 1.34 higher methylation in sexually abused subjects than in nonsexually abused individuals (Table 9.1). These data suggest that early life events may permanently impact the hypothalamic-pituitary-adrenal axis through epigenetic alterations of the GR/NR3C1 (Perroud et al. 2011). Others have analyzed GR/NR3C1 methylation in lymphocytes of women with eating disorder, childhood abuse, or BPD (Steiger et al. 2013). They reported a significant 1.4- to 1.5-fold increased methylation at two out of 18 CpG sites adjacent of exon 1C in women with BPD and bulimia nervosa compared to subjects without BPD (Table 9.1 and Fig. 9.1). No parallel effects owing to childhood abuse were observed (Steiger et al. 2013). Martin-Blanco et al. have analyzed the methylation of GR/NR3C1 (exon 1F) in blood isolated from 281 BPD subjects (Martin-Blanco et al. 2014). They reported a significant positive correlation between GR/NR3C1 methylation and childhood maltreatment (e.g., physical abuse or emotional neglect) (Table 9.1 and Fig. 9.1). Absence of employment was also associated with increased GR/NR3C1 methylation (Martin-Blanco et al. 2014).

Fig. 9.1
figure 1

Genomic organization of the analyzed NR3C1/GR exons at chromosome 5q31.3. CpGs are show as vertical lines and significant aberrantly methylated CpGs are marked with asterisks. Transcription start sites are depicted with arrows and genomic positions are indicated. Graphics were generated with the python.vs.cobra program (https://launchpad.net/python.vs.cobra)

Full size image

Methylation of the brain derived neurotrophic factor (BDNF) gene has also been analyzed in psychiatric disorders, including BPD (Perroud et al. 2013; Mitchelmore and Gede 2014; Thaler et al. 2014). In Fig. 9.2, we have indicated genomic regions of BDNF that were investigated. In peripheral blood cells from 115 BPD subjects and 52 healthy controls, exon I and IV of BDNF were examined for DNA methylation (Perroud et al. 2013). Increased DNA methylation was observed in both exons of BDNF, compared to healthy controls (Table 9.1 and Fig. 9.2). In BPD individuals, a decrease in methylation levels was observed in responders to a 4-week psychotherapy course, whereas nonresponders showed increased methylation (Perroud et al. 2013). Thaler et al. have analyzed the methylation of BDNF in women with bulimic eating syndrome (Thaler et al. 2014). They report that subjects which BPD, eating disorder (bulimia nervosa), or childhood abuse exhibit increased level of methylation at specific CpGs of the exon IV promoter region (Table 9.1 and Fig. 9.2). They observed that patients with BPD and bulimia nervosa had significant higher methylation levels at four out of 27 CpG sites compared to 32 non-eating disorder women (Thaler et al. 2014). Methylation of the dopamine D2 receptor (DRD2) was also analyzed in women with BPD and eating disorder (Groleau et al. 2014). Increased methylation of DRD2 was revealed in eight subjects with BPD and compared to 19 women with no eating disorder and no BPD (Table 9.1). Subjects with eating disorder and without eating disorder did not differ in the mean methylation level of the DRD2 promoter (Groleau et al. 2014).

Fig. 9.2
figure 2

Genomic organization of the analyzed exon I and IV of BDNF at chromosome 11p14.1. CpGs are shown as vertical lines and analyzed CpGs are marked with x. Thaler et al. (2014) observed that patients with BPD and bulimia nervosa had significant higher methylation levels at four (asterisks) out of 27 CpG sites compared to non-eating disorder women. Transcriptions start sites are depicted with arrows and genomic positions are indicated. Graphics were generated with the python.vs.cobra program (https://launchpad.net/python.vs.cobra)

Full size image

In summary, these data suggest that aberrant methylation of disease-relevant genes (e.g., GC/NR3C1 and BDNF) occurs in peripheral blood of BPD patients, and these methylation marks may serve as candidate tool to identify potentially affected BPD subjects.

4 Epigenetic Alterations Detected by High Throughput Technology in Borderline Personality Disorder

Microarray (chip) technology is a high throughput assay to identify disease-related candidate genes. For microarray-based methylation analysis, two different methylation specific-bead chips from Illumina are available: the HumanMethylation27k bead chip (>27,000 CpG sites) and the second generation HumanMethylation450k bead chip (>450,000 CpG sites) (Bibikova et al. 2006; Sandoval et al. 2011). We have performed 27k methylation bead chip assay with bisulfite converted DNA from blood of 26 BPD patients and 11 controls (Teschler et al. 2013). As expected, our results show that no huge difference in methylation levels between BPD patients and controls are found (Teschler et al. 2013). Genome wide methylation levels of blood samples of BPD patients and control samples are rather similar. However, on several gene-specific CpG sites, bead chip technology and quantitative bisulfite pyrosequencing showed a significantly increased methylation of amyloid beta (A4) precursor protein-binding family A member 2 and 3 (APBA2: 1.1-fold and APBA3: 1.1-fold), potassium voltage-gated channel KQT-like subfamily member 1 (KCNQ1: 1.5-fold), MCF.2 cell line derived transforming sequence (MCF2: 1.1-fold), and ninjurin 2 (NINJ2; 1.2-fold) in BPD patients (Table 9.1). For the CpG sites of GATA4 and HLCS, an increase in DNA methylation was observed but was only significant in the bead chip assay (Teschler et al. 2013). These results show a significant 1.26-fold average increase in methylation at the analyzed gene-associated CpG sites in the blood of BPD patients compared to controls (Teschler et al. 2013). It is interesting to note that three of these genes (APBA2, APBA3, and NINJ2) are correlated with Alzheimer’s disease.

Others have utilized the HumanMethylation450k bead chip to analyze methylation level in DNA isolated from peripheral blood leukocytes in 96 BPD subjects suffering from a high level of child adversity and 93 subjects with major depressive disorder (MDD) and a low rate of child maltreatment (Prados et al. 2015). They report significant differential methylation of IL17RA, miR-124-3, KJCNQ2, EFNB1, OCA2, MFAP2, RPH3AL, WDR60, CST9L, EP400, A2ML1, NT5DC2, FAM163, and SPSB2 either in BPD compared with MDD or in relation to the severity of childhood maltreatment (Prados et al. 2015). Aberrant methylation of one CpG site of miR124-3 was also analyzed by pyrosequencing in BPD and MDD subjects; however, differential methylation was insignificant by this assay (Prados et al. 2015). The microRNA miR-124 promotes neuronal differentiation by triggering brain-specific alternative pre-mRNA splicing and could be involved in the regulation of GR/NR3C1 (Makeyev et al. 2007; Prados et al. 2015).

5 Altered Histone Modifications in Borderline Personality Disorder

Aberrant histone modifications represent another important epigenetic alteration that may cause changes in expression pattern of disease-related genes in borderline personality disorder. However, histone modifications were not investigated in BPD up to now. The main method to analyze histone modifications or histone variants is crosslinking chromatin in tissue samples and to utilize specific antibodies to precipitate and to quantify chromatin changes. For obvious reasons, relevant fresh tissue samples are rather difficult to obtain from human subjects. Candidate genes to investigate of aberrant histone modification or variants are BDNF, HTR2A, and GR/NR3C1 (Perroud et al. 2011; Mitchelmore and Gede 2014; Paquette and Marsit 2014).

6 Conclusions

Increased DNA methylation of several neuropsychiatric genes occurred in blood of BPD patients and was also linked to child abuse or clinical severity (e.g., GR/NR3C1) (Perroud et al. 2011; Martin-Blanco et al. 2014). A number of studies have suggested a role for environmentally mediated aversive events in the development of BPD and have found an association between the diagnosis of BPD and psychotraumatization during childhood (Herman et al. 1989; Famularo et al. 1991; Goldman et al. 1992; Paris et al. 1994; Waller 1994; Silk et al. 1995). In a family environment, early maltreatment (childhood abuse and neglect) within a family environment may be particularly important in producing long-term epigenetic changes (Gunnar and Quevedo 2007). In animal models, postnatal maternal care has been linked to epigenetic alteration via DNA methylation (Kaffman and Meaney 2007; McGowan et al. 2011). In the blood of BPD subject, aberrant methylation of several neuropsychiatric genes was revealed, and these findings are summarized in Table 9.1. In future work, it will be interesting to verify this methylation changes in other tissues and to reveal novel BPD-specific biomarkers.