S. Wagner, PhD, Department of Psychiatry and Psychotherapy, University Medical Center Mainz, Untere Zahlbacher Strasse 8, D-55131 Mainz, Germany. E-mail:email@example.com
Impulsive aggression belongs to the key features of borderline personality disorder (BPD). In the development of BPD, serious life events are known to play a major role. Acute and chronic stress has been suggested to inhibit hippocampal brain-derived neurotrophic factor (BDNF) synthesis and to mediate neural plasticity in response to adverse social experiences. Recently it has been reported that the frequency of violent suicide attempts is higher in adult suicide attempters reporting severe childhood sexual abuse and carrying the Val66Val genotype of the BDNF Val66Met polymorphism. In this study we analysed modulating effects of BDNF Val66Met polymorphism on the effects of physical maltreatment, rape and childhood sexual abuse on impulsive aggression. One hundred and fifty-nine BPD patients from Germany and of Caucasian descent were included. Impulsive aggression was assessed by the Buss-Durkee-Hostility Inventory (BDHI). Childhood sexual abuse accounted for 23.6% of the variance of BDHI sum score. Childhood sexual abuse decreased BDHI sum score in BDNF Val/Val carriers but not in Met carriers. In contrast to previous findings this study analysing a specific gene × environment interaction in BPD patients suggests a decreasing effect of childhood sexual abuse on impulsive aggression in BPD patients, particularly in BDNF Val/Val carriers. The interrelations between serious life events, impulsive aggression and the BDNF Val66Met polymorphism as well as their implication for BPD are far from understood and require further investigations.
In a recent study, Perroud and colleagues (2008) reported in this journal that the frequency of violent suicide attempts is higher in adult suicide attempters reporting severe childhood sexual abuse and carrying the Val66Val genotype of the brain-derived neurotrophic factor (BDNF) Val66Met polymorphism. This finding might be of particular importance for studies investigating patients with borderline personality disorder (BPD), because up to 71% of BPD patients report the experience of severe childhood trauma (Lieb et al. 2004) and ‘self-injury/suicide attempts' is established as diagnostic criterion for BPD (DSM-IV). BPD is a severe mental disorder with a complex symptomatology and a high mortality rate because of suicide (Clarkin & Posner 2005). Impulsive aggression belongs to its key clinical features (Fossati et al. 2004). There is evidence that impulsive aggression plays an important role for the suicide risk of BPD patients, that is, BPD patients who attempted or completed suicide exhibited more impulsive aggression than BPD patients who did not attempt suicide (Brodsky et al. 1997; McGirr et al. 2007).
In the etiology of BPD and its predominant clinical feature impulsive aggression, a general disease model of interacting genetic and environmental factors (G × E) has been proposed (Lieb et al. 2004). Regarding environmental factors, literature points to the association of severe childhood trauma with features of BPD (McLean & Gallop 2003). Specifically, the percentage of sexual abuse-related events was significantly higher in subjects with BPD than in patients with major depression or healthy controls (Horesh et al. 2008).
Acute and chronic stress has been suggested to inhibit hippocampal BDNF synthesis and to play an important role in mediating neural plasticity in response to adverse social experiences (Duman & Monteggia 2006). In many studies, the occurrence of serious life events has been linked to alterations of the serotonergic system (Caspi et al. 2003; Clarke et al. 1996; Higley et al. 1996) and BDNF was reported to be involved in the regulation and growth of the 5-HT neurons during childhood (Aquilera et al. 2009; Jiang et al. 2005). Therefore, the 5-HT dysfunction in depression and other stress-related disorders like BPD could be the result of a decreased expression of BDNF in limbic regions (Post et al. 2007). The BDNF gene is located on chromosome 11p13–14 (Maisonpierre et al. 1991). The 5′ pro-BDNF sequence (Egan et al. 2003) contains a functional G196A single nucleotide polymorphism (SNP), which results in a valine to methionine change at position 66. It has been shown in vitro that BDNF secretion was reduced in 66Met- compared with Val66-containing neurons as a result of abnormal intracellular packaging (Chen et al. 2004; Egan et al. 2003).
We are not aware of a specific report on the interaction of BDNF and impulsive aggression in BPD patients yet. A first study in our laboratory investigating the interaction between the BDNF Val66Met polymorphism and the HTR1B gene showed a gene–gene interaction with an increased susceptibility for BPD in patients with BDNF 66Met allele in the group of HTR1B A-161 allele carriers (Tadic et al. 2009).
Previous studies suggest that (1) BPD patients have higher Buss-Durkee-Hostility Inventory (BDHI) scores than patients with other personality disorders or healthy controls (Fossati et al. 2004; Tritt et al. 2005), (2) childhood sexual abuse decreases impulsive aggression in BPD patients (Wagner et al. 2009; Sack 2004) and (3) childhood sexual abuse increases the frequency of adult violent suicide attempts in BDNF Val66Val carriers (Perroud et al. 2008). Taken together, one could expect that the occurrence of serious life events increases impulsive aggression in BPD patients with BDNF Val66Val genotype. The purpose of the present study was to determine whether the BDNF Val66Met polymorphism modulates the effects of serious life events on impulsive aggression in patients with BPD.
All participants gave their written informed consent to participate in the study after a complete and extensive description. All study components were approved by the local ethical committees of the Landesärztekammer Rheinland-Pfalz and are compliant with the Code of Ethics of the World Medical Association (Declaration of Helsinki). One hundred and fifty-nine patients (110 females, 49 males) of Caucasian descent with DSM-IV BPD (American Psychiatric Association 2004) were recruited at the Department of Psychiatry in Mainz, Germany. Patients were not included if one or more of the following criteria were present: (1) dementia or any other organic brain syndrome with obvious cognitive impairment, (2) doubts in the patients' ability to give informed consent (e.g. patients suffering from depression with psychotic symptoms), (3) a current diagnosis of a psychotic disorder, (4) the patient was involved in a legal procedure at the time of study participation and (5) the patient was not of Caucasian descent.
The German version of the Structured Clinical Interview for DSM-IV axis II disorders (Wittchen et al. 1996a) was applied by trained research assistants to each patient. For the evaluation of DSM-IV axis I disorders, the computerized DSM-IV version of the Munich-Composite International Diagnostic Interview (M-CIDI) (Wittchen et al. 1998) was applied. Impulsive aggression was assessed by the BDHI (Buss & Durkee 1957). The BDHI has a long history of use in aggression research (Coccaro et al. 2007). It is a forced-choice self-report questionnaire consisting of 75 items with a two-point scale (1 = ‘yes’ and 0 = ‘no’). Serious life events were assessed by means of the PTSD-section included in the M-CIDI. This M-CIDI/PTSD-section comprises eight kinds of serious life events with a five-point scale (−2 = ‘answer denied’, −1 = ‘do not know’, 0 = ‘no serious life event in past', 1 = ‘no’ and 5 = ‘yes’). The different kinds of serious life events were experience of war, physical maltreatment, rape, childhood sexual abuse, natural disaster, severe accident, captivity/hostage-taking and witness of a serious life event. Because in previous studies physical maltreatment, rape and childhood sexual abuse were strongly associated with the development of BPD (Golier et al. 2003; Horesh et al. 2008; McLean & Gallop 2003) and the frequencies of the other serious life events in our sample were very small, we focused on these three variables and excluded the other from further analysis. In addition, the cumulative number of these serious life events was summed up for each subject.
DNA was obtained from venous blood using standard techniques. The BDNF G196A SNP (rs6265) was genotyped using a Sequenom (San Diego, CA, USA) chip-based matrix assisted laser desorption/ionisation (MALDI-TOF) mass spectrometry platform. In brief, polymerase chain reaction (PCR) and extension reactions were designed using Massarray version 3.1 design software, and were carried out using 2.5 ng of template DNA. Unincorporated nucleotides in the PCR product were deactivated using shrimp alkaline phosphatase. The primer extension products were then cleaned and spotted onto a SpectroChip with a Massarray nanodispenser. The chips were scanned using a mass spectrometry workstation (Massarray compact analyser, Sequenom) and the resulting spectra were analysed and genotypes were determined using the Sequenom spectroTYPER-RT software.
All analyses were performed using spss 15.0. First, the frequencies of independent variables were analysed. Differences in BDHI sum scores between subjects with and without serious life events were examined by t-tests for independent variables. Differences in BDHI sum scores between the three BDNF genotype groups were calculated with anova. Because the BDNF Met66Met genotype group was small, BDNF genotype was dichotomized as 66Met carriers and Val66Val carriers. To examine possible modulating effects of the BDNF Val66Met polymorphism, we applied hierarchical regression analyses separately for 66Met carriers and Val66Val carriers. Differences in BDHI sum score between the BDNF genotype groups with different parameter values in the serious life event measures were tested with t-test for independent variables or anova. P-values were adjusted for multiple testing with Bonferroni correction. Taking into account four tests (effect of physical maltreatment, rape, childhood sexual abuse and the number of serious life events on BDHI sum score), significance was set at P <.0125. We calculated the increment in variance (R2), because it describes the percentage of the dependent variable which is determined by the independent variable(s).
BPD subjects (mean age = 33 ± 9.5 years) had a mean BDHI sum score [(±standard deviation (SD)] of 45.9 (±11.0). This BDHI sum score is comparable with other studies in BPD patients (Fossati et al. 2007; Tritt et al. 2005). Concerning serious life events, 40% of the participants reported the occurrence of physical maltreatment, 38% of childhood sexual abuse and 31% of rape. The mean cumulative number (±SD) of these serious life events was 1.1 ± 0.9 (range: 0–3). Ninety-six patients (60%) were BDNF Val/Val carriers, 57 (35%) Val/Met carriers and eight (5%) BDNF Met/Met carriers. The distribution of genotype frequencies of the BDNF Val66Met polymorphism did not show any divergence from Hardy–Weinberg equilibrium (P > 0.1; df = 1). BDNF Val66Val and 66Met carriers did not differ in their mean BDHI sum score (±SD) (mVal/Val = 45.2 ± 11.0; mMet = 47.1 ± 11.0; t159 = 1048, P = 0.296). Concerning serious life events, there were no significant differences between BDNF Val66Val and 66Met carriers (data not shown; P > 0.10 for each analysis).
Childhood sexual abuse accounted for 23.6% of the variance of BDHI sum score (F1,159 = 9,411; P = 0.003). Possible modulating effects of the BDNF Val66Met polymorphism on the relation between serious life events and BDHI sum score were tested with linear regression procedures for BDNF Val66Val and 66Met carriers (Table 1). In BDNF Val66Val carriers, childhood sexual abuse accounted for 28.6% of the variance of BDHI sum score (F1,94 = 8,375; P = 0.005). In BDNF 66Met carriers, serious life events had no relevant effects on BDHI sum score.
Table 1. Effects of SLEs on BDHI sum score in BDNF Val66Val and 66Met allele carriers with BPD
†P-values in this table are descriptive and can not be regarded as significant at any level.
Childhood sexual abuse
Cumulative number of SLEs
In order to further investigate the relation between the relevant serious life event and the BDHI sum score, we analysed in detail the effect of childhood sexual abuse on mean BDHI sum score stratified by BDNF Val66Val and 66Met carriers. In BDNF Val66Val carriers, subjects with the occurrence of childhood sexual abuse had lower mean BDHI sum scores than those without childhood sexual abuse (t94 = −2894; P = 0.005; see Fig. 1). In BDNF 66Met carriers, subjects with and without childhood sexual abuse did not differ in their mean BDHI sum score (t63 = −1066; P = 0.290).
Our analysis of the relation between serious life events and impulsive aggression incorporating the BDNF Val66Met polymorphism showed that childhood sexual abuse had pronounced decreasing effects on BDHI sum score in BDNF Val66Val, but not in 66Met allele carriers. At first glance, this result is counterintuitive and not consistent with the results from the study of Perroud et al. (2008), who reported a higher frequency of violent suicide attempts in suicide attempters reporting severe sexual abuse and carrying the BDNF Val66Val genotype compared with BDNF Met66Met and Val66Met carriers. An explanation for our finding might be that patients who experienced serious life events, especially severe childhood trauma, perceive the environment being insecure and characterized by unknowable risks (Sack 2004). To cope with the feared risks and to defend themselves from suspected assaults the persons concerned of serious life events might develop a high necessity to control their behaviour and social interactions (Sack 2004). As a result, the necessity to control might reduce impulsive aggression. The effects of adverse environmental factors on impulsive aggression have mostly been investigated in subjects with antisocial personality disorder (ASPD) and reports in BPD patients are absent as yet. The studies in ASPD patients show that childhood abuse is associated with an increased risk to develop aggressive behaviour and APSD in adulthood (Frazzetto et al. 2007; Luntz & Spatz Widom 1994; Reif et al. 2007). However, studies comparing impulsive aggression in subjects with BPD and ASPD reported that BPD and ASPD patients show different kinds of impulsive aggression. ASPD is characterized by physical aggression and outwardly directed behaviour patterns, whereas BPD is characterized by the emotional components of impulsive aggression, that is, irritability and self-destructive behaviour patterns (Fossati et al. 2007; Tritt et al. 2005). Taken together, impulsive aggression in ASPD and BPD patients is not comparable. Therefore, the increasing effect of serious life events on impulsive aggression in ASPD patients is not transferable to BPD patients. However, the result of this study can not serve as evidence about a causal relationship between the occurrences of serious life events and reduced impulsive aggression, that is, it remains unclear whether impulsive aggression decreased after the occurrence of serious life events or already was decreased in a particular patient group before the occurrence of the serious life events. If this would be true, lower impulsive aggression would predispose to a greater risk for victimization and perhaps a diminished ability to adapt to or recover from such events. To clarify the relation between serious life events and impulsive aggression, longitudinal studies are essential.
Nevertheless, our result is in line with studies showing females with the occurrence of serious life events developing autoaggressive attitudes and female BPD patients tending to inwardly direct their emotions (Johnson et al. 2003, McCormick et al. 2007) as well as with our previous finding of childhood sexual abuse decreasing impulsivity in BPD patients (Wagner et al. 2009). Perroud and colleagues investigated suicide attempters suffering from a bipolar disorder, major depression or schizophrenia and related psychotic disorders. The results in these subjects might not be transferable to BPD patients. The association between impulsive aggression, serious life events and the BDNF Val66Met polymorphism as well as the development from childhood to adulthood is very complex and a variety of adaptive processes could appear in this progress. Literature points to the importance of protective factors and interindividual differences, implicating that some subjects experiencing serious life events do not develop mental disorders (Egle et al. 1997), whereas other subjects may develop a variety of possible phenotypes in result of the occurrence of serious life events. Thus, it seems possible that a particular patient group (66Val carriers with bipolar disorder, major depression, schizophrenia) might develop a tendency for violent suicide attempts in result of the occurrence of serious life events, whereas another patient group (66Val carriers with BPD) respond to the occurrence of serious life events with a decrease of impulsive aggression. Taken together, the interrelations between serious life events, impulsive aggression and the BDNF Val66Met polymorphism as well as their implication for BPD are far from understood and require further investigations.
It is important that childhood sexual abuse but not rape was associated with lower BDHI sum scores. An explanation for this finding might be that stressors early in development are more likely to affect the development of personality traits than those later in life (Schmahl et al. 2002). In our sample and by definition, subjects were significantly younger at the occurrence of childhood sexual abuse than at the occurrence of rape. Another explanation might be that the duration of a stressor is important for the development of a specific phenotype (Schmahl et al. 2002). One could suppose that single traumatic events like rape do not lead to alterations in impulsive aggression, whereas a long term exposure to stressors such as childhood sexual abuse could affect impulsive aggression.
A limitation of the study was that potential mediating factors such as social support or the parent–child relationship were not investigated. This is relevant, because the consequences of serious life events in an individuals' life are regulated by many different factors such as social support, other environmental factors and memories about the serious life events (Jonzon & Lindblad 2006). Further, the PTSD-section of the M-CIDI did not assess the duration and continuity of serious life events. However, the M-CIDI is a widely used international instrument to diagnose mental disorders and allows for a reliable and valid assessment of symptoms, syndromes and diagnoses of DSM-IV mental disorders. The advantage of the M-CIDI is its high objectivity and interrater reliability (Wittchen et al. 1996b).
In conclusion, this is the first investigation on a modulating effect of the BDNF Val66Met polymorphism on the relation between serious life events and impulsive aggression in BPD patients. We found that childhood sexual abuse was negatively associated with impulsive aggression in BDNF Val66Val carriers. In order to corroborate our findings, replications in independent populations addressing the limitations of our study are necessary taking into account the risk of false positive results in a single association study (Ioannidis et al. 2001). Future studies should extend our findings involving gene variants of the BDNF pathway and further neurotrophic factors. Further, it should be of particular interest, whether the BDNF Val66Met is associated with the regulation of other phenotypes of BPD.
A.T. received two grants from the University of Mainz faculty of medicine (MAIFOR 2007 and Neuroscience Focus Program 2008). We are grateful to all participants as well as to our co-operating clinical partner (Drs Guth and Krafczyk at the Clinic for Psychiatry in Alzey; Dr Lehr at the Center for Social Psychiatry in Eltville) for study support. Further we are grateful to Amelie Elsäßer (IMBEI, University of Mainz) for here helpful comments on statistics.