Self-injury is a complex and poorly understood behavior observed in people with psychopathology or neurodevelopmental disorders (NDD). Despite the differences in etiology and progression of these distinct disease domains, it is possible that overlapping molecular pathways underlie the expression of self-injurious behaviors (SIBs). This review outlines the similarities and differences at the behavioural and molecular level, where SIBs in both conditions may involve opioid, nucleoside, and dopamine signalling. These points of convergence have important implications for treatment and research of SIB in both populations.
Key points
- •
Self-injury (SI) is a harmful behavior present in persons with neurodevelopmental disorders (NDDs) such as Lesch-Nyhan syndrome, and those with psychopathological conditions such as major depressive disorder or borderline personality disorder.
- •
There appears to be a genetic underpinning to SI, given its prevalence in monogenic disorders and heritable psychiatric conditions.
- •
There is convergent evidence regarding the functions of SI in NDD and psychopathology.
- •
Between NDD and psychopathology, the roles of neurotransmitter systems in self-injurious behaviors (SIBs) converge on the opioid signaling system, diverge with respect to serotonergic signaling, and show mixed effects on dopaminergic and nucleoside signaling.
- •
Further study of commonalities between SI in psychopathology and NDDs will increase our understanding of this challenging behavior and may lead to improved treatment of SIBs.
Introduction
Self-injurious behaviors (SIBs) are observed in people with neurodevelopmental disorders (NDDs) and those with psychopathological conditions, although it is unclear whether the biological bases of SIBs in these 2 populations are similar. Evidence of similarities between the SIBs observed in patients with psychopathology and NDDs at the behavioral and molecular level, reviewed and analyzed here, suggests that the 2 fields of research could establish a synergistic relationship. Results from defined molecular models of genetic disorders could inform drug development and the basic biology of self-injury (SI) in psychiatric populations, while results from clinical research taking advantage of the larger populations with psychopathological disorders could inform behavioral and pharmacological treatment of NDD patients.
Why might SI in the NDD population inform the underlying causes of SI in a psychiatric, nonintellectually disabled population? SIB in NDDs is often associated with particular syndromes, many of which are linked to disorders caused by mutations in single genes or deletions of small parts of a chromosome, and can be studied in animal models or in vitro. The best example of this is Lesch-Nyhan disease (LND) whereby mutations in HPRT1 are associated with SIBs in almost all cases. As genetics are likely an important part of the complex etiology of SIBs in psychiatric populations, investigating the genes and molecular pathways that are altered in disorders like LND may reveal important information about molecular biology of SIBs in psychiatric populations.
SIBs are broadly defined as any nonnormative behavior performed with the intent of causing physical self-harm without the intent to die. Although diverse terminology is used to describe these behaviors, SIB is often used when describing NDD populations, whereas nonsuicidal self-injury (NSSI) is used when describing behaviors in psychiatric or community samples. For simplicity, SIB is used throughout this article when referring to SI in either population. A wide variety of definitions of SIBs are also used, and behaviors considered self-injurious in some studies are not in others. In part this accounts for the wide diversity of prevalence estimates observed, discussed in greater detail later. The Diagnostic and Statistical Manual of Mental Disorders (5th edition) (DSM-5) defines NSSI in part as “intentional self-inflicted damage to the surface of his or her body of a sort likely to induce bleeding, bruising, or pain (eg, cutting, burning, stabbing, hitting, excessive rubbing), with the expectation that the injury will lead to only minor or moderate physical harm.” This definition is used in this review.
There is not always a clear diagnostic boundary between NDD and psychopathology, in that persons with NDDs may have comorbid psychiatric disorders and subjects presenting primarily with psychopathology may have undiagnosed mild or moderate intellectual disability (ID). For example, a large population-based study in Scotland found that 40.9% of patients with ID met clinical diagnostic criteria for mental illness, and 15.7% met DSM-IV-TR (DSM 4th edition, text revision) diagnostic criteria. In addition, patients presenting with SI or other related behavioral problems are more likely to have psychiatric conditions than controls matched for level of disability. In other words, not all subjects with SIB have a diagnosed psychopathology.
Introduction
Self-injurious behaviors (SIBs) are observed in people with neurodevelopmental disorders (NDDs) and those with psychopathological conditions, although it is unclear whether the biological bases of SIBs in these 2 populations are similar. Evidence of similarities between the SIBs observed in patients with psychopathology and NDDs at the behavioral and molecular level, reviewed and analyzed here, suggests that the 2 fields of research could establish a synergistic relationship. Results from defined molecular models of genetic disorders could inform drug development and the basic biology of self-injury (SI) in psychiatric populations, while results from clinical research taking advantage of the larger populations with psychopathological disorders could inform behavioral and pharmacological treatment of NDD patients.
Why might SI in the NDD population inform the underlying causes of SI in a psychiatric, nonintellectually disabled population? SIB in NDDs is often associated with particular syndromes, many of which are linked to disorders caused by mutations in single genes or deletions of small parts of a chromosome, and can be studied in animal models or in vitro. The best example of this is Lesch-Nyhan disease (LND) whereby mutations in HPRT1 are associated with SIBs in almost all cases. As genetics are likely an important part of the complex etiology of SIBs in psychiatric populations, investigating the genes and molecular pathways that are altered in disorders like LND may reveal important information about molecular biology of SIBs in psychiatric populations.
SIBs are broadly defined as any nonnormative behavior performed with the intent of causing physical self-harm without the intent to die. Although diverse terminology is used to describe these behaviors, SIB is often used when describing NDD populations, whereas nonsuicidal self-injury (NSSI) is used when describing behaviors in psychiatric or community samples. For simplicity, SIB is used throughout this article when referring to SI in either population. A wide variety of definitions of SIBs are also used, and behaviors considered self-injurious in some studies are not in others. In part this accounts for the wide diversity of prevalence estimates observed, discussed in greater detail later. The Diagnostic and Statistical Manual of Mental Disorders (5th edition) (DSM-5) defines NSSI in part as “intentional self-inflicted damage to the surface of his or her body of a sort likely to induce bleeding, bruising, or pain (eg, cutting, burning, stabbing, hitting, excessive rubbing), with the expectation that the injury will lead to only minor or moderate physical harm.” This definition is used in this review.
There is not always a clear diagnostic boundary between NDD and psychopathology, in that persons with NDDs may have comorbid psychiatric disorders and subjects presenting primarily with psychopathology may have undiagnosed mild or moderate intellectual disability (ID). For example, a large population-based study in Scotland found that 40.9% of patients with ID met clinical diagnostic criteria for mental illness, and 15.7% met DSM-IV-TR (DSM 4th edition, text revision) diagnostic criteria. In addition, patients presenting with SI or other related behavioral problems are more likely to have psychiatric conditions than controls matched for level of disability. In other words, not all subjects with SIB have a diagnosed psychopathology.
Epidemiology of self-injurious behaviors
Prevalence estimates of SIBs vary widely within and between psychopathology and NDD populations. In persons with NDDs the prevalence of SIB is disorder dependent, and estimates from populations with mixed NDD diagnoses range from 1.7% to 41%. In some forms of NDD (see later discussion), nearly 100% of patients show SIB. However, a 2003 meta-analysis examining risk factors for SIB in NDD suggested that degree of ID, autism, and deficiencies in communication are the strongest risk markers of SI, and are thus more prevalent in NDD populations with these attributes. A second meta-analysis found that SIBs were increased in NDD subjects with urinary incontinence, pain related to cerebral palsy, chronic sleep problems, or visual impairment. SIBs can be documented as early as 6 months, and tend to be fully expressed by 5 years of age. It is interesting that unlike other measures of aggression, SIBs do not decline with age in NDD populations.
In the non-NDD population, SIB appears to be particularly prevalent in younger people. In a random sample of American adults, Klonsky reported a lifetime prevalence of SIB of 4.8% in those older than 30, and of 18.9% in those 30 years or younger, while the Child & Adolescent Self-harm in Europe (CASE) Study reported an 8% lifetime incidence of SI in a random sample of the population. Longitudinal studies in youth and college students have shown that current or prior depression and parental depression are strong predictors of SIB in both age groups. Current depressive symptoms, diagnosis with an internalizing disorder (such as depression or borderline personality disorder), and diagnosis with an eating disorder were significantly associated with SIBs in another large cohort of university undergraduates.
Description of self-injurious behaviors in specific disorders
Lesch-Nyhan Disease
LND is caused by a mutation in HPRT1 , which encodes hypoxanthine-guanine phosphoribosyltransferase (HPRT), an enzyme required for the salvage of the purines hypoxanthine and guanine, and is estimated to occur in 1 in 380,000 live births. The amount of functional HPRT protein varies depending on the type of mutation (eg, a single DNA base change or a complete deletion of the gene) seen in LND patients, and a strong correlation exists between residual HPRT activity and disease severity. NDD without SIB generally occurs in between 1.5% and 8% HPRT1 activity, and SI occurs in nearly 100% of patients with less than 1.5% activity. Of note, there is variation in disease expression including SIBs, even in members of the same family with the identical HPRT1 mutation, suggesting environmental effects and an interaction with genetic background.
Patients with LND demonstrate a wide variety of SIBs. Among the most common are biting of the lips and fingers, sticking out an arm, leg, or head when passing through a doorway, and head banging, but people with LND appear to use whatever form of SIB is available to them. LND is also associated with increased aggression toward others, expressed through hitting, spitting, scratching, and swearing. Parents of LND patients report that SIBs increase during times of stress and boredom, and independent observation of boys younger than 5 years with LND documented increased SIBs during times of reduced social contact. An immediate prophylactic measure for children with LND is thus to increase social contact and reduce boredom.
Smith-Magenis Syndrome
Smith-Magenis syndrome (SMS) is a developmental disorder associated with abnormal facial development, hearing loss, expressive speech delay, short stature, brachydactyly, sleep disturbances, and variable levels of ID, believed to occur in from 1:15,000 to 1:25,000 live births. SMS is usually caused by deletions in the 17p11.2 region. However, mutation of the gene RAI1, a transcription factor whose function remains largely unknown, is sufficient to cause most of the observed phenotypes, including SIBs. Similarly to LND, the prevalence of SIB approaches 100%. The forms of SIB most commonly demonstrated in SMS are biting, hitting, picking at fingers or toenails (onychotillomania), and the insertion of foreign objects into body orifices (polyembolokoilamania). SMS patients also demonstrate a characteristic “self-hugging,” tightly wrapping their arms around their torso and squeezing, which appears to be benign and exacerbated by happiness.
Cri du Chat Syndrome
Cri du chat syndrome (CdCS) is a developmental disorder named for a distinctive cat-like cry observed in many infants with the disorder. It is caused by a large deletion on chromosome 5p, although the genes responsible for the behavioral symptoms have not yet been confirmed. CdCS is characterized by intellectual disability, limited language development, sleep disturbances, hypersensitivity to auditory stimuli, and SIB, with prevelance estimates ranging from 1:15,000 to 1:50,000 live births.
SIB occurs in 76.8% –92% of patients with CdCS, and is predominantly expressed as hitting the head with another body part, hitting the head with an object, vomiting and rumination, scratching, and self-biting. CdCS patients also display aggressive behaviors toward others, including hair pulling, hitting, and throwing objects, as is seen in LND.
Borderline Personality Disorder
Borderline personality disorder (BPD) is a psychopathology characterized by pervasive instability of affect, interpersonal relationships, identity, and self-image, and is a significant risk factor for SIB. Indeed one of the diagnostic criteria for BPD is SI or attempted suicide, and prevalence estimates for these behaviors are 17% to 80% and 46% to 92%, respectively, in this population. People with BPD engage in self-cutting and burning at a higher rate than other populations engaging in SIB but also use biting, banging, and needle sticking, and are more likely to endorse suicide prevention and self-punishment as reasons for SIB.
Major Depressive Disorder
Major depressive disorder (MDD) is strongly associated with SIBs, and is the risk factor most consistently associated with SIB in both referred and community samples of adolescents and adults. Studies of depression have observed rates of SIBs (recorded as NSSI) of between 9% and 30%, while another study found that more than 70% of hospitalized self-injurers had depression. In the context of depression, SIB most commonly expresses as cutting, scratching, and burning, and has been associated with negative self-esteem and heightened emotional reactivity.
Conceptualization of self-injurious behaviors
SIB in non-NDD populations is predominantly conceptualized using the 4-factor model proposed by Nock and Prinstein. In this model, SIB is thought of along 2 dichotomous scales: automatic (ie, intrapersonal) versus social (ie, interpersonal), and positive (ie, increasing a desired stimulus) or negative (ie, decreasing an aversive stimulus); thus, stopping negative emotions is an example of an automatic negative function, whereas increasing attention from others would be a social positive function. Results from studies using interviews and surveys have largely supported this framework, although it has been debated whether these functional groups delineate distinct clinical populations. The most common motivation for SIBs in community samples and patients with psychopathologies appears to be autonomic negative, but most people endorse multiple reasons for SIBs. Some commonly cited reasons are “to cope with feelings of depression” (83%) or “distraction from unwanted memories” (62%), “expressing frustration” (73%), and “self-punishment” (50%). Following SIB participants tend to report relief, or a decrease in negative emotions, but an increase in self-aware emotions such as shame and guilt.
Several studies have suggested that non-NDD people with SIB have a heightened response to stressful situations, which can be reduced by imagining SI. Interestingly this stress reduction has also been observed in self-injuring macaques, in which self-biting was seen to significantly reduce an accelerated heartbeat in a subpopulation of monkeys. Not all studies have supported these findings. Kaess and colleagues did not see evidence of increased heart rate in a sample of 14 self-injuring patients compared with 14 normal controls when subjected to standard psychosocial stressors. However, self-injuring populations self-report greater psychological stress from these tests, even without an increase in physiological response.
The operant model of SIB, which is commonly used for subjects with NDDs, classifies the rationale for SIBs along similar axes to the 4-factor model proposed for subjects without an NDD; however, the autonomic positive and negative categories are collapsed into one. In one large study, 152 NDD patients were assessed for social positive, social negative, and autonomic reinforcement of SIBs. Behavioral therapy using one or all of these models reduced SIB in 95% of patients, and suggested that 38% of subjects used SIB for social negative reasons, 26% for social positive, and 25% for autonomic. A more recent review of research has supported the idea that social functions are the predominant mode of reinforcement for SIB.
SIB functions also correlate with the severity of disability, with people with a mild to moderate NDD exhibiting more outwardly directed aggression and challenging behaviors, and people with more severe NDD exhibiting more self-directed aggression and SIB. Subsequent research has suggested that people with severe NDD are more likely to use SIBs to obtain tangible items and escape social demands more often than people with mild to moderate ID.
Fig. 1 provides a summary of some differences between NDD SIB and psychiatric NSSI.
Neurotransmitters implicated in self-injurious behaviors
Dopamine
The involvement of the dopamine (DA) system has been most strongly implicated in SIB through its role in the pathogenesis of LND. Postmortem brains from patients with LND show up to a 60% decrease in DA concentrations compared with controls. This decrease seems to be the result of a deficiency in DA synthesis rather than decrease in the number of dopaminergic neurons. These reductions in DA synthesis are accompanied by large increases in cell-surface expression of D1 and D2 receptors in the putamen and caudate.
Direct evidence for the involvement of the dopaminergic system in psychopathology-associated SIB has been limited. However, depressed suicide attempters, some of whom had the intent to die and so do not fall into strict SIB criteria used here, have significantly lower levels of homovanillic acid (HVA), one of the primary metabolites of DA, in their cerebrospinal fluid and urine than normal controls or affective controls. These findings and others have led to the proposal of an addiction model of SIB, which suggests that SI leads to the release of DA and the stimulation of the reward system. SIB has several characteristics of dependent behavior, including a requirement of increased severity or frequency to achieve the same effect and perseverance despite negative consequences ; however, cravings for SIB seem to be more context dependent than the cravings observed in substance addiction.
Serotonin
The serotonin (5-HT) pathway has been suspected to play a role in SIB, largely because of its involvement in impulsivity and aggression, which have been associated with SIB, and its well-established connection to depression and suicide. However, the relevance of self-report impulsivity has been questioned, and findings on experimental measures of impulsivity have been inconsistent. Patients receiving treatment with selective serotonin reuptake inhibitors (SSRIs) show little change in SIBs.
Trials using SSRIs to limit SIBs in NDD have yielded mixed results. Fluoxetine has been demonstrated to reduce SIB in autism but citalopram has not, raising questions of off-target effects. Risperidone, an atypical antipsychotic that antagonizes DA and 5-HT receptors, has also been repeatedly shown to reduce challenging behaviors, including SIB in patients with NDDs, autism, or LND.
Nucleoside Signaling
Signaling by the neuromodulatory purine nucleoside adenosine (Ade) has been hypothesized to play in important role in the pathology of LND. Ade receptors are decreased in lymphoblasts from LND patients, and the elevated hypoxanthine concentrations present in these cells have been shown to interfere with the normal transport of adenosine. Although HPRT dysfunction disrupts many neurological pathways, the lack of HPRT function seems to be requisite for the SIB phenotype. For instance, although the dopaminergic system has been implicated in SIBs in LND, dopaminergic dysfunction may not suffice to cause SI, as is evidenced by the absence of SIBs in patients with tyrosine hydroxylase deficiencies or mutations in dopamine receptors.
Ade has not been studied in SIB in the context of psychopathology per se; however, adenosine signaling though the A1 and A2A receptors has been associated with depression and anxiety, and regular consumption of caffeine, an adenosine receptor antagonist, has been associated with increased anxiety but a reduction in depression and suicide attempts. Adenosine antagonists might provide an intriguing avenue for the treatment of SIBs in both psychiatric and NDD populations.
Endogenous Opioids
Reduction in pain sensitivity, which is frequently reported by people engaging in SIB, implies involvement of the endogenous opioid (EO) system. One model suggests that people who engage in SIBs have lower baseline levels of EOs, and that SIB causes the activation of the EO system. This process leads to relief from negative emotions, which is the most common reason reported for engaging in the SIB. Consistent with this hypothesis, patients with BPD were shown to have increased μ-opioid receptor availability during a baseline state and increased occupancy during induced sadness. In a second study, cerebrospinal fluid from group of subjects who repeatedly engaged in SIBs was compared to that of a diagnostically matched control group, and was found to contain significantly lower levels of the EOs β-endorphin and met-enkephalin. Together, these studies support the idea that people with psychopathology show low baseline levels of EO and may engage in SIB, in part to stimulate this system.
Reduced pain sensitivity is less consistent in patients with NDD, in whom SIBs may in fact be associated with increased nociception. A systematic review of 10 randomized controlled trials of opioid antagonists to treat SIB in patients with NDD (where ID ranged from mild to profound, with 40% having an autism diagnosis) revealed that 8 of the 10 reported positive results and opioid antagonist treatments reduced episodes of SI in 50% of the subjects. Opioid signaling may be a convergence point for SIBs for persons with an NDD or psychopathology; however, the dynamics of this signaling, and the baseline states, may differ between these populations.
Interactions Between Signaling Pathways
When considering complex behaviors such as SIB it is important to remember that the neurotransmitter systems do not act independently, but interact with each other in a myriad of ways, many of which we do not yet fully understand. For instance adenosine A2A receptors form heterodimers with DA D2 receptors, and they reciprocally oppose each other’s functions, including D2-mediated enkephalin release. Adenosine signaling through both the A1 and A2A receptors is analgesic and reduces symptoms of opiate withdrawal, while opiates may increase the degradation of adenosine. Meanwhile the EO system bidirectionally modulates the dopaminergic system ( Fig. 2 ). This level of interconnectivity indicates that one neurotransmitter system may alter its signaling to compensate for another. Although this is likely to be beneficial in many instances it is possible that overcompensation, perhaps caused by genetic predisposition, or deficits in multiple systems, could drive the SIB phenotype, which would suggest that combinations of pharmacological treatments may be more effective than any one alone.
Full access? Get Clinical Tree
