Use of the Cross-Translational Model to Study Self-Injurious Behavior in Human and Nonhuman Primates

Abstract

Nonsuicidal self-injurious behavior occurs in the general human population, particularly among teenagers and young adults. Some rhesus macaques also develop self-injurious behavior (SIB) as adolescents or young adults. In both of these cases, the development of harmful behaviors is idiopathic, only coming to the attention of physicians or veterinarians after the disorder is established. Thus, a combination of retrospective, statistical, and empirical procedures are used to understand this disorder. Here, we identify concordances between macaques and humans across five different levels of analysis—(1) form and prevalence, (2) etiology, (3) triggering events, (4) function/maintenance, and (5) therapeutic intervention—and show the value of the cross-translational model (macaques to humans and humans to macaques) in understanding this phenomenon. Substantial concordance is present with respect to the range of severity, the presence of early life stress exposure, and emotional dysregulation. In the macaque model, additional information is available on the hypothalamic–pituitary–adrenal axis stress response system, possible genetic involvement, and the immediate contextual situations that appear to trigger or exacerbate SIB episodes. In contrast, considerably more information is available from human studies on the effectiveness of various treatment regimens. Veterinarians have drawn on this information to explore these therapeutic interventions in monkeys. We expect that models of SIB will continue to have cross-translational impact as scientists and practitioners move from preclinical to clinical research and treatment.

Introduction

Self–injurious behavior (SIB) is a significant human health problem affecting many different populations and for which there is no widely effective treatment (Nock 2010). SIB has been linked to severe intellectual disabilities, autism, genetic defects, and certain psychiatric disorders in which early childhood trauma is often a key feature (Dubo et al. 1997; Favazza 1998; Oliver et al. 2009; Simeon and Favazza 2001). However, SIB also occurs in the general population, and disturbing trends suggest that it is on the rise among teenagers and young adults (Fortune and Hawton 2005; Hasking et al. 2008; Rodham and Hawton 2009). An overview of 50 studies from 2005 to 2010 reveals an average lifetime prevalence rate of 18% (Muehlenkamp et al. 2012). In college students, estimated prevalence of at least one past wounding event ranges 7–34% (Gollust et al. 2008; Gratz et al. 2002; Hasking et al. 2008; Tresno et al. 2012; Whitlock et al. 2006). SIB in adolescents and college students has also been linked to early life trauma (Whitlock et al. 2008), substance abuse (particularly nicotine; Riala et al. 2009), and concurrent depressive and anxiety disorders (Gollust et al. 2008).

The diversity of conditions that elicit SIB suggests that there are subtypes of this disorder with different etiologies and developmental trajectories. In humans, the term SIB is generally restricted to individuals with profound cognitive impairments, autism spectrum disorder, fragile X syndrome, and genetic defects (e.g., Lesch Nyhan and Prader-Willi syndromes). In these populations, SIB is highly ritualized and repetitive, largely independent of emotional context, and tends to develop early in life, often within the first year (Kurtz et al. 2012). In some of these conditions, there is evidence of brain abnormalities (autism spectrum disorder and Fragile X syndrome; Wolff et al. 2013).

Self-injurious behavior occurring in normative populations has been termed nonsuicidal self-injury (NSSI); however, this definition may be controversial, inasmuch as suicidal ideation is present in some cases (Rao et al. 2008). NSSI is a heterogeneous disorder with at least two subtypes, a compulsive form of self–inflicted wounding (e.g., hair plucking, scratching, nail biting) and a reactive form, typically manifested as cutting or burning in response to emotional stress (Simeon and Favazza 2001). Individuals with the reactive form appear to develop a maladaptive emotion regulation strategy in which tension associated with aversive emotional states is temporarily reduced by cutting or burning and ultimately maintained by a process of negative reinforcement (Chapman et al. 2006; Gratz and Roemer 2008; Kamphuis et al. 2007; Klonsky 2009; Mikolajczak et al. 2009; Nock and Mendes 2008). Both forms of NSSI tend to develop in early or late adolescence, although onset is not restricted to these developmental stages. In the past, NSSI was viewed primarily as a symptom of different mood disorders, and treatment was predicated on the disorder and not the symptom. Because NSSI can also occur without any apparent underlying psychopathology, it is now under consideration in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders as a distinct diagnostic entity (In-Albon et al. 2013).

SIB also occurs in many animal species, ranging from feather plucking in birds (Garner et al. 2006) and lick granuloma in dogs (Rapoport et al. 1992) to self-inflicted wounding in nonhuman primates (Novak 2003). The above are examples of idiopathic models of SIB in which the etiology is largely unknown and can only be assessed using a combination of statistical, retrospective, and empirical procedures. In contrast, a number of experimental models have been developed in rodents. SIB can be induced by administration of the dopamine agonist pemoline (Muehlmann et al. 2008); by BayK 8644, an activator of voltage-regulated (L-type) calcium channels (Jinnah et al. 1999); and by neonatal chemical lesioning of the dopaminergic system with 6-hydroxydopamine (Breese et al. 2005).

In this article, we focus on the idiopathic development of SIB in rhesus macaques and assess the utility of the rhesus macaque model for the development of NSSI in the general human population. The occurrence of NSSI in humans is also an example of an idiopathic model wherein researchers and physicians are confronted with the challenges of uncovering the causes, identifying the triggering events, and developing effective treatments. Given these and other similarities described herein, the prospect for gaining significant information and applying it in either direction (from the monkey model to humans and from the human model to monkeys) is substantial. We use the term “cross-translational model” to refer to this process.

Rhesus Macaque Natural Life Histories

Rhesus macaques (Macaca mulatta) are commonly used in biomedical research because of their close genetic relationship to humans and because of similarities in physiology, anatomy, and behavior. Rhesus macaques belong to the Old World monkey genus Macaca and are native to South, Central, and Southeast Asia. They have the widest geographic range of any nonhuman primate species and thrive in many different environments, ranging from remote forests to agricultural regions, towns, and large cities. Rhesus macaques have been labeled weed macaques in reference to their unusual ability to coexist with humans in many different habitats (Richard et al. 1989).

Rhesus macaque social structure is highly complex, multilayered, and hierarchical. Monkeys typically live in troops that can vary considerably in size (i.e., 20 to 200 individuals), with the core of the troop consisting of multigenerational matrilines of female macaques and their female offspring (Berman 1983; Lindburg 1971; Melnick et al. 1984; Teas et al. 1980). This emphasis on females is related to a sex bias in philopatry, in which female macaques stay in their natal troop generally throughout life, whereas male macaques typically emigrate. The emigration period is associated with heightened mortality, but eventually male macaques that survive work their way into a different troop (Berard 1989). Macaques have complex social repertoires involving both highly affiliative interactions (e.g., prolonged social grooming bouts) and highly volatile aggressive altercations that can result in serious injury and even death (e.g., Southwick et al. 1965; Teas et al. 1982). Each troop is essentially a closed society, and members show xenophobia by reacting aggressively to unfamiliar monkeys that approach them (Southwick et al. 1974). Because of their highly complex social nature and their ability to adapt to diverse environments (traits characteristic of humans), rhesus macaques may be an ideal model for understanding causes and circumstances surrounding the development of abnormal behavior such as SIB.

Levels of Analysis in the Study of SIB in Rhesus Macaques

Both humans and monkeys can develop idiopathic self-injurious behavior that typically comes to the attention of physicians or veterinarians only after the disorder is well established. Thus, considerable detective work involving observation and evaluation of various physiological systems is necessary to understand these cases and develop effective treatments. We use a systems approach to examine SIB in monkeys from five different perspectives that reflect increasingly complex yet interdependent levels of analysis. These are (1) description, (2) etiology, (3) contextual factors (i.e., triggering events), (4) function/maintenance, and (5) therapeutic intervention. At each level, the value of the cross-translational model (monkeys to humans and humans to monkeys) is assessed.

Description (Form, Sex Differences and Prevalence Rates)

Laboratory-housed rhesus macaques can spontaneously develop a syndrome of self-inflicted wounding that may require veterinary treatment (Novak 2003; Novak et al. 2012). Typically, SIB is manifested as self-directed biting of arms and legs, which can produce tissue damage ranging from skin abrasions and punctures to deep lacerations transecting the skin and the underlying tissues. Serious wounding events are relatively rare, with months separating episodes; however, mild biting may occur on a daily basis in some monkeys. Humans occasionally bite themselves but more generally use implements (e.g., scissors, knives, and cigarette lighters) to produce tissue damage. As is the case with rhesus macaques, the damage ranges from skin abrasions and superficial cuts to more serious burns and deep lacerations (Rao et al. 2008).

There is conflicting evidence as to whether this disorder is sex biased in rhesus macaques. At two facilities, male macaques were more likely to exhibit SIB (Gottlieb et al. 2013; Lutz et al. 2003), whereas at other facilities no sex bias was present (Bellanca and Crockett 2002). Because of the sexual dimorphism of canine teeth in rhesus macaques, male macaques have a much greater potential to inflict serious wounds on both themselves and others than do female macaques. In humans, initial findings suggested a much higher prevalence in females than males, but more recent evidence shows that this difference is markedly affected by age, being most highly skewed in 10- to14-year-olds and declining thereafter until it is nearly equivalent in 25- to 49-year-olds (Hawton and Harriss 2008). In another study, distinctive for its use of random sampling methods, NSSI was unrelated to sex (Klonsky 2011). Similar to the sex differences observed in rhesus macaques, men tend to engage in more serious forms of self-injury and are more likely to require medical treatment than women (Andover et al. 2012; Fujioka et al. 2012; but see Whitlock et al. 2008).

SIB typically has been studied in rhesus macaques housed with indirect (visual and auditory) but not direct physical contact with conspecifics, a common biomedical research housing environment of the recent past. The focus on this form of housing ensured that injuries were self-inflicted and not socially inflicted. Studies over the last decade have revealed a 10–12% prevalence rate in singly housed monkeys that is highly concordant across facilities (Bayne et al. 1992; Bellanca and Crockett 2002; Lutz et al. 2003; Rommeck et al. 2009). However, SIB has also been reported in socially housed monkeys in zoo environments (Novak et al. 2006) and in Japanese macaques living in natural environments (Grewal 1981). In the latter case, wrist biting was observed without serious penetrating wounds. The number of affected individuals in these social environments is relatively low, but the aforementioned prevalence rate may slightly underestimate the actual rate of SIB at research facilities, where some monkeys are housed singly and others are housed in social groups. Compared with college students, in which NSSI prevalence rates vary 7–34%, the prevalence rate in rhesus macaques falls within the range but appears to be slightly lower than that reported for humans.

Etiology

Many diverse and potentially interacting causes have been suggested for the occurrence of SIB in rhesus macaques. In the past, the greatest emphasis was placed on environmental factors, inasmuch as SIB was typically observed in monkeys housed singly. However, research has shown that, although housing environment matters, it is not the only factor that contributes to SIB (Novak 2003). Only a small percentage of monkeys housed singly subsequently develop this disorder (Lutz et al. 2003), suggesting that the majority of monkeys may be more resilient to socially restrictive environments. Instead, recent work has focused on the multiple risk factors that predispose some monkeys to develop behavioral pathology under these conditions.

Early Life Stress

Exposure to early life stress appears to be one of the salient risk factors for the development of SIB in rhesus macaques. SIB has been associated with restrictive early rearing experiences and with early removal from a social living environment. The two most common restrictive rearing environments are peer rearing, in which infants are reared continuously with each other, and surrogate peer rearing, in which infants are reared with an inanimate surrogate and given 2 hours of daily peer exposure (Dettmer et al. 2012). At one facility in which both rearing conditions are studied and compared directly with maternal rearing, all monkeys remain in their rearing groups until 8 months of age. They are then housed in large mixed-rearing social groups consisting of 25 to 50 individuals. Although peer rearing and surrogate peer rearing monkeys exhibit species-typical behavior and appropriate parental behavior (Novak et al. 1992; Roma et al. 2006; see Sackett et al. 2002 for similar findings in pigtailed macaques), they show altered emotional regulation, which includes anxious behavior and higher levels of fear and aggression, as compared with maternally reared infants (Dettmer et al. 2012; Harlow and Harlow 1962). However, only the surrogate peer rearing condition is a significant predictor of SIB at this facility, occurring in 28% (n = 23 of 80) monkeys as compared with 0% (n = 0 of 84) in peer-reared monkeys and 1% (n = 2 of 181) in maternally reared monkeys (Lutz et al. 2007). Additionally, SIB vulnerability is associated with how social these monkeys were in their 2-hour play periods. Surrogate-peer-reared monkeys that eventually developed SIB showed significantly lower levels of social contact during their 2-hour play periods compared with noninjuring monkeys (Lutz et al. 2007). The authors suggest that some deficit or limitation in social functioning combined with surrogate peer rearing appears to tip the scale toward SIB. However, more recent findings also reveal increased vulnerability to SIB in peer-reared monkeys when compared with monkeys reared with a mother and peers in outdoor cages (Gottlieb et al. 2013).

SIB vulnerability is not limited to rhesus macaque infants reared in restrictive environments. Social separation in maternally reared juvenile monkeys can increase SIB vulnerability, an effect which is age dependent. Maternally reared monkeys that were individually housed at an early age (approximately 14 months on average) were much more likely to develop SIB than monkeys that remained with companions until 25 months of age on average (Lutz et al. 2003). Other strong predictors of SIB include the length of individual cage housing and number of veterinary procedures (Lutz et al. 2003). Our findings also agree with those reported for pigtailed and rhesus macaques at other facilities (Bellanca and Crockett 2002; Lutz et al. 2007; Rommeck et al. 2009), indicating that these risk factors cut across species and facility.

Similar vulnerabilities to those described in monkeys are often associated with the development of NSSI in humans. Self-injuring behavior is frequently associated with early life stress, often in the form of child abuse or neglect (Tang et al. 2013; Tresno et al. 2012). In a review of 59 research articles, early childhood trauma was identified retrospectively as a significant correlate of NSSI in 58 of these studies (Fliege et al. 2009). A strong relationship also exists between NSSI and poor emotional regulation, with individuals reporting more frequent and escalating negative affect (Gratz 2006; Gratz and Chapman 2007; Paivio and McCulloch 2004; Victor and Klonsky 2013). In one study, individuals with NSSI who were exposed to a distressing situation displayed heightened arousal, difficulty in handling distress, and weak social problem solving skills compared with noninjuring control subjects (Nock and Mendes 2008).

Hypothalamic–Pituitary–Adrenocortical Axis

Adverse early experience and lifetime trauma have been linked to long–lasting changes in peripheral (i.e., hypothalamic–pituitary–adrenocortical [HPA] axis) and central (i.e., brain corticotropin-releasing factor) stress response systems in both monkeys (Clarke et al. 1998; Coplan et al. 1996) and humans (Davidson et al. 2004). Because adverse experiences are important risk factors in the development of SIB in nonhuman primates (Lutz et al. 2003), we examined possible abnormalities in HPA axis functioning in maternally reared macaques with this disorder. Monkeys with a history of self–inflicted wounding showed an attenuated plasma cortisol response to the mild stress of blood sampling compared with nonwounding control monkeys (Tiefenbacher et al. 2000). Moreover, stress–induced cortisol levels were inversely related to rates of self–directed biting. Subsequent studies investigating the mechanisms underlying this blunted stress response found evidence for reduced adrenocortical sensitivity in the SIB group (Tiefenbacher et al. 2004). These findings suggest that SIB in some rhesus macaques is associated with complex and persistent changes in HPA activity that are related to both the outcome (i.e., wounding) and the expression (i.e., self–directed biting) of the behavioral pathology. These findings do not address the issue of cause and effect—namely, whether altered HPA axis activity increases vulnerability to SIB or is a consequence of SIB.

Despite the strong relationship between early life stress and NSSI in humans (Briere and Gil 1998; Zoroglu et al. 2003), considerably less information is available on HPA axis activity in humans with this disorder. However, a recent report (Kaess et al. 2012) suggests that the HPA axis may be hyporesponsive. In response to the Trier Social Stress Test, adolescent girls with NSSI showed a blunted response at 15 and 40 minutes after the stress experience as compared with noninjuring control subjects.

Genetic Risk Factors

Differential responsiveness to stressful events is important in the development of SIB in macaques, and this effect may be mediated by genetic factors. Recent research has examined the role of several genetic polymorphisms associated with anxiety and impulse disorders in humans in conferring enhanced vulnerability to SIB in macaques. Genotyping of monkeys with and without SIB was performed to characterize polymorphisms in the serotonin transporter gene (5-HTT), the mu-opioid receptor gene (OPRM1) (Kelly et al. 2009), and the tryptophan hydroxylase-2 gene (TPH2) (Chen et al. 2010). The 5-HTT polymorphism consists of a long and short allele distinguished by differences in the gene's promoter region. In humans, the loss-of-function short allele is associated with increased risk for impulsivity and suicidal behavior (Arias et al. 2001). However, the presence of the 5-HTT short allele was unrelated to SIB in rhesus macaques (Tiefenbacher, Newman, et al. 2005).

In contrast, the OPRM1 polymorphism appears to show more promise. Functionally similar nonsynonymous mu-opioid receptor (OPRM1) single nucleotide polymorphisms occur in humans (A118G) and rhesus macaques (C77G). The gain-of-function G allele confers increased affinity for β-endorphin in vitro and has been associated with both increased distress responses to maternal separation in rhesus macaque infants and increased restraint of infants by rhesus macaque mothers (Barr et al. 2008; Higham et al. 2011). The authors hypothesize that G allele carriers experience increased opioid reward in response to affiliation. In monkeys, the G allele was significantly associated with severe pathological behavior that includes SIB. Additionally, macaques with SIB showed reduced baseline plasma β-endorphin levels compared with control macaques (Tiefenbacher, Novak, et al. 2005), raising the possibility that self-biting is maintained, at least in part, by enhanced sensitivity to endogenously released opioid peptides such as β-endorphin (but see Crockett et al. 2007 for elevated levels of plasma β-endorphin levels in macaques with severe abnormal behavior, including SIB). Severe pathological behavior was also associated with lower levels of hair and salivary cortisol as compared with behaviorally normal monkeys, a finding consistent with previous research that demonstrated lower plasma cortisol levels in a different population of monkeys with SIB compared with nonwounding control monkeys (Kelly et al. 2009).

Lastly, SIB in monkeys has also been associated with polymorphisms in the TPH2 gene (Chen et al. 2010). TPH2 codes for the enzyme that synthesizes serotonin in central serotonergic neurons, and loss-of-function mutations of TPH2 have been associated with major depression and other affective disorders in humans (Harvey et al. 2004; Zhang et al. 2005). In rhesus macaques, the distribution of rhTPH2 5'-FR haplotypes differed significantly between monkeys with and without SIB and were differentially associated with HPA axis activity.

Although some preliminary research suggests that both the OPRM-1 and TPH2 genes are involved in the expression of SIB in macaques, there has been relatively little study of the genetic factors that contribute to NSSI in humans (but see Bresin et al. 2013). With recent developments in high-throughput DNA microarrays, deep sequencing (RNA-seq), and epigenetics, new techniques are now coming online that will more fully address how genetic and epigenetic processes interact with environmental factors to regulate behavioral output and physiologic responsiveness in humans and nonhuman primates with self-injurious behavior.

Contextual Factors/Triggering Events

Once acquired, SIB is an episodic behavior, occurring presumably in response to some event or situation. In macaques, mild self-directed biting (i.e., yielding abrasions or minor cuts) tends to occur across the day and appears to be associated, in part, with the activity of animal technicians. Mild biting was highest in the morning during routine cleaning and feeding, lowest at noon when the animal technicians were having lunch, and intermediate during the afternoon feeding-only period (Novak 2003). Under some conditions, however, biting can escalate to severe wounding, and it is essential to identify the features of the environment that cause the escalation. Understanding these immediate contextual conditions may lead to an effective redesign of the environment and modification of husbandry practices so as to prevent or reduce future occurrences of wounding.

Not surprisingly, differential response to stress events appears to be a significant factor in triggering wounding episodes in macaques with a preexisting syndrome of SIB. A retrospective analysis of colony records combined with caretaker reports indicated that three different types of events tended to precede a wounding episode using a 2-day window. These were (1) relocation of the monkey either within the same room or to a different room, (2) the introduction of a new animal into the room, or (3) exposure to a minor medical procedure (Lutz et al. 2003; Novak 2003). It should be noted that in 25% of the cases, no obvious triggering event could be identified. SIB monkeys showed increased sensitivity inasmuch as most monkeys (e.g., those without SIB) showed only brief reactions to these events.

Although stress exposure appears to trigger episodes of SIB, this linkage is generally based on retrospective evidence. A recent study has shown that significant stress exposure can exacerbate SIB both in terms of the number and intensity of these episodes in adult monkeys. An administratively mandated move of macaques with and without SIB to a new building provided a unique opportunity to assess the effects of relocation on the expression of SIB (Davenport et al. 2008). The new environment was expected to be stressful because of substantial changes in cage sizes, noise levels, and animal densities as compared with the housing situation in which the macaques had lived for 5 or more years. Indeed, the relocation resulted in significant rises in both plasma and salivary cortisol concentrations that persisted for at least 2 months under the new housing conditions. Furthermore, hair samples obtained 3 months before the move, 4 months after relocation, and 12 months later revealed a significant elevation of cortisol concentrations during the relocation phase as compared with the other two time points. In addition to these physiologic indicators, macaques with SIB showed pronounced increases in self-directed biting, which persisted for more than a year. The change in biting rate was also associated with significant sleep disturbances in the SIB monkeys but not in control monkeys. These data are the first to support a causal link between exposure to major life stress events and self-injurious behavior.

In humans, most of the research on NSSI in humans has focused on identifying risk factors such as altered emotional regulation and self-dissatisfaction. However, relatively little research has examined the events that precede episodes of cutting and burning to identify the more immediate triggers of such activity. In one recent study, individuals diagnosed with NSSI experienced heightened negative affect for a prolonged period, which peaked at the time of self-inflicted wounding and then gradually diminished over the next several hours (Armey et al. 2011). Others have noted that individuals may ruminate on the act of self-injury before engaging in it (Nock et al. 2009).

Function/Maintenance

Self-injurious behavior is a seeming paradox. The act of cutting or burning is considered aversive to many, and yet some individuals regularly engage in these activities and find it difficult to stop. A number of models have been proposed to explain the maintenance of self-injurious behavior. The first two models have been derived from studies of individuals with severe intellectual disabilities, genetic defects, and psychiatric disorders, but they might also apply to humans with NSSI. These include the pain model, in which individuals with SIB show altered pain perception, and the social interaction model, in which SIB is used either to elicit attention or avoid social demands. The remaining two models have been studied in NSSI—namely, the emotional cascade model, in which NSSI creates a distraction, essentially short-circuiting the increased negative emotions created by excessive rumination (Selby et al. 2013), and the anxiety/tension-reduction model, in which NSSI reduces escalating feelings of anxiety. In this latter model, the act of cutting and burning appears to reduce negative affect, thereby creating a cycle of behavior maintained by negative reinforcement (Chapman et al. 2006; Haines et al. 1995; Kemperman et al. 1997; Klonsky 2009). In some cases, the act itself is not sufficient; seeing blood is important for tension relief (Glenn and Klonsky 2010). The proposed models are not mutually exclusive or exhaustive. As noted herein, the pain and the social interaction model have been studied primarily in individuals with severe intellectual disabilities, and the emotional cascade model is difficult to test in animals.

The focus in this review is on the potential application of the anxiety/tension-reduction model to monkeys with SIB. In macaques, this model predicts that self-biting occurs in response to anxiety and that measures of arousal should increase before biting episodes and decrease shortly thereafter. Additionally, biting should be differentially affected by anxiolytic and anxiogenic compounds.

Behavioral and physiological findings in macaques support a role for anxiety in the expression of SIB. Monkeys with a history of SIB show increased behavioral reactivity to normal everyday events, and during self–biting episodes, they exhibited heart rate changes that are consistent with a role for SIB in tension reduction (Novak et al. 2006). Monkeys were monitored for heart rate using remote telemetry and simultaneously videotaped. Videotapes were used to identify spontaneously occurring biting episodes (typically mild), and heart rate activity was examined immediately before, during, and after such episodes and compared with baseline heart rate activity. In the minutes before biting, heart rate was significantly elevated. Heart rate increased further during the act of biting and then declined to baseline within a few minutes.

If SIB is an anxiety disorder in macaques, then anxiolytic compounds should decrease biting behavior whereas anxiogenic compounds should increase biting behavior. Administration of the benzodiazepine (BDZ) diazepam, an anxiolytic compound, resulted in significant reductions in self–directed biting and the incidence of wounding in monkeys. Surprisingly, however, the treatment worked in only half of monkeys; diazepam actually increased the wounding incidence in the remainder of the animals (Tiefenbacher, Fahey, et al. 2005). Colony records indicated that the positive responders spent significantly more time housed individually and had experienced a greater number of minor veterinary procedures than the negative responders. Both of these variables have previously been shown to be risk factors for the development of SIB (Lutz et al. 2003). A similar bimodal response was observed when monkeys were administered the anxiogenic BDZ partial inverse agonist FG7142 (Major et al. 2009). Half of the SIB monkeys exhibited a dose-dependent increase in mild self-biting, whereas there was no effect on self-biting in the remaining monkeys. These findings suggest the existence of at least two subpopulations of macaques with SIB, one of which has a form of reactive SIB in which the disorder arises as a consequence of certain lifetime experiences and is responsive to compounds that alter anxiety. The other population is currently uncharacterized.

Therapeutic Intervention

Treatments in Humans

Despite all that is known about NSSI in humans and SIB in macaques, effective treatment of this disorder in both human and nonhuman primates remains elusive. In humans, there is no universally accepted treatment for NSSI, and indeed, individuals in the general population are unlikely to seek treatment (Deliberto and Nock 2008; Gollust et al. 2008). As a result, many treatment studies have been conducted using clinical populations with comorbid conditions (e.g., severe mental retardation, autism, depression, genetic disorders), making interpretation of efficacy difficult. Treatments generally fall into two categories: psychotherapy and/or pharmacotherapy, the latter of which is most relevant to nonhuman primates. A number of pharmacological agents have been used to treat NSSI, most of which target either the opioid system (e.g., the opioid antagonist naltrexone) or the serotonergic system (e.g., selective serotonin reuptake inhibitors [SSRIs]). The use of SSRIs in treating NSSI is complicated by the fact of increased risk of suicide, particularly among teenagers (Hetrick et al. 2007).

A number of studies support the view that naltrexone produces a clinically significant reduction in NSSI in humans (see reviews by Sandman 2009; Symons et al. 2004). This effect is thought to be mediated by altering the reward value of potentially painful experiences, and a recent study provides support for this general hypothesis. Benedetti and colleagues (2013) exposed two groups of subjects to ischemic arm pain. Both groups were informed about the pain, but one of the groups received additional information explaining the pain in a positive context (i.e., it would improve muscle function). Pain tolerance was significantly higher in the latter group, and this tolerance was partially blocked by naltrexone (Benedetti et al. 2013). If NSSI is “reinforcing” or associated with greater tolerance of pain, then naltrexone should block this effect and reduce NSSI in individuals with the disorder. In a review of 27 studies that examined the efficacy of naltrexone in ameliorating SIB in developmentally disabled humans, 80% of individuals with SIB showed significant improvement during naltrexone treatment. This effect was most pronounced in male subjects (Symons et al. 2004). Additionally, buprenorphine, an opioid partial agonist, led to marked reductions in severe NSSI in humans, an effect which continued into a post-treatment period. Positive results were observed in 3 female subjects and 1 male subject (Norelli et al. 2013).

Treatment in Monkeys

Given the partial success of pharmacotherapy in humans, it is not surprising that a similar approach has been adopted to ameliorate SIB in nonhuman primates, and at this level of analysis, the translation has occurred mostly from human research/clinical findings to macaques. Ideally the objective is to find a treatment that can effectively result in a cure such that subsequent removal of treatment drug does not produce relapse. Most of the pharmacotherapeutic treatments focus on serotonergic drugs such as SSRIs, opioid antagonists, or BDZs. Moderate efficacy has been reported for serotonergic compounds in reducing SIB in rhesus macaques (Fontenot et al. 2005). Male rhesus macaques treated with either the SSRI fluoxetine or the serotonin 1A receptor agonist buspirone exhibited significantly lower rates of self-biting during treatment weeks 1 to 8 compared with a baseline (no treatment) pretest; unfortunately, this effect did not continue through treatment weeks 9 to 12. In a second study, fluoxetine was found to be superior to the combined serotonin-norepinephrine reuptake inhibitor venlafaxine, but the monkeys were not evaluated after the treatment was over (Fontenot et al. 2009). Serotonin levels can also be manipulated by adding the precursor L-tryptophan to the diet. Tryptophan was highly effective in reducing SIB in monkeys in a short 21-day dosing period; however, the monkeys relapsed after the treatment was removed (Weld et al. 1998). Tryptophan also was effective in promoting wound healing in bush babies. The size of wounds decreased during the dosing period, but the bush babies were not evaluated after treatment (Watson et al. 2009). As mentioned previously, treatment with the BDZ diazepam was effective in reducing both wounding and biting but only in a subset of monkeys with a history of heightened stress exposure, and as with the previous studies, there was no assessment of possible relapse during a post-treatment period (Tiefenbacher, Fahey, et al. 2005).

More recent studies suggest a promising role for opioid antagonists in treating SIB in macaques. A relationship between SIB and the opioid system is suggested by findings of lowered resting levels of beta-endorphin-like immunoreactivity in blood plasma and met-enkephalin-like immunoreactivity in cerebral spinal fluid in monkeys with SIB compared with noninjuring control monkeys (Tiefenbacher, Newman, et al. 2005). Indeed, treatment with extended-release naltrexone markedly reduced the time monkeys engaged in self-directed biting by 54%, an effect that was maintained for at least 4 weeks after treatment (Kempf et al. 2012).

Environmental Manipulations in Monkeys

A substantial effort has been directed to identifying possible environmental changes as a form of treatment for SIB in macaques. At best, the results are mixed. SIB appears to be resistant to simple environmental changes, ranging from increased enrichment in the form of puzzle feeders (Novak et al. 1998) to 6-fold increases in cage size (Kaufman et al. 2004). Similarly, there has been little impact of positive reinforcement training in the form of increased human contact on SIB (Baker et al. 2009). However, housing monkeys in outdoor pen environments has been shown to reduce the rates of biting behavior in monkeys with SIB regardless of whether they are housed individually or in pairs (Fontenot et al. 2006). Subsequent studies have shown that the process of pairing monkeys with a history of SIB does not increase self-biting (Baker et al. 2013). Indeed, pairing SIB male monkeys with female monkeys actually led to decreases in SIB, but this strategy of opposite sex pairing may require pregnancy prevention (e.g., vasectomy; Weed et al. 2003). Cage configurations apparently also play a role in SIB. Monkeys housed on the lower racks of two-tier caging systems were 1.4 times more likely to self-bite compared with those housed on top racks. Additionally, the risk of self-biting decreased by 4% for every cage positioned between a subject and the colony room entrance (Gottlieb et al. 2013). Whether repositioning monkeys in a room environment can serve as an effective treatment for monkeys with SIB is unknown at this time. Such a strategy has to be considered in the context of a previous finding that relocation may trigger wounding episodes and exacerbate SIB symptoms (Davenport et al. 2008).

SIB in Nonhuman and Human Primates and Brain Pathology

Thus far, we have considered the idiopathic development of SIB in rhesus macaques as a potential cross-translational model for the development of NSSI in humans. The restricted focus on NSSI rather than the SIB associated with genetic disorders, autism, and severe developmental disabilities is based on the fact that these human disorders are often associated with extensive brain pathology, including white matter abnormalities in children (Widjaja et al. 2008) and reduced cortical folding in individuals with severe cognitive deficits (Zhang et al. 2010).

Currently, there is no evidence to suggest that macaques with idiopathic SIB have brain abnormalities. Behaviorally, they perform as well or better than control monkeys on complex object permanence tasks (De Blois and Novak 1998), but to date, no one has examined their brain morphology. However, a relationship between SIB and brain pathology has been noted in one clinical case of a pigtailed macaque and in one form of extreme early social deprivation—namely, isolation rearing that was studied for a brief period of time in the 1960s. In the clinical case, unremitting SIB was associated with significant brain damage that included loss of cortical structure in both gray and white matter, enlarged cerebral ventricles, neuronal loss, and astrogliosis (Bielefeldt-Ohmann et al. 2004). Isolation rearing of rhesus macaques during the first several months of life has been shown to produce profound behavioral and cognitive deficits as well as changes in the striatum, in particular the neurochemical architecture of this region (Martin et al. 1991). Because of the lack of concordance in many features (e.g., etiology) and the severity of the impact on social development, isolation rearing is not considered a viable or ethical model for SIB in developmentally delayed human populations.

Evaluation of the Cross-Translational Model

Is the development of idiopathic SIB in rhesus macaques a model for the occurrence of specific types of NSSI in humans? We have shown that substantial concordance exists between these conditions with respect to the range of severity, the presence of early life trauma and emotional dysregulation, and reasonably similar responses to various pharmacotherapeutic manipulations. In the monkey model, we now have considerable evidence regarding a role for the HPA axis and the stress response system in this disorder. Existing evidence in humans with NSSI similarly suggests a dysregulation in HPA function. However, there are also important differences in our current understanding of SIB in monkeys compared with NSSI in the human population. For example, there is growing evidence for a contribution of genetic differences to the development of SIB and a deeper understanding of the more immediate triggers of self-directed biting than is currently available for humans with NSSI. On the other hand, the human literature provides more detailed information on the contributions of early life trauma, the nature of emotional dysregulation, and a more thorough examination of pharmacotherapy. Indeed, primate veterinarians and researchers have made use of the human clinical data to develop treatments for SIB in rhesus macaques and possibly other nonhuman primate species. Thus, going forward we anticipate that the rhesus macaque models of SIB will continue to have important cross-translational value in bridging the gap between preclinical and clinical research and therapy.