1
Biological Markers of Psychosis: Cortisol Levels, Daily Life
Stress and Psychotic Symptoms in Patients with 22q11.2
Deletion Syndrome
Peter Saalbrink, University of Amsterdam – June 15, 2018
Research Project 1 Peter Saalbrink, 6102794
Supervisor: Esther van Duin, MSc Co-assessor: dr. Jan Booij
MSc in Behavioral Neuroscience, Brain and Cognitive Sciences, University of Amsterdam
Abstract: 22q11.2 deletion syndrome (22q11DS) is associated with several psychiatric
disorders, amongst which anxiety and psychotic disorders. Patients with 22q11DS experience increased daily life stress, although it remains unknown whether a causal relationship between these characteristics exists. Here, we present the first study of diurnal cortisol levels in adults with 22q11DS. We aimed to uncover the association between perceived daily life stress, HPA axis reactivity as measured by salivary cortisol levels, and psychiatric symptoms in 22q11DS patients. We studied 27 adults with 22q11DS and 24 healthy controls (HC) for 6 days using the experience sampling method (ESM) combined with cortisol sampling. Using a independent t-test, we found that adults with 22q11DS have significantly lower mean cortisol levels than HC. In addition, using a multilevel regression model, we found indication that perceived stress and psychiatric symptoms were both negative predictors for cortisol levels in 22q11DS patients. Our results point to permanent alterations to the HPA axis by chronic overactivation in adults with 22q11DS, resulting in an over-sensitization of the stress response. Finally, we provide a possible mechanism for the long-term effect of stress in 22q11DS patients.
Keywords: 22q11.2 deletion syndrome; 22q11DS; cortisol; daily life stress; DiGeorge
syndrome; experience sampling method; HPA axis; psychosis; schizophrenia; stress; stress reactivity; velocardiofacial syndrome (VCFS).
2 Contents Introduction ... 4 A Deletion at Chromosome 22q11.2 ... 4 Deficits in 22q11DS ... 5 Psychiatric Symptoms ... 5
Daily Life Stress ... 5
Psychosis and Schizophrenia ... 5
Neural Substrates of Deficits in 22q11DS ... 6
Cortisol and the HPA Axis ... 6
Cortisol Follows a Diurnal Rhythm ... 7
Alterations to the HPA Axis in 22q11DS ... 7
Long-Term Exposure to Cortisol Leads to Psychiatric Problems ... 8
Dopamine, COMT, and the PFC ... 9
Interaction Between These Systems ... 10
Aim, Hypotheses and Expected Results ... 11
Methods... 13
Subjects ... 13
Patients ... 13
Healthy Controls ... 13
Study Procedure ... 13
Salivary Cortisol and Experience Sampling... 13
Clinical Assessment and Questionnaires... 14
Medication Use ... 15 Statistical Analysis ... 15 Cortisol Means ... 16 Cortisol Models ... 16 Results ... 18 Missing Data ... 18 Sample Characteristics ... 18 Cortisol Values ... 19 Cortisol Means ... 20
3
Cortisol Models ... 22
Null Model ... 22
Event Score Model ... 24
Clinical Symptoms Model ... 25
Discussion ... 28
Explanation of Results ... 28
Decreased Cortisol Levels in Adults with 22q11DS ... 28
Increased Negative Affect Towards Daily Stressors in 22q11DS ... 29
Relationship Between Symptoms and Cortisol in 22q11DS ... 29
Disrupted Mechanisms ... 29
Disruption of the HPA Axis: CRF Receptor Downregulation ... 30
Disruption of PFC Dopamine: Catecholamine Autotoxicity ... 30
Suggestions for Improvement ... 31
Suggestions for Future Research ... 32
Conclusion ... 33 References ... 34 Supplementary Data ... 44 Cortisol Values ... 44 Cortisol Means ... 45 Cortisol Models ... 47 Null Model ... 47
Event Score Model ... 48
4
Introduction
A Deletion at Chromosome 22q11.2
The 22q11.2 deletion syndrome (22q11DS) is a copy number variant disorder that involves a microdeletion at the 11.2 locus on the long arm of chromosome 22. The syndrome, in the past referred to as DiGeorge syndrome or velocardiofacial syndrome (VCFS), often occurs de novo.
The patient population shows patterns of congenital cardiac disease, velopharyngeal insufficiency, hypocalcemia, immune disorders, and aberrant facial features. Moreover, 22q11DS is associated with anxiety and mood disorders, cognitive impairments and low IQ, aberrant experiences of daily life stress, and a sense of overcharging (Bassett et al., 2011; Fung et al., 2015). Recently, these impairments have been associated with elevated cortisol levels during childhood and adolescence in these patients (Jacobson, Bursch, & Lajiness-O'Neill, 2016).
Furthermore, the 22q11.2 deletion is the highest genetic risk factor for psychosis (Ripke et al., 2014; Ripke et al., 2013; Stefansson et al., 2009; Stefansson et al., 2008), and 22q11DS patients are therefore considered a high-risk population for the development of psychosis (Jonas, Montojo, & Bearden, 2014; Schneider et al., 2014a; Schneider et al., 2014b), with psychosis risk increasing with age (Hooper et al., 2013). Several predictors for the development of psychotic illness in these patients have been discovered, amongst which decreased IQ (Hooper et al., 2013; Vorstmann et al., 2015).
Recently, more biomarkers for psychosis in 22q11DS are being identified. For example, brain development in 22q11DS – or more specifically, synaptic pruning in the right frontal region during adolescence – is severely disrupted, resulting in deviations in cortical thickness (Ramanathan et al., 2017). Indeed, recent work suggests that cortical thickness is a biomarker for psychosis (Dazzan, 2014; Díaz-Caneja et al., 2015; Fond et al., 2015; Ramanathan et al., 2017).
In the current research study, we studied cortisol levels and stress reactivity in adults with 22q11DS using a daily sampling method. Furthermore, we aimed to show whether salivary
5 cortisol levels are a predictor of increased risk for psychosis in adults with 22q11DS. Finally, we propose a possible mechanism for alterations in the HPA axis in 22q11DS patients. First, we will discuss the role of stress and other psychiatric symptoms exhibited by this population (i.e., anxiety and psychosis), and the neural substrates that mediate these impairments in social cognition.
Deficits in 22q11DS
Symptoms of 22q11DS include several cognitive deficits, amongst which impairments in IQ, attention, mental flexibility, working memory, verbal working memory, and executive network efficiency. Additionally, patients with 22q11DS show several social deficits (Norkett, Lincoln, Gonzalez-Heydrich, & D’Angelo, amongst which impairments in communication, processing facial expressions, facial responses, and affective responses, as well as shyness and social withdrawal. Furthermore, 22q11DS is often comorbid with several psychiatric disorders.
Psychiatric Symptoms
Psychiatric disorders in patients with 22q11DS may include psychotic disorders, anxiety disorders, attention deficit hyperactive disorder, autism spectrum disorders, obsessive compulsive disorder, affective disorders, and bipolar disorder, amongst others.
Daily Life Stress
Comorbid with these symptoms, or perhaps even as a result, patients with 22q11DS experience high amounts of stress, perceived stress, and anxiety, even from minor stressors in daily life. Increased stress reactivity in 22q11DS may result from stressful experiences from childhood onwards, including surgery and other medical issues, experiences emerging from low IQ and impaired social cognition, or parental anxiety (Beaton & Simon, 2011).
Psychosis and Schizophrenia
As already mentioned over the previous paragraphs, patients with 22q11DS are at high risk for the development of psychotic disorder. Early-onset psychosis is relatively common amongst patients with 22q11DS, and one in every four to five patients will have developed
6 this mental disorder, usually during late adolescence or early adulthood. Furthermore, schizophrenia is often comorbid with schizoaffective disorder in 22q11DS.
Neural Substrates of Deficits in 22q11DS
Cortisol and the HPA Axis
In response to stressors, the steroid hormone cortisol, a glucocorticoid, is secreted by the adrenal gland, part of the hypothalamic-pituitary-adrenal (HPA) axis. Reactivity to stressors, as measured by cortisol levels, reflects activity of the HPA axis. In subjects that often experience stress from daily life stressors, activity of the HPA axis, and thus stress reactivity, is altered (Zorn et al., 2017). For example, patients with major depression or psychotic disorder have elevated cortisol levels during the day and show an increased cortisol stress response (Mondelli et al., 2010; Peeters, Nicolson, & Berkhof, 2004), while cortisol levels are lowered and reactivity is blunted in patients with anxiety disorder or posttraumatic stress disorder (Walker, O’Connor, Schaefer, Talbot, & Hendrickx, 2011; Yehuda, Teicher, Trestman, Levengood, & Siever, 1996). HPA axis reactivity to pharmacological and psychological stressors depends on current life adversity, age, and genetic factors. The HPA axis is programmed early in life, and functioning of the HPA axis is modulated by multiple factors such as early life stress. However, it is unknown whether adults with 22q11DS have altered stress reactivity.
Normal functioning of the HPA axis is crucial for a healthy response to stress. When a stressor is present, corticotropin-releasing factor (CRF), released by the hypothalamus, triggers the release of corticotropin (ACTH) by the anterior pituitary, which in turn promotes cortisol secretion by the adrenal gland, increasing blood glucose levels. Cortisol secretion is then moderated by the negative feedback it exhibits on the hypothalamus and pituitary. Cortisol levels rise rapidly but transiently, providing an adequate and dynamic mechanism to cope with stressors (Lehrner, Daskalakis, & Yehuda, 2016). However, prolonged periods of elevated cortisol levels, caused by long-term exposure to stress, result in negative effects. Excessive activation of the HPA axis may trigger alterations in the brain that can lead to the development of psychiatric disorders, as we will show over the next paragraphs.
7 Cortisol Follows a Diurnal Rhythm
In healthy subjects, cortisol levels follow a diurnal rhythm, peaking early in the morning right before waking up and then gradually decreasing as the day progresses to reach their lowest point in the evening right before going to bed, only interrupted by a small stagnation around lunchtime (Nicolson, 2008). However, the diurnal slope of cortisol is disrupted whenever the subject has a stressful experience. Approximately 15 minutes after a stressor, salivary cortisol levels will be transiently increased. For subjects under adverse circumstances (e.g., childhood maltreatment or trauma), basal cortisol levels can be elevated, and the diurnal rhythm and the cortisol stress response can be severely disrupted (i.e., either exacerbated or blunted; Tarullo & Gunnar, 2006).
Cortisol sampling can provide information on a subject’s risk of developing psychiatric disorders, but cortisol sampling alone is not sufficient to assess reactivity to daily life stressors. Because cortisol levels are dependent on time of day as well as other external factors, it is important to carefully register those factors when studying cortisol levels. This includes an assessment of the subject’s current activity as well as self-reported experience of stress from the activity and social company at the time of sampling. Because cortisol follows a diurnal rhythm, it is essential to collect multiple samples while interrupting the subject’s activity as least as possible to assess cortisol levels and reactivity of the HPA axis during daily life. To these ends, the experience sampling method (ESM) has been developed by Myin-Germeys and collegues (Myin-Germeys, Delespaul & Van Os, 2003; Myin-Germeys et al., 2009; Myin-Germeys & Van Os, 2007), and has been used in combination with cortisol sampling successfully (Collip et al., 2011; Hernaus et al., 2015; Holtzman et al., 2013; Lardinois, Lataster, Mengelers, Van Os, & Myin-Germeys, 2011; Lardinois et al., 2009; Lataster et al., 2011; Peeters et al., 2004; Reininghaus et al., 2016; Tarullo & Gunnar, 2006; Van Winkel, Stefanis & Myin-Germeys, 2008).
Alterations to the HPA Axis in 22q11DS
Although little is known about the long-term effects of the 22q11.2 deletion on the HPA axis (Beaton & Simon, 2011), the effects have recently been studied in children with 22q11DS (Jacobson et al., 2016; Sanders, Hobbs, Stephenson, Laird, & Beaton, 2017). It has been
8 found that children with 22q11DS show elevated basal cortisol levels (Jacobson et al., 2016; Sanders et al., 2017) and an exacerbated cortisol stress response (Sanders et al., 2017).
Long-Term Exposure to Cortisol Leads to Psychiatric Problems
Furthermore, 22q11DS is associated with several anxiety disorders (Schneider et al., 2014a), thought to result from alterations to the HPA axis caused by long-term exposure to stress or childhood stressful events. Indeed, in the general population, more stressful life events during childhood lead to a less pronounced cortisol increase, possibly indicating increased resilience to stress (Armbruster et al., 2012). Moreover, extreme stress, for example resulting from childhood trauma or abuse, has been shown to result in HPA axis alterations, leading to increased risk for several psychiatric disorders, amongst which psychotic illness (Charmandari, Tsigos, & Chrousos, 2005; Faravelli et al., 2012; Fries, Hesse, Hellhammer, & Hellhammer, 2005; Janssen et al., 2004; Nicolson, 2008; Posener et al., 2000; Read, Van Os, Morrison, Ross, 2005; Tarullo & Gunnar, 2006; Varese et al., 2012). The severity of the childhood stressful events, in turn, predict cortisol levels and severity of symptoms in psychiatric patients (Yehuda & Charney, 1993).
In psychosis, reactivity to daily life stress is increased and the cortisol response is altered (Collip et al., 2011; Jansen et al., 1998; Lardinois et al., 2009; Mondelli et al., 2010; Zorn et al., 2017). Patients with schizophrenia or psychotic disorder show a blunted cortisol awakening response (Mondelli et al., 2010) and a blunted cortisol response to psychosocial stressors (Jansen et al., 1998). Furthermore, patients show elevated overall diurnal cortisol levels (Mondelli et al., 2010). Deviations from the diurnal cortisol slope predict intensity of psychotic experiences on the patient level, although no group difference with healthy controls has been found (Collip et al., 2011; Lardinois et al., 2009; Mondelli et al., 2010). However, the exact association and its direction between increased cortisol reactivity and the intensity of psychotic experiences remains unknown.
We have shown that several psychiatric disorders (e.g., major depression, psychotic disorders, anxiety disorders, and posttraumatic stress disorder) are associated with deviations in cortisol levels as opposed to the general population, resulting from sensitization or dysregulation of the HPA axis (Charmandari et al., 2005; Fries et al., 2005; Posener et al., 2000; Yehuda, 2002; Yehuda et al., 1996). For example, lower cortisol levels in patients with
9 posttraumatic stress disorder are suggested to be caused by disruption of the CRF receptor (Justice et al., 2015). Conversely, altered cortisol levels or an altered cortisol stress response both have been identified as a risk factor for the development of these psychiatric disorders (De Girolamo & McFarlane, 1996; Van der Kolk, 1997; Resnick, Yehuda, Pitman, & Foy, 1995).
Dopamine, COMT, and the PFC
The increased genetic risk for psychotic illness in 22q11DS patients may be mediated through a loss of control of the prefrontal dopamine system over stress-induced subcortical dopaminergic activity. Dopaminergic activity in the prefrontal cortex (PFC) is associated with psychotic reactivity to daily life stress (Gevonden et al., 2014; Hernaus et al., 2013; Lataster et al., 2011). Indeed, 22q11DS patients show psychotic reactivity to daily life stress (Beaton & Simon, 2011). Furthermore, it has recently been discovered that the reward system, which relies on dopaminergic neurotransmissions (Da Silva Alves et al., 2011), is impaired in patients with schizophrenia (Da Silva Alves et al., 2013) and in 22q11DS patients (Van Duin et al., 2016).
The impairment of the dopamine system found in 22q11DS patients can be partly explained by hemizygosity of the catechol-O-methyltransferase (COMT) gene. COMT is crucially involved in the breakdown of prefrontal dopamine (Tunbridge, Harrison, & Weinberger, 2006). The COMT gene is part of the 22q11.2 microdeletion found in 22q11DS patients. Therefore, these patients suffer from low enzymatic COMT (Gothelf et al., 2014). This haploinsufficiency of COMT results in disrupted dopaminergic neurotransmission (Boot et al., 2008).
Moreover, this effect is amplified in patients with a COMT Val158Met polymorphism. COMT Met hemizygosity results in a reduction of COMT mRNA, protein and enzyme activity of 40% (Chen et al., 2004) in addition to the reduction of 50% caused by the hemizygous deletion of the gene (Gothelf et al., 2014). COMT hemizygosity, and especially the Val158Met polymorphism, increases susceptibility for psychosis through a loss of prefrontal dopamine function (Boot et al., 2011a; Boot et al., 2011b; Hernaus et al., 2013; Van Duin et al., 2016).
10 Interestingly, individuals with this polymorphism, both 22q11DS patients and HC, show an improvement in prefrontal cognitive function under normal circumstances (Bearden et al., 2004), but not under stress (Stein, Newman, Savitz, & Ramesar, 2006). However, there has been no clear dissociation between these polymorphisms and the risk of developing psychiatric disorders.
Thus, although the effect of stress on alterations of the HPA axis and the resulting psychiatric problems in 22q11DS patients may be mediated through the prefrontal dopamine system (Beaton & Simon, 2011), how remains unclear.
Interaction Between These Systems
Dopamine dysfunction may alter the HPA axis in 22q11DS patients through loss of control by the prefrontal dopamine system (Beaton & Simon, 2011). Activity of the prefrontal dopamine system and the HPA axis are closely related. For example, healthy children with a COMT Val158Met polymorphism show a higher cortisol stress response (Armbruster et al., 2012).
We have shown in the previous paragraphs that reactivity of the HPA axis is altered in subjects that often experience daily life stress and psychotic symptoms. Conversely, reactivity to daily life stress is increased in psychotic patients as a result of an altered HPA axis (Mondelli et al., 2010). This underlies observations of increased psychotic reactivity to stress at the behavioral level (Hernaus et al., 2015) and elevated cortisol levels and increased cortisol reactivity to daily life stress in psychotic patients (Collip et al., 2011). However, little is known about the functioning of the HPA axis in 22q11DS patients. Studying the diurnal rhythm of cortisol helps discovering the mechanisms of psychosis.
We have shown in the previous paragraphs that the increased risk of developing mental disorders resulting from childhood stress, or even extreme stressful events during adulthood, may be mediated through alterations to the HPA axis as well as to the prefrontal dopamine system. Therefore, we aimed to show whether this association could also be studied in 22q11DS patients.
11
Aim, Hypotheses and Expected Results
Individuals with 22q11.2 deletion syndrome are at increased genetic risk of developing psychotic illness (Jonas et al., 2014; Schneider et al., 2014a; Schneider et al., 2014b). This research study aimed to propose a model for explaining psychotic symptoms in 22q11DS patients, through self-reported experience of stress and diurnal salivary cortisol levels, as well as other predictors. As a result, we report the first study of cortisol levels in adults with 22q11DS, to the knowledge of the authors.
First, we tested whether diurnal salivary cortisol levels were altered in 22q11DS patients when compared to healthy controls. We expected that cortisol levels were higher in adults with 22q11DS, as they are in children with 22q11DS (Jacobsen et al., 2016; Sanders et al., 2017). Elevated cortisol levels would reflect increased activity of the HPA axis, in accordance with findings of increased comorbidity with psychiatric symptoms such as anxiety disorders in this patient population.
Furthermore, we tested whether cortisol levels would be associated with psychiatric symptoms in adults with 22q11DS. We expected to find an exacerbated relationship between cortisol levels and symptoms in patients, i.e. that increased cortisol levels would be associated with increased occurrence of psychotic symptoms. Altered cortisol levels possibly reflect changes to the HPA axis, which in turn affects the dopamine system in basal and prefrontal brain regions. Studying stress reactivity and its contribution to psychiatric symptoms in 22q11DS could therefore provide an explanation for the increased risk for psychosis in 22q11DS patients.
Finally, we tested whether adults with 22q11DS showed altered reactivity to daily life stressors. We expected that patients would show increased negative affect towards daily life stressors. Additionally, we examined whether self-reported experiences of stress were a predictor for salivary cortisol levels. We expected that 22q11DS patients would show increased HPA axis reactivity to daily life stressors as measured by diurnal salivary cortisol levels, evidencing of a biomarker for the divergent symptoms and impairments in this group. Several studies have found effects of numerous types of medication on salivary cortisol levels. The suppressive effect of antipsychotics (either typical or atypical), psychoactives (e.g. selective serotonin reuptake inhibitors [SSRIs], methylphenidate), and other drugs (e.g.
12 oxazepam, [oral] contraceptives) on cortisol levels has been well established (Granger, Hibel, Fortunato, & Kapelewski, 2009; Hibel, Granger, Kivlighan, Blair, & Family Life Project Investigations, 2006; Mondelli et al., 2010; Nicolson, 2008). Furthermore, factors such as mood and anxiety disorders or major depression, as well as the presence of early life adverse events affect long-term reactivity of the HPA axis and thus cortisol levels (Peeters et al., 2004; Tarullo & Gunnar, 2006; Yehuda et al., 1996). Additionally, other factors also affect salivary cortisol levels, such as COMT Val158Met polymorphism, occurrence of childhood trauma, intelligence and cognitive performance, social and occupational functioning, and self-reported stress. Therefore, these factors are all assessed in this research study and used as a confounder to check for any effects on cortisol levels.
13
Methods
Subjects
All participants were aged between 18 and 65 years and had to be mentally competent to give informed consent. Informed consent was only obtained after participants were fully informed on the study procedure.
Patients
27 adults with the 22q11.2 deletion syndrome (DS) were recruited through the Dutch 22q11DS family association (Stichting Steun 22q11) and their newsletter, and the 22q11DS inpatient clinics of the University Hospital Maastricht, University Medical Centre Utrecht and University Hospital Leuven. The 22q11.2 deletion had to be confirmed by Fluorescence In Situ Hybridisation (FISH), Multiplex Ligation-dependent Probe Amplification (MLPA), or micro-array analysis.
Healthy Controls
24 healthy volunteers (HC) who were matched for age and gender with patients were selected. No healthy controls were on medication.
Study Procedure
The protocol was approved by the Ethics Committee of Maastricht University. The study was conducted at the homes of participants, at the Amsterdam Medical Center, and at the University Hospital Maastricht.
Salivary Cortisol and Experience Sampling
All subjects carried an electronic momentary assessment technology device used for experience sampling (ESM) (PsyMate, Maastricht University, the Netherlands; Myin-Germeys, Birchwood, & Kwapil, 2011) over a period of 6 consecutive days, in which participants were required to live their normal daily life. Each day, the PsyMate went off 10 times between random time intervals during awake hours (07h30-22h30). At each beep, participants first filled out a questionnaire assessing emotional state, social company, and current activity, including an event score.
14 Directly at the end of each questionnaire, participants collected a saliva sample using a cotton swab (Salivette, Sarstedt, the Netherlands), which were stored in a salivette tube in their home freezers. Collection times were registered through the PsyMate. After the 6-day period, samples were kept at -20°C until analysis at Dresden University of Technology in Germany. Cortisol was analysed from the saliva samples in duplicate using radio-immunoassays. Tracer solution Cortisol 3-CMO coupled with 2-[125I]-histamine and antibodies for Cortisol 3-CMO-BSA was used (Collip et al., 2011).
Cortisol values above 60 nmol/L were removed. Participants with less than three days of at least three cortisol samples per day were excluded from further analyses. Cortisol values were log-transformed to reduce skewness of distribution (Collip et al., 2011).
Clinical Assessment and Questionnaires
During testing, a number of questionnaires were administered. Dutch versions were used for all questionnaires. All questionnaires were administered before the ESM period.
General demographic characteristics were registered for all participants. These included smoking, drinking, and other addictive behaviour, which was used to assign participants to an addicted and a non-addicted group.
Total intelligence quotient (TIQ) was assessed using the Dutch abbreviated version of the Wechsler Adult Intelligence Scale (WAIS-III-NL) (Canavan, Dunn, & McMillan, 1986; Wechsler, 1997) in patients and the Dutch Adult Reading Test (DART) in healthy controls. The Mini International Neuropsychiatric Interview (M.I.N.I. 5.0.0) was used to assess DSM-IV diagnosis of patients (Overbeek, Schruers & Griez, 1999). 22q11DS patients that had an additional diagnosis were assigned to a psychiatry group to discern them from other patients. Healthy controls were assumed to have no diagnosis and thus formed a third group.
The Beck Depression Inventory (BDI-II) was used to assess depressive symptoms in patients (Beck, Steer, & Brown, 1996). The Brief Psychiatric Rating Scale (BPRS) (Faustman & Overall, 1999) was used to assess symptoms of depression in healthy controls. The State-Trait Anxiety Inventory (STAI) was used to measure anxiety in both groups (Spielberger, Gorsuch, & Lushene, 1970).
15 Adverse events during childhood were assessed using the 28-item version of the Childhood Trauma Questionnaire (CTQ) (Bernstein et al., 2003), which consists of five subscales (i.e., physical abuse, emotional abuse, physical neglect, emotional neglect, and sexual abuse), a minimization/denial factor (MacDonald et al., 2016), and a total score. A short version of the Retrospective Bullying Questionnaire was used in both patients and healthy controls to measure bullying experienced in the past (Schäfer et al., 2004).
Psychotic symptoms in patients were assessed using the Positive and Negative Symptom Scale (PANSS) (Kay, Fiszbein & Opfer, 1987). In addition to the regular subscales of the PANSS (i.e., Positive symptoms, Negative symptoms, and General Psychopathology), five other subscales were derived from the item scores and used for analyses: negative, positive, disorganized, excited, and anxiety/depression (Emsley, Rabinowitz, & Torreman, 2003). Prodromal symptoms of psychosis were assessed in patients using the 16-item version of the Prodromal Questionnaire (PQ-16) (Loewy et al., 2005; Ising et al., 2012). In addition to a subscore for prodromal symptoms, the distress associated with these symptoms was also scored.
Medication Use
Medication use was assessed in all participants. All participants were divided amongst a medication and a no-medication group. Participants in the medication group were using medication with a known effect on cortisol levels. Participants in the no-medication group were using only medication without a known effect on cortisol levels, or no medication at all. For some analyses, participants were further subdivided into one of five different groups, based on medication: patients on either typical or atypical antipsychotics, participants using other psychoactive medication (e.g. selective serotonin reuptake inhibitors (SSRIs), methylphenidate), medication use with suppressive effects on cortisol levels (e.g. oxazepam, (oral) contraceptives), a group using medication without known effects on cortisol levels, and finally a group without any form of medication use.
Statistical Analysis
To study the effects of the 22q11.2 deletion and its clinical features on salivary cortisol of patients, several statistical analyses were performed using R version 3.3.2 for Windows (R
16 Core Team, 2016). As described above, only participants with at least three days of at least three cortisol samples per day were included in analyses. Logarithmical transformation of the cortisol values was performed to reduce skewness of distribution of cortisol values and to normalize the data, and to provide a linear trend of log-cortisol values over time (Peeters et al., 2004). Using the sample time, cortisol samples were assigned to one of three moments: morning (07h30-12h30), afternoon (12h30-17h30), and evening (17h30-22h30).
Cortisol Means
Independent sample t-tests were performed to assess differences in log-cortisol values between patients and healthy controls, between males and females in both groups, and between users and non-users of medication amongst patients. Means of log-cortisol were compared between groups on the whole-day level as well as morning, afternoon and evening independently. To assess the nature of the diurnal rhythm of cortisol, mean differences between moments within groups were also calculated using independent samples t-tests. No paired samples t-tests could be used due to the random nature of the sampling method (i.e., sample sizes across moments were different).
Cortisol Models
To further investigate the diurnal pattern of salivary cortisol, multiple linear regression was performed on log-cortisol values. Due to the complexity of the measurements, which were nested within days and within subjects, a multilevel mixed model was selected containing a day- and a subject-level, to account for variance at the level of the individual and for variance at the level of the moment of the day (Lardinois et al., 2009). Random effects at the beep-level were accounted for by including random sample time intercepts and slopes for each level. Fitting of fixed effects included confounding factors gender, age, IQ, medication use, psychiatric diagnosis, and addiction. Models were fitted for both groups independently and both groups together. Every model was additionally estimated without IQ as confounder. In a first model, no additional predictors were included to assess the effects of the 22q11.2 deletion on salivary cortisol values. Secondly, the event score from the ESM questionnaire was entered as a predictor. In a third step, clinical symptoms were used as predictors in the first model using Bullying, CTQ, PQ, PANSS, and STAI questionnaire scores. Cases with missing data for any of the predictors were omitted from analyses (Little & Rubin, 2014).
17 Wald test z-scores with corresponding p-values were calculated (i.e., dividing the estimated effect by its standard error) for each predictor using regression coefficients. Tests for F-scores and log-likelihood ratios were performed to assess performance of models.
18
Results
Missing Data
Study subjects responded to a beep 2415 times to fill out an ESM questionnaires, of which 1951 included a cortisol sample. One sample was removed because it was over 60 nmol/L, and 31 samples were removed because they were from participants with less than three days of at least three cortisol samples per day. This resulted in a data set of 1919 cortisol samples from 51 subjects (27 DS patients and 24 controls).
Sample Characteristics
See Table 1 for an overview of demographic characteristics of participants, and Table 2 for an overview of medication use amongst patients. In addition to the comparison of means between groups, a correlation between IQ and gender was found in the patients group (rs=.420, p<.05), but no mean IQ difference between the sexes (t(10.1)=2.17, n.s.) (Figure 1). Furthermore, there was a correlation between age and gender (rs=.320, p<.05) and a mean age difference between sexes (t(31.0)=2.36, p<.05) in the entire study population, caused by a correlation (rs=.537, p<.01) and a mean age difference between sexes (t(14.5)=2.96, p<.01) in the healthy controls group, but not in patients (Figure 2).
22q11DS HC Total t (df) p N 27 24 51 Gender 33.3% M 66.7% F 29.2% M 70.8% F 31.4% M 68.6% F 0.31 (48.6) 0.75 Age (± SD) 34.1 (± 9.8) yrs 38.9 (± 13.4) yrs 36.4 (± 11.8) yrs 1.44 (41.7) 0.16 IQ (± SD) 78.3 (± 10.4) 106 (± 8.4) 91.3 (± 16.8) 10.5 (48.5) < 0.001 Table 1. Overview of characteristics of participants.
19 Figure 1. Age (left) and IQ (right) differences between genders between research groups. There is a correlation (rs=.537, p<.01) and a mean age difference between sexes (t(14.5)=2.96, p<.01) in the healthy controls group, but not in patients. There is a mean IQ difference between research groups (t(48.5)=10.5, p<.001). Despite a significant correlation (rs=.420, p<.05), there is no mean IQ difference between genders within groups (t(10.1)=2.17, n.s.).
Medication (N = 13) Antipsychotic Psychoactive Suppressive
Risperdal Amitriptyline Betamethason
Zyprexa Concerta Deso 30
(contraceptive) Paroxetin (2) Flixonase Priadel Omeprazol Sertraline Oxazepam Sipralexa Strattera Table 2. Overview of medication use of participants.
Cortisol Values
A logarithmic transformation of the cortisol values normalized the data at the beep-level (Figure S1), and provided a linear trend of cortisol values over time (Figure S2).
20
Cortisol Means
Mean log-cortisol values of patients (M=0.936 nmol/L, SD=0.844) and healthy controls (M=2.01 nmol/L, SD=0.811) differed significantly over the whole day (t(1905)=28.41, p<.0001) as well as for other parts of the day (Table 3; Figure 2). For both groups, cortisol values decreased over the day, resulting in significant differences in mean log-cortisol values within groups across moments of day (Table S1; Figure 2).
There was a significant difference in mean log-cortisol values between female and male subjects in the healthy controls group across the whole day (Table S2; Figure 2), although not in the afternoon only (12h30-17h30). There was no significant difference in mean log-cortisol values across the whole day between genders within the patients group, except in the evening only (17h30-22h30).
There was a significant difference in mean log-cortisol values between users of medication that are known to affect cortisol levels (Table S3; Figure 3). Log-cortisol values of medication users (M=0.937 nmol/L, SD=0.757) were significantly lower than those of non-medication users (M=1.56 nmol/L, SD=0.991) over the whole day (t(405.4)=11.88, p<.0001) as well as for other parts of the day. The differences in mean log-cortisol values between groups of users of different types of medication were bigger during the evening.
DS HC t df p
Morning 1.57 2.61 19.71 565.4 <.0001
Afternoon 0.989 2.13 27.08 656.3 <.0001
Evening 0.379 1.35 16.32 656.1 <.0001
Day 0.936 2.01 28.41 1905 <.0001
Table 3. Mean differences of log-cortisol values between research groups for each moment of day, including significance from a t-test.
21 Figure 2. Mean differences between research groups (top) and between sexes (bottom) of log-cortisol values at each ESM beep (left) and at each moment of day (right).
Figure 3. Mean differences in log-cortisol values between medication users and non-medication users (left), and between users of different types of non-medication (right).
22
Cortisol Models
All models consisted of a subject-level and a day-level. All models included gender, age, IQ, medication use, psychiatric diagnosis, and addiction as confounding factors. Medication and psychiatric diagnosis groups did not differ within the control group (i.e., the value was always ‘0’), and were not included as confounders in models for healthy controls only. Including confounders in the models resulted in a worse fit than a model without confounders in all cases, as assessed by log-likelihood ratio tests.
Null Model
The null model predicted the effect of the 22q11.2 deletion on salivary cortisol values (Figure 4). No additional predictors were included in this model, resulting in the following model formulation:
Log-cortisol = 5.09 - 1.45 * Test group - 0.136 * Sample time + 0.00938 * Sample time * Test group ± 0.338
See Table 4 for fixed effects and Table S4 for random effects. Test group was a significant predictor (t=-4.29, p=.0001), indicating that the 22q11.2 deletion has a negative effect on log-cortisol values. Sample time was also a significant predictor (t=-0.136, p<.0001), indicating that log-cortisol values decrease over the day. When comparing two separate models that are fitted to the two different test groups (Table S5), it was found that the intercepts were significantly different (z=-4.30, p<.0001), whereas the slopes did not, confirming these findings. There was no significant interaction between test group and sample time, indicating that the effect of the 22q11.2 deletion on log-cortisol values does not change during the day.
23
Groups Coefficient Estimate SE t-value p-value
DS (N=941) Intercept 3.00 0.586 5.12 <.0001 Sample time -0.126 0.00942 -13.4 <.0001 HC (N=978) Intercept 6.93 0.703 9.86 <.0001 Sample time -0.136 0.00756 -18.0 <.0001 both (N=1919) Intercept 5.09 0.566 8.98 <.0001 Sample time -0.136 0.00870 -15.6 <.0001 Test group -1.45 0.338 -4.29 0.0001 Interaction 0.00938 0.0121 0.774 0.439
Table 4. Fixed effects for the null model including both research groups.
24 Figure 5. Linear fits of the regression models for the patients group.
Event Score Model
The event score model predicted the added effect of the ESM event score on salivary cortisol values (Figure 5 and Figure S5), resulting in the following model formulation:
Log-cortisol = 5.03 + 0.00471 * Event score - 1.37 * Test group - 0.131 * Sample time + 0.0321 * Event score * Test group ± 0.0149
See Table 5 for fixed effects and Table S6 for random effects. Event score was not a significant predictor, indicating that there was no main effect of event score on log-cortisol values. Test group (t=-4.30, p=.0001) and sample time (t=-22.0 p<.0001) were still significant predictors. There was no significant interaction between test group and event score, although a trend was visible when comparing the slopes of the two separate models that are fitted to the two different test groups (Table S7).
25 When a model was fit for the 22q11DS patients group, it was found however that there was a significant main effect of event score on log-cortisol values (t=2.39, p=.0171), indicating that a higher event score (i.e., for a more pleasant event) results in elevated salivary cortisol levels. This resulted in the following model formulation for 22q11DS patients:
Log-cortisol = 3.00 + 0.0365 * Event score - 0.127 * Sample time ± 0.0153
When comparing the event score model to the null model using a log-likelihood ratio test, it was found that the fits of these models were not significantly different, and in the case of the healthy control group the null model was even a significantly better fit (Table S8).
Group Coefficient Estimate SE t-value p-value
DS (N=940) Intercept 3.00 0.605 4.96 <.0001 Event score 0.0365 0.0153 2.39 0.0171 Sample time -0.127 0.00931 -13.6 <.0001 HC (N=974) Intercept 6.96 0.714 9.74 <.0001 Event score 0.00187 0.0495 0.0377 0.970 Sample time -0.137 0.00886 -15.4 <.0001 both (N=1914) Intercept 5.03 0.575 8.75 <.0001 Event score 0.00471 0.0149 0.316 0.752 Sample time -0.131 0.00596 -22.0 <.0001 Test group -1.37 0.317 -4.30 0.0001 Interaction 0.0321 0.0206 1.55 0.120
Table 5. Fixed effects for the event score model including both research groups.
Clinical Symptoms Model
The clinical symptoms model predicted the added effect of the questionnaire scores on salivary cortisol values (Figure 5 and Figure S5), resulting in the following model formulation:
Log-cortisol = 5.68 - 0.0163 * STAI - 0.00125 * CTQ - 0.00893 * Bullying - 1.53 * Test group - 0.134 * Sample time + 0.00380 * Sample time * Test group ± 0.00913
26 See Table 6 for fixed effects and Table S9 for random effects. STAI score was a significant predictor for log-cortisol values in a model for both groups (t=-2.90, p=.0066), indicating that participants with greater anxiety had lower log-cortisol values. CTQ and Bullying scores were not a significant predictor in this model, indicating that there was no main effect of these scores on log-cortisol values. Test group (t=-4.42, p=.0001) and sample time (t=-14.7, p<.0001) were still significant predictors. There was no significant interaction between test group and sample time. When comparing two separate models that are fitted to the two different test groups – although the number of predictors in these models was not equal – (Table S10), it was found that the intercepts and slopes were both not significantly different. In the model for healthy controls only, none of the questionnaire scores for clinical symptoms were significant predictors. In the model for 22q11DS patients only, additional predictor PANSS score was significant (t=-2.35, p=.0385), mainly driven by the score on the Negative subscale. PQ, as well as other questionnaire scores, was not significant, but a trend was visible for STAI and CTQ scores. This resulted in the following model formulation for 22q11DS patients:
Log-cortisol = 5.55 - 0.0287 * PANSS + 0.0107 * PQ - 0.0144 * (STAI + CTQ) - 0.0131 * Bullying - 0.130 * Sample time ± 0.00984
When comparing the clinical symptoms model to the null model using a log-likelihood ratio test, it was found that the fits of these models were significantly different (χ2=14.9, p=.00187) in favor of the null model, also when comparing the two separate models for the two different test groups (Table S11).
27
Group Coefficient Estimate SE t-value p-value
DS (N=845) Intercept 5.55 0.862 6.43 <.0001 Sample time -0.130 0.00984 -13.2 <.0001 PANSS -0.0287 0.0122 -2.35 0.0385 PQ 0.0107 0.0483 0.221 0.829 STAI -0.0144 0.00788 -1.83 0.0949 CTQ -0.0144 0.0119 -1.21 0.250 Bullying -0.0131 0.0269 -0.485 0.637 HC (N=861) Intercept 6.48 0.882 7.35 <.0001 Sample time -0.134 0.00825 -16.3 <.0001 STAI -0.0230 0.0120 -1.91 0.0765 CTQ 0.00316 0.00676 0.468 0.647 Bullying -0.00320 0.0207 -0.155 0.879 both (N=1706) Intercept 5.68 0.614 9.25 <.0001 Sample time -0.134 0.00913 -14.7 <.0001 Test group -1.53 0.346 -4.42 0.0001 STAI -0.0163 0.00562 -2.90 0.0066 CTQ -0.00125 0.00601 -0.209 0.836 Bullying -0.00893 0.0168 -0.533 0.598 Interaction 0.00380 0.0128 0.230 0.767
28
Discussion
Using a daily life sampling method, we tested whether cortisol levels were altered in adult 22q11DS patients. We found that 22q11DS patients had lower cortisol levels than healthy controls. Furthermore, we confirmed that experiences of stress result in lower cortisol levels in 22q11DS patients, while we found that presence of psychotic symptoms, anxiety, and childhood trauma resulted in lower cortisol values in 22q11DS patients, although these latter effects were not significant.
Explanation of Results
Decreased Cortisol Levels in Adults with 22q11DS
We hypothesized that the constant experience of stress in 22q11DS patients would alter the HPA axis. We expected to find elevated salivary cortisol levels using daily life sampling in this population. However, our results indicate the opposite effect: cortisol levels were lower in 22q11DS patients than in healthy controls.
In several studies assessing salivary cortisol in patients with psychotic illness only antipsychotics are included as confounding factors, because antipsychotic medications reduce HPA activity and thus decrease cortisol levels (e.g., Mondelli et al., 2010; Walker, Mittal, & Tessner, 2008). However, other types of medication affect cortisol levels as well (Granger et al., 2009). In the present study, these drugs are categorized and used as confounders. Therefore, our results provide a robust base to draw conclusions on cortisol levels in 22q11DS. In any case, the effects of medication on HPA axis activity and cortisol levels should not be taken lightly. The sex differences in mean cortisol values we found are explained by the influence of sex hormones and the use of contraceptives (Zorn et al., 2017). Our results may suggest the possibility of a dissociation between the stress response and the cortisol response. However, this is unlikely, because we found no evidence of a divergent cortisol response to stress in patients; only mean cortisol values are lower in patients. A possible mechanism is discussed in the next section of this chapter.
29 Increased Negative Affect Towards Daily Stressors in 22q11DS
Our ESM results showed that 22q11DS patients rated events negatively more often than HC did. Differences in affect towards events during daily life sampling suggest more negative feelings after minor stressors in 22q11DS patients. Furthermore, patients showed a tendency to ascribe lower scores for negative experiences. Indeed, recent work shows that 22q11DS patients exhibit chronic anxiety from childhood onwards (Jacobson et al., 2016; Sanders et al., 2017).
Relationship Between Symptoms and Cortisol in 22q11DS
We found no definitive association between psychiatric symptoms and cortisol values in 22q11DS patients. A possible reason for this is a mismatch between the timeframe of cortisol sampling, which was continuous by nature, and the timeframe over which symptoms were assessed, which always occurred before the cortisol sampling and thus does not necessarily reflect the symptoms at the time of cortisol sampling (Mondelli et al., 2010).
However, chronic anxiety and anhedonia in 22q11DS, supported by our ESM data, together with lower cortisol levels point to metabolic or immunologic problems within the HPA axis and serotonergic, dopaminergic, and noradrenergic systems. Indeed, environmental programming of neural systems can contribute substantially to the development of stable individual differences in HPA responsivity to stressful stimuli (Meaney et al., 1996).
Disrupted Mechanisms
Our results indicate that people with 22q11DS experience higher self-reported negative affect to small stressors in daily life, whilst showing lower mean cortisol levels than HC. This is a divergent pattern from the relationship between affective responses and adrenocortical stress responses in HC (Jacobs et al., 2007). Here, we argue that this finding results from an over-sensitization of the HPA axis in 22q11DS patients.
In posttraumatic stress disorder and psychotic major depression, chronic overactivation of the HPA axis (i.e., allostatic load) gives rise to hypocortisolism (Charmandari et al., 2005; Fries et al., 2005; Nicolson, 2008; Posener et al., 2000; Tarullo & Gunnar, 2006). In patients with first episode psychosis, increased cortisol levels demonstrate basal overactivity of the HPA
30 axis (Ryan, Sharifi, Condren, & Thakore, 2004). We suggest that this long-term effect of stress (McEwen & Gianaros, 2011) is permanently present in adults with 22q11DS as well. Elevated cortisol levels are an indicator of increased risk of psychosis (Collip et al., 2011; Lardinois et al., 2009). However, our results show that 22q11DS patients, a group equally at risk of psychosis, are characterized by allostatic-load-induced hypocortisolism, thus showing the opposing effect: lowered cortisol levels. Herein lies an important conclusion for clinicians that are monitoring patients for risk and development of psychosis. When using cortisol levels to assess, for instance, psychosis risk, it is important to dissociate individuals who suffered from long-term exposure to stress, such as adults with 22q11DS, from individuals who did not.
Disruption of the HPA Axis: CRF Receptor Downregulation
The permanent long-term effect of stress resulting in hypocortisolism, present in 22q11DS, might be explained through several pathways. The first possibility is that cortisol secretion is suppressed through a disruption of the noradrenergic system of the adrenal gland.
Recent studies show lowered cortisol levels in PTSD patients (Lehrner et al., 2016). Because catecholamine and CRF levels are elevated in PTSD, normally stimulating cortisol production in HC, we argue here that this might be caused by disruption of the CRF receptor. CRF hypersecretion is a consequence of extreme stressful events (Heim et al., 2000; Lehrner et al., 2016), and might lead to downgrading of CRF receptor. The CRF receptor mediates fear and anxiety responses (Takahashi, 2001). Thus, catecholamine and CRF systems mediate the association between long-term exposure to stress and attenuated cortisol levels and increased risk for psychosis in adults with 22q11DS.
Disruption of PFC Dopamine: Catecholamine Autotoxicity
Dopamine system dysfunction in 22q11DS patients, caused by COMT haploinsufficiency, could cause a downregulation of both dopamine and cortisol levels. COMT haploinsufficiency leads to a disruption of the HPA axis (Oswald, McCaul, Choi, Yang, & Wand, 2004). This is intensified in patients with a Val158Met polymorphism, which leads to increased cortisol responses in HC (Oswald et al., 2004).
31 We found decreased cortisol levels in adults with 22q11DS, while recent studies show increased cortisol levels in children with 22q11DS (Jacobson et al., 2016; Sanders et al., 2017). Previous findings showed increased prefrontal dopamine in children with 22q11DS (Bearden et al., 2004), but decreased dopaminergic function in adults with 22q11DS (Boot et al., 2008; De Koning et al., 2012; Van Duin et al., 2016). Furthermore, recent findings from patient studies show disrupted cortisol and dopamine regulation, resulting in a fundamental pathology in the molecular clock underlying circadian rhythms (Breen et al., 2014), and, concerning dopamine dysfunction, to catecholamine autotoxicity (i.e., cytotoxicity; Bisaglia, Greggio, Beltramini, & Bubacco, 2013; Goldstein, Kopin, & Sharabi, 2014), as described by the neural diathesis-stress model (Walker & Diforio, 1997). Taken together, these findings point to overactivity of the catecholamine system in patients during childhood, eventually leading to toxicity, resulting in a breakdown of the involved systems. Thus, we conclude that 22q11DS is a developmental disorder that severely disrupts neurotransmissions in the catecholaminergic system.
Suggestions for Improvement
Although the results of this research study show robust group effects on mean cortisol levels, as well as other significant predictors, a few arguments against the validity of the results could be formulated and will be invalidated here.
Cortisol levels follow a diurnal rhythm, as described previously. In analyses, linear models are not truly representative for the diurnal rhythm of cortisol; instead, this rhythm is usually modelled using a sinusor. This could be achieved by adding a fourth-degree polynomial to the model (Peeters et al., 2004). We have not included this in our models, because we found robust group differences in the first part of our analyses. When looking into more detail at the specifics of the diurnal rhythm of cortisol in 22q11DS patients, however, the model that is used should be extended.
Cortisol values are dependent of a lot of factors, some of which cannot be assessed through ESM or any of the other methods used in this study. ESM only displays the tip of the iceberg of patients’ daily life experiences; however, it helps beginning to uncover the exact workings of stress and the development of psychiatric disorders in 22q11DS patients.
32 Additionally, failure to find significance of other predictors might also be explained. As described in the first section of this chapter, many of the other predictors used in this study were assessed before the ESM testing period. Questionnaire scores from before the ESM testing period do not necessarily represent the subject during ESM testing. This could be improved by assessing psychiatric symptoms and other factors not only before, but also during and after ESM testing. Furthermore, subscales of some of the questionnaires could be used in addition to the scores on the general scales.
Finally, the question whether to include IQ as a confounder stands open to debate, as lower IQ is a well-established feature in 22q11DS, and might thus not reflect a between-groups confounding factor in addition to the group factor. However, low IQ might result in low ability for self-reflection, and might therefore be included as a confounder. Indeed, IQ, as well as social-behavioral problems, serve as a predictor for the development of further psychiatric disorders in 22q11DS patients (Hooper et al., 2013).
Suggestions for Future Research
In addition to the steps described in the previous paragraph, we propose a few directions for the future of 22q11DS research. Much still is unclear on what constitutes the association between psychiatric symptoms and lowered cortisol values in 22q11DS patients, and how these are related to changes in the dopamine system. Animal models could further help in understanding the developmental disabilities of 22q11DS (Guna, Butcher & Bassett, 2015). Finally, clinical trials with different types of medication could help further uncover the role of cortisol in the development of psychiatric disorders in 22q11DS patients. For example, mifepristone, a drug that reduces cortisol secretion, could be administered to children with 22q11DS to prevent alterations to the HPA axis that mediate the development of psychiatric disorders (Flores et al., 2006; Jacobson et al., 2016).
33
Conclusion
This research study shows that adults with the 22q11.2 deletion syndrome have lower cortisol levels than healthy individuals, whilst showing increased reactivity to daily life stress. This points to an over-sensitization of the HPA axis in these patients, possibly mediated by CRF receptor downregulation and exacerbated by catecholamine autotoxicity in the prefrontal dopamine system.
34
References
Armbruster, D., Mueller, A., Strobel, A., Lesch, K. P., Brocke, B., & Kirschbaum, C. (2012). Children under stress–COMT genotype and stressful life events predict cortisol increase in an acute social stress paradigm. International Journal of Neuropsychopharmacology, 15(9), 1229-1239.
Bassett, A. S., McDonald-McGinn, D. M., Devriendt, K., Digilio, M. C., Goldenberg, P., Habel, A., ... & Swillen, A. (2011). Practical guidelines for managing patients with 22q11.2 deletion syndrome. The Journal of pediatrics, 159(2), 332-339.
Bearden, C. E., Jawad, A. F., Lynch, D. R., Sokol, S., Kanes, S. J., McDonald-McGinn, D. M., ... & Zackai, E. (2004). Effects of a functional COMT polymorphism on prefrontal cognitive function in patients with 22q11.2 deletion syndrome. American Journal of Psychiatry, 161(9), 1700-1702.
Beaton, E. A., & Simon, T. J. (2011). How might stress contribute to increased risk for schizophrenia in children with chromosome 22q11. 2 deletion syndrome?. Journal of neurodevelopmental disorders, 3(1), 68.
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Beck depression inventory.
Bernstein, D. P., Stein, J. A., Newcomb, M. D., Walker, E., Pogge, D., Ahluvalia, T., ... & Zule, W. (2003). Development and validation of a brief screening version of the Childhood Trauma Questionnaire. Child abuse & neglect, 27(2), 169-190.
Bisaglia, M., Greggio, E., Beltramini, M., & Bubacco, L. (2013). Dysfunction of dopamine homeostasis: clues in the hunt for novel Parkinson's disease therapies. The FASEB Journal, 27(6), 2101-2110.
Boot, E., Booij, J., Zinkstok, J., Abeling, N., De Haan, L., Baas, F., ... & Van Amelsvoort, T. (2008). Disrupted dopaminergic neurotransmission in 22q11 deletion syndrome. Neuropsychopharmacology, 33(6), 1252.
Boot, E., Booij, J., Abeling, N., Meijer, J., da Silva Alves, F., Zinkstok, J., ... & Van Amelsvoort, T. (2011). Dopamine metabolism in adults with 22q11 deletion syndrome, with and without schizophrenia–relationship with COMT Val108/158 Met polymorphism, gender and symptomatology. Journal of Psychopharmacology, 25(7), 888-895.
35 Boot, E., Booij, J., Zinkstok, J. R., Baas, F., Swillen, A., Owen, M. J., ... & Van Amelsvoort, T. A. (2011). COMT Val158Met genotype and striatal D2/3 receptor binding in adults with 22q11 deletion syndrome. Synapse, 65(9), 967-970.
Breen, D. P., Vuono, R., Nawarathna, U., Fisher, K., Shneerson, J. M., Reddy, A. B., & Barker, R. A. (2014). Sleep and circadian rhythm regulation in early Parkinson disease. JAMA neurology, 71(5), 589-595.
Canavan, A. G. M., Dunn, G., & McMillan, T. M. (1986). Principal components of the WAIS–R. British Journal of Clinical Psychology, 25(2), 81-85.
Charmandari, E., Tsigos, C., & Chrousos, G. (2005). Endocrinology of the stress response 1. Annu. Rev. Physiol., 67, 259-284.
Chen, J., Lipska, B. K., Halim, N., Ma, Q. D., Matsumoto, M., Melhem, S., ... & Egan, M. F. (2004). Functional analysis of genetic variation in catechol-O-methyltransferase (COMT): effects on mRNA, protein, and enzyme activity in postmortem human brain. The American Journal of Human Genetics, 75(5), 807-821.
Collip, D., Nicolson, N. A., Lardinois, M., Lataster, T., Van Os, J., & Myin-Germeys, I. (2011). Daily cortisol, stress reactivity and psychotic experiences in individuals at above average genetic risk for psychosis. Psychological medicine, 41(11), 2305-2315.
Da Silva Alves, F., Bakker, G., Schmitz, N., Abeling, N., Hasler, G., Van der Meer, J., ... & Van Amelsvoort, T. (2013). Dopaminergic modulation of the reward system in schizophrenia: a placebo-controlled dopamine depletion fMRI study. European neuropsychopharmacology, 23(11), 1577-1586.
Da Silva Alves, F., Schmitz, N., Figee, M., Abeling, N., Hasler, G., Van der Meer, J., ... & Van Amelsvoort, T. (2011). Dopaminergic modulation of the human reward system: a placebo-controlled dopamine depletion fMRI study. Journal of psychopharmacology, 25(4), 538-549.
Dazzan, P. (2014). Neuroimaging biomarkers to predict treatment response in schizophrenia: the end of 30 years of solitude? Dialogues in clinical neuroscience, 16(4), 491.
De Girolamo, G., & McFarlane, A. C. (1996). The epidemiology of PTSD: a comprehensive review of the international literature.
de Koning, M. B., Boot, E., Bloemen, O. J., van Duin, E. D., Abel, K. M., de Haan, L., ... & van Amelsvoort, T. A. (2012). Startle reactivity and prepulse inhibition of the acoustic
36 startle response are modulated by catechol-O-methyl-transferase Val158 Met polymorphism in adults with 22q11 deletion syndrome. Journal of Psychopharmacology, 26(12), 1548-1560.
Díaz-Caneja, C. M., Pina-Camacho, L., Rodríguez-Quiroga, A., Fraguas, D., Parellada, M., & Arango, C. (2015). Predictors of outcome in early-onset psychosis: a systematic review. npj Schizophrenia, 1, 14005.
Emsley, R., Rabinowitz, J., & Torreman, M. (2003). The factor structure for the Positive and Negative Syndrome Scale (PANSS) in recent-onset psychosis. Schizophrenia research, 61(1), 47-57.
Faravelli, C., Sauro, C. L., Godini, L., Lelli, L., Benni, L., Pietrini, F., ... & Ricca, V. (2012). Childhood stressful events, HPA axis and anxiety disorders. World journal of psychiatry, 2(1), 13.
Faustman, W. O., & Overall, J. E. (1999). Brief Psychiatric Rating Scale.
Flores, B. H., Kenna, H., Keller, J., Solvason, H. B., & Schatzberg, A. F. (2006). Clinical and biological effects of mifepristone treatment for psychotic depression. Neuropsychopharmacology, 31(3), 628.
Fond, G., D’Albis, M. A., Jamain, S., Tamouza, R., Arango, C., Fleischhacker, W. W., ... & Meyer-Lindenberg, A. (2015). The promise of biological markers for treatment response in first-episode psychosis: a systematic review. Schizophrenia bulletin, sbv002.
Fries, E., Hesse, J., Hellhammer, J., & Hellhammer, D. H. (2005). A new view on hypocortisolism. Psychoneuroendocrinology, 30(10), 1010-1016.
Fung, W. L. A., Butcher, N. J., Costain, G., Andrade, D. M., Boot, E., Chow, E. W., ... & García-Miñaúr, S. (2015). Practical guidelines for managing adults with 22q11.2 deletion syndrome. Genetics in Medicine.
Gevonden, M. J., Myin-Germeys, I., Van den Brink, W., Van Os, J., Selten, J. P., & Booij, J. (2014). Psychotic reactions to daily life stress and dopamine function in people with severe hearing impairment. Psychol Med, 1-10.
Goldstein, D. S., Kopin, I. J., & Sharabi, Y. (2014). Catecholamine autotoxicity. Implications for pharmacology and therapeutics of Parkinson disease and related disorders. Pharmacology & therapeutics, 144(3), 268-282.
37 Gothelf, D., Law, A. J., Frisch, A., Chen, J., Zarchi, O., Michaelovsky, E., ... & Weizman, A. (2014). Biological effects of COMT haplotypes and psychosis risk in 22q11.2 deletion syndrome. Biological psychiatry, 75(5), 406-413.
Granger, D. A., Hibel, L. C., Fortunato, C. K., & Kapelewski, C. H. (2009). Medication effects on salivary cortisol: Tactics and strategy to minimize impact in behavioral and developmental science. Psychoneuroendocrinology, 34(10), 1437-1448.
Guna, A., Butcher, N. J., & Bassett, A. S. (2015). Comparative mapping of the 22q11.2 deletion region and the potential of simple model organisms. Journal of Neurodevelopmental Disorders, 7(1).
Heim, C., Newport, D. J., Heit, S., Graham, Y. P., Wilcox, M., Bonsall, R., ... & Nemeroff, C. B. (2000). Pituitary-adrenal and autonomic responses to stress in women after sexual and physical abuse in childhood. Jama, 284(5), 592-597.
Hernaus, D., Collip, D., Lataster, J., Ceccarini, J., Kenis, G., Booij, L., ... & Myin-Germeys, I. (2013). COMT Val 158 Met genotype selectively alters prefrontal [18F] fallypride displacement and subjective feelings of stress in response to a psychosocial stress challenge. PLoS One, 8(6), e65662.
Hernaus, D., Collip, D., Lataster, J., Viechtbauer, W., Myin, E., Ceccarini, J., ... & Myin-Germeys, I. (2015). Psychotic reactivity to daily life stress and the dopamine system: A study combining experience sampling and [18F] fallypride positron emission tomography. Journal of abnormal psychology, 124(1), 27.
Hibel, L. C., Granger, D. A., Kivlighan, K. T., Blair, C., & Family Life Project Investigators. (2006). Individual differences in salivary cortisol: Associations with common over-the-counter and prescription medication status in infants and their mothers. Hormones and Behavior, 50(2), 293-300.
Holtzman, C. W., Trotman, H. D., Goulding, S. M., Ryan, A. T., Macdonald, A. N., Shapiro, D. I., ... & Walker, E. F. (2013). Stress and neurodevelopmental processes in the emergence of psychosis. Neuroscience, 249, 172-191.
Hooper, S. R., Curtiss, K., Schoch, K., Keshavan, M. S., Allen, A., & Shashi, V. (2013). A longitudinal examination of the psychoeducational, neurocognitive, and psychiatric functioning in children with 22q11.2 deletion syndrome. Research in Developmental Disabilities, 34(5), 1758-1769.
38 Ising, H. K., Veling, W., Loewy, R. L., Rietveld, M. W., Rietdijk, J., Dragt, S., ... & van der Gaag, M. (2012). The validity of the 16-item version of the Prodromal Questionnaire (PQ-16) to screen for ultra high risk of developing psychosis in the general help-seeking population. Schizophrenia bulletin, 38(6), 1288-1296.
Jacobs, N., Myin-Germeys, I., Derom, C., Delespaul, P., Van Os, J., & Nicolson, N. A. (2007). A momentary assessment study of the relationship between affective and adrenocortical stress responses in daily life. Biological Psychology, 74(1), 60-66.
Jacobson, D., Bursch, M., & Lajiness-O'Neill, R. (2016). Potential Role of Cortisol in Social and Memory Impairments in Individuals with 22q11.2 Deletion Syndrome. Journal of Pediatric Genetics, 5(03), 150-157.
Jansen, L. M., Gispen-de Wied, C. C., Gademan, P. J., De Jonge, R. C., Van der Linden, J. A., & Kahn, R. S. (1998). Blunted cortisol response to a psychosocial stressor in schizophrenia. Schizophrenia research, 33(1), 87-94.
Janssen, I., Krabbendam, L., Bak, M., Hanssen, M., Vollebergh, W., Graaf, R. D., & Van Os, J. (2004). Childhood abuse as a risk factor for psychotic experiences. Acta Psychiatrica Scandinavica, 109(1), 38-45.
Jonas, R. K., Montojo, C. A., & Bearden, C. E. (2014). The 22q11.2 deletion syndrome as a window into complex neuropsychiatric disorders over the lifespan. Biological psychiatry, 75(5), 351-360.
Justice, N. J., Huang, L., Tian, J. B., Cole, A., Pruski, M., Hunt, A. J., ... & Zheng, H. (2015). Posttraumatic stress disorder-like induction elevates β-amyloid levels, which directly activates corticotropin-releasing factor neurons to exacerbate stress responses. Journal of Neuroscience, 35(6), 2612-2623.
Kay, S. R., Fiszbein, A., & Opfer, L. A. (1987). The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophrenia bulletin, 13(2), 261.
Lardinois, M., Lataster, T., Mengelers, R., Van Os, J., & Myin-Germeys, I. (2011). Childhood trauma and increased stress sensitivity in psychosis. Acta Psychiatrica Scandinavica, 123(1), 28-35.
Lardinois, M., Nicolson, N., Lataster, T., Sulon, J., Mengelers, R., Van Os, J., & Myin-Germeys, I. (2009). Stress and cortisol in subjects at risk for psychosis: an experience sampling study. Why stress causes psychosis and psychosis causes stress.
