Norman Cousins Lecture. Mechanisms of cytokine-induced behavioral changes: psychoneuroimmunology at the translational interface

Work in our laboratory has focused on the mechanisms by which cytokines can influence the brain and behavior in humans and non-human primates. Using administration of interferon (IFN)-alpha as a tool to unravel these mechanisms, we have expanded upon findings from the basic science literature implicating cytokine-induced changes in monoamine metabolism as a primary pathway to depression. More specifically, a role for serotonin metabolism has been supported by the clinical efficacy of serotonin reuptake inhibitors in blocking the development of IFN-alpha-induced depression, and the capacity of IFN-alpha to activate metabolic enzymes (indolamine 2,3 dioxygenase) and cytokine signaling pathways (p38 mitogen activated protein kinase) that can influence the synthesis and reuptake of serotonin. Our data also support a role for dopamine depletion as reflected by IFN-alpha-induced changes in behavior (psychomotor slowing and fatigue) and regional brain activity, which implicate the involvement of the basal ganglia, as well as the association of IFN-alpha-induced depressive-like behavior in rhesus monkeys with decreased cerebrospinal fluid concentrations of the dopamine metabolite, homovanillic acid. Neuroimaging data in IFN-alpha-treated patients also suggest that activation of neural circuits (dorsal anterior cingulate cortex) associated with anxiety and alarm may contribute to cytokine-induced behavioral changes. Taken together, these effects of cytokines on the brain and behavior appear to subserve competing evolutionary survival priorities that promote reduced activity to allow healing, and hypervigilance to protect against future attack. Depending on the relative balance between these behavioral accoutrements of an activated innate immune response, clinical presentations may be distinct and warrant individualized therapeutic approaches.     

Interferon-alpha effects on diurnal hypothalamic-pituitary-adrenal axis activity: relationship with proinflammatory cytokines and behavior

Interferon (IFN)-alpha has been used to investigate pathways by which innate immune cytokines influence the brain and behavior. Accordingly, the impact of IFN-alpha on diurnal secretion of hypothalamic-pituitary-adrenal (HPA) axis hormones was assessed in 33 patients eligible for treatment with IFN-alpha plus ribavirin for hepatitis C. In addition, the relationship between IFN-alpha-induced HPA axis changes and proinflammatory cytokines and behavior was examined. Plasma ACTH and cortisol as well as tumor necrosis factor (TNF)-alpha, interleukin-6 and their soluble receptors, were measured hourly between 0900 and 2100 hours at baseline and following approximately 12 weeks of either no treatment (n=13) or treatment with IFN-alpha/ribavirin (n=20). Plasma IFN-alpha was also measured at each visit. Depression and fatigue were assessed using the Montgomery-Asberg depression rating scale and the multidimensional fatigue inventory. Compared to no treatment, IFN-alpha/ribavirin administration was associated with significant flattening of the diurnal ACTH and cortisol slope and increased evening plasma ACTH and cortisol concentrations. Flattening of the cortisol slope and increases in evening cortisol were correlated with increases in depression (r=0.38, P<0.05 and r=0.36, P<0.05, respectively) and fatigue (r=0.43, P<0.05 and r=0.49, P<0.01, respectively). No relationship was found between immune and HPA axis measures, although increases in plasma IFN-alpha, TNF-alpha and soluble TNF-alpha receptor2 were independently correlated with behavioral endpoints. These data indicate that chronic exposure to innate immune cytokines may contribute to the altered diurnal HPA axis activity and behavior found in medically ill individuals. However, given the lack of correlation between HPA axis and immune measures, the mechanism by which chronic cytokine exposure influences HPA axis function remains to be determined.     

ACTH in plasma and CSF in the rhesus monkey

Adrenocorticotrophic hormone (ACTH) and other biosynthetically related peptides are found both in the brain and peripherally, but the function and regulation of these substances differ in the brain and in the periphery. It has been suggested that measurement of peptide hormones in the cerebrospinal fluid (CSF) might provide information relevant to the diagnosis and pathophysiology of neurological and psychiatric illnesses. We report experiments using a rhesus monkey model to evaluate parameters affecting CSF ACTH concentrations. We found that CSF ACTH concentrations follow a diurnal rhythm that is markedly different from that in plasma, concentrations of ACTH in monkey CSF, but not in plasma, increased significantly after 4 days of social separation, and CSF ACTH concentrations did not change after hypophysectomy. These results suggest that CSF ACTH concentrations reflect the activity of brain and not peripheral ACTH systems.     

Cerebrospinal fluid measures of neurotransmitter changes associated with pharmacological alteration of the despair response to social separation in rhesus monkeys

Social separation is a risk factor for major depressions that can be modeled in nonhuman primates. Changes in central monoamine neurotransmission are also likely to be involved in major depression. This study examined the relationship between separation-induced depressive-like behavior and central monoamine neurotransmitter changes in rhesus monkeys. Measures of cerebrospinal fluid (CSF) norepinephrine (NE), 5-hydroxyindoleacetic acid (5HIAA), and homovanillic acid (HVA) were used to assess the neurobiological impact of social separation and drug treatments alone or in combination. alpha-Methyl-p-tyrosine exacerbated, and fusaric acid ameliorated, the depressive-like response to separation. Probenecid-induced accumulations of HVA and 5HIAA reflected changes in behavior, but were not consistently affected by drug treatment. In contrast, pretreatment CSF NE was comparatively stable across repeated sampling, and drug-induced changes in this measure were correlated with changes in behavior. Low CSF NE, whether drug-induced or naturally occurring, was associated with a more severe depressive-like response to separation.     

Effects of reserpine on the social behavior of rhesus monkeys

Reserpine was administered daily by intubation for 81 days to 3 rhesus monkeys to investigate a possible relationship with depression. Their behavior during the experimental period was compared to their behavior before and after the drug period, as well as to that of a control group of 3 monkeys given water instead of reserpine. Experimental findings were as follows: 1) Reserpine caused significant behavioral changes in the rhesus monkey. These changes included decreases in visual exploration and locomotion, and increases in self-huddling, posturing, and tremor; and 2) the behavioral effects of repeated daily dosage were not cumulative nor was tolerance developed by the drugged monkeys. There was no visible effect 15 hours after each daily drug administration. The behavioral alterations were not detectable more than 15 hours following each daily dose. It was noted that the data were insufficient to justify the conclusion that reserpine administration produced depression in the monkey subjects, even though it did have marked behavioral effects on them.     

Long-term effects of early social isolation in Macaca mulatta: changes in dopamine receptor function following apomorphine challenge.

The hypothesis that early social isolation results in long-term alterations in dopamine receptor sensitivity was tested using older adult rhesus monkeys. Isolated and control monkeys were challenged with apomorphine (0.1 and 0.3 mg/kg), and the drug effects on spontaneous blink rate, stereotyped behavior, and self-injurious behavior were quantified using observational measures. Monoamine metabolites were quantified from cisternal CSF by HPLC-EC, prior to pharmacological challenge. Isolated and control monkeys did not differ in CSF concentrations of HVA, 5-HIAA, or MHPG. At the higher dose, apomorphine significantly increased the rate of blinking, the occurrence of whole-body stereotypies, and the intensity of stereotyped behavior (as measured by observer ratings) in isolated monkeys. The frequency of occurrence of self-injurious behavior was too low to allow for meaningful comparisons. These significant differences in response to apomorphine challenge support the hypothesis that long-term or permanent alterations in dopamine receptor sensitivity, as assessed by drug challenge, are a consequence of early social deprivation.     

Role of the primate orbitofrontal cortex in mediating anxious temperament.

BACKGROUND: Excessive behavioral inhibition during childhood marks anxious temperament and is a risk factor for the development of anxiety and affective disorders. Studies in nonhuman primates can provide important information related to the expression of this risk factor, since threat-induced freezing in rhesus monkeys is a trait-like characteristic analogous to human behavioral inhibition. The orbitofrontal cortex (OFC) and amygdala are part of a circuit involved in the processing of emotions and associated physiological responses. Earlier work demonstrated involvement of the primate central nucleus of the amygdala (CeA) in mediating anxious temperament. This study assessed the role of the primate OFC in mediating anxious temperament and its involvement in fear responses. METHODS: Twelve adolescent rhesus monkeys were studied (six lesion and six control monkeys). Lesions were targeted at regions of the OFC that are most interconnected with the amygdala. Behavior and physiological parameters were assessed before and after the lesions. RESULTS: The OFC lesions significantly decreased threat-induced freezing and marginally decreased fearful responses to a snake. The lesions also resulted in a leftward shift in frontal brain electrical activity consistent with a reduction in anxiety. The lesions did not significantly decrease hypothalamic-pituitary-adrenal (HPA) activity or cerebrospinal fluid (CSF) concentrations of corticotrophin-releasing factor (CRF). CONCLUSIONS: These findings demonstrate a role for the OFC in mediating anxious temperament and fear-related responses in adolescent primates. Because of the similarities between rhesus monkey threat-induced freezing and childhood behavioral inhibition, these findings are relevant to understanding mechanisms underlying anxious temperament in humans.     

Recombinant human interferon-alpha does not alter reward behavior, or neuroimmune and neuroendocrine activation in rats

Recombinant human interferon-alpha (IFN-alpha) induces depression, and neuroendocrine and neuroimmune activation, in a significant number of patients undergoing treatment for viral illnesses (e.g., hepatitis C), yet these effects have not been consistently reproduced in rodents. As such, we sought to determine the effects of acute or chronic IFN-alpha treatment on basic reward and immobility in the forced swim test (FST), neuroendocrine and neuroimmune activation, and monoamine turnover in brain. In the first experiment, male Wistar rats (N = 7/group) treated with human recombinant IFN-alpha (100,000 IU/kg, i.p.), as compared to saline, did not exhibit alterations to rate of sucrose pellet self-administration or total reinforcers obtained, corticosterone release, plasma IL-6 release, IL-1beta or IL-6 mRNA expression in hippocampus, or monoamine turnover in prefrontal cortex, striatum, nucleus accumbens, or amygdala. However, acute IFN-alpha decreased body weight and produced a trend toward reduced food consumption in the home cage 2 h after injection. In the second experiment, Wistar rats (N=4/group) were subjected to a chronic treatment regimen of saline or IFN-alpha (100,000 IU/kg, i.p.) once daily for 14 consecutive days. The data reveal that animals exposed to chronic IFN-alpha exhibited similar amounts of time immobile and similar latencies to primary immobility in the FST as compared to saline-treated controls. Chronic IFN-alpha did not induce corticosterone release, plasma TNF-alpha, or IL-6 release. Tissue monoamine analysis revealed that chronic IFN-alpha reduced DA levels in prefrontal cortex, and decreased 5-HT levels and increased 5-HT turnover in amygdala. In the third experiment, Wistar rats (N = 4/group) were exposed to either acute or chronic pegylated IFN-alpha (pegIFN-alpha: 3.25, 10 or 75 mg/kg, i.p.) at one of several time points from 1 h to 23 days. The data reveal that neither acute nor chronic pegIFN-alpha induced corticosterone release. Overall, the current report demonstrates that neither acute nor chronic IFN-alpha induced depressive-like behavior and neither IFN-alpha nor peg-IFN-alpha was capable of inducing neuroendocrine or neuroimmune activation. Despite the neurochemical alterations observed in the chronic treatment regimen, the data indicate that recombinant human IFN-alpha does not produce a robust model of depressive-like behavior in rodents.     

Effects of beta-carboline on fear-related behavioral and neurohormonal responses in infant rhesus monkeys.

We examined the effects of the inverse benzodiazepine agonist ethyl-beta-carboline-3-carboxylate (beta-CCE) on behavioral, hormonal, and neurochemical responses in infant rhesus monkeys exposed to fearful situations. Our paradigm elicits three distinct adaptive patterns of defensive behavior. From previous work, we hypothesized that behaviors induced by attachment bond disruption are predominantly mediated by opiate systems, whereas behaviors induced by the threat of attack are mediated by benzodiazepine systems. When beta-CCE (0, 125, 250, and 500 micrograms/kg) was administered immediately after maternal separation, the 500 micrograms/kg dose increased freezing and the 250 and 500 micrograms/kg doses reduced environmental exploration. Test conditions produced increased plasma ACTH and cortisol concentrations and increased cerebrospinal fluid (CSF) concentrations of MHPG and DOPAC; beta-CCE did not further affect these metabolites. A dose of 1000 micrograms/kg of beta-CCE increased CSF concentrations of DOPAC and MHPG in infants left with their mothers. During test conditions, it further increased CSF MHPG (but not DOPAC) concentrations, and reduced cooing while increasing freezing and barking and other hostile behaviors. Our results thus confirm that benzodiazepine systems mediate threat-related behaviors and suggest that coos, which were thought to predominantly reflect the degree of distress during separation, can be modulated by the infant's level of fear. beta-CCE also activated stress-related pituitary-adrenal hormonal systems and brain norepinephrine (NE) and dopamine (DA) systems. These effects occurred when animals remained undisturbed in their home cages with their mothers, suggesting that benzodiazepine receptors directly modulate brain NE and DA systems.     

Cerebrospinal fluid corticotropin-releasing hormone levels are elevated in monkeys with patterns of brain activity associated with fearful temperament.

BACKGROUND: Asymmetric patterns of frontal brain activity and brain corticotropin-releasing hormone (CRH) systems have both been separately implicated in the processing of normal and abnormal emotional responses. Previous studies in rhesus monkeys demonstrated that individuals with extreme right frontal asymmetric brain electrical activity have high levels of trait-like fearful behavior and increased plasma cortisol concentrations. METHODS: In this study we assessed cerebrospinal fluid (CSF) CRH concentrations in monkeys with extreme left and extreme right frontal brain electrical activity. CSF was repeatedly collected at 4, 8, 14, 40, and 52 months of age. RESULTS: Monkeys with extreme right frontal brain activity had increased CSF CRH concentrations at all ages measured. In addition, individual differences in CSF CRH concentrations were stable from 4 to 52 months of age. CONCLUSIONS: These findings suggest that, in primates, the fearful endophenotype is characterized by increased fearful behavior, a specific pattern of frontal electrical activity, increased pituitary-adrenal activity, and increased activity of brain CRH systems. Data from other preclinical studies suggests that the increased brain CRH activity may underlie the behavioral and physiological characteristics of fearful endophenotype.