The widely-used anti-viral drug interferon-alpha induces depressive- and anxiogenic-like effects in healthy rats

Abstract

Interferon-alpha (IFN-α) is a naturally occurring human cytokine that is a key therapy in the treatment of several viral diseases and cancers. However, treatment can produce significant neuropsychiatric and neurotoxic adverse events, including depression and anxiety. Here we investigated the effects of a clinically-comparable treatment regime of IFN-α on depressive- and anxiety-like behaviour in rats; and also examined frontal-cortical and hippocampal BDNF levels. Rats treated with IFN-α for four weeks showed significant increases in depressive- and anxiety-like behaviour. Further experimental investigation revealed that hedonic dysregulation (stronger preference for a sweet solution) did not emerge until the second week of treatment, and became more persistent as treatment progressed. No significant IFN-α-induced changes in BDNF levels were found. These results indicate that the affective deficits seen in patients may be modelled in healthy animals. This model may represent a novel tool to investigate the extent of and mechanisms underlying the IFN-α psychiatric syndrome.

Introduction

Isaacs and Lindenmann in 1957 were the first to observe that virally-infected cells could develop a protein substance, interferon (IFN), that could provide protection against further viral infection [22]. While long-term administration of IFN-α is now a mainstay of treatment for a variety of viral diseases, cancers and other disorders, therapy is frequently hindered by the precipitation of debilitating physical and mental effects (for a comprehensive list see [49]). As soon as six to eight hours after initial IFN-α administration somatic effects (such as flu-like symptoms) can emerge [12], but usually retreat within two to three weeks and do not generally necessitate treatment cessation. However, disturbing CNS phenomena become apparent later on in the course of treatment and continued IFN-α therapy can have substantial problems because of the emergence of such adverse neuropsychiatric events.

During the investigation of cancer-associated asthenia, Adams et al., [1] were among the first to document the psychiatric effects of IFN-α treatment. Vast evidence has accumulated in the intervening years detailing a wide range of CNS disturbances occurring with IFN-α therapy. These include depression (as discussed below) with suicidal ideation [16], impaired concentration [30], seizures [16], memory impairments [39], personality change [39], emotional liability [45], return of pre-existing or development of new phobias or compulsive behaviours [36], psychoses [16], delusions [55], aggression [23], anxiety [2], [32], insomnia [18], [30], and motor impairments [39]. CNS effects have even been reported to persist up to one year after IFN-α discontinuation [19], [37].

IFN-α-associated depressive effects consistently emerge as the most commonly documented psychiatric adverse events and can lead to discontinuation of treatment and consequential negative impact on health and social gain. The literature has been repetitively reviewed but briefly, the incidence of depression has been reported as being anything from 3.4 to 40.7% [21], [40], [42], [50]. These issues highlight the need for comprehensive study of the drug in a healthy population, and the situation is thus one where animal models are potentially important and useful tools in elucidating the neurobiological mechanisms of action of IFN-α.

Early studies showed that human IFN-α can have significant physiological effects in rodents (for example, [4], [5]), and while several subsequent studies have documented additional effects of human recombinant IFN-α in rodents, there are noteworthy problems within the literature: small sample size [31], exceptionally high doses of the drug [14], [54], possible confounding factors (for example, stress arising from single housing [47] or prolonged water deprivation [47], [48]), routes of administration not usually used in humans [25], [33], [34], [56], the use of inbred rat strains [31], observation of physiological rather than psychological effects [14], [27], and acute or (what may be) sub-chronic dosing regimes [11], [33], [34], [48].

We therefore set about developing a clinically-relevant model of the IFN-α psychiatric syndrome in the laboratory. We tested the effects of four weeks systemic treatment with IFN-α on depressive- and anxiety-like behaviour in normal, healthy male Wistar rats using well-defined behavioural tests. We hypothesized that this long-term treatment with IFN-α would produce an affective syndrome that mimics the human phenomenon.

Firstly, we investigated possible effects of IFN-α on depressive- and anxiety-like behaviour in the rat. Depression was assessed in terms of two parameters, anhedonia and despair. Anhedonia is the loss of interest in pleasure, and the DSM-IV lists this as a requirement criterion in diagnosing major depressive disorder. In experimental terms, anhedonia is measured as a decrease in sensitivity to reward, or an impaired capacity to respond to rewarding stimuli. Despite its lack of calorific value, saccharin is said to be 300 times sweeter than sugar, and is considered to be a pleasurable or “hedonic” signal to rats. Normal rats naturally show a significant preference for a saccharin solution over water (saccharin preference test, SPT). It is thought that anhedonic rats do not show preferences for sweet solutions because of a lack of ability to sense reward, or decreased sensitivity to reward.

The forced swim test (FST) is believed to assess behavioural “despair”. When rodents are forced to swim in an environment from which there is no escape, they will vigorously attempt escape but soon will readily adopt a characteristic immobile posture, making only movements necessary to keep their heads above water (“float time”). Float time in the FST is significantly decreased by anti-depressant drugs, and this test has been cited as the most widely-used screen for such compounds [10]. In this instance, however, we investigated if the converse would hold true, that is, that a putative depressive intervention (IFN-α treatment) would increase float time in the test.

Using the elevated plus maze apparatus (EPM), the anxiogenic effects of IFN-α were investigated. This maze test relies on the innate conflict between exploration of a novel area and avoidance of its aversive features that is experienced by rodents. While a number of anxiety-related behaviours can be measured during EPM testing, the most reliable measures of anxiety-related behaviour are both time spent in open arms, and the number of entries made into open arms, since these parameters are sensitive to both anxiolytic and anxiogenic drug treatment [17]. The number of closed arm entries is considered to be the best measure of locomotor activity.

Having considered the induction of a putative affective syndrome in IFN-α-treated animals in Experiment 1, we then sought to investigate the time-course of the emergence and persistence of such effects. The aim of the second experiment (Experiment 2) was two-fold: our primary goal was to assess at what point during treatment depressive-like effects emerged; we also investigated the impact of IFN-α treatment on the expression of brain-derived neurotrophic factor (BDNF: see below).

In patients, it is believed that the IFN-α syndrome is two-phase, with somatic symptoms lasting for the first few weeks of treatment and psychiatric effects presenting from about six to eight weeks into treatment. If this is also the case in rodents, with affective symptoms taking a number of weeks to emerge, it may go towards explaining why some groups have failed to find behavioural effects of IFN-α. The SPT was thus employed since repeated administration of the test does not produce practice effects. Preference for the sweet solution was recorded prior to treatment, and then monitored throughout treatment, allowing us to build a week-by-the week profile of emergent hedonic effects. Given the patient pattern, we predicted that IFN-α treatment may not have immediate effects, but that animals would lose preference for the saccharin solution as weeks progressed. Based on the results of Experiment 1, we hypothesized that IFN-α animals should show no preference by week four of treatment.

The neurotrophins are potent regulators of neuronal survival and function during development and recent advances also have pointed to a possible role for brain-derived neurotrophic factor (BDNF) in the pathophysiology of affective disorders. Decreases in levels of BDNF mRNA and its specific receptor, receptor tyrosine kinase B (TrkB) in post-mortem brain tissue (notably in the prefrontal cortex and hippocampus) and decreases in BDNF in serum of suicide subjects have been reported [15], [26]. Selective serotonin reuptake inhibitors (SSRI's) have been used with good success in the treatment [21] and prevention [29], [51] of IFN-α-induced depression in patients, and changes in BDNF expression that are effected by these drugs were proposed some time ago to be involved in their therapeutic action [13]. Depressive-like effects during IFN-α treatment could therefore be associated with alterations in BDNF expression. To investigate this molecular candidate for affective dysregulation in IFN-α-treated animals, we also tested the effects of our IFN-α regime on frontal cortical and hippocampal levels of BDNF.