Effects of Tryptophan Depletion on Brain Processing of Emotions in Patients With Mood Disorders

Major depressive disorder (MDD) has been associated with abnormally reduced function of central serotonergic systems by various types of evidence. One instructive paradigm for investigating the relationship between serotonergic function, cognition and depression has involved the mood response to tryptophan depletion (TD), achieved by oral loading with all essential amino acids excepting the 5-HT precursor, tryptophan. Subjects who are depressed show altered behavioral and neural response to affectively valenced stimuli, similar to those seen under TD.

More, specifically, increased response latencies to happy words have been noted in both depressed patients and healthy controls following TD using the Affective Go/No go (AGNG) test, in which subjects are required either to respond or inhibit a response to a series of emotionally valenced words.

The first round of our study employed fMRI imaging of BOLD response and arterial spin labeling to investigate response to affectively valenced words under TD in two groups, remitted MDD (rMDD) patients and healthy controls with no family history of depression. We have thus far examined the effects of TD depletion in our healthy control group. We found that emotional information processing was indeed altered by 5-HT depletion.

A behavioral bias toward positive stimuli was attenuated following depletion, which was accompanied by increased hemodynamic responses during the processing of emotional words in several subcortical structures, including the ventral striatum, hippocampal cortex, anterior insula, superior temporal gyrus and posterior cingulate. Further, inter-individual differences in tryptophan depletion-elicited anxiety correlated positively with caudate bias toward negative stimuli. On the basis of these data, we suggest that the bias towards positive stimuli might represent a normal suppression of attention to negative stimuli; a process which is abrogated in conditions, such as depression, which are characterized by a serotonergic dysfunction. In the second round of this study, we are examining whether such biases also occur beneath the level of conscious awareness.

One of the weaknesses of these data is that the patients were consciously aware of the verbal stimuli that were presented to them and likely possess learned, individualized associations to the verbal stimuli in question. We have developed a conditioning task which obviates this problem through the presentation of emotionally-valenced visual stimuli below the threshold of conscious awareness.

Since embarking on the first two phases of this study, a new paradigm has been developed which may provide further insight into affective behavioral biases. The Split Reversals task is a pavlovian reversal-learning task which allows the assessment of responses to expected and unexpected punishment and reward cues. Previous research with this task has demonstrated that reduction of serotonin in healthy individuals removes a disproportionate inability to predict punishment at baseline. In other words, serotonin reduction removes a positive bias on this task. . This is consistent with the observed removal of a positive bias on the AGNG following TD. However, on the split reversals task this positive bias is revealed by an inability to process negative stimuli at baseline, which is improved, rather than a bias towards positive stimuli which is diminished, by TD. As such, the new task provides a different direction from which to consider the emotional biases observed in affective disorders. We have suggested that the bias on the new task reflects a serotonin-mediated suppression of responses to negative stimuli in healthy individuals which, in turn, promote resilience to affective disorders. Serotonin reduction via TD hence removes this bias. That this task may therefore be sensitive to cognitive biases resulting from resilience mechanisms.

Similarly, fear potentiated startle which has been shown to be sensitive to the symptoms of anxiety has been shown to be reduced by the elevation of 5-HT via two week citalopram treatment. This is consistent with reduced processing of negative information following 5-HT increase (Grillon et al 2009). We predict that reducing 5-HT via TD will have the opposite effect and increase the duration of fear potentiated startle processing. This would provide an indirect link between reduced 5-HT and anxiety disorders. As such, we include this task in a new phase of this protocol.

In addition these behavioural TD findings, this task has been neuroimaged (fMRI) in healthy individuals. As predicted, the striatum was shown to be involved in reward processing, whilst the amygdala was shown to be involved in punishment processing. A subproject of this study will therefore attempt to extend the prior findings from the split reversals task (and reconcile them with the AGNG findings) by examining healthy individuals completing the task whilst undergoing both serotonin manipulation (TD) and fMRI. This subproject will be initially restricted to healthy individuals and should provide us with further insight into the neural substrates involved in serotonergic inhibition of aversive processing highlighted by the first phase of this study. This will also provide a basis from which to consider potential future study with the same task in MDD subjects undergoing TD.

There may well be a genetic basis to these phenomena. It is known that the functional variant in the promoter region of the serotonin transporter gene (5-HTTLPR) is partly responsible for determining how individuals respond to TD. In fact, recently reported that carriers of the short 5-HTTLPR allele showed impaired recognition of fearful faces following TD.