A zebrafish model of glucocorticoid resistance shows serotonergic modulation of the stress response

Abstract

One function of glucocorticoids is to restore homeostasis after an acute stress response by providing negative feedback to stress circuits in the brain. Loss of this negative feedback leads to elevated physiological stress and may contribute to depression, anxiety, and post-traumatic stress disorder. We investigated the early, developmental effects of glucocorticoid signaling deficits on stress physiology and related behaviors using a mutant zebrafish, gr(s357), with non-functional glucocorticoid receptors (GRs). These mutants are morphologically inconspicuous and adult-viable. A previous study of adult gr(s357) mutants showed loss of glucocorticoid-mediated negative feedback and elevated physiological and behavioral stress markers. Already at 5 days post-fertilization, mutant larvae had elevated whole body cortisol, increased expression of pro-opiomelanocortin (POMC), the precursor of adrenocorticotropic hormone (ACTH), and failed to show normal suppression of stress markers after dexamethasone treatment. Mutant larvae had larger auditory-evoked startle responses compared to wildtype sibling controls (gr(wt)), despite having lower spontaneous activity levels. Fluoxetine (Prozac) treatment in mutants decreased startle responding and increased spontaneous activity, making them behaviorally similar to wildtype. This result mirrors known effects of selective serotonin reuptake inhibitors (SSRIs) in modifying glucocorticoid signaling and alleviating stress disorders in human patients. Our results suggest that larval gr(s357) zebrafish can be used to study behavioral, physiological, and molecular aspects of stress disorders. Most importantly, interactions between glucocorticoid and serotonin signaling appear to be highly conserved among vertebrates, suggesting deep homologies at the neural circuit level and opening up new avenues for research into psychiatric conditions.

Keywords: SSRI; anxiety; depression; glucocorticoid receptors; stress; zebrafish.

Figure 1

A simplified schematic of the vertebrate HPA stress pathway in normal and gr^s357 mutant zebrafish in which the glucocorticoid receptor is non-functional. Lines with arrowheads indicate feed forward excitation and lines with blocked heads indicate feedback inhibition. Regular font signifies neural structures and italicized font signifies diffusible signaling molecules. Glucocorticoid feedback from the periphery to the brain normally inhibits further brain stress activity by transcriptional suppression of genes including those encoding CRH and ACTH (left, solid lines). In gr^s357 mutant zebrafish, the glucocorticoid receptor is non-functional, preventing negative feedback (right, dashed line), and leading to prolonged release and higher levels of stress intermediaries (bold font). Abbreviations: CRH, Corticotropin releasing hormone; ACTH, Adrenocorticotropic hormone; HPA, Hypothalamic-pituitary-adrenal axis; PVN, Paraventricular nucleus.

Figure 2

Representative photographs of a VBA+ (left) and VBA− (right) larva from the same clutch after exposure to the VBA stimulus on day 5 post-fertilization. The dark fish were reproducibly genotyped and shown to be homozygous gr^s357 mutants. Non-dark siblings were either homozygous wildtype or heterozygotes. Scale bar = 1 mm.

Figure 3

Whole body cortisol measurements in 7 day old zebrafish larvae. Cortisol measurements in 7 day old gr^s357 mutants and wildtype siblings with and without fluoxetine treatment. Each group consisted of 24 larvae processed together, split into two duplicate samples, and tested concurrently. Values shown are transformed from ELISA absorbance measurements. Error bars are SEMs of the two duplicate samples per group.

Figure 4

RNA in situ hybridization for pomc transcript in 5 day old larvae. Results are presented from a 2 × 2 experimental design testing the effects of genotype (rows) and betamethasone 17-valerate treatment (columns) on pomc expression. Two representative larvae per group are shown (N = 5 per group). pomc reactivity is localized to the pituitary region in all groups, but in gr^s357 larvae expression appeared stronger and showed no suppression after treatment with betamethasone 17-valerate.

Figure 5

Spontaneous activity in gr^s357 and wildtype larvae. (A) Average hourly activity (±standard error) across a 24 h spontaneous activity recording (N = 48 per group). Both groups showed a circadian activity rhythm with a peak at ZT13, but wildtype larvae had a higher overall activity level. (B) Activity as a function of subjective light cycle phase. Both groups had significantly higher activity during the subjective light phase. (C) Fluoxetine effects on spontaneous swimming activity (N = 24 per group). Fluoxetine increased activity in gr^s357 but had no effect in the wildtype group. ^*P < 0.5.

Figure 6

Startle responses in wildtype (N = 17), untreated gr^s357 (N = 17), and fluoxetine treated gr^s357 larvae (N = 14). (A) A comparison of startle magnitude (±SEM) in untreated wildtype and gr^s357 larvae. Mutants had significantly larger responses than wildtype across all trials. Habituation was evident both within blocks and across blocks. (B) A comparison of startle magnitude in mutant larvae with and without fluoxetine treatment. Fluoxetine resulted in significantly lower startle responses in mutants, but the groups did not differ on measures of habituation. (C) Habituation within block shown as the slope of the regression line through the 5 trails on each block, and as an average of all 25 startle trials. Mutants had steeper habituation slopes (more habituation) in 4 out of 5 blocks, and across all 25 trials, reflecting the larger startle responses of mutants on the earliest trials of each block. This graph illustrates a trend in the data, but effect of genotype on habituation was not statistically significant. (D) Percent of group responding across all 25 trials. Untreated mutants responded significantly more frequently than wildtype groups or fluoxetine treated mutants (^*P < 0.05).