Effect of the matrix metalloproteinase inhibitor doxycycline on human trace fear memory

Abstract

Learning to predict threat is of adaptive importance, but aversive memory can also become disadvantageous and burdensome in clinical conditions such as post-traumatic stress disorder. Pavlovian fear conditioning is a laboratory model of aversive memory and thought to rely on structural synaptic reconfiguration involving matrix metalloproteinase (MMP) 9 signalling. It has recently been suggested that the MMP9-inhibiting antibiotic doxycycline, applied before acquisition training in humans, reduces fear memory retention after one week. This previous study used cued delay fear conditioning, in which predictors and outcomes overlap in time. However, temporal separation of predictors and outcomes is common in clinical conditions. Learning the association of temporally separated events requires a partly different neural circuitry, for which the role of MMP9 signalling is not yet known. Here, we investigate the impact of doxycycline on long-interval (15 s) trace fear conditioning in a randomised controlled trial with 101 (50 females) human participants. We find no impact of the drug in our pre-registered analyses. Exploratory post-hoc analyses of memory retention suggested a serum level-dependent effect of doxycycline on trace fear memory retention. However, effect size to distinguish CS+/CS- in the placebo group turned out to be smaller than in previously used delay fear conditioning protocols, which limits the power of statistical tests. Our results suggest that doxycycline effect on trace fear conditioning in healthy individuals is smaller and less robust than anticipated, potentially limiting its clinical application potential.

Significance statement

The inhibition of matrix metalloproteinase-9 attenuates memory consolidation and subsequent recall in a delay cue conditioning paradigm. However, it is currently unclear whether this is also the case for other learning scenarios that rely on a different neurocircuitry. We test this hypothesis in human trace fear conditioning which employs remote cues and is additionally dependent on hippocampus involvement. We find that doxycycline does not reduce fear retention in our pre-registered analyses. Exploratory analyse might potentially suggest a weak effect of doxycycline on trace fear memory retention.