Embodied theories of emotion posit that the processing of emotional stimuli re-enacts exteroceptive, interoceptive, and prior affective states congruent with the emotion. Activating an exteroceptive, interoceptive, or prior affective state congruent with a particular emotion should consequently anticipate perception of that emotion. Previous research found that an increased heart rate facilitates the processing of faces expressing the emotion that interoception is congruent with (fear), in comparison to those expressing an incongruent emotion (disgust) or neutrality. Here, we investigate whether the (top-down) prior affective context in which the emotion simulation is situated also interacts with (bottom-up) interoceptive heart rate manipulation to facilitate emotional processing of exteroceptive stimuli. Participants categorized faces with happy, fearful, and neutral expressions after their interoceptive state (heart rate: increased or normal) and contextual affective state (primes: positive, negative, and neutral) were manipulated. We hypothesized that increased heart rate would facilitate the processing of fearful and happy faces, specifically for congruent affective priming, i.e. negative or positive, respectively. We found that: a) negative affective primes facilitated the processing of fearful faces, over and above the facilitating effect of increased heart rate reported in previous research, and b) no effect of either top-down or bottom-up factors, nor their interaction, in the domain of positive emotions. It remains to be established whether this asymmetry depends on the specificity of negative emotions, the quality of our manipulations, or other factors.