Interesting. Digging through literature, I find these abstracts:

http://www.sciencedirect.com/science/article/pii/S0018506X03002368

An alarm pheromone released from stressed conspecifics evokes behavioral and autonomic responses in rats. We have previously reported that male Wistar rats show behavioral changes including increased sniffing, walking and rearing, and decreased resting as well as exaggerated response of body temperature to a novel environment [known as stress-induced hyperthermia (SIH)] when they are exposed to an alarm pheromone released from other male rats receiving foot shocks. The purpose of the present study was to examine the role of testosterone in the production and release of the alarm pheromone using these behavioral and autonomic responses in recipient rats. Three groups of alarm pheromone donors were presented, namely, intact males, castrated males, and testosterone-implanted castrated males. The effects of the alarm pheromone on the autonomic responses did not differ among the three groups, regardless of the donor's steroidal milieu, whereas behavioral responses were altered by castrating the donor males and the effects were restored by testosterone implantation. These results suggest that the alarm pheromone released from stressed male rats can be classified into at least two categories according to the androgen dependency of their production and/or release.

http://www.sciencedirect.com/science/article/pii/0031938488902879

Exposure to bedding taken from the soiled home cage of an isolated male resident elicited a significant increase in the nociceptive responses of male deer mice, Peromyscus maniculatus artemisiae, from mixed sex pairs. The analgesia induced by exposure to the male scent was insensitive to the opiate antagonist, naloxone, and was blocked by either pre- or post-olfactory exposure injections of the benzodiazepine antagonist, Ro15-1788, or agonist, diazepam. This non-opioid analgesia was of brief duration (15–30 min) and rapid onset, being evident after 1 min of exposure to the olfactory cues. Bedding treated with the novel odor of peppermint also induced analgesia in the deer mice. This analgesia was opioid mediated, being blocked by naloxone and insensitive to the benzodiazepine manipulations. Exposure to either fresh bedding, or the soiled bedding of another mixed sex pair of deer mice, had no significant effect on nociception. These results indicate that exposure of male deer mice to the olfactory cues associated with a potentially threatening individual (dominant/aggressive isolated male) elicits an analgesic response that involves alterations in the activity of benzodiazepine systems.

The last paragraph of the study:

The concentrations of androstenone and androstadienone used in this study are several orders of magnitude above those observed in human sweat. However, presentation of androstadienone in such concentrations increased cortisol levels in women, via activation of the hypothalamus. Hypothalamic activation, however, is not specific to androstadienone, as a chemically and biologically unrelated but perceptually similar molecule (Polysantol) has the same effect15, pointing toward perceptual (i.e., olfactory) quality as the key underlying factor. In fact, we observed analgesic effects upon the presentation of three different odorants, with one (3M2H) being chemically unrelated to the others. It is therefore impossible that we observed a substance-specific 'pheromone'-like effect. Our data suggest instead that an odor evoked by a cocktail of chemicals within the body secretions of isolated males (except cage mates) produce stress in rodents, leading to SIA. The effects produced by human males are simply due to the invariant structure of these chemicals across mammalian species. Although it is short lasting, stress caused by male experimenters may represent a confound of much existing animal research, extending even to nonbehavioral studies in which tissues were obtained from live rodents euthanized by either male or female personnel. Our findings strongly suggest that standard laboratory practice should account for experimenter sex when investigating any phenomenon possibly affected by stress.