Characteristics of midbrain control of spinal nociceptive neurons and nonsomatosensory parameters in the pentobarbital-anesthetized rat.

1. GABAergic mechanisms in the midbrain periaqueductal gray (PAG) have been proposed to control the activity of descending antinociceptive systems and defensive behavior. Here, the effect of neuronal disinhibition by gamma-aminobutyric acid (GABAA) receptor blockade at midbrain sites on spinal neuronal responses to noxious and innocuous skin stimulation was quantitatively characterized. It was compared with the effect of direct neuronal excitation by glutamate microinjections or electrical stimulation at identical sites. Changes in mean arterial blood pressure and other nonsomatosensory responses were also assessed. 2. Responses of 101 multireceptive lumbar spinal dorsal horn neurons to noxious radiant skin heating (50 degrees C, 10 s), innocuous skin brushing, and electrical stimulation of primary afferent A- and C-fibers were recorded in deeply pentobarbital-anesthetized rats. The mean blood pressure was continuously monitored in one carotid artery, and other nonsomatosensory parameters, such as frequency and depth of spontaneous respiration and contractions of abdominal and facial muscles, were classified according to their relative intensity into five groups. 3. A fine, multibarrel glass pipette was constructed for monopolar electrical stimulation and microinjection of the GABAA receptor antagonist bicuculline (40, 200, or 400 pmol), or glutamate (10-50 nmol), or Fast Green dye in 50 or 100 nl at identical sites in the midbrain. 4. Bicuculline microinjections into discrete regions of the PAG selectively abolished spinal neuronal responses to noxious skin stimulation but did not affect brush-evoked responses or responses to electrical A-fiber stimuli. This antinociception was often, albeit not necessarily, accompanied by tachypnoea and abdominal and facial muscle contractions and changes--mostly increases--in mean arterial blood pressure. Injections into other areas of the PAG and adjoining ventral tegmentum (VT) were less effective. The vast majority of injection sites in the lateral tegmentum (LT) were ineffective. 5. Glutamate microinjections at midbrain sites to detect areas of origin of descending antinociceptive neurons were characterized by a high incidence (greater than 50%) of false-negative results, as bicuculline was shown to be effective at numerous glutamate-insensitive sites. Glutamate microinjections into some sites of the PAG and adjoining VT reduced, but did not abolish, spinal neuronal responses to noxious skin heating. Injections into the LT were ineffective. 6. The efficacy of electrical stimulation at midbrain sites on spinal nociceptive responses had no predictive value for the effect of glutamate or bicuculline.(ABSTRACT TRUNCATED AT 400 WORDS)