We have shown that the effect of acupuncture analgesia (AA) and mo- rphine analgesia could be dose-dependently attenuated by intracerebrove- ntricular (icv) injection of cAMP

and potentiated by cGMP. In the present study dibutyryl derivatives of cAMP or cGMP were injected into the subara- chnoid space of the spinal cord to see their effects on AA. Intrathecal catheterization was performed under ether anesthesia. polye- thylene tube of 0.61 mm o. d. was introduced through foramen magnum into the subarachnoid space

reaching the level of thoracolumbar junction. Measurement of tail-flick latency started 4 h after surgery. No significant fluctuation of tail-flick latency was noticed in a period of 60 min follo- wing the intrathecal administration of 50 μg/10 μl of dbcAMP

20μg/10 μl of dbcGMP or 10μl of artificial CSF. In electroacupuncture (EA) experiments

rats were given intrathecal injection of CSF

followed 15 min later by a session of EA at Zusanli and Sanyinjiao points (2～15 Hz

with increasing voltage from 1V to 2.5V in 30 min). Tail-flick latency was assessed every 10min. The percentage increases of latency in the 3 measurements during the EA period were ave- raged as the average EA analgesia. One hour later

the EA experiment was repeated with intrathecal injection of dbcAMP instead of CSF. The differe- nce in average EA aualgesia between dbcAMP and CSF sessions in individual animals was taken for statistical analysis. It was shown that EA effect was decreased by 30±7%

58±12% and 75±12% in groups of rats injected with 12.5

25 and 50μg of dbcAMP respectively (P<0.01 in each group

n=7～11). On the contrary

an augmentation of AA by 13±7% (p>0.05)

25±10% (p<0.05) and 29±10%±(p<0.05) was observed in groups of rats injected with 10

20 and 40μg of dbcGMP (n=10～13). The results indicate a marked attenuation of AA by intrathecal injection of dbcAMP and a mo- derate potentiation of AA by dbcGMP. A similar effect on AA was thus shown for cyclic nucleotides both in the brain and in the spinal level.