Effect of Normal and Reversed Sleep-Wake Cycles upon Nyctohemeral Rhythmicity of Plasma Thyrotropin: Evidence Suggestive of an Inhibitory Influence in Sleep

The relation of nyctohemeral variation in plasma TSH to sleep-wake cycles was examined in 10 normal young men who had their sleep polygraphically monitored and their blood sampled every 20 min for 24,36, or 48 h periods. Studies of normal sleepwake cycles in which sleep was allowed from the usual bedtime to 0630 h totalled 21 nights (night = 1840-0620 h) and their corresponding 16 days. TSH was measured by a sensitive RIA. On 17 nights, the mean nightly TSH significantly exceeded that of the day's and, on 18 nights, clear nyctohemerally maximal peaks in TSH were seen in the 2100-0100 h interval. Greater amplitude, duration and rhythmic repetition over several nights distinguished 2100-0100 h maxima from a background of persistent briefly episodic release. These nyctohemeral peaks were pre-sleep maxima, as rises uniformly began, and on 15 nights, the peaks occurred prior to the onset of sleep. The peaks clustered within the 30 min just before (12 nights) or after (3 nights) entry into sleep. TSH release then declined across sleep. Other evidence suggestive of an inhibitory influence in sleep upon TSH release was that sleep began early on the 3 nights without clear 2100-0100 h TSH maxima and that the mean 2100-0100 h TSH peak was significantly reduced when sleep began prior to the usual 2300-0000 h interval and significantly increased when the onset of sleep was delayed or postponed. After a 24 h baseline, 4 men underwent phase-reversal of their sleep-wake cycles for 48 h, in which sleep was shifted to the 1100-1830 h interval. On the first wakeful night of reversal, the 2100-0100 h peak began normally, but, in the absence of sleep, the enhanced TSH release then simply continued across this night, delaying achievement of the nyctohemeral maxima. On the second wakeful night of reversal, the maximum in mean TSH lay in the same 0400-0600 h interval as that of first reversal night, and the mean 2100-0100 h peak was no longer evident. The TSH of the second 24 h of reversal also was significantly reduced, suggestive of a negative feedback effect of enhanced release of the first reversal day. No shift of basal pre-sleep TSH peaks to the 0900-1300 h interval or of sleep-enhanced TSH release was seen during reversal. Thus, despite the persistence of TSH's nyctohemeral rhythmicity across acute sleep-wake reversal, its pattern changed significantly in relation to shifts in sleep. We currently view these results as consistent with the origin of TSH's nyctohemeral rhythmicity in a circadian mechanism whose expression is subject to modulation by the inhibitory influences of feedback and sleep.