Maria Vittoria Ambrosini

The sequential hypothesis on sleep function. I. Evidence that the structure of sleep depends on the nature of the previous waking experience.

The sequential hypothesis on sleep function assumes that the information gathered by brain during the waking period is processed during sleep in two main steps occurring during synchronized sleep (SS) and, eventually, during paradoxical sleep (PS). To verify the main consequences of the hypothesis, i.e., (1) that SS is involved in brain information processing; and (2) that the structure of sleep is dependent on the nature of the previous waking experience, an experiment was designed involving rats exposed to a training session (two-way active avoidance) but failing to learn (NL), and rats left in their home cages in the same training room (C). The structure of sleep, determined by EEG techniques in the postacquisition period (3 hr), was different in NL rats in comparison to C rats, chiefly because SS episodes were markedly longer in the former group. A more detailed analysis indicated that, in NL rats, SS episodes not followed by PS increased their duration first, while those followed by PS became longer in the second half of the sleep period. Comparable results were obtained in the comparison of NL and C subgroups deprived of PS at the end of the acquisition period by chlomipramine treatment. The data support the sequential hypothesis and provide evidence for a primary role of SS in brain information processing.

The sequential hypothesis of sleep function. III: The structure of postacquisition sleep in learning and nonlearning rats

EEG methods were used to examine the structure of postacquisition sleep in learning (L) and nonlearning (NL) rats previously exposed to a session of two-way active avoidance training, and in control rats (C) left in their home cages. In agreement with literature data, the number and total amount of paradoxical sleep (PS) episodes were higher in L rats than in NL rats. In addition, significant differences between L and NL rats concerned the episodes of synchronized sleep followed by wakefulness or by PS (SS-W and SS-PS, respectively). The average duration and related parameters of SS-W episodes, and the average duration, number, amount and related parameters of SS-PS episodes increased in NL and L rats in comparison with C rats. Longer SS-W episodes occurred early in NL and L rats, but the effect lasted longer in NL rats. On the other hand, the increments concerning SS-PS episodes occurred earlier, were more pronounced and laster longer in L rats. The results support a role of SS in brain information processing, as envisaged by the sequential hypothesis on the role of sleep. They suggest, furthermore, that memory traces lacking adaptive value may be destabilized and cleared away during SS-W and SS-PS episodes, while the remaining memory traces may be retained and eventually stored again in more integrated form during SS-PS and PS episodes, respectively.

Effect of sleep on cerebral DNA synthesized during shuttle-box avoidance training

Female Wistar rats weighing 200 g were implanted with cortical electrodes and two intraventricular cannulae. Five days later they were given 3H-thymidine and exposed to shuttle-box training for four hours. They were then left free to sleep in the following three hours during which their EEG activity was recorded. In comparison with control animals (C), learning (L) and non-learning (NL) rats exhibited an increase in SS. In comparison to the EEG recording made the previous day, all animal groups displayed an increase in SS, but only NL rats suffered a decrease in PS(%). The specific radioactivity of DNA measured in several brain regions was tendentially lower in NL rats, but significance was achieved only in the cerebellum in the comparison between NL rats and C rats. No change occurred in liver. More marked and significant decrements in the DNA specific radioactivity of all brain regions were observed in the subgroup of NL rats displaying relatively high values of PS time in comparison to the analogous subgroups of C and L animals. Comparable decrements were present with regard to the subgroup of NL rats endowed with relatively low PS time. Less widespread and more limited changes were observed in the concentration of acid-soluble radioactivity. In addition, several significant correlations were detected by Spearmans analysis among behavioral, biochemical and sleep parameters. The results are consistent with the interpretation that the selective decrease in brain radioactive DNA observed in NL rats reflects a loss of DNA synthesized during the training period. The loss is related to the amount of post-training PS and is associated to a lengthening of the mean duration of PS episodes. It may be concluded that the loss of newly-synthesized brain DNA reflects the elimination of molecules associated with neural information devoid of adaptive value.

The sequential hypothesis on sleep function. II. A correlative study between sleep variables and newly synthesized brain DNA

The information acquired by brain during wakefulness (W) may be processed in two sequential steps occurring during synchronized sleep (SS) and paradoxical sleep (PS), respectively. On the assumption that brain molecules synthesized during the acquisition step undergo a comparable sleep processing, we have designed an experiment aimed at the verification of the sequential hypothesis. Groups of adult female Wistar rats received [3H-methyl] thymidine by intraventricular injection 30 min before being exposed to a 4 hr session of a two-way active avoidance training. Animals failing to achieve the learning criterion were further allowed a period of 3 hr during which they were left free to sleep, or were deprived of PS or of total sleep. Control rats were similarly treated, but were left in their home cages in the same training room during the period of acquisition. The results of correlative study among behavioral, sleep and biochemical variables demonstrate that the specific radioactivity of DNA in the cerebral cortex, cerebellum and brainstem is correlated with several variables of postacquisition sleep, mostly SS parameters. The correlations depend on the previous waking experience of the rats. The data substantiate the two main consequences of the hypothesis, i.e., (1) the involvement of SS in brain information processing; and (2) the dependence of the operations performed by the sleeping brain on the nature of the previous waking experience. The results also provide some insight into the kind of processing which occurs in the sleeping brain.

Differential changes in Cu, Zn and Mn superoxide dismutase activity in developing rat brain and liver.

The aim of our study was to assess the pattern of copper and zinc-containing superoxide dismutase (Cu, ZnSOD) and manganese-containing superoxide dismutase (MnSOD) activity from embryonic life to senescence in rat brain and liver. The two isoenzymes showed different profiles in the two organs examined. In particular, the cerebral MnSOD activity profile suggests a primary role during differentiation of this enzymatic form.

Synthesis of brain DNA during acquisition of an active avoidance task

The possible involvement of cerebral DNA synthesis in the learning process was examined in rats injected intracerebrally with 3H thymidine. During the period of incorporation (4.5 hr) one rat was trained to an active avoidance task while a second animal was kept in the same experimental room. In comparison with control rats paired to learning animals, the concentration of PCA-soluble radioactivity and of radioactive DNA of the cerebral cortex increased in all animal groups, i.e., control rats paired to non-learning animals, learning rats and non-learning rats. No change occurred in liver. In the cerebral cortex the slope of the regression line obtained by plotting the concentration of radioactive DNA versus the concentration of PCA-soluble radioactivity was lower in learning rats than in the group of pooled control animals. A comparable effect was noted in the hippocampus. In non-learning animals a similar decrease was present in the cerebral cortex and in cerebellum. In addition, it was found that in learning animals the percent incorporation was inversely related to the total number of avoidances only in the cerebral cortex. In non-learning rats a similar inverse relationship was present in the cerebral cortex and in cerebellum. In the former region the regression line of learning rats was shifted upwards in comparison with the regression line of non-learning animals. These results are interpreted to indicate that the incorporation of 3H-thymidine into cerebral DNA is directly related to the level of stress and is increased by learning.

Permanent brain ischemia induces marked increments in hsp72 expression and local protein synthesis in synapses of the ischemic hemisphere

Transient focal ischemia induced in rat brain by occlusion of the middle cerebral artery (MCAo) elicits a generalized induction of the 72 kDa heat-shock protein (hsp72) heralding functional recovery. As this effect implies activation of protein synthesis, and local systems of protein synthesis are present in brain synapses, and may be analyzed in preparations of brain synaptosomes, we evaluated hsp72 expression and protein synthesis in synaptosomal fractions of spontaneously hypertensive rats (SHRs) subjected to permanent MCAo. SHRs were randomly divided in ischemics and sham controls, anaesthesia controls and passive controls. Focal ischemia was induced under chloral hydrate anaesthesia by unilateral permanent MCAo. Protein synthesis was determined by [35S]methionine incorporation into synaptosomal proteins from ischemic and contralateral cortex/striatum, and from cerebellum. Hsp72 expression was measured in the same fractions by immunoblotting. Our data demonstrate that under these conditions synaptic hsp72 markedly increases in the ischemic hemisphere 1 and 2 days after MCAo, progressively declining in the following 2 days, while no significant change occurs in control rats. In addition, in the ischemic hemisphere the rate of synaptic protein synthesis increases more than two-fold between 1 and 4 days after MCAo, without showing signs of an impending decline. The present data provide the first demonstration that synaptic protein synthesis is massively involved in brain plastic events elicited by permanent focal ischemia.

The sequential hypothesis of sleep function. IV. A correlative analysis of sleep variables in learning and nonlearning rats

Female adult rats were trained for a two-way active avoidance task (4 h), and allowed free sleep (3 h). Control rats (C) were left in their home cages during the acquisition period. Dural electrodes and an intraventricular cannula, implanted one week in advance, were used for EEG recording during the period of sleep and for the injection of [3H]thymidine at the beginning of the training session, respectively. Rats were killed at the end of the sleep period, and the DNA-specific activity was determined in the main brain regions and in liver. Correlations among sleep, behavioral and biochemical variables were assessed using Spearmans nonparametric method. In learning rats (L), the number of avoidances was negatively correlated with SS-W variables, and positively correlated with SS-PS variables (episodes of synchronized sleep followed by wakefulness or paradoxical sleep, respectively) and with PS variables. An inverse pattern of correlations was shown by the number of escapes or freezings. No correlations occurred in rats unable to achieve the learning criterion (NL). In L rats, the specific activity of brain DNA was negatively correlated with SS-W variables and positively correlated with SS-PS variables, while essentially no correlation concerned PS variables. On the other hand, in NL rats, comparable correlations were positive with SS-W variables and negative with SS-PS and PS variables. Few and weak correlations occurred in C rats. The data support a role of SS in brain information processing, as postulated by the sequential hypothesis on the function of sleep. In addition, they suggest that the elimination of nonadaptive memory traces may require several SS-W episodes and a terminal SS-PS episode. During PS episodes, adaptive memory traces cleared of nonadaptive components may be copied in more suitable brain sites.

Induction of heat shock protein 70 reduces the alteration of striatal electrical activity caused by mitochondrial impairment

Since mild thermal stress seems to exert neuroprotection via induction of heat-shock protein 70 kDa (hsp70), we tested whether hsp70 would preserve striatal bioelectrical activity under conditions of mitochondrial impairment. Corticostriatal slices from rats that had undergone mild thermal stress were exposed to either rotenone or 3-nitropropionic acid (3-NP), that selectively inhibits mitochondrial complex I and complex II, respectively. Rotenone is utilized to obtain an experimental model of Parkinsons disease while 3-NP replicates Huntingtons disease phenotype in experimental animals. The cerebral hsp70 increase did not alter field potential amplitude of the slices but partially protected them against rotenone-induced neurotoxicity. Similarly, induction of hsp70 had also a partial neuroprotective effect on the neurotoxicity caused by 3-NP on striatal field potential. Since rotenone and 3-NP treatments mimic the mitochondrial impairment and oxidative stress that contribute to development of Parkinsons disease and Huntingtons disease, these data suggest that induction of hsp70 might represent a possible neuroprotective mechanism against the pathophysiological chain of events implicated in these neurodegenerative disorders.

Sequential Hypothesis of sleep function. V. Lengthening of post-trial SS episodes in reminiscent rats

Rats failing to learn a two-way active avoidance task during the training session were tested for performance the following day. One group of rats maintained its low level of avoidances (non improving or NI rats), while the remaining rats dramatically improved their avoidance score (improving or I rats). EEG recording during the posttrial period demonstrated significant variations in the sleep structure of I rats, in comparison with NI rats. The main change consisted in an increase in the average duration of the episodes of slow wave sleep followed by wakefulness or by paradoxical sleep. These variations occurred in the third hour of the posttrial period, while an increment in the amount of PS was observed in the sixth hour. In I rats, but not in NI rats, comparable variations emerged from the comparison of baseline sleep (determined the day before training) with posttrial sleep. The data are in agreement with the main postulate of the sequential hypothesis of sleep function which attributes a primary role to slow wave sleep in the processing of newly acquired memories.