Physiological response to food restriction in adult rats

Adult rats (>P60) were subjected to a widely used protocol of FR consisting of a fasting regimen on alternate days with maintenance of the composition of the diet in terms of vitamins, minerals and proteins. In other words, the animals had no access to food for a full day, every other day, whereas they were allowed to eat ad libitum on the intervening days. After 4 weeks of this dietary regimen (P60–P90), food-restricted rats (FR rats) showed a reduction of about 22% in the daily food assumption (Fig. 1a). The body weight of FR rats was lower than controls (CTLs), the reduction being present since the first week of diet. At the end of the fourth week of FR, mean body weight of restricted rats decreased by about 10% with respect to CTLs (Fig. 1b).

Figure 1: Effects of food restriction on body physiology and stress behaviour in food-restricted rats. (a) Food intake was measured in the last 10 days of diet. Food intake of food-restricted rats (FR; n=6; mean daily food intake 16.15±1.47 g) was significantly lower with respect to controls (CTL, n=6; mean daily food intake 20.67±1.52 g; two-way ANOVA P<0.001). (b) Body weight measured during the whole period of dietetic regimen showed a significant reduction in the FR animals (solid line) with respect to CTLs (dashed line) since the first week of diet. At the end of treatment, mean body weight of FR rats was 112.06±21.56% of the initial value, whereas mean body weight of CTLs was 126.31±14.55% of their initial value. The difference between the two groups was always statistically significant (FR n=6; CTL n=6; two-way ANOVA P<0,001). (c) Serum content of corticosterone in FR animals and CTLs. Samples from each animal were collected at the same time of the day, that is, at the beginning of the active phase (1700 hours). In FR rats (n=6), corticosterone levels were significantly higher with respect to CTLs (n=10) at the end of a fasting day (383.75±77.5 versus 205.09±17.05 ng ml−1; t-test P=0.026), whereas they returned to normality after 1 day of feeding (n=7; 155.10±34.25 ng ml−1; feeding versus CTLs t-test P=0.173). (d) In the elevated plus maze test, FR rats (n=6) spent the same time in open arms with respect to CTLs (n=6; 58±10 versus 37±7 s; t-test P=0.123). Error bars represent s.e.m.; asterisks indicate statistical significance. Full size image

Most studies on FR have shown that a reduction in the caloric intake is accompanied by an increase in the serum levels of corticosterone10,11,12,13, a body homeostatic response to the need for enhanced glicogenolysis. In FR rats, we observed almost a 100% increase in the blood content of corticosterone. Noteworthy, this increase was present only at the end of a fasting day, whereas corticosterone levels returned to normality after 1 day of feeding (Fig. 1c). The elevation of plasma corticosterone indicates that FR could cause a period of stress in the animals. Nevertheless, it has been suggested that FR represents only a low-intensity stressor, as the FR-induced elevation of plasma-free corticosterone is generally smaller compared with other stressing experimental protocols, such as restraint, which causes a rapid, marked elevation of plasma corticosterone and a disregulation of the hypothalamic–pituitary–adrenal axis14. Accordingly, we found that FR rats did not show any increase in mean daily locomotor activity with respect to CTLs and, when subjected to the elevated plus maze test (a validated method to measure anxiety-like behaviours in rodents15), they displayed a performance very similar to the CTL group (Fig. 1d).

Restoration of visual cortex plasticity in adult rats

To assess whether reducing the caloric intake could restore visual cortex plasticity in adulthood, we used two classical models of visual system plasticity: the OD shift of cortical neurons after MD and the recovery from amblyopia after reverse suture (RS). The effectiveness of these two manipulations in the rat is restricted to the CP.

To test whether 7 days of MD were effective in shifting the OD distribution (binocularity) of adult FR rats, the animals were monocularly deprived during the last week of the dietetic regimen. We evaluated OD distribution in the binocular portion of the primary visual cortex (V1) contralateral to the occluded eye by measuring the contra-ipsi (C/I) visual evoked potential (VEP) ratio, that is, the ratio between the responses elicited by the stimulation of the two eyes (see Methods). The C/I VEP ratio is in the range of 2–3 in adult normal animals, whereas it is strongly reduced (∼1) in rats subjected to MD during the CP.

As expected, 1 week of MD did not affect the C/I VEP ratio in the visual cortex of CTL adult animals, confirming the absence of OD plasticity in adult rats (C/I VEP ratio=2.81±0.21; n=5; Fig. 2a). When MD was performed in FR rats, however, we observed a marked OD shift in favour of the non-deprived eye, reflected by a significant decrease of the C/I VEP ratio (C/I VEP ratio=1.29±0.21; n=8; t-test, P<0.001; Fig. 2a). We also measured VA of both eyes by using VEP recordings in the visual cortex contralateral to the closed eye and, also in this case, we found that the responsiveness to MD was reinstated in FR rats. The VA of the deprived eye, indeed, was significantly reduced with respect to the open eye in FR rats (0.86±0.05 versus 1.04±0.04 c deg−1; n=6; paired t-test P=0.027; Fig. 2b), whereas the VA of the two eyes was comparable in CTL animals (0.92±0.03 versus 0.95±0.04 c deg−1; n=6; paired t-test P=0.385).

Figure 2: Restoration of visual cortex plasticity in animals. (a) OD was measured after 7 days of monocular deprivation in food-restricted rats (MD-FR, C/I VEP ratio=1.29±0.21; n=8) and in controls (MD-CTL, C/I VEP ratio=2.81±0.21; n=5; t-test, P<0.001). (b) Similarly, VA of the deprived eye and of the fellow non-deprived eye was evaluated both in MD-FR rats (0.86±0.05 versus 1.04±0.04 c deg−1; n=6; paired t-test P=0.027) and in MD-CTL animals (0.92±0.03 versus 0.95±0.04 c deg−1; n=6; paired t-test P=0.385) after 7 days of MD. (c) Analysis of binocularity was carried out in the visual cortex contralateral to the long-term deprived eye through VEP ratio measurement in food-restricted, reverse-sutured rats (RS-FR, C/I VEP ratio=2.08±0.12; n=7) and in controls (RS-CTL, C/I VEP ratio=1.21±0.14; n=6; t-test, P<0.001). (d) Recovery of VA in the formerly deprived eye was evaluated in RS-FR rats (1.01±0.05 versus 1.02±0.07 c deg−1; n=5; paired t-test P=0.851) and in RS-CTL animals (0.72±0.03 versus 1.03±0.02 c deg−1; n=4; paired t-test P=0.004). Error bars represent s.e.m.; asterisks indicate statistical significance. Full size image

We next evaluated whether this potential for plasticity could be exploited for restoring normal visual functions in adult amblyopic animals. Rats rendered amblyopic by long-term MD starting from P21 were subjected to FR protocol in adulthood (1 month starting at P60) and RS was performed during the last 2 weeks of diet. In adult FR rats, binocularity (C/I VEP ratio=2.08±0.12; n=7) and VA of the amblyopic eye (VA of long-term deprived eye=1.01±0.05 c deg−1, VA of not occluded eye=1.02±0.07 c deg−1; n=5; paired t-test P=0.851) exhibited a complete rescue. In contrast, we did not observe any sign of recovery in CTL animals (C/I VEP ratio=1.21±0.14; n=6; VA of long-term deprived eye=0.72±0.03 c deg−1; VA of not occluded eye=1.03±0.02 c deg−1; n=4; paired t-test P=0.004; Fig. 2c,d).

FR regulates inhibition levels in the adult visual cortex

Inhibition is well known to have a crucial role in the expression and regulation of neural plasticity during development and adulthood: there is evidence, indeed, that the inhibitory tone is essential for the time course of critical period in the visual cortex16 and it has been demonstrated that strategies decreasing inhibition reactivate developmental levels of plasticity in the adult brain17. We quantified the extracellular levels of γ-aminobutyric acid (GABA) and glutamate in the binocular visual cortex of FR and CTL animals using in vivo brain microdialysis and high-performance liquid chromatography (HPLC). Extracellular GABA levels were significantly reduced in the visual cortex FR rats with respect to CTLs (0.51±0.12 versus 2.94±0.33 pmol μl−1; FR n=7; CTLs n=6; t-test P<0.001; Fig. 3a), whereas no differences in glutamate levels were detected (Supplementary Fig. S1).

Figure 3: Reduction of intracortical inhibition in the visual cortex and increased synaptic plasticity following food restriction. (a) Measurement of GABA release in the visual cortex of food-restricted rats (FR=0.51±0.12 pmol μl−1; n=7; t-test P<0.001) with respect to controls (CTL=2.94±0.33 pmol μl−1; n=5) using in vivo brain microdialysis. (b) GAD65 expression was quantified through immunohistochemistry in the visual cortex of FR rats (n=6) and in CTLs (n=6; t-test P=0.001). (c) Average time course of layer II–III field potential (FP) amplitude before and after θ-burst stimulation of white matter (WM). White circles represent averaged field potential from FR rats (n=7 slices, 5 animals; two-way RM ANOVA, baseline versus the last 20 min of post-theta, P=0.025), whereas filled circles represent data from CTLs (n=7 slices, 5 animals; two-way RM ANOVA, baseline versus the last 20 min of post-theta, P=0.218). The difference between post-theta response in FR animals and CTLs was also evaluated (two-way ANOVA in the last 20 min of post-theta, P<0.001). A dot is the result of the average of three different measurements, each taken every 30 s. Waveform examples are represented. Vertical calibration bar, 100 μV; horizontal calibration bar, 2 ms. (d) LTP in the CA1 subfield of the hippocampus was induced through stimulation of Schaffer collaterals and time course of FP slope was recorded before and after high-frequency stimulation. Results from FR rats (white circles; n=6 slices, 5 animals; two-way RM ANOVA, baseline versus the last 20 min of post-theta, P=0.004) and CTLs (filled circles; n=4 slices, 3 animals; two-way RM ANOVA, baseline versus the last 20 min of post-theta, P=0.049) are illustrated. The difference between LTP in FR animals and CTLs was also measured (two-way ANOVA in the last 20 min of post-theta, P<0.001). A dot is the result of the average of four different measurements, each taken every 30 s. Examples of waveforms from hippocampal slices are represented. Vertical calibration bar, 500 μV; horizontal calibration bar, 2 ms. Error bars represent s.e.m.; asterisks indicate statistical significance. Full size image

To further analyse the effects of FR on the GABAergic system, the expression of GAD65 in the visual cortex of FR and controls animals was assessed by means of semi-quantitative immunohistochemistry. GAD65 is the 65-kDa isoform of the enzyme responsible for GABA synthesis (glutamic acid decarboxylase), whose expression is more prominent at synaptic level18. We found that puncta-ring immunoreactivity (see Methods) of GAD65 was significantly reduced in FR rats as compared with CTLs (t-test P=0.001; Fig. 3b).

Potentiation of cortical and hippocampal synaptic plasticity

Maturation of inhibition has been proposed to reduce synaptic plasticity in the visual cortex by acting as a gate, which filters the level and pattern of activity that layer IV, the major thalamo-recipient layer, is able to relay to supragranular layers19,20,21. Indeed, it has been demonstrated that the induction of long-term potentiation (LTP) in the layer II–III of the cortex through stimulation of the white matter (WM-LTP) declines with age and contrasting inhibition by acute application of GABA receptor antagonists on visual cortical slices in adulthood enhances WM-LTP22.

Given these results, we tested whether the reduction of inhibition observed in adult FR rats was sufficient to restore WM-LTP in the visual cortex. As expected, LTP induction was absent in CTL animals, whereas visual cortical slices taken from FR rats showed a robust potentiation of field potential (FP) amplitudes after WM stimulation (FR n=7 slices, 5 animals; CTLs n=7 slices, 5 animals; two-way analysis of variance (ANOVA) FR versus CTLs P<0.001; Fig. 3c).

Noteworthy, FR also increased LTP expression induced by Schaffer collateral stimulation in the CA1 hippocampal subfield (FR n=6 slices, 5 animals; CTLs n=4 slices, 3 animals; two-way ANOVA FR versus CTLs P<0.001; Fig. 3d), demonstrating that the facilitation of synaptic plasticity was not confined to the visual cortex. No difference in the excitability of FR slices with respect to CTLs was found in either visual cortex or hippocampus (Supplementary Fig. S2).

Benzodiazepine treatment prevents restoration of plasticity

It has been recently demonstrated that pharmacological treatment with mercaptopropionic acid or picrotoxin is effective in reducing inhibition and restoring plasticity to the visual cortex of adult rats23. These results were comparable to those obtained with FR (Supplementary Fig. S3).

To analyse how crucial was the role of decreased inhibition for the effects obtained in FR rats, we artificially strengthened GABAergic transmission in a separate group of FR animals by cortical infusion via osmotic minipumps of the benzodiazepine agonist diazepam. Diazepam enhances GABA receptor-mediated current, and is a common tool used for enhancing inhibition in the cerebral cortex16. We measured binocularity in monocularly deprived FR rats treated with diazepam by recording C/I VEP ratio in the visual cortex contralateral to the deprived eye. Diazepam infusion prevented the decrease of C/I VEP ratio due to FR (2.03±0.14; n=6; t-test on rank P=0.02; Fig. 4a). Moreover, VA of the deprived eye was normal in FR animals intracortically infused with diazepam and did not differ from VA of the open eye (VA of the deprived eye=0.95±0.05 c deg−1; VA of the open eye 1.00±0.02 c deg−1; t-test P=0.505; Fig. 4b). These results demonstrate that the decrease of intracortical inhibition has a pivotal role for the restoration of visual cortex plasticity induced by FR.

Figure 4: Restoration of plasticity is prevented by benzodiazepines treatment. (a) The C/I VEP ratio of control (CTL) animals was compared with the VEP ratio of FR rats treated with diazepam and monocularly deprived for one week (FR+Diaz, 2.03±0.14; n=6; t-test P=0.02). (b) Visual acuity of the deprived eye and of the open eye in FR+Diaz rats (0.95±0.05 versus 1.00±0.02 c deg−1; paired t-test P=0.505). Error bars represent s.e.m.; asterisk indicates statistical significance. Full size image

Corticosterone mimics FR effects on visual cortex plasticity

It has been hypothesized that exposure to mild stressors activates homeodynamic pathways of maintenance and repair triggering beneficial effects on brain functions24. Indeed, although high levels or chronic stress have been associated with reduced synaptic plasticity in the hippocampus, a mild or acute stress has been demonstrated to increase LTP induction25. Given its well-known involvement in stress response, corticosterone could emerge as a candidate for regulating processes of experience-dependent plasticity triggered by FR. We addressed this issue pharmacologically mimicking the every other day elevation of blood corticosterone caused by our protocol of FR. We treated a group of rats with a low dose of corticosterone in drinking water on alternate days for 4 weeks, as previously reported in rodents26,27,28, and we tested visual cortex plasticity evaluating the effects of 1 week of MD on OD distribution of cortical neurons. We found a significant shift of C/I VEP ratio in favour of the open eye in the visual cortex of MD rats treated with corticosterone (C/I VEP ratio=1.22±0.07; n=5; one-way ANOVA P<0.001; post hoc Holm–Sidak test CTL versus FR P<0.001; CTL versus Cort P<0.001; Cort versus FR P=0.826; Fig. 5a), as well as a decrease of the deprived-eye VA (VA of the deprived eye=0.75±0.02 c deg−1; VA of the open eye=1.00±0.02; paired t-test P<0.001; Fig. 5b).

Figure 5: Corticosterone treatment restores plasticity in adult rats. (a) Assessment of C/I VEP ratio in control (CTL) and in adult MD rats treated with corticosterone on alternate days (Cort, C/I VEP ratio=1.22±0.07; n=5; one-way ANOVA P<0.001; post hoc Holm–Sidak test CTL versus FR P<0.001; CTL versus Cort P<0.001; Cort versus FR P=0.826). (b) Visual acuity of the two eyes in MD animals treated with corticosterone (VA of the deprived eye=0.75±0.02 c deg−1; VA of the open eye=1.00±0.02; paired t-test P<0.001). (c) Western blot analysis was used to evaluate GAD65 expression in adult animals treated with corticosterone on alternate days with respect to controls (n=5 Cort; n=11 CTL; t-test P=0.015). Error bars represent s.e.m.; asterisks indicate statistical significance. Full size image

Interestingly, western blot analysis revealed a reduction of GAD65 levels in the visual cortex of rats treated with corticosterone (n=11 CTL; n=5 Cort; t-test P=0.015; Fig. 5c), suggesting that the enhancement of corticosterone levels due to FR could be involved in the modulation of intracortical excitation–inhibition balance.

Extracellular matrix remodelling and neurotrophin expression

Restoration of plasticity in the adult has been often associated with increased levels of neurotrophins and remodelling of extracellular matrix29,30,31,32. We tested whether FR affected brain-derived neurotrophic factor (BDNF) content in the adult visual cortex by means of immunohistochemistry and western blot. No difference in BDNF was detectable between the two groups using both methods (Fig. 6a,b). Similarly, immunohistochemistry for Wisteria Floribunda (WFA, an effective marker of CSPGs which are major components of PNNs) did not show any difference between FR rats and CTLs (Fig. 6c). Thus, we conclude that a short-term protocol of FR is unable to induce changes in BDNF expression or perineuronal nets composition.

Figure 6: Levels of BDNF protein and WFA in the visual cortex of adult food-restricted rats and controls. (a) Immunohistochemistry for BDNF in food-restricted rats (FR; cell density 949±46 cell per mm2; n=6) and control (CTL; cell density 888±38 cell per mm2; n=6; t-test P=0.330). (b) Analysis of BDNF protein by means of western blot in FR animals (n=4) and control rats (n=4; t-test P=0.145). (c) Density of WFA-labelled cells was measured in the two experimental groups by means of immunohistochemistry (cell density in FR rats 47±5 cell per mm2, n=6; cell density in controls 45±6 cell per mm2, n=6; t-test P=0.812). Error bars represent s.e.m. β-tub, β-tubulin. Full size image

FR promotes histone acetylation in V1 and hippocampus

Finally, we analysed the level of histone H3 acetylation in the visual cortex and hippocampus of adult animals subjected to FR using Lys 9 acetyl-H3 (AcH3)-specific antibodies in western blot experiments. Histone post-translational modifications regulate chromatin susceptibility to transcription with high levels of histone acetylation on a specific DNA segment being generally correlated with increased transcription rates33,34. Visual experience activates histone acetylation in the visual cortex during the critical period, but this capacity is downregulated in adult animals35,36,37. We found that FR regimen induced a pronounced increase in H3 acetylation in the visual cortex (FR n=6; CTLs n=6; t-test P=0.048). A similar effect was evident also in the hippocampus (FR n=8; CTLs n=8; t-test P=0.046; Fig. 7). More specifically, we reported that the enhancement of histone acetylation in FR rats was in the range of 50–60% in the visual cortex, whereas it was of about 100% in the hippocampus.