An estimated 82 % of pregnant women worldwide consume a lower-than-recommended dietary dose of Zn; this rate may approach 100 % in developing countries(Reference Caulfield, Zavaleta and Shankar1). Zn is one of the most abundant divalent metal ions in the central nervous system and is mainly stored in the hippocampus, which is important for maintaining cognitive function(Reference Ni2). The developing nervous system is a prime target for the disruptive effects of Zn deficiency, as the brain undergoes its most rapid period of maturation during fetal life. Studies have shown a correlation between maternal Zn status and neonatal and infant behaviour and cognitive function(Reference Kirksey, Wachs and Yunis3).
Though many studies have shown that learning and memory dysfunction can be induced by Zn deficiency in utero, the mechanism underlying this effect is still obscure(Reference Frederickson, Suh and Silva4, Reference Colvin, Davis and Nipper5). Recently, the relationships between changes in cell signalling pathways and long-term potentiation (LTP) in the nervous system as well as learning and memory function have become the focus of research. Neural plasticity, which is reflected by LTP, is considered to be the electrophysiological basis for learning and memory(Reference Malik and Chattarji6). Notably, vesicular Zn is released as a neuromodulator into the synaptic cleft to modify different types of receptors during synaptic transmission. Studies involving dietary depletion of Zn or the Zn chelators dithizone(Reference Lu, Taverna and Ackerley7) and Ca-EDTA(Reference Li, Hough and Frederickson8, Reference Takeda, Suzuki and Tamano9) have suggested that endogenous Zn is required for the induction of LTP in CA1 and CA3 pathways. It is not known whether synaptic plasticity in the medial perforant path–dentate gyrus (MPP–DG) pathway is also affected by Zn deficiency.
Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family that is activated by Ca/calmodulin-dependent protein kinase II α (α-CaMKII) and plays important roles in synaptic plasticity as well as learning and memory(Reference Bramham and Messaoudi10). Interestingly, Zn has been shown to be an important modulator of this pathway, facilitating the maturation of BDNF from pro-BDNF through the activation of Zn-dependent matrix metalloproteinases(Reference Hwang, Park and Choi11).
α-CaMKII, one of the major CaMKII isoforms in the brain, also plays key roles in synaptic plasticity and learning and memory function(Reference Cho, Bhatt and Elgersma12). Although several studies have shown that Zn can modulate the activity of CaMKII(Reference Lengyel, Fieuw-Makaroff and Hall13), to our knowledge, no studies have evaluated changes in α-CaMKII levels resulting from Zn deficiency. Hence, in the present study, we postulated that the cognitive deficits and synaptic obstacles observed in Zn-deficient rats result from a perturbation in α-CaMKII/BDNF signal transduction. In addition, little information is available regarding the effects of Zn supplementation on Zn deficiency-induced impairments in cognitive function and synaptic plasticity. Therefore, we investigated the possible mechanisms of learning and memory impairment in this context by analysing synaptic plasticity in the MPP–DG pathway and α-CaMKII/BDNF expression levels in the hippocampus in both Zn-deficient and Zn-supplemented rats.
Materials and methods
Animals, diets and tissue samples
The present study complied with the Guide for the Use and Care of Laboratory Rats and was administered under the auspices of the Animal Resource Services of Xinhua Hospital, affiliated with the Medical School of Shanghai Jiaotong University, which is accredited by the Chinese Association for the Accreditation of Laboratory Animal Care. A total of twenty-four virgin Sprague–Dawley rats (220–250 g) were obtained from a commercial source (Bikei Animal Company). After consumption of a standard, non-purified diet (Bikei Animal Dietary) for a 5 d acclimatisation period, the rats were impregnated. The pregnant rats were randomly divided into four groups on the first day of pregnancy: control (CO) group; pair-fed (PF) group; Zn-deprived (ZD) group; orally Zn-supplemented (ZS) group. Each group consisted of six pregnant rats. After birth, each dam's litter was culled to eight according to weight. The ZD and ZS groups received an AIN-93G-based diet containing 2 mg Zn/kg diet ad libitum throughout pregnancy and lactation. Both the CO and PF groups were fed the basal diet adequately supplemented with zinc sulphate at a dose of 25 mg Zn/kg, but each PF dam received only the daily average amount of food consumed by its ZD paired dam. After weaning (day 21), two male pups of similar weight were removed from each litter, so that each group contained twelve male pups. Pups in the CO and PF groups continued to receive the control diet, and pups in the ZD group continued to receive the Zn-deficient diet. However, the ZS pups were switched to the adequate-Zn diet (25 mg Zn/kg). All the four groups were supplied with deionised water ad libitum. The weight of the dams and offspring were monitored throughout the experiments. When the pups were 35 d old, they were tested using the Morris water maze and subsequently anaesthetised and decapitated. Serum samples were collected, and serum Zn levels were determined by atomic absorption spectrophotometry (Thermo M6). Resected hippocampal tissues were immediately frozen at − 196°C and then stored at − 80°C.
Morris water maze test
Beginning on postnatal day 35, the rats received 5 d of training to test their capacity for learning and memory acquisition using the Morris water maze. It is possible that Zn deficiency had led to a failure of the development of sensory systems(Reference Grahn, Paterson and Gottschall-Pass14). Thus, an invisible platform was used in the Morris water maze. The rats were trained to swim to reach a platform in a circular pool (160 cm diameter × 50 cm height) located in a test room. The pool was filled with water (22–25 °C) to a depth of 16 cm. A movable circular platform, 12 cm diameter, was submerged 2 cm below the surface of water, which was opaque after being mixed with prepared Chinese ink (Shanghai Chinese Ink Factory) as described previously(Reference Zhang, Li and Cao15). Briefly, for the place navigation test (spatial learning acquisition), each rat was subjected to two trials per d for four consecutive days. Each trial consisted of placing the rat in the water so that it faced the wall of the pool at one of the four starting locations (north, east, south and west) in a random order. The rat was allowed to search the platform for a maximum of 120 s. If the rat did not find the platform in 120 s, it was gently lifted up and placed onto the platform for 20 s before being returned to the cage. The escape latency (the duration before finding the platform) and swim paths were automatically recorded by a video/computer system. The escape latency, path length and swim speed were recorded as indices of learning and memory capacity.
Electrophysiological recordings and long-term potentiation induction
From each group, six pups were subjected to LTP induction. Hippocampal slices were prepared as described previously(Reference Cao, Cui and Feng16). Briefly, the rats were anaesthetised with sodium pentobarbital (2 %) and decapitated. The brains were rapidly removed, and the hippocampi were dissected in ice-cold oxygenated (95 % O2 and 5 % CO2) artificial cerebrospinal fluid. Approximately 400 μm transverse slices were prepared on a tissue slicer and then incubated in a recovery chamber filled with oxygenated artificial cerebrospinal fluid. The slices were allowed to recover for at least 1·5 h before recordings were attempted. For electrophysiological recordings, the slices were transferred into a perfusion chamber continuously superfused with artificial cerebrospinal fluid at a rate of 2·5 ml/min and maintained at 30°C while recording. All the experiments were conducted on slices maintained in vitro for 2–8 h. For extracellular stimulation of the MPP–DG, a monopolar tungsten electrode was placed in the middle molecular layer of the dentate gyrus. Extracellular field excitatory postsynaptic potentials were recorded with a glass microelectrode (2–10 MΩ resistance) filled with pontamine sky blue and situated in the middle molecular layer of the dentate gyrus. Responses were evoked using single-pulse stimuli of a fixed duration (0·1 ms), delivered at 30 s intervals. The optimal recording location for each rat was determined as the point at which the largest response could be obtained with a minimal amount of current. Only those slices that produced field excitatory postsynaptic potentials of 1 mV or higher in amplitude were selected for the experiments. Medial pathway responses were confirmed on the basis of a depression of field excitatory postsynaptic potentials elicited with paired pulses spaced 40–100 ms apart. The stimulation intensity that produced 30 % of the maximal evoked amplitude was chosen for the test pulse and tetanic stimulation. Test stimuli (0·1 ms pulse width) were delivered every 30 s before tetanisation. Once a stable baseline recording was obtained, LTP was induced by modifying a protocol (high-frequency stimulation: two trains, 500 ms each, 100 Hz within the train, repeated every 30 s).
Western blot analysis
At the end of the experiments, total protein was extracted from the rat hippocampi using a protein lysis kit (Beyotime). Using SDS–PAGE, 30 μg of protein were separated and then transferred onto polyvinylidene difluoride membranes. After blocking in 5 % fat-free milk for 3 h, the membranes were incubated with primary antibodies, including rabbit anti-BDNF (1:1500; Chemicon), rabbit anti-α-CaMKII (1:200; Santa Cruz) and mouse anti-actin (1:10 000; Kang Chen) overnight. On the next day, the membranes were washed and incubated with horseradish peroxidase-conjugated goat anti-rabbit (1:5000) or horseradish peroxidase-conjugated goat anti-mouse (1:5000) antibodies for 1 h. Immunolabelled protein bands were detected using an enhanced chemiluminescence system. Films were digitised using a scanner, and the relative optical density of the bands was determined with Scion Image software 4.03 (Scion Corp.).
Statistical analysis
All data were analysed using ANOVA. The level of significance was set at P< 0·05. Statistical analysis was performed using SPSS for windows (version 13.0; SPSS, Inc.). All data are presented as the means and standard deviations.
Results
Zinc deficiency affected serum zinc concentrations and growth of the offspring
A general growth inhibition was observed in rats fed the Zn-deficient diet. In the ZD group, the dams had poor appetites after being fed the Zn-deficient diet for approximately 12–15 d and subsequently showed growth retardation. On the day of weaning (day 21), the final gain in body weight of the ZD dams (12·26 (sd 1·45) g) was significantly lower than that of the CO (70·50 (sd 5·96) g) and PF (61·67 (sd 9·09) g) dams. On day 35, the ZD pups exhibited a significantly lower weight gain than the CO and PF pups (ZD pups 35·75 (sd 2·95) g, CO pups 105·50 (sd 9·74) g and PF pups 85·35 (sd 8·38) g). These changes in growth, combined with the serum Zn status of these rats, indicated that we had successfully established a model of significant Zn deficiency. Pups in the ZS group had better appetites after 4–5 d of Zn supplementation and showed an accelerated weight gain. At the end of the experiments, serum Zn concentrations and body weight gain were similar between the ZS and PF pups (P>0·05; Table 1).
Effects of zinc deficiency and supplementation on serum zinc and body weight gain in rats (Mean values and standard deviations; number of rats)
CO, control group; PF, pair-fed group; ZD, Zn-deprived group; ZS, Zn-supplemented group.
* Mean value was significantly different from that of the CO and PF groups (P< 0·05).
† Mean value was not significantly different from that of the CO group (P>0·05).
‡ Mean value was significantly different from that of the CO group (P< 0·05).
§ Mean value was not significantly different from that of the ZS group (P>0·05).
∥ Mean value was significantly different from that of the ZS group (P< 0·05).
¶ Maternal weight gain represents the period from conception to weaning. Offspring weight gain represents that from birth to 35 d.
Learning and memory test
The swim speeds among the four groups were not significant (P>0·05), which meant that intact motor capability was not impaired by Zn deficiency. Compared with the rats in the CO and PF groups, those in the ZD group had poor orientation abilities. Their swimming traces were distracted, as they took more time to find the hidden platform (escape latency) and swam a longer distance to get there (P< 0·05). The indices of the ZS group were similar to those of the CO and PF groups (P>0·05). With increased training, both the escape latency and path length gradually declined in all the groups. However, starting on the third training day, the improvements in the ZD group were not as apparent as those in the PF, CO and ZS groups (P< 0·05) (Fig. 1).
Curve for (a) escape latency, (b) path length and (c) swim speed calculated for each rat on each training day. Values are means of twelve rats per group. CO (
), control group; PF (
), pair-fed group; ZD (
), zinc-deprived group; ZS (
), orally zinc-supplemented group. (A colour version of this figure can be found online at http://www.journals.cambridge.org/bjn)
Synaptic plasticity in the medial perforant path–dentate gyrus pathway
To detect changes in the hippocampal LTP, the LTP of the MPP–DG pathway was compared among the four groups. As shown in Fig. 2 and Table 2, the high-frequency stimulation evoked significantly larger LTP in the CO and PF slices than in the ZD slices, suggesting that Zn deficiency depressed LTP. In addition, there was a significant difference between the ZD and ZS samples, suggesting that Zn supplementation might improve the LTP depression caused by Zn deficiency during fetal life. However, the ZS samples showed significantly less LTP than the CO and PF samples, indicating that Zn supplementation did not completely restore the impaired LTP.
Hippocampal long-term potentiation (LTP) in zinc-deprived (ZD, ) and zinc-supplemented (ZS, ) rats. The LTP induced by tetanic stimulations in the ZD slices was significantly smaller than that of the control (), pair-fed () and ZS slices (P< 0·05). fEPSP, field excitatory postsynaptic potentials. (A colour version of this figure can be found online at (http://www.journals.cambridge.org/bjn).
Effects of zinc deficiency and supplementation on long-term potentiation in offspring rats (Mean values and standard deviations)
fEPSP, field excitatory postsynaptic potentials; CO, control group; PF, pair-fed group; ZD, Zn-deprived group; ZS, Zn-supplemented group.
* Mean value was significantly different from that of the CO and PF groups (P< 0·05).
† Mean value was significantly different from that of the CO and PF groups (P< 0·05).
‡ Mean value was significantly different from that of the ZD group (P< 0·05).
Zinc deficiency suppressed calmodulin-dependent protein kinase II α/brain-derived neurotrophic factor signalling in the hippocampus
Because α-CaMKII/BDNF signalling is closely involved in LTP formation, we decided to examine whether α-CaMKII or BDNF in the hippocampus was affected by Zn nutrition status. As shown in Fig. 3, both α-CaMKII and BDNF levels were markedly decreased in the hippocampus of the ZD rats than in that of the CO and PF rats. Quantitative analysis revealed that α-CaMKII and BDNF protein levels in the ZD rats were significantly lower than those measured in the CO and PF rats (P< 0·05). Rats that received Zn supplementation after weaning showed increased levels of α-CaMKII and BDNF relative to the ZD rats; however, their expression was still significantly reduced relative to the levels measured in the CO and PF rats (P< 0·05), suggesting that Zn supplementation did not completely restore α-CaMKII and BDNF protein expression.
Effect of zinc deficiency on the hippocampal expression of calmodulin-dependent protein kinase II α (α-CaMKII) and brain-derived neurotrophic factor (BDNF) in rats. (a) Western blot analysis of hippocampi obtained from the control (CO), pair-fed (PF), zinc-deficient (ZD) and zinc-supplemented (ZS) rats after incubation with antibodies against α-CaMKII and actin. (b) Densitometry of α-CaMKII expression was normalised to actin. Values are means (n 6), with standard deviations represented by vertical bars. * Mean value was significantly different from that of the CO, PF and ZS groups (P< 0·05). † Mean value was significantly different from that of the CO and PF groups (P< 0·05). (c) Western blot analysis of hippocampi obtained from the CO, PF, ZD and ZS rats after incubation with antibodies against BDNF and actin. (d) Densitometry of BDNF expression was normalised to actin. Values are means (n 6), with standard deviations represented by vertical bars. * Mean value was significantly different from that of the CO, PF and ZS groups (P< 0·05). † Mean value was significantly different from that of the CO and PF groups (P< 0·05).
Discussion
In the present study, we used dietary Zn deficiency (2 mg Zn/kg diet) and Zn supplementation rat models to examine changes in learning and memory behaviour as well as LTP-related signalling pathways in the hippocampus. Because the effects of Zn deprivation could be ascribed to reduced food intake rather than to Zn deficiency, one group of controls (PF group) was pair-fed with the same amount of food consumed by the Zn-deprived group. The PF rats showed no changes in the parameters assessed, indicating that the effects were specific to Zn deficiency and not the consequence of a general reduction in food intake.
Growth retardation may be induced by the inhibition of protein synthesis and increased catabolic response to Zn deficiency(Reference Giugliano and Millward17). After weaning, the pups were fed an oral Zn-abundant diet, and as a result serum Zn concentrations and body weight improved significantly and were similar to those of the PF group on day 35. These results are consistent with results reported by others(Reference Halas, Hunt and Eberhardt18, Reference Golub, Keen and Gershwin19), suggesting that oral supplementation with Zn after weaning could improve the growth of pups.
The results of the Morris water maze test revealed that the Zn-deficient rats exhibited defects in memory behaviour. This is consistent with previous reports showing that dietary Zn deficiency appears to damage learning and memory processes in the hippocampus(Reference Takeda, Fuke and Minami20). Because intracellular Zn is necessary for many enzyme activities and protein functions associated with signal transduction and gene expression(Reference Vallee and Falchuk21), it is reasonable to speculate that Zn deficiency-induced impairments of hippocampus-dependent learning and memory functions might be caused by disruptions in the LTP-related signalling cascade. With our rat model, we identified a new mechanism whereby dietary Zn deficiency impairs learning and memory function by disrupting the α-CaMKII/BDNF signalling pathway in the hippocampus. Furthermore, this is the first study to report LTP-related signalling pathway in rats subjected to Zn supplementation after experiencing Zn deficiency during gestation and lactation. The present results demonstrated that while body weight and memory behaviour were normal in the ZS rats, neither LTP nor α-CaMKII/BDNF protein levels had recovered to control levels by the end of the experiments.
α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptors have been reported to elicit N-methyl-d-aspartic acid receptor activation and result in the influx of Zn (as well as Ca) into postsynaptic neurons, a process that is believed to be involved in LTP induction(Reference Takeda, Fuke and Minami20, Reference Weiss, Koh and Christine22). LTP in the hippocampus is considered to be an important tool for elucidating the molecular mechanisms of learning and memory(Reference Miyamoto23, Reference De Roo, Klauser and Muller24). Ca enters the postsynaptic neurons through AMPA and N-methyl-d-aspartic acid receptors and Ca channels and then binds with calmodulin (CaM) to form a Ca–CaM complex(Reference Miyamoto23, Reference Hudmon and Schulman25). The activated Ca–CaM complex combines with CaMKII, resulting in its autophosphorylation(Reference Hudmon and Schulman25). CaMKII activity is essential for normal N-methyl-d-aspartic acid receptor-dependent forms of LTP and spatial learning and memory(Reference Hudmon and Schulman25) as well as for plasticity in other regions of the brain(Reference Hardingham, Glazewski and Pakhotin26). Interestingly, Zn can directly activate CaMKII, independently of Ca and CaM. Zn can also affect CaMKII activity by modulating the binding of the Ca–CaM complex to CaMKII(Reference Lengyel, Fieuw-Makaroff and Hall13). There are four isoforms of CaMKII, α, β, γ and δ. Accumulating evidence suggests that different CaMKII isoforms are involved in different neural functions based on their characteristic cellular localisation patterns. The α-CaMKII protein is highly enriched at synapses, and a significant amount of α-CaMKII mRNA is localised in dendrites(Reference Wells, Dong and Quinlan27). Recent experimental evidence indicates that almost 83 % of α-CaMKII found in dendrites during normal brain function is synthesised locally(Reference Miller, Yasuda and Coats28). α-CaMKII is known to play key roles in synaptic plasticity, hippocampal place cell stability and spatial learning. Mice with a mutation in the α-CaMKII gene show hippocampal LTP deficits and learning and memory abnormalities in hippocampal tasks(Reference Elgersma, Fedorov and Ikonen29). The effect of Zn deficiency on α-CaMKII has not been explored previously. Our data showed that chronic Zn deficiency reduced α-CaMKII protein levels in the hippocampus.
BDNF, a member of the neurotrophin family, is highly expressed in neurons throughout the central nervous system(Reference McAllister, Katz and Lo30). BDNF promotes the survival of a wide variety of neurons and regulates axonal and dendritic growth, thereby accounting for synaptic plasticity in the central nervous system(Reference Lindholm, Castren and Berzaghi31). Previous studies have reported that membrane depolarisation-triggered Ca2+ influx through L-type voltage-dependent Ca2+ channels induces an increase in BDNF mRNA expression in cultured neurons and that this effect is blocked by the inhibitors of CaM(Reference Zafra, Lindholm and Castern32). These data suggested that BDNF lies downstream of the CaM/CaMKII signalling pathway and can be activated by CaMKII. Interestingly, the present study showed that the Zn deficiency-induced down-regulation of BDNF is mediated by an α-CaMKII-independent signalling pathway. Collectively, the down-regulation of both α-CaMKII and BDNF by Zn deficiency might in part explain the observed impairments in learning and memory.
The few studies on the effects of Zn supplementation on cognitive recovery in Zn-deprived animals have reported conflicting results. Halas et al. (Reference Halas, Hunt and Eberhardt18) showed that Zn nutritional insults during the critical period of cerebral growth were not reversed by subsequent Zn supplementation. However, the present study found that Zn supplementation restored the cognitive defects observed in the Zn-deficient rats, as reported by Piechal et al. (Reference Piechal, Blecharz-Klin and Pyrzanowska33). The discrepancy in these results may be attributed to age differences in the offspring undergoing the cognitive tests. Importantly, Zn supplementation in children also confirmed the present results: Zn supplementation in undernourished children improves their developmental quotients, activity patterns and neuropsychological functions(Reference Gardner, Powell and Baker-Henningham34). Our finding that Zn supplementation is beneficial for reversing hippocampal dysfunction seems to further substantiate the pivotal role played by synaptic Zn in shaping the physiological neurotransmission of the hippocampus, a crucial brain region. Notably, the fact that Zn supplementation only partially restored the protein expression of α-CaMKII and BDNF indicated that while the impairments in key molecules caused by in utero Zn deficiency may be irreversible, the cognitive symptoms can be restored.
In conclusion, the present study has shown that α-CaMKII/BDNF signalling is altered in a Zn-deprived model known to affect synaptic plasticity and cognitive function. Furthermore, Zn supplementation restored the cognitive dysfunctions induced by maternal Zn deficiency. Nevertheless, Zn supplementation could not completely restore LTP or α-CaMKII/BDNF expression to control levels. These findings will help to clarify the mechanism of Zn deficiency-induced cognitive impairments. In addition, the regulation of α-CaMKII/BDNF signalling pathway by Zn nutriture may represent an important new strategy for the prevention and treatment of neurodegenerative diseases.
Acknowledgements
The present study was funded by the Chinese National Natural Science (no. 30600496) and the Shanghai Committee of Science and Technology, China (no. 114119a1600). XiaoDan Y. was responsible for the conception, design, molecular experiments and data analysis and drafted the manuscript. Y. B., T. R. and XiaoGang Y. were responsible for animal feeding and element measurements. All authors approved the final version. There are no conflicts of interest.
