The neuropeptide VIP and its receptors are expressed in regions of the brain implicated in the control of learned behaviors including the hippocampus, cortex, amygdala and hypothalamus [e.g. [4
]]. In these regions, VIP is most commonly expressed in GABAergic interneurons [e.g. [40
]] that may use this peptide to communicate with specific post-synaptic targets. However, it is equally plausible that VIP functions more as a paracrine signal acting at sites more distant than just the adjacent postsynaptic neurons. While the physiological actions of this neuropeptide have not been extensively studied, VIP regulates synaptic transmission [e.g. [7
]] and intrinsic membrane currents [e.g. [9
]]. Thus, this neuropeptide can be a potent modulator of neural activity and function in specific circuits in the adult nervous system. Given this anatomical distribution and potential physiological functions in both the developing and mature nervous system, the loss of VIP may well have been expected to have global influences on learning and memory. However, the present study found that the deficits in the VIP-deficient mice were quite selective. Deficits were not observed in foot shock-evoked fear behavior. This demonstrates that the basic sensory and motor processes controlling this behavior are intact in the VIP-deficient mice. Similarly, the shock-evoked corticosterone response was intact in these mutant mice. Given that VIP is expressed in the adrenals and the circadian system that regulate corticosterone secretion, we felt that it was important to confirm that this aspect of the stress response was functioning in the VIP-deficient mice. Finally, we observed no deficits in the VIP-deficient mice in the acquisition of fear conditioning or in the recall measured at 24-hrs after training. Clearly, the circuitry involved in the acquisition of contextual fear conditioning is either not regulated by VIP or there is compensation for the loss of this neuropeptide in these transgenic mice.
The recall of the contextual fear conditioning was affected by the loss of VIP. These recall deficits were seen in three independent experiments including those mice only tested for recall a single time (Fig. ), those tested daily (Fig. ), and those tested every 12-hrs (Fig. ). These results fit with previous pharmacological studies that have raised the possibility that VIP plays a physiological role in the modulation of learning and memory. For example, administration of VIP into rat hippocampus following training on a passive avoidance task induced recall in rats tested 24-hrs later, while administration of a VIP antagonist enhanced retention [14
]. Other studies have also found evidence that the central administration of VIP impairs recall in passive avoidance learning [13
]. Spatial learning as measured by the water maze was also impaired by the ventricular administration of this neuropeptide [15
]. In contrast, one study reported that the intracerebral administration of a VIP receptor antagonist, but not VIP itself, inhibited performance on the Morris water maze [18
]. In addition, learning deficits in mice carrying a chimeric VIP-diptheria toxin gene have been reported [19
]. These transgenic mice lost about 20% of VIP content as measured by radioimmunoassay and exhibited learning deficits as measured by the Morris water maze. Taken together, these studies suggest that abnormally high or low levels of VIP can interfere with the acquisition and recall of specific learned behaviors.
The VIP-deficient mice used in the present study are a traditional transgenic model in which the gene coding for VIP has been inactive throughout development. VIP is expressed early in the fetal brain [44
] with VIP binding sites abundant on the floor plate of the neural tube [45
]. While not extensively studied in central neurons, VIP is a neurotrophic factor that can regulate neural growth, migration, and process formation [reviewed by [46
]] and, through these developmental mechanisms, influence the neural circuits involved in learning and memory functions. A recent study examined the consequences of pharmacologically blocking VPAC receptors during embryogenesis and examining potential cognitive deficits in the adult offspring [30
]. Male, but not female, treated mice exhibited deficits in contextual fear conditioning and social behavior. The selective set of behavioral deficits coupled with the gender difference led these authors to propose these mice as a model for the behavioral deficits of autism. Like the VIP-deficient mice, these mice treated developmentally with the VPAC antagonist did not show deficits in acquisition or in recall measured 24-hrs after training. The male offspring did show recall deficits when measured 48-hrs after training. Thus, the memory deficits observed in the transgenic animals in the present study could well be due to a loss of VIP early in development.
As described before, there is strong evidence that VIP and the VPAC2
R are critical for the normal functioning of the circadian system [reviewed by [49
]]. Together, these data indicate that VIP and VPAC2
R are critical for the generation of behavioral rhythms in mice and that the deficit occurs at the level of the SCN. Furthermore, the VPAC2
R-deficient mice even lose the daily rhythms in clock gene expression that are thought to lie at the heart of the machinery for the generation of daily rhythms [33
]. Given the overwhelming data that the loss of VIP influences the generation of circadian oscillations, the finding that these mutants showed no apparent deficiency in the diurnal and circadian regulation in recall was unexpected. This disassociation was particularly striking in the case of VIP-deficient mice that exhibited arrhythmic locomotor activity. The 6 hr interval in testing may have occluded small changes in peak recall time, which most likely tracks with the intrinsic free-running period of the animals. However, the frequency of testing was sufficient to show that there remains an optimal time of day, which tracks with the same time of day as the initial training exercise, even in mice with arrhythmic wheel-running activity. This is consistent with the phenomenon of "time-stamping" of learned behavior as observed in hamsters [51
] and by us in mice [37
]. What's more, this time-stamp phenomenon may be independent of the SCN. Lesioning the SCN of rats does not affect the time-stamped training of a T-maze reward task [53
]. Previous work has found that these behaviorally arrhythmic VIP-deficient mice also exhibit arrhythmic electrical activity rhythms when measured at the level of the SCN [35
]. We now show that these arrhythmic mice exhibited clear rhythms in recall and these rhythms were extremely similar to those exhibited by the behaviorally rhythmic mice. Thus, the loss of circadian function in these arrhythmic VIP-deficient mice had no obvious impact on the rhythms in recall suggesting that a rhythmic SCN may not be necessary for the circadian rhythm in recall. In recent years, it has become clear that many of the "clock genes" are expressed outside of the SCN [e.g. [54
]]. This raises the possibility that oscillators outside of the SCN may drive the rhythms in recall of learned behaviors. While we did not directly address this possibility in the present study, we did examine the expression of one clock gene, Period2
, in the hippocampus. We found that levels of this gene continue to show circadian differences in the hippocampus of VIP-deficient mice (Fig. ). This type of rhythmic expression is at least consistent with the possibility that extra-SCN rhythms in gene expression may be directly tied to the rhythms in recall observed in fear-conditioned mice.
In mammals, neurons in the hippocampus [[38
], present data], olfactory bulb [56
], SCN [e.g. [55
]], and other brain regions [54
] have now been shown to generate oscillations in circadian gene expression. This pool of data contributes to the view that the circadian system is comprised of multiple oscillatory components, with the role of the SCN being a master timer synchronizing these disparate cell populations [e.g. [57
]]. With this new view of clock genes and circadian organization, it becomes critical to determine the tissue-specific function of these genes. Logically, local rhythms in clock gene expression could serve to control the temporal program of gene expression and physiology specific to the hippocampus, as recently demonstrated in liver [58
]. However, unfortunately, we still do not know much about the functional significance of clock gene expression outside of the SCN. The VIP-deficient mice with weakened rhythms in SCN electrical activity may represent an advantageous model to explore coupling between different circadian oscillators. Our data with contextual fear conditioning raise the possibility that these rhythms in recall or memory are only weakly coupled to core time-keeping mechanisms in the SCN.