Experimental and clinical evidence supports the hypothesis that the loss of basal forebrain (BF) cholinergic neurons and the consequent reduction of acetylcholine (ACh) synthesis and release significantly contribute to the cognitive impairment of aging disorders, such as mild cognitive impairment (MCI) and Alzheimer’s disease (AD) [1–3]. Acetylcholinesterase inhibitors (AChE-Is) such as donepezil prevent the hydrolysis of the residual ACh in the brain and represent the best pharmacological tool to attenuate cognitive disturbances in patients with mild to moderate AD [4, 5]. AChE-Is are currently used as a symptomatic treatment to improve or at least maintain central cholinergic function [6, 7].

To date, besides the research of new drugs able to combat age-related cognitive decline, the protection of neurons from damage and death associated with neurodegenerative disorders is a major challenge in neuroscience. The concept of neuroprotection has found increasing acceptance in neurology during the past decade and includes interventions aimed to slow or even halt the progress of neuronal degeneration. Interestingly, there is growing evidence that, beyond allowing alleviation of cognitive symptoms, AChE-Is produce effective neuroprotection [7]. In fact, it has been shown that AChE-Is protect against glutamate excitotoxicity, neuronal damage and amyloid β (Aβ) neurotoxicity. Furthermore, many studies have shown that they induce upregulation of nicotinic ACh receptors (nAChRs) [8–18]. Importantly, α₄ and α₇ nAChRs play a crucial role in AChE-I-mediated neuroprotection, mainly through the involvement of the phosphatidylinositol 3-kinase (PI3K) pathway [8, 17]. Unfortunately, few in vivo studies have examined AChE-I neuroprotective action [7]. Although a lot of studies have demonstrated the symptomatic effects of donepezil in models of aging and dementia [19–25], a handful of studies have distinguished symptomatic from neuroprotective effects by administering donepezil only before (not during) behavioral testing [12–14, 26, 27]. Namely, injecting donepezil before a hypoxic insult has been shown to alleviate hypoxia-induced neurodegeneration and behavioral impairment [12], and, similarly, administrating donepezil before Aβ injection was demonstrated to block lipid peroxidation and learning deficits [13]. In both studies, donepezil neuroprotective effects appeared to be mediated by the activation of the σ₁ receptor, a protein involved in modulation of intracellular Ca²⁺ mobilization, oxidative stress and neurotransmitter response. Furthermore, donepezil pre-treatment significantly prevented isoflurane-induced cholinergic degeneration and spatial memory impairment in aged mice [27] and attenuated okadaic acid–induced memory impairment, mitochondrial dysfunction and apoptotic cell death [26]. Donepezil pretreatment also prevented streptozotocin (STZ)-induced memory deficits, oxidative stress, mitochondrial dysfunction and caspase-3 activity through the specific involvement of nAChRs [14].

In the light of these studies, it seemed interesting to assess the neuroprotective properties of long-term pre-lesion donepezil treatment in a rat model of BF cholinergic depletion induced by 192-immunoglobulin G (IgG)-saporin (Sap) injection. The resulting permanent and selective Sap-dependent loss of cholinergic BF neurons mimics neuropathological features and cognitive impairments associated with MCI and AD. In fact, Sap selectively causes death of cholinergic cells by inhibiting ribosomal protein synthesis when it is taken up into cells expressing the low-affinity neurotrophin receptor p75 [28]. In the past two decades, the availability of Sap allowed studying the role of BF cholinergic system in several cognitive functions and its implications in aging and dementia [29].

In the present study, we focused on the neuroprotective action of donepezil by investigating the influence of donepezil pre-treatment on cognitive deficits and neuronal impairment induced by intraparenchymal Sap injections into the medial septum (MS) and nucleus basalis magnocellularis (NBM). To achieve this aim, cognitive performance and caspase-3 activity levels of cholinergically depleted rats pre-treated with donepezil or saline were compared with those of pre-treated sham-lesioned animals. Cognitive function was analyzed by means of a wide battery of behavioral tests, including elevated plus maze (EPM), open field with objects (OF), radial arm maze (RAM), sociability and preference for social novelty test (PSNT) and fear conditioning (FC) with ultrasonic vocalization (USV) recording. The low-frequency USVs reflect a negative affective state [30, 31] and are positively correlated with the aversiveness of the situation [32]. After behavioral testing, neurodegeneration was analyzed by measuring caspase-3 activity in the hippocampus and neocortex, projection areas of the lesion sites. In fact, caspase-3 is the main effector caspase, whose localized activation can trigger synaptic loss, causing cognitive and behavioral deficits [33], whereas strong activation leads to switching on of the apoptotic cascade. Because the increase in caspase-3 activity has been proposed as an early neurodegenerative event in AD progression [34–36], its quantification might be useful in evaluating the efficacy of neuroprotective pharmacological treatment.

As ACh reduction is a morphofunctional hallmark in AD, AChE-Is, such as donepezil, are used to improve AD-related cognitive deterioration [7]. Besides symptomatic effects, many in vitro studies have shown that AChE-Is might exert neuroprotective action against neurotoxic agents, such as glutamate and Aβ plaques [8–11, 15–18]. Unfortunately, only a few in vivo experiments have confirmed AChE-I pre-lesion neuroprotective action [12–14, 26, 27]. Interestingly, the present study demonstrates a marked neuroprotective effect of donepezil against neurodegeneration and cognitive impairment induced by cholinergic depletion in rats. Donepezil pretreatment did not induce any biochemical or behavioral modification in sham-lesioned rats, which is in accord with the findings of Saxena et al. [14] and Su et al. [27]. We previously demonstrated that healthy rats treated with donepezil during behavioral testing showed enhanced memory and explorative functions in RAM and OF, as well as reduced anxiety levels in EPM, but they did not exhibit any improvement in spatial span, motivational levels and associative learning [41]. Thus, although donepezil treatment during testing is able to improve cognitive performance even in unlesioned animals, donepezil treatment before testing (pre-treatment) exerts neuroprotective action only in the presence of lesion or insult.

Cholinergic depletion per se (Sal-Sap group) increased hippocampal and neocortical caspase-3 activity and impaired working memory performance in RAM and spatial change detection in OF, altered social discrimination performance in PSNT and reduced USV production during FC, but it did not affect anxiety levels in EPM and FC acquisition. The behavioral results in our study are in line with those reported in previous studies of memory deficits [20, 29, 42, 43] and impaired production of USVs [38, 40, 43] and no effects on anxiety levels following 192 IgG-Sap lesion [20, 44]. Our present is the first one that explains caspase-3 activity in the target areas of cholinergic projections. Our results demonstrate that donepezil pretreatment (Don-Sap group) was able to counteract neurodegeneration associated with cholinergic depletion. In fact, hippocampal and neocortical caspase-3 levels observed in Don-Sap rats were significantly decreased in comparison to Sal-Sap. This result is in line with previous studies using donepezil pre-treatment in different experimental paradigms. In a glutamate excitotoxicity study, donepezil pre-treatment significantly prevented caspase-3 activation, contributing to internalization and downregulation of N-methyl-D-aspartate (NMDA) receptors through α₇ nAChR stimulation [15]. Moreover, following STZ lesioning, donepezil pre-treatment significantly prevented hippocampal and neocortical caspase-3 activity [14]. The reduction of hippocampal and neocortical caspase-3 activity by donepezil pre-treatment following cholinergic depletion adds further evidence to donepezil neuroprotective action. This effect could be ascribed to the up-regulation of nAChRs and to the consequent activation of the anti-apoptotic nAChR/PI3K/Akt/Bcl2 pathway, as previously hypothesized [8, 17]. Namely, several studies have demonstrated the involvement of nAChRs in donepezil-induced neuroprotection. In fact, in contrast to the AChE-Is galantamine and tacrine, prolonged donepezil administration induces not only a functional up-regulation but also an increased expression of nAChRs [9, 17]. Importantly, nAChR up-regulation by donepezil is able to activate intracellular anti-apoptotic secondary messenger systems, such as the nAChR/PI3K pathway, that protect neurons against neuron death [17]. Furthermore, the neuroprotective effects of donepezil appear to be mediated by σ₁ receptor activation that result in intracellular modulation of Ca²⁺ mobilization and activation of phospholipase C (PLC)/protein kinase C (PKC) transduction pathways, attenuating the effects of mitochondrial or endoplasmic reticulum dysfunction [12, 13].

The neuroprotective action of donepezil pre-treatment on hippocampal and neocortical neurodegeneration could account for the improved RAM and OF spatial performance of the pre-treated lesioned animals. Namely, in RAM, Don-Sap rats exhibited a lower number of errors and a higher spatial span than Sal-Sap rats, reaching the mnesic performance levels of sham-lesioned rats.

Beneficial effects of donepezil treatment on spatial working memory have been demonstrated in many experimental models. In fact, donepezil administration reduced RAM errors in aged or scopolamine-treated animals [45, 46]. Working memory improvement has been reported in aged mice treated with donepezil or a nicotinic agonist α₇ (starting one week before behavioral training) [23]. In addition, in a rat model of cholinergic depletion, donepezil administration (starting 1 week before and lasting 9 weeks after lesioning) mitigated procedural and working memory deficits in RAM [20]. In hypoxia and Aβ toxicity models, donepezil pre-treatment prevented spatial working memory deficits [12, 13].

Even in OF, although cholinergic depletion per se (Sal-Sap group) significantly affected spatial discrimination abilities, donepezil pre-treatment (Don-Sap group) prevented the spatial memory deficits. Consistent with the previously described impaired spatial change and preserved novelty recognition following Sap lesions [20, 42, 43, 47], all lesioned rats detected the presence of the novel object, just as sham-lesioned rats did. In fact, though spatial change detection implies efficient spatial mapping, novelty detection implies recognition of the salient object, thus minimizing the effect of the spatial and contextual factors [38, 48, 49]. Interestingly, the preserved abilities in detecting spatial change of Don-Sap rats reveal the specific neuroprotective effect of donepezil on a subtle spatial deficit induced by the cholinergic lesion and are in agreement with donepezil neuroprotective properties on spatial mnesic performances in different brain damage models [12–14, 26, 27].

In all animals, the shock exposure induced a clear freezing response during FC training, and the conditioned freezing response was retained during context and tone tests. Although the lesion did not alter freezing response, it decreased the concomitant production of 22-kHz USVs. Interestingly, donepezil pre-treatment during FC training just tended to rescue the defect of fear-associated USVs, which were nearly totally suppressed in Sal-Sap rats. The 22-kHz USVs are part of the rodent defensive repertoire and are emitted in response to negative experiences, such as foot-shock, presence of predators or startling noise exposure [30–32]. These USVs are closely associated with freezing response to actual or potential threats and might serve as “alarm calls” for conspecificity. There is converging evidence that Sap lesions selectively impair the functionality of neural systems regulating USVs [38, 40, 43]. Overall, our data are in agreement with the previously described disassociation between intact freezing response and impaired USV production in cholinergically depleted rats [38]. Intriguingly, donepezil pre-treatment (Don-Sap group) appeared to mitigate the USV suppression of lesioned rats during FC training. Thus, our data corroborate the added value of USV measurement, which provides additional information about the emotional state of the animals in situations of coping with inescapable aversive stimuli.

Also, social interactions and recognition of conspecificity are fundamental and adaptive components of the behavioral repertoire of numerous species. Similarly to social transmission of food preference, social recognition memory, which relies on the functional activation of hippocampal and prelimbic cortical circuitry [50], is impaired in the early phases of AD. Clinical trials have demonstrated that donepezil enhances AD patients’ social interaction, engagement and interest [51]. On these bases, in the present research, we investigated for the first time donepezil pre-treatment effects on social skills using the three-chamber test. During sociability tests, all animals showed a clear social motivation. In fact, all of them preferred the conspecific stranger, as they spent more time and traveled longer distances in the chamber containing the social stimulus. These results fit with those obtained in the sociability test described by Riedel et al. [50], but they seem to be in disagreement with those obtained by Savage et al. [47]. On the basis of use of a social interaction test, these authors reported a significant reduction in active interactions following selective cholinergic depletion of the neocortex in rats. Notably, at variance with the social interaction test, sociability does not allow any direct contact with the social stranger. Thus, in the present study, the indirect social interactions and the scarce salience of the empty compartment featuring sociability appear to have facilitated the exploration of the stranger by the lesioned rats.

During PSNT, though sham rats showed a clear preference for the novel stranger, the cholinergically depleted animals did not exhibit overt social novelty. This result has to be interpreted as a specific social recognition memory deficit (as described by Riedel et al. [50]), because the OF lesioned rats did recognize the novel object, which is in line with the findings reported by Savage et al. [47]. Although Riedel et al. [50] reported that donepezil administration succeeded in rescuing social memory scopolamine-induced deficits, donepezil pre-treatment failed to prevent PSNT deficits in lesioned rats in our present study. The different results could be explained by methodological differences, such as treatment time (acute vs. pre-treatment), behavioral protocols (short-term vs. long-term memory) and cholinergic manipulations (scopolamine vs. Sap). Notably, hippocampal and neocortical cholinergic deafferentation by Sap results in massive dysregulation of other neurotransmitter systems, such as dopaminergic and glutamatergic ones, which is known to contribute to social discrimination [46, 52, 53].

Our current research reveals that donepezil pre-treatment is able to reduce hippocampal and neocortical caspase-3 activity, thus preventing neuron degeneration, and to exert beneficial effects on specific behavioral deficits induced by cholinergic depletion. As indicated in previous studies, donepezil neuroprotective effects might be mediated by many protective mechanisms, such as nAChR upregulation and activation of the nAChR/PI3K pathway [8, 15, 17] and the σ₁ receptor/PLC/PKC pathway [12, 13]. Such effects result in a reduction of neurotoxicity linked to NMDA receptor–mediated Ca²⁺ influx, oxidative stress and caspase-3 activity [14, 15]. Furthermore, donepezil exerts a protective action against Aβ toxicity [7]. In fact, in AD patients, Aβ plaques colocalize with nAChRs and α₄ and α₇ nAChR expression is reduced. In CA1, the α4-nAChR activation causes γ-aminobutyric acid (GABA) release from interneurons that inhibit pyramidal neurons, whereas the activation of α₇ nAChRs in pyramidal neurons results in Ca²⁺ influx, presynaptic neurotransmitter release and postsynaptic depolarization [7]. Thus, Aβ interferes with nAChR activity, and the consequent increase of glutamate and decrease of GABA induce glutamate excitotoxicity and excitation–inhibition imbalance, altering the fine-tuning of hippocampal firing [6]. In AD patients, increase of caspase-3 activity has been reported in the hippocampal and neocortical postsynaptic density fractions [54]. Also, in the Tg2576 AD mouse model, the increase of caspase-3 in hippocampal postsynaptic compartment leads to alteration of synaptic plasticity and dendritic spine loss. In parallel, the Sap-induced cholinergic depletion mimicking Aβ interference might cause an alteration of the excitation–inhibition balance and produce excitotoxic damage in hippocampal and neocortical neurons, which accumulate active caspase-3. On these bases, the question is: How can donepezil neuroprotection be exerted? By inhibiting AChE activity, the donepezil pre-treatment might reduce the GABAergic alterations and prevent glutamatergic excitotoxicity and the excitation–inhibition imbalance. In this way, caspase-3 accumulation would be decreased, contributing to the maintenance of hippocampal and neocortical functioning.

The present results show for the first time that donepezil pre-treatment is able to slow down the memory deficits induced by cholinergic depletion and to reduce caspase-3 accumulation in hippocampal and neocortical areas. Although further studies deepening understanding of the molecular mechanisms of donepezil neuroprotective action are needed, the present results are promising and may lead to the development of novel strategies for prevention and therapy of neurodegenerative diseases.