Improvement of memory for context by inhibition of caspase-1 in aged rats

Impaired learning and memory is a common pathologic feature associated with numerous neurologic disorders. There is strong evidence that central inflammation contributes significantly to the progression of several neurodegenerative diseases as well as to the ageing process. For example, in aged rats an increase in interleukin‐1β (IL‐1β) is implicated in the decline of synaptic plasticity in the hippocampus and impaired performance on cognitive tasks such as contextual fear conditioning. IL‐1β is a proinflammatory cytokine initially synthesized in an inactive precursor form that is cleaved by caspase‐1 to generate the biologically mature form. In the present study, cleavage of IL‐1β was chronically inhibited using a specific caspase‐1 inhibitor (Ac‐YVAD‐CMK; 10 pmol) in both aged (22 month) and young (4 month) rats. Both groups received Ac‐YVAD‐CMK for 28 days intracerebroventricularly through a brain infusion cannula connected to an osmotic minipump. On day 20 the animals were trained in contextual fear conditioning, and memory for context was tested on day 22. Chronic infusion of a specific caspase‐1 inhibitor in aged rats ameliorated age‐related increases in hippocampal IL‐1β and improved memory for context.

Keywords: ageing; caspase; inflammation; interleukin‐1β; learning and memory; tumour necrosis factor α

Interplay between the immune system and the central nervous system underlies the neurophysiologic changes that occur with ageing ([33]; [15]). Cytokines are polypeptides normally expressed at low levels in healthy tissue and that are rapidly induced in response to trauma or immune challenge. Brain levels of certain cytokines increase as a function of age even in the absence of a pathologic stimulus. For example, there is a progressive increase in interleukin (IL)‐1β levels in the brain and in microglia activation in neurologically intact ageing patients; ([11]; [35]; [41]). IL‐6 levels are also increased in the ageing mouse brain ([42]). We previously demonstrated that tumour necrosis factor‐alpha (TNFα) is dramatically increased in the cerebellum of aged rats compared with young rats, and such an increase can be prevented through a diet rich in antioxidants ([17]). Inflammatory cytokines disrupt normal physiology and contribute to age‐dependent deficits in behavioural function ([27]; [29]; [23]; [45]). IL‐1β is one of the most studied cytokines in regard to cognition. Peripheral administration of IL‐1β induces diverse cognitive‐behavioural effects, including decreased exploratory behaviour, decreased locomotor activity, inhibition of sexual behaviour, sleep promotion, anorexia and anxiogenic reactions ([37]; [28]; [7]; [44]; [20]). In particular, IL‐1β is linked to hippocampal‐dependent cognitive processes in both physiologic and pathophysiologic contexts. IL‐1β exerts its effect by binding to a high affinity receptor, the IL‐1 type 1 receptor ([33]; [8]; [9]). The IL‐1 type 1 receptor is constitutively expressed in the hippocampus ([12]; [4]; [34]), an area of the brain that has a critical role in memory and learning. The highest density of IL‐1β binding sites is in the dentate gyrus of the hippocampus ([12]; [38]). These findings suggest that the effects of IL‐1β are specific to hippocampal‐dependent memory processes. Additionally, IL‐1β is involved in neuronal plasticity as revealed by the fact that IL‐1β impairs long‐term potentiation (LTP), a model system for the neural mechanism underlying hippocampal‐dependent memory in aged rats ([22]; [2]). Moreover, mice with targeted deletion of the IL‐1 type 1 receptor gene exhibit impairments in memory and plasticity ([18]), providing support for a physiologic role of IL‐1β in cognition. Furthermore, central administration of IL‐1 receptor antagonist, an endogenous molecule that binds to IL‐1 type 1 receptors to block the binding and biologic activity of IL‐1, impairs the maintenance of LTP ([6]) and fear‐conditioning potentiation ([31], [29]) and performance in young rats in the Morris water maze ([45]). Despite extensive evidence that IL‐1 is required for normal learning and memory processes, a large body of literature indicates that at pathophysiologic levels IL‐1β is detrimental to hippocampal‐dependent learning and memory processes. The most direct evidence that IL‐1β impairs hippocampal‐dependent memory comes from studies demonstrating that central administration of low doses of IL‐1β in young rats impairs contextual fear conditioning ([25]; [31], [29]; [3]). Deficits in learning and memory in fear conditioning and Morris water maze paradigms also occur in rats peripherally injected with lipopolysaccharide ([30]), a component of gram‐negative bacterial cell walls and a potent inducer of IL‐1β both centrally and peripherally. We previously demonstrated that contextual fear conditioning is impaired in aged rats and administration of a nonsteroidal anti‐inflammatory drug, sulindac, reverses the impairment; in addition, sulindac decreased IL‐1β levels ([26]).

Thus, our hypothesis for the present study is that physiologic levels of IL‐1β are critical for normal memory formation in the hippocampus and increased IL‐1β levels observed with age might underlie age‐related impairments in hippocampal‐dependent learning and memory. Therefore, blocking IL‐1β production in young rats should impair learning and memory, while in aged rats blocking IL‐1β should improve learning and memory. Although several studies have examined the involvement of IL‐1β in a fear conditioning paradigm in young animals, no studies have examined hippocampal‐dependent memory deficits in aged animals in response to the inhibition of endogenously increased levels of IL‐1β. IL‐1β modulates the expression of other cytokines, thereby initiating a cascade of molecular changes ([10]). Thus, to determine whether cytokines other than IL‐1β are indirectly involved in the beneficial effects of a caspase‐1 inhibitor on memory for context, we performed a cytokine assay for simultaneous quantification of IL‐1α, IL‐1β, IL‐2, IL‐6, IL‐10, interferon (IFN)γ and TNFα. Finally, because caspase‐1 directly affects caspase‐3 activity ([14]), its inhibition could reduce caspase‐3 enzymatic activity, which might be partially responsible for the observed changes in behaviour. Thus, we analysed the effect of caspase‐1 inhibition on the enzymatic activity of caspase‐3 in both young and aged rats.

All experiments were performed in accordance with the National Institute of Health Guide for the Use of Laboratory Animals.

This study was conducted on male Fischer 344 rats (NIA contract colony, Harlan Sprague Dawley, Indianapolis, IN, USA). Rats were pair‐housed and maintained in environmentally controlled chambers on a 12 : 12‐h light : dark cycle at 21 ± 1 °C, and provided with food and water ad libitum. We used groups of eight aged (20‐month) or young (4‐month) Fischer 344 rats. There were four subgroups: aged caspase‐1 inhibitor (Ac‐YVAD‐CMK), aged controls, young Ac‐YVAD‐CMK and young controls. The rats were implanted with an intracerebroventricular brain infusion cannula in the left lateral ventricle and a subcutaneous osmotic minipump (Alzet, model 2004) for 28 days. Prior to implantation, pumps were primed by incubation in sterile saline for 48 h at 37 °C. For implantation, rats were anaesthetized with isoflurane and placed in a stereotaxic frame. A guide cannula (28‐gauge) was stereotaxically implanted into the left ventricle (AP, −1.0; ML, 1.6; DV, −3.5 mm) and connected to the osmotic minipump, which was inserted subcutaneously. Minipumps were weighed before implantation and at the end of the experiment to ensure complete delivery of their contents. Ac‐YVAD‐CMK (10 pmol in 200 µL; Calbiochem, La Jolla, CA, USA) was infused through the cannula via the minipump, delivering Ac‐YVAD‐CMK at a rate of 0.25 µL/h, with a total volume of 200 µL. The infusion began on the day of surgery and continued for 28 days. Control animals received the same amount of 0.6% DMSO (Sigma Aldrich, St Louis, MO, USA) in saline.

On day 20, the rats were trained in the contextual fear conditioning task. On day 22, the animals were tested in the training context and in an altered context with and without the conditioned stimulus (CS; details below) tone. The contextual fear conditioning task tests two forms of memory, contextual (hippocampus‐dependent) and auditory‐cue‐specific. In this task, animals learn to associate both a specific cue (noise) and nonspecific cue (training context) with a mild footshock. The rats develop fear, which is observed as behavioural immobility (freezing) for both the auditory‐cue‐specific and the training context. We used previously published methods to examine contextual fear conditioning in young and aged rats ([26]). Briefly, rats were placed in a box (30.5 × 24.1 × 21 cm, Medical Associates, St Albans, VT, USA) with a grid floor (4.8‐mm‐diameter rods, spaced 1.6 cm apart) connected to a scrambled constant current source (Medical Associates). Immediately prior to placing each rat in the box, the box was cleaned with 3% acetic acid, which functioned as a specific odourant for the training context. Two consecutive training blocks were administered. Each training block was 180 s long with a 30‐s, 85‐dB 3 kHz CS tone and a 2‐s, 0.5‐mA footshock (unconditioned stimulus; US). The CS tone and US coterminated at the end of the training block. All rats reacted to the footshock by jumping. The rats remained in the training box for 30 s following the second training block. Retention was tested 2 days after training by placing the rat in the same apparatus, again using 3% acetic acid as an odourant. Training was performed for 5 min without presentation of the CS tone or US. Two to three hours later, the rats were placed in an altered context chamber (the same box, but with the floor covered with dark Plexiglas and two stationary objects placed in the box; cleaned with 3% ammonium hydroxide as a novel odourant) for 6 min during which the CS tone was administered for the final 3 min. Freezing was quantified automatically using a video‐based conditioned fear testing system (FreezeFrame, Actimetrics Software, Evanston, IL, USA). The software package allowed for simultaneous visualization of the animal's behaviour and adjustment of a 'freezing threshold' that defined the behaviour as freezing or not freezing. The freezing threshold was confirmed separately for each animal. The experimenter assessing the freezing behaviour was blind to the assigned treatment groups to avoid bias. Each trial was divided into 1‐s bins. Freezing was defined as no movement above the freezing threshold for the entire duration of each bin. Total freezing was summed over the length of the trial. Freezing was presented as the percentage time spent freezing (100 × time spent freezing/total time). Three measures of freezing were examined: freezing to the training context associated with the shock, freezing to the altered context, and freezing to the CS tone.

To avoid nonspecific and stress‐related alterations in the gene expression of proinflammatory cytokines, the animals were killed 7 days after completing behavioural testing. The rats were decapitated, brains were removed, and the brain regions were dissected. Hippocampal tissues were dissected from both hemispheres and collected separately. All tissues were frozen in liquid nitrogen and stored at −80 °C until assayed. Tissues from both the right and left hemisphere were randomly used for each analysis. Three animals were excluded due to illness acquired during the study.

Homogenization of tissues and determination of total protein and enzyme‐linked immunosorbent assay (ELISA) were performed on the same day to avoid repetitive thawing of samples. Hippocampal tissues were mechanically homogenized, using an ultrasonic cell disrupter, in ice‐cold lysis buffer 1, 1 : 5 w/v, containing HEPES, 25 mm, pH 7.4; [3‐(–cholamidopropyl) dimethyl‐ammonio]1‐propanesulphonate, 0.1%; MgCL2, 5 mm; EDTA, 1.3 mm; EGTA, 1 mm; pepstatin, aprotinin and leupeptin, 10 µg/mL each; and phenylmethylsulphonyl fluoride, 1 mm. Sonicated samples were centrifuged at 9300 g at 4 °C for 30 min and supernatant was collected in pyrogen‐free tubes and kept at 4 °C until the assay was performed (≈ 1 h). Total protein concentration from the supernatant was determined using a Bradford protein assay (BIO‐RAD Laboratories, Hercules, CA, USA). The ELISA for rat IL‐1β was performed using a commercially available kit (Bio‐Source International, Camarillo, CA, USA) following the manufacturer's protocol. The detection limit was 3 pg/mL.

Cleavage of the caspase‐1 substrate YVAD‐AFC (Alexis Biochemical, San Diego, CA, USA) and the caspase‐3 substrate DEVD‐AFC (BD Bioscience, Palo Alto, CA, USA) to their fluorescent products was used as a measure of caspase‐1 and caspase‐3 activity. The hippocampus of each hemisphere was separately dissected and homogenized in 400 µL of lysis buffer 2 (in mm: HEPES, 25; MgCl2, 5; dithiothreitol, 5; EDTA, 5; phenylmethylsulphonyl fluoride, 2; and proteinase inhibitors cocktail, 0.4 µL, pH 7.4). The homogenate was centrifuged at 21 000 g for 30 min at 4 °C. The supernatant (50 µL) was added to 50 µL of 2 × the kit reaction buffer with either 5 µL of YVAD‐AFC (caspase‐1 substrate, 1 mm; final concentration 50 µm) or 5 µL of DEVD‐AFC (caspase‐3 substrate, 1 mm; final concentration 50 µm) and incubated in a water bath at 37 °C for 3 h. Fluorescence was assessed using a fluorometer with a 400‐nm excitation filter and a 505‐nm emission filter.

The Bio‐Plex Cytokine Assay (BIO‐RAD) is a Luminex bead‐based immunoassay consisting of premixed beads coated with target capture antibody, which allows the simultaneous quantification of a panel of cytokines. The rat cytokine assay for simultaneous quantification of IL‐1α, IL‐1β, IL‐2, IL‐6, IL‐10, IFNγ and TNFα was performed according to the manufacturer's protocol. The detection limit of the multiplex cytokine assay was < 10 pg/mL for each cytokine.

The behavioural data were analysed with a three‐way repeated‐measures anova over the three testing paradigms. Post hoc between‐task within‐subjects comparisons were made with two‐tailed paired Student's t‐tests. The biochemical data were analysed with 2 × 2 Age × Treatment factorial anova. Two values > 2 SD from the population mean were excluded from analysis as statistical outliers. A P‐value of < 0.05 was considered statistically significant. Statistical analyses were performed using StatView (version 5.0; SAS, Cary, NC, USA).

Freezing to the training context, CS tone, and altered context 2 days after training are shown in Fig. 1. The results revealed several main effects and interactions. Specifically, there was a significant main effect of test condition (F1,24 = 103.85, P < 0.001), with greater freezing shown in response to context and auditory cue than to alternate context. The main effect of age was also statistically significant (F1,24 = 8.04, P = 0.009), with younger rats showing more freezing than older rats. The main effect of caspase‐1 inhibition was not statistically significant. Several interactions were also statistically significant, including the test condition by age (F1,24 = 5.58, P = 0.027) and the three‐way interaction for test condition × age × caspase‐1 inhibition (F1,24 = 5.43, P = 0.028). Post hoc analysis demonstrated that aged control rats froze significantly less to the training context than did young control and young Ac‐YVAD‐CMK rats (all P < 0.05). There was no difference between aged Ac‐YVAD‐CMK rats and young controls, and there was a significant difference between aged Ac‐YVAD‐CMK rats and aged controls (t‐test, P < 0.05). Aged rats froze significantly less to the CS tone than did the young rats (P < 0.02). Caspase‐1 inhibition had no effect on freezing to the CS tone in either young or aged rats. Young controls, young Ac‐YVAD‐CMK and aged Ac‐YVAD‐CMK froze significantly more to the training context than to the altered context (paired t‐test; all P < 0.01). Young controls and young Ac‐YVAD‐CMK froze significantly more to the CS tone than to the altered context (paired t‐test; all P < 0.05).

Graph: 1 (A) Chronic administration of caspase‐1 inhibitor improves age‐dependent deficits in contextual hippocampal memory. Twenty‐two days of intracerebroventricular infusion of Ac‐YVAD‐CMK (10 pm) ameliorated age‐related deficits in contextual fear conditioning in aged animals. Behavioural freezing was measured in animals 48 h after training. In contextual fear conditioning, in which hippocampal‐dependent memory (freezing to the original context) and CS tone‐dependent memory (freezing to the auditory cue) were examined, there was an age‐related deficit in hippocampal‐dependent memory. Aged control animals exhibited significantly less freezing to the original context than young controls. Aged Ac‐YVAD‐CMK rats froze more that the aged controls. There was no difference in percentage freezing between young control and young Ac‐YVAD‐CMK. Two to three hours after exposure to the training context, animals were placed in a novel environment for 6 min. During the first 180 s there was no tone. (B) Aged animals froze significantly more to the CS tone than did young animals. (C) In the altered context there was no significant difference between any of the groups. Data are expressed as mean percentage freezing during context and altered context. *P < 0.05.

The effects of age and caspase‐1 inhibition on hippocampal caspase‐1 activity and IL‐1β are shown in Fig. 2. There was a significant main effect of age on caspase‐1 activity (F1,18 = 4.95. P < 0.05), but no main effect of treatment (F1,18 = 3.69. P = 0.07; Fig. 2A). Caspase‐1 activity was significantly higher in aged control rats than in young controls (P < 0.02). Caspase‐1 activity was decreased in the aged treatment group when compared with the aged control group (P = 0.02). Caspase‐1 inhibition did not alter caspase‐1 activity in young animals. While there were no significant main effect of age (F1,21 = 0.72, ns) or treatment (F1,21 = 2.95, ns) on IL‐1β levels, there was a significant age × treatment interaction (F1,21 = 5.05, P < 0.05; Fig. 2B). IL‐1β levels in aged controls were significantly higher than in the young controls (P < 0.05). Caspase‐1 inhibition blocked this age‐related increase in IL‐1β levels (P < 0.02). There was no effect of caspase‐1 inhibition on IL‐1β levels in young rats. Consistent with the literature, caspase‐1 activity, and consequently hippocampal levels of IL‐1β, were not altered in the young Ac‐YVAD‐CMK animals ([40]), and there was no corresponding change in freezing to the context.

Graph: 2 Effects of intracerebroventricular infusion of Ac‐YVAD‐CMK on caspase‐1 activity and IL‐1β concentration, in hippocampus homogenate of aged and young rats. (A) Caspase‐1 activity was determined by measuring cleavage of the caspase‐1 substrate YVAD‐AFC to its fluorescent product. Caspase‐1 enzymatic activity was higher in aged rats than in young rats. Infusion of Ac‐YVAD‐CMK for 22 days significantly reduced caspase‐1 enzymatic activity in aged rats but not in young rats. (B) There was a corresponding age‐dependent increase in hippocampal IL‐1β levels in aged rats. Twenty‐two days of Ac‐YVAD‐CMK (10 pmol) infusion restored IL‐1β levels to control levels. There was no effect of Ac‐YVAD‐CMK observed in the hippocampal IL‐1β levels in young rats. Data are expressed as mean ± SEM; **P < 0.01, *P < 0.05 compared to young controls; †P < 0.05 compared to aged controls.

The effects of age and caspase‐1 inhibition on hippocampal TNFα and IL‐10 levels are shown in Fig. 3. There was no main effect of age on TNFα levels (F1,18 = 2.90, ns), and there was no significant effect of treatment (F1,18 = 1.02, ns; Fig. 3A). There was a significant age × treatment interaction (F1,18 = 8.77, P < 0.01). Aged control rats had a significant increase in TNFα (59%, P < 0.01) compared to young controls. Caspase‐1 inhibition blocked the age‐related increase in TNFα (P < 0.05) in aged Ac‐YVAD‐CMK rats, which were not different from young controls. There was no effect of caspase‐1 inhibition on TNFα levels in the young Ac‐YVAD‐CMK rats. Age had no significant effect (F1,17 = 2.44, ns) on IL‐10 levels, but there was a significant effect of treatment (F1,17 = 6.11, P < 0.05; Fig. 3B). There was no significant age × treatment interaction (F1,17 = 3.58, P = 0.08); however, a strong trend was apparent. Individual post hoc comparisons revealed a significant decrease in IL‐10 in aged controls rats compared with young controls. (P < 0.05). Aged treated rats, however, had significantly higher levels of IL‐10 when compared with aged controls (P < 0.01). The other cytokines measured in the Multiplex ELISA assay were below the limits of detection.

Graph: 3 Infusion of Ac‐YVAD‐CMK decreased the TNFα and increased IL‐10 concentrations in hippocampal homogenate of aged rats. (A) The simultaneous analyses of a panel of cytokines, including IL‐1α, IL‐1β, IL‐2, IL‐6, IL‐10, IFNγ and TNFα, indicated that there was an age‐dependent increase in TNFα in the hippocampus of aged rats, and an age‐dependent decrease in IL‐10. The other cytokines measured in the Multiplex ELISA assay were below the limits of detection. (A) Ac‐YVAD‐CMK significantly reduced the increased levels of TNFα observed in aged rats, to the levels of young rats. (B) The levels of IL‐10 were significantly increased in the aged Ac‐YVAD‐CMK rats. Inhibition of caspase‐1 did not affect the levels of TNFα and IL‐10 in young animals. *P < 0.05, **P < 0.01 compared to young controls; †P < 0.05, ‡P < 0.01 compared to aged controls.

The effects of caspase‐1 inhibition on hippocampal caspase‐3 activity are shown in Fig. 4. There was a significant effect of age on caspase‐3 activity (F1,17 = 8.74, P < 0.01) but no effect of treatment (F1,15 = 1.81, ns). There was no age × treatment interaction (F1,17 = 1.30, ns). Aged controls had significantly higher caspase‐3 activity than did young controls (P < 0.05). The aged Ac‐YVAD‐CMK rats were not significantly different from young controls (ns) or aged controls (P = 0.06). There was no effect of caspase‐1 inhibition on caspase‐3 activity in young Ac‐YVAD‐CMK rats.

Graph: 4 Hippocampal caspase‐3 activity was increased in aged control rats. Given that caspase‐1 is able to directly process pro‐caspase‐3 to its active form, we analysed the effect of the inhibition of caspase‐1 on caspase‐3 hippocampal activity by cleavage of the caspase‐3 substrate DEVD‐AFC to its fluorescent product. Caspase‐3 activity was higher in aged control rats than in young controls. Treatment with Ac‐YVAD‐CMK did not significantly reduce caspase‐3 enzymatic activity in the young or aged hippocampus. Data are expressed as mean + SEM; *P < 0.05 compared to young controls.

The present study examined whether the increase in IL‐1β observed with age is a critical factor that leads to impaired hippocampal‐dependent learning and memory. To examine this hypothesis, the irreversible caspase‐1 inhibitor Ac‐YVAD‐CMK was used to block caspase‐1 activity, thereby inhibiting synthesis of the active mature form of IL‐1β, in young and aged rats. Caspase‐1 activity was inhibited in aged rats as evidenced by the decrease in hippocampal caspase‐1 activity. There was no effect of Ac‐YVAD‐CMK on caspase‐1 activity or IL‐1β levels in young rats, suggesting that young rats were able to compensate for the effects of the inhibitor. Aged control rats had impaired memory for the training context compared to young controls. Chronic inhibition of caspase‐1 activity ameliorated this age‐related memory deficit. Consistent with the hypothesis that elevated hippocampal IL‐1β levels impair memory in aged rats, the aged control rats had higher hippocampal levels of IL‐1β than did the young controls. Caspase‐1 inhibition in aged rats reduced hippocampal IL‐1β levels, approaching the levels of young controls. Consistent with these results, [40]) reported a correlation between an age‐related LTP impairment and increased caspase‐1 activity and IL‐1β concentrations in the hippocampus from aged rats. Moreover, they revealed that lipopolysaccharide‐induced inhibition of LTP is mediated by the activation of caspase‐1. Together these findings strongly support the hypothesis that IL‐1β is involved in hippocampal‐dependent memory processes as well as in neural plasticity. The results of the present study support a causal link between caspase‐1 activity and age‐related alterations in cognition. This relationship might possibly be a consequence of the age‐dependent increase in IL‐1β, which modulates hippocampal cognitive processes ([29]). It is also possible, however, that more than one mechanism is involved in the observed effect. In fact, in conjunction with the decrease in IL‐1β in the hippocampus of aged rats, there was a parallel inhibition of TNFα expression. Furthermore, there was an increase in the anti‐inflammatory cytokine IL‐10. The possibility that the above changes result from a downstream effect caused by a decrease in caspase‐1 activity is supported by the fact that IL‐1β is one of the main promoters of the immune cascade, which results in an exacerbated inflammatory response ([10]). In a similar study in which Ac‐YVAD‐CMK was centrally administered to test the hypothesis that caspase‐1 inhibition reduced neurodegeneration following ischemic insult ([32]), the same downstream effects were observed with the exception that IL‐10 levels were not altered. This discrepancy with the present study is probably due to the different conditions (dose and brain region examined) used in the studies. The decrease in TNFα as well as the increase in IL‐10 might have a role in the beneficial effects exerted by Ac‐YVAD‐CMK on age‐dependent memory loss. For example, IL‐10 attenuates the inhibitory effects of IL‐1β on LTP potentiation ([19]). IL‐1β stimulates the production of TNFα and IL‐6 while simultaneously decreasing IL‐10 levels. Thus, the Ac‐YVAD‐CMK‐induced decrease in IL‐1β decreases TNFα and IL‐6, leading to an increase in IL‐10, which in turn decreases IL‐6 ([43]).

Neural apoptosis is implicated in cognitive deficits. For example, mice transgenic for the bcl‐2 gene exhibit enhanced learning and have an increased cell number in the brain due to blockade of apoptosis ([5]). In addition, mice lacking DNA fragmentation factor 45 have more cells in the dentate gyrus and perform better than control mice in the Morris water maze, a hippocampal‐dependent paradigm ([36]). TNFα can induce apoptosis by activating caspase‐3 and/or caspase‐8, and its apoptotic effects are modulated by IL‐1β ([13]; [21]; [4]). In contrast, IL‐10 decreases apoptotic markers, by decreasing caspase‐3 activity ([1]). Thus, we cannot exclude the possibility that the improvement we observed in contextual memory in the present study is due to the direct action of TNFα or IL‐10 on the apoptotic pathway as there was a trend for decreased capsase‐3 activity in the aged Ac‐YVAD‐CMK‐treated rats (P = 0.06). The caspase‐1 inhibitor might directly inhibit caspase‐3; however, a high concentration of caspase‐1 inhibitor does not affect caspase‐3 activity in vitro ([16]). Given the fact that caspase‐1 directly processes pro‐caspase‐3 to its mature form ([39]) it might be more likely that a trend for reduced caspase‐3 activity is a downstream effect of caspase‐1 inhibition. In support of this, IL‐1β stimulates caspase‐3 activity in hippocampal tissue of aged rats and IL‐10 inhibits the activity of apoptotic protein, including caspase‐3 ([1]; [24]). Further experiments are required to examine the role of caspase‐3 and apoptosis in the effects observed here. Moreover, inhibition of caspase‐1 blocked the initiation of inflammatory processes and, simultaneously, might have interfered with the activation of the apoptotic pathway. One caveat is that previous research indicates that the ELISA used in this study does not differentiate between pro‐ and mature IL‐1β; however, IL‐1β can only be produced by pro‐IL‐1, indicating that the changes observed are a consequence of altered caspase‐1 activity. These data are the first evidence of the involvement of caspase‐1 in a hippocampal‐dependent memory task.

• Abbreviations
• Ac‐YVAD‐CMK caspase‐1 inhibitor
• CS conditioned stimulus
• ELISA enzyme‐linked immunosorbent assay
• IFN interferon
• IL interleukin
• LTP long‐term potentiation
• TNFα tumour necrosis factor‐alpha
• US unconditioned stimulus