Address correspondence and reprint requests to Dr. C.E. Stafstrom at Department of Neurology, H6-528, University of Wisconsin, 600 Highland Avenue, Madison, WI 53792, U.S.A. E-mail: firstname.lastname@example.org
Present address of Dr. Sayin: Institute of Forensic Sciences, Istanbul University, Cerrahpasa, Istanbul, Turkey.
Summary: Purpose: Seizures in the developing brain cause less macroscopic structural damage than do seizures in adulthood, but accumulating evidence shows that seizures early in life can be associated with persistent behavioral and cognitive impairments. We previously showed that long-term spatial memory in the eight-arm radial-arm maze was impaired in rats that experienced a single episode of kainic acid (KA)-induced status epilepticus during early development (postnatal days (P) 1–14). Here we extend those findings by using a set of behavioral paradigms that are sensitive to additional aspects of learning and behavior.
Methods: On P1, P7, P14, or P24, rats underwent status epilepticus induced by intraperitoneal injections of age-specific doses of KA. In adulthood (P90–P100), the behavioral performance of these rats was compared with that of control rats that did not receive KA. A modified version of the radial-arm maze was used to assess short-term spatial memory; the Morris water maze was used to evaluate long-term spatial memory and retrieval; and the elevated plus maze was used to determine anxiety.
Results: Compared with controls, rats with KA seizures at each tested age had impaired short-term spatial memory in the radial-arm maze (longer latency to criterion and more reference errors), deficient long-term spatial learning and retrieval in the water maze (longer escape latencies and memory for platform location), and a greater degree of anxiety in the elevated plus maze (greater time spent in open arms).
Conclusions: These findings provide additional support for the concept that seizures early in life may be followed by life-long impairment of certain cognitive and behavioral functions. These results may have clinical implications, favoring early and aggressive control of seizures during development.
Epilepsy is a common neurologic disorder that occurs much more frequently in children than in adults (1). Many data support the idea that prolonged or frequent seizures in young animals and patients lead to later cognitive deficits, which can often be subtle [e.g., (2–11)]. In both laboratory animals and humans, the brain's susceptibility to seizures is increased during specific developmental epochs, yet the behavioral consequences of seizures tend to be less pronounced when seizures occur early in life than in adulthood. Reasons for these age dependencies are not fully understood but are clearly complex (12–14).
To investigate the long-term consequences of epilepsy early in life, several animal models have been developed (3,6,7,10,13,15). Most of these models use chemoconvulsants to induce acute seizures, mimicking status epilepticus. Later, a variety of tests assessing behavioral and cognitive function, seizure susceptibility, and histologic evidence of neuronal damage is applied. For example, kainic acid (KA) has been used to investigate the behavioral consequences of status epilepticus at various ages. In adult animals, KA causes an epilepsy syndrome similar to human temporal lobe epilepsy, with mesial temporal sclerosis, spontaneous seizures, and significant deficits in learning and memory (16). In developing animals compared with adults, however, the long-term effects of KA-induced status epilepticus are less severe in terms of macroscopic structural damage and behavioral impairments (17–21). When administered intraperitoneally at ages younger than about postnatal day (P) 20 in rats, KA does not cause significant hippocampal cell loss or recurrent seizures, and synaptic reorganization (mossy fiber sprouting) is less pronounced than that in adults, even if the seizures are more severe in the pups (19). These observations are not restricted to the KA model; seizures induced by kindling, pilocarpine, and flurothyl are likewise not associated with significant synaptic reorganization, as assessed by mossy fiber sprouting (22,23). Nevertheless, seizure-induced behavioral impairments may occur in later life in response to status epilepticus during early development (24–26).
Depending on the method of seizure induction, dose, age of administration, and behavioral testing, investigators have found varied effects of early seizures on subsequent learning and behavior (25–29). We recently provided additional evidence for long-term adverse cognitive changes after KA-induced status epilepticus in early development (30). KA seizures were evoked during early postnatal development (P1–P24). A seizure between P1 and P14 was associated with life-long impairment in spatial learning and memory in the radial-arm maze. These rats also exhibited impaired induction of hippocampal long-term potentiation (LTP). Seizure-induced alteration of synaptic plasticity also was manifested as a long-term reduction in susceptibility to kindled seizure development and enhanced paired-pulse inhibition in the dentate gyrus (30). Such chronic behavioral effects of early seizures on hippocampal-based memory demonstrate the influence of activity-dependent and seizure-induced plasticity during development on subsequent cognitive function in adulthood. Together, these findings suggest that seizures early in development are not as harmless as was once thought (24).
The present study addressed the behavioral effects of KA-induced status epilepticus at various stages of hippocampal development (31), focusing on whether early-life seizures differentially alter several hippocampal-dependent behavioral and cognitive processes, including short- and long-term spatial learning and memory, anxiety, and exploratory activity.
MATERIALS AND METHODS
In these experiments, seizures were induced by the systemic administration of KA on one of the following postnatal days: P1, P7, P14, or P24. Rats were allowed to reach adulthood (>P90) and were then evaluated with behavioral tests. Separate groups of rats were used for the radial-arm maze, water maze, and elevated plus maze/open-field tests. All procedures involving animals were approved by the Research Animal Care Committee of the University of Wisconsin.
Administration of kainic acid
KA was dissolved in saline and administered intraperitoneally to male rats at dosages of 1 to 2 mg/kg (P1, P7), 2 to 4 mg/kg (P14), and 8 to 10 mg/kg (P24) according to previous studies, which demonstrated that these doses reliably induce electrographic and behavioral seizures at these ages (20,30). Male littermates, either uninjected or injected with 0.9% (wt/vol) NaCl, served as controls. Seizures were observed and their characteristics noted over a 4- to 8-h period. Both sham-treated controls and animals that received KA were then returned to their home cages, in pairs (P24) or with their dams (P ≤14).
Radial-arm maze testing
To assess the short- and long-term effects of early postnatal seizures on spatial memory in adulthood, subsets of the adult rats that experienced seizures on P1, P7, P14, or P24 were evaluated in a modified eight-arm radial maze beginning on P90 to P100. Before the experiment was started, the rats were calorically restricted until they lost 15% of their body weight, to ensure adequate motivation in this appetitive task (32). Before the experiment, the rats were put into the radial-arm maze to habituate for 5 min, to learn to navigate the apparatus and find and eat food pellets (bait) hidden in recesses at the ends of four arms of the maze. Arm entry was defined as placement of all four feet into an arm of the maze.
The test consisted of six trials, 30 min apart, for a total testing time of ∼3 h. Small food pellets were placed at the end of the same arms of the radial-arm maze that were baited during the habituation trial. In addition, visual clues (large geometric shapes) were attached at the entrance of each arm in which a pellet was located, to assist the rats in learning the task. The rat was placed at the center of the radial-arm maze and allowed to explore the maze until the four baits were consumed. The time to consume the four pellets (latency), total number of entries into the arms, and number entries into unbaited arms (reference errors) were recorded. The same four arms were baited each trial.
A different subset of rats was tested in a modified water maze (33), designed to assess visual spatial learning (acquisition) and memory (retention). A circular steel tank (117-cm diameter) was filled with water (26 ± 1°C) to a depth of 25 cm. The water was made opaque by addition of ∼100 ml evaporated milk. The room was illuminated by overhead lights, and visual cues around the room (furniture, shelves) were kept constant from day to day. Four points on the perimeter of the pool were designated north (N), east (E), south (S) and west (W), thus dividing the pool into four quadrants (NW, NE, SE, SW). An 8 × 8-cm Plexiglas platform, onto which the rat could escape, was positioned in the center of one of the quadrants, 1 cm below the water surface. Because the opaque water precluded visual determination of platform position, rats had to use other means to locate it (i.e., distant visual cues).
On day 1 of training, each rat was placed in the pool for 60 s without the platform present; this free swim enabled the rat to become habituated to the training environment. On days 2 to 5, rats were trained for 24 trials (six trials per day) to locate and escape onto the submerged platform. For each rat, the quadrant in which the platform was located remained constant, but the point of immersion into the pool varied between N, E, S, and W in a quasirandom order for the 24 trials, so that the rat would not be able to predict the platform location from the point at which it was placed into the pool. The latency from immersion into the pool to escape onto the platform was recorded for each trial by one observer, while another observer manually mapped the route taken by the rat to reach the platform. On mounting the platform, the rat was given a 30-s rest, after which the next trial commenced. If the rat did not find the platform in 120 s, it was manually placed on the platform for a 30-s rest.
Four hours after the final training trial on day 5, the spatial bias in the rats' search pattern was tested to assess memory for platform location (“probe trial”). The probe test assessed the hypothesis that control rats would spend more time searching in the quadrant in which the platform was located than would KA-treated rats. The platform was removed from the pool, and the rat was placed back into the pool at the point opposite the previous platform location. The rat was allowed 60 s of free swimming, and its swim route was recorded manually. The distance swum in each quadrant was calculated by tracing the swim path by using a digital curvimeter. These swimming distances are proportional to the time spent in each quadrant, because swimming speed was equivalent in each group.
Open-field and elevated plus maze
In another subset of rats that was not assessed on the radial-arm maze or water maze, locomotor activity and exploratory behavior were tested by using an open-field apparatus. The purpose of this test was to determine whether the control and experimental groups differed in baseline locomotor activity. The open-field test board consisted of a wooden box 60 × 60 × 35 cm, with four holes of 4 cm in diameter symmetrically spaced on the floor (34). The floor of the box was divided into nine equal-sized squares, allowing an observer to count the number of squares visited during a given period of time. For testing, adult rats (P90–P100) were placed on the board for 5 min. The number of head dips into the holes (exploration), squares visited (locomotor activity), and rearings were recorded.
After the open-field test, anxiety was tested by using the elevated plus maze. The elevated plus maze consisted of two open arms 50 × 10 cm at right angles to two covered (closed) arms, 50 × 40 × 10 cm. The maze was elevated to a height of 50 cm above the ground, forming an aversive stimulus to animals in the open arms, as described previously (35,36). The animal was placed in the center of the elevated plus maze with its head facing an open arm. Animals were tested for 5 min. Entry into a particular arm was defined as the placement of all four feet into that arm. The relative time spent in open versus closed arms is a measure of anxiety, with anxious rats preferring closed arms. Anxiolytic drugs increase the time spent in open arms, whereas anxiogenic drugs increase the time spent in closed arms (35,37).
Radial-arm maze, water maze, open-field maze, and elevated plus maze data were analyzed with analysis of variance, by using either analysis of variance (ANOVA) or the Kruskal–Wallis test on ranks, depending on whether the data were normally distributed (thereby generating variance ratios of F or H, respectively). For ANOVA, either one or two factors were used, with repeated measures when appropriate. For post hoc testing, we used the Student–Newman–Keuls test or Bonferroni test. Significance was defined as p < 0.05 for all comparisons.
In KA-treated rats of all ages, status epilepticus developed, with a sequence of clinical features that differed somewhat by age (20,30). P1 and P7 rats initially became immobile with loss of limb tone and ataxia. They then exhibited intermittent hyperactivity with rhythmic “bicycling” movements of all extremities, opisthotonic arching, and tonic limb extension. Previous studies with implanted hippocampal electrodes showed that such clinical signs represent electrographic ictal activity (30). In older rats (P14, P24), KA-induced seizures consisted of initial immobility and staring. At P14, repetitive scratching movements and rudimentary wet dog shakes (WDSs) were seen. At P24, WDSs were prominent, followed by facial clonus, masticatory automatisms, and increased salivation; tonic–clonic movements of one or more limbs, rearing, and falling then ensued. At all ages, seizures (status epilepticus) persisted intermittently for 2 to 6 h. The status epilepticus was not terminated pharmacologically. Later, spontaneous seizures were not monitored specifically, but none was observed during routine handling and behavioral testing.
The performance of rats in the radial-arm maze was judged by the latency to find and consume the four baits, the total number of arm entries, and the number of arm entry “errors.” Entering an arm without bait present is referred to as a reference memory error, whereas entering an arm in which the bait had already been consumed is referred to as a working memory error. The number of total arm entries is a reflection of anxiety; a more anxious rat would tend to remain stationary and explore the maze less (fewer arm entries). Therefore each of these criteria measures a different aspect of radial-arm maze performance.
All groups of rats learned to find and consume the four baits in each trial (Fig. 1). As a function of trial number, the latency to reach criterion (find and eat all four baits) was significantly longer in rats that received KA (Fig. 1). By post hoc testing, differences were significant when comparing the first trial with the fifth and sixth trials, for each age group. It took significantly more time for the KA-treated P1, P7, P14, and P24 rats to find the baits than for the controls.
In addition, in each age group, fewer reference errors (entries into unbaited arms) were committed over time (i.e., with consecutive trials; Fig. 2). Post hoc analysis showed that, for each age group, by the third or fourth trial, a significant difference existed between controls and experimentals in terms of the total number of reference errors, with the animals that experienced seizures committing more errors (Fig. 2).
Previous studies using the radial-arm maze showed that anxious rats are less active and avoid exploring arms of the maze (38,39). Therefore to rule out an effect of anxiety on radial-arm maze performance between control and experimental groups, the total number of arm entries (baited, unbaited, previously baited) was compared (Fig. 3). On the first trial, no difference was found between controls and any KA-treated group as to the total number of arm entries (p < 0.05; Fig. 3, First Trial). Therefore anxiety was not a factor limiting initial arm exploration. By the sixth trial, controls and each KA-treated group made significantly fewer total arm entries compared with the first trial (i.e., they each learned the task, as described earlier). However, on the sixth trial, the control animals had significantly fewer total arm entries than each KA-treated group (p < 0.05; Fig. 3, Sixth Trial). These findings suggest that differences in total arm entries from the first trial to the sixth trial cannot be explained by a difference in anxiety between the groups.
During the habituation trial, the rats swam randomly around the pool, with no preference for a particular quadrant. The acquisition of place learning is shown in Fig. 4A, in which escape latencies are plotted against day of training. All rats learned to find the hidden platform and escape onto it, as indicated by the progressive decrease in mean escape latency over the 4 training days in each group. The escape latencies for the experimental groups that received KA on P1, P7, or P24 differed significantly from untreated controls (by experimental groups: F= 14.9, p < 0.0001; by days of testing, F= 42.0, p < 0.0001; Group × Days interaction, F= 4.5, p = 0.0009). By post hoc testing, the performance of the P7 and P24 KA-treated groups differed significantly from controls, whereas P1 KA-treated rats performed similar to controls. Although the performance of the rats that received KA on P7 reached control values by testing day 4, their overall learning was much slower than controls, as evidenced by results on testing days 1, 2, and 3, each of which differed significantly from controls on post hoc testing.
On the probe test (Fig. 4B), the distance spent swimming in the target quadrant (proportional to the dwell time spent in that quadrant) differed between controls and each experimental age group (P1, P7, P24; F= 6.4; p < 0.002 by one-way ANOVA with post hoc Student–Newman–Keuls test). In Fig. 4C, the probe test results for the P24 KA group is compared with those of controls. For controls, the distance swum in the target quadrant differed from that in each other quadrant (F= 26.0; p < 0.0001), whereas for rats treated with KA on P24, swimming distances did not differ between quadrants (F= 0.97; p = 0.42). Similar results were found with the P1 and P7 groups (data not shown). Therefore each KA-treated group showed deficits in spatial bias by spending equivalent time in all quadrants rather than dwelling in the target quadrant.
Elevated plus maze
Anxiety scores were lower (indicating greater anxiety levels) in all age groups (P1, P7, P14, P24) among rats that had experienced KA–induced status epilepticus compared with the controls that did not receive KA (Fig. 5A; F= 26.7; p < 0.0001; post hoc Bonferroni test, p < 0.05). These findings demonstrate that increased anxiety existed in all experimental groups, and they spent less time in the open arms, compared with the controls.
On open-field testing, no difference was seen in exploratory activity (number of head dippings) or locomotor activity (number of squares visited and rearings) between age groups or between controls and rats that had experienced KA seizures (Fig. 5B and C). Therefore the differences in anxiety scores between the control and experimental groups were not due to an alteration of locomotor or exploratory activity and may therefore be a consequence associated with the seizure history.
These results add to the accumulating evidence that seizures during early brain development can result in life-long behavioral and cognitive deficits. In this study, status epilepticus induced by the glutamate receptor agonist KA early in life was associated with significant learning and memory deficits and persistent anxiety in adulthood. As summarized in Table 1, KA seizures on P1, P7, P14, and P24 (radial-arm maze) or P7 and P24 (water maze) were associated with both short- and long-term impairments in spatial learning and memory, respectively, both of which are hippocampus-dependent functions. The finding that controls and rats that received KA on P1 did not differ on water maze testing, whereas P1 rats demonstrated impaired radial-arm maze learning, suggests that the latter test may be more sensitive for seizure-induced behavioral dysfunction. Furthermore, rats that experienced status epilepticus at these early ages exhibited heightened levels of anxiety as adults (elevated plus maze), again correlating with abnormal hippocampal/limbic system function. The lack of major hippocampal structural damage in other studies using similar protocols (18,19,30) suggests that seizures early in life can have long-term cognitive consequences despite the absence of overt cellular damage. Although some synaptic reorganization may occur after multiple postnatal seizures (40), its functional consequences are unclear and may differ by age. It is already known that KA–induced status epilepticus in adult rats causes a marked disruption of spatial learning and memory, far greater than the changes seen in developing rats (21,41).
Table 1. Summary of results
KA, kainic acid.
Short-term spatial learning and memory
Animals with KA seizures early in life (P1, P7, P14, P24): – took longer to learn food location – made more reference errors (entries into unbaited arms) – no difference was found in performance as a function of age at which seizure occurred.
Long-term spatial learning and memory
Animals with KA seizures early in life (P1, P7, P24): – took longer to learn platform location (except P1) – spent less time swimming in target quadrant (probe test) – deficits varied inversely with age at the time of status epilepticus: P24 > P7 > P1.
Elevated plus maze
Animals with KA seizures early in life (P1, P7, P14, P24):- – were more anxious (preferred closed arms).
Animals with KA seizures early in life (P1, P7, P14, P24): – exhibited no difference from controls in head dips, squares visited, or number of rears (therefore anxiety differences were not explainable on basis of differences in locomotion).
These experiments confirm and extend those of Lynch et al. (30). In that study, a different version of the radial-arm maze was used, with the protocol requiring that rats learn the location of the food pellet over several days of testing, until a criterion was reached (consuming all food pellets in the first five arm entries). The task was considered “learned” once the animal achieved five consecutive criterion performances. Here, we used a modified radial-arm maze protocol, in which all learning and testing occurred over multiple trials on a single day. In this abbreviated training condition as well, KA seizures early in development were associated with impaired learning of the task and more reference errors compared with those in controls. However, compared with the previous data (30), our results in the radial-arm maze had no relation to age. That is, the latency to criterion, number of reference errors, and total errors were similar, regardless of the age at which status epilepticus occurred. Similarly, in the elevated plus maze, anxiety levels were increased to a similar degree in each KA-treated group. Therefore the fact that status epilepticus occurred had a greater impact than the age at which it occurred.
In this study, we used an additional measure of hippocampal integrity, the water maze, a well-validated test of spatial learning and memory (33,42,43). In this task, rats that were treated with KA on P7 or P24 learned the platform location slower than did controls, with a rank order of performance (worst to best): P24 > P7 > P1, controls. Therefore performance varied inversely with the age at which KA status epilepticus occurred. On the probe (spatial bias) component of the water maze, rats are allowed to swim freely without a platform present, to assess their preference (memory) for platform location in the target quadrant. Each KA-treated group (P1, P7, P24) spent significantly less time swimming in the target quadrant than controls, suggesting that the seizure groups had poorer memory of platform location.
Compared with previous studies using the Morris water maze (21), the current study detected more evidence for memory impairment as a consequence of seizures at P5, P10, and P20. The reasons for these discrepancies are uncertain, but accumulating evidence indicates that behavioral and learning deficits occur as a consequence of early-life status epilepticus. The persistence of increased anxiety from early-life status epilepticus suggests that affect is influenced as well as cognition.
Prolonged seizures caused by other etiologies [e.g., corticotropin-releasing hormone (44)] early in development also are associated with later impairment of cognitive function such as water-maze learning, and the deficits appear to be progressive over time (6). Lithium-pilocarpine status epilepticus at P16 or P20, but not at P12, was associated with cognitive impairment in the water maze in early adulthood (P55) (11).
In addition to prolonged seizures early in life leading to later cognitive abnormalities, even brief, recurrent seizures at young ages also can be detrimental (4,8,45,46). Recurrent pentylenetetrazol (PTZ) seizures in early development (P10–P14) caused significant spatial deficits in the water maze when the animals were tested on P35 and P60 (10).
The choice of model and assessment tool is critical, because different results have been obtained with different methods. The limitations of rodents as an experimental model for epilepsy were discussed recently (47,48). Furthermore, detailed knowledge of the chosen behavioral test is crucial before attributing a cognitive deficit to an experimental paradigm (25,42,43,49).
In summary, on multiple measures of hippocampus-based cognitive function, rats that experienced KA seizures during early development had persistent deficits as adults: radial-arm maze, long-term version (30), radial-arm maze, abbreviated version (present study); and water-maze acquisition learning and spatial memory (present study). These results suggest that seizures disrupt some aspect of hippocampal function during the early, vulnerable “critical period” of brain development, with deficits observable long afterward (in adulthood). Early-life KA seizures reduced long-term potentiation, increased paired-pulse inhibition in the dentate gyrus, and reduced the susceptibility to kindled seizures as adults (30). The increased inhibition was thought to be related to the loss of plasticity caused by the excessive neuronal activity of the seizure. Therefore the dentate gyrus appears to be a site of long-term cellular alterations induced by abnormal patterns of postnatal neural activity. It is likely that other hippocampal and extrahippocampal areas also are involved in these maladaptive plastic changes after early-life seizures. Indeed, each of the behavioral tests we used reflected more than just hippocampal function; rather, each also requires integration of several neural systems (31,50,51).
Together, these studies demonstrate that disruption of normal neural activity by seizures during early postnatal development produces deficits in a variety of cognitive measures and behaviors in adulthood. As such damage occurs even before the full maturation of hippocampal circuits (52), it is apparent that subsequent developmental events are not sufficient to overcome the adverse effects of early postnatal seizures. As substantial evidence indicates that status epilepticus in the developing brain does not produce the overt macroscopic damage seen in adults, the observation of long-term behavioral and memory deficits in these experiments supports the view that the term seizure-induceddamage also should include adverse functional consequences. These observations may have clinical implications for cognitive and memory dysfunction associated with epilepsy during development.
Acknowledgment: We are grateful for grant support from the National Institutes of Health (NINDS RO1 25020 to T.P.S.), the Epilepsy Foundation (to U.S.), the Charlie Foundation (to C.E.S.), and Parents Against Childhood Epilepsy (to C.E.S.).