The main findings of our study, in which immature and juvenile rats had KA-induced SE, were the following: First, the expression of a number of cytokine mRNAs was induced in both juvenile (P21) and immature (P9) rats after SE; second, the time course and levels of increase in cytokine mRNA expression after SE varied according to the age of rats; third, glial cells exhibited an active appearance in P21 juvenile rats subacutely (3 d) after SE, especially microglia within the CA1 region, while in P9 immature rats glial cells retained their resting appearance. These topics are discussed more thoroughly below.
Expression of cytokines after SE
It has previously been shown in developing hippocampus that IL-1β mRNA expression is augmented in P15 and P21 rats but not in P9 rats 4 h after KA-induced SE, while TNF-α and IL-6 mRNA expression are increased only in P21 rats [
6]. Our recent results are in accordance with these findings and further show that the induction of IL-1β and TNF-α mRNA expression is more prolonged (8 and/or 24 h after SE) in P21 rats and, although to a lesser extent, also occurs in P9 rats. Furthermore, in a previous study in which excitotoxic lesions were induced via NMDA injection in the right sensorimotor cortex of P9 rats, IL-1β and TNF-α protein expression was induced in sensorimotor cortex, corpus callosum, and internal capsule 4 h - 24 h after injection [
21]. However, to our knowledge, until now there has been no evidence of any SE-induced cytokine gene expression in P9 rats. Our current results suggest that IL-1β and TNF-α mRNA expression is increased more rapidly and is more pronounced after SE in the more mature juvenile hippocampus compared to a delayed and weak induction in the immature hippocampus. Thus, the time-course and intensity of augmented IL-1β and TNF-α mRNA expression after SE in P21 rat hippocampus resembles that of adult rats [
2], which implies that this response is already mature in juvenile rats. It is noteworthy that, compared to controls, the increases in IL-1β protein expression 8 h after KA injection in P21 rats are much lower than the increases in IL-1β mRNA expression at the same time-point. However, the selected time-point probably does not represent the peak in IL-1β protein expression after KA, which is likely to occur later. Furthermore, the result does confirm the tendency of increased mRNA expression to be translated into an increase at the protein level, albeit not on a one-to-one basis. As the mRNA expression levels of many other cytokines were also highly upregulated, it is reasonable to suggest that these are followed by increases in their corresponding protein expressions, although these increases might be significantly lower and occur at later time-points than the mRNA changes.
In addition to neurodegeneration, the classical pro-inflammatory cytokines have been suggested to contribute to epileptogenesis [
4]. For example, transgenic mice overexpressing IL-6 showed increased sensitivity to seizures [
22], and chronic IL-1β expression has recently been associated with the development of spontaneous limbic seizures after prolonged febrile seizures in P11 rats [
23]. Conversely, the IL-1 receptor antagonist (IL-1Ra) mediates anticonvulsant effects in adult rodents in a number of limbic seizure models (i.e. intrahippcompal KA and bicuculline injection, and electrical kindling) [
24]. The initial pronounced upregulation of pro-inflammatory cytokine mRNAs seen in juvenile rats in our current study might therefore be an early factor promoting epileptogenesis. While these cytokines can be produced by both glial cell types after seizures, there could be some subtle differences between the glial cell types as to which cytokines they produce, at least in certain seizure models. For example, 12 h after soman-induced seizures in adult rats, IL-1β was observed to be expressed by activated microglia, whereas IL-6 was expressed by astrocytes in the hippocampus, piriform cortex and thalamus [
25].
mRNA expression of the anti-inflammatory cytokine IL-10, which has so far been less studied in both immature and mature brain after SE, was augmented in both age groups. In an earlier study in P9 rats, in which excitotoxic brain injury was caused by an intracerebral NMDA injection, IL-10 and its receptor were upregulated in glial cells, suggesting a protective role for IL-10 in excitotoxicity [
12]. In our current study, IL-10 expression was increased in concert with the pro-inflammatory cytokines IL-1β and TNF-α, which is in line with the earlier observation that IL-10 expression may be induced simultaneously with pro-inflammatory cytokines in the brain following an insult [
26]. Although the pathways of cytokine expression regulation are intricate, the finding that induction of IL-10 mRNA expression was quite moderate and short-lived compared to the more pronounced and prolonged induction of IL-1β and TNF-α mRNA in P21 rats favours the idea that pro-inflammatory cytokine expression might overcome IL-10 expression after SE at this age. In contrast, in P9 rats, IL-10 mRNA expression showed a more pronounced induction than that of IL-1β or TNF-α, which suggests that, at this age, IL-10 might inhibit activated microglial cells and thus suppress expression of pro-inflammatory cytokines after SE more effectively than in P21 animals. Furthermore, microglial cells expressing IL-10 have been shown to attenuate neurodegeneration in hippocampal cultures when apopotosis is promoted with NMDA [
27]. This enables the hypothesis that microglia might self-limit their response via negative feedback by producing IL-10 and thus can have either pro-inflammatory or anti-inflammatory functions, depending on circumstances.
Also MMP-9 mRNA expression increased in both age groups after SE. In
in vitro conditions, MMP-9 has been shown to be produced by activated microglia [
28], and cytokines are proposed to regulate MMP-9 activity in astrocytes [
29]. In an earlier study, increased MMP-9 activity was observed 8 h after KA-induced seizures in adult rats, and this was linked with neuronal death as confirmed in KA-treated organotypic hippocampal slice cultures (prepared from P11 rats and cultured for 14 days) [
30]. In our current study, MMP-9 mRNA expression increased in both P21 and P9 rats and, as previous studies have shown that P9 rats do not exhibit any detectable neuronal damage from KA-induced SE [
6,
8], this suggests that MMP-9 mRNA upregulation after SE alone is not sufficient to induce neuronal damage in the P21 rat hippocampus. MMP-9 might have a more central role in epileptogenesis, as MMP-9 knockout mice show lower sensitivity to epileptogenesis after pentylenetetrazole kindling-induced seizures, while the sensitivity of transgenic rats overexpressing MMP-9 is higher [
31].
GDNF mRNA expression is augmented in the adult rat hippocampus 6-24 h after KA-induced SE [
16]. In our study in developing rats, GDNF was the only studied cytokine whose mRNA was increased exclusively in P21 rats after SE. In another earlier study, in which SE was induced with KA in adult rats, adenoviral-vector-delivered GDNF, introduced before KA injection, increased Bcl-2 expression and reduced the number of apoptotic cells in the CA3 and dentate gyrus (DG) regions of the hippocampus compared to the rats, which did not receive GDNF [
32]. This suggests that GDNF has a neuroprotective role in KA-induced SE. Furthermore, GDNF is proposed to suppress seizures in temporal lope epilepsy models (e.g. kindling) in adult rats [
33]. Our results indicate that GDNF upregulation after SE is a mechanism that has reached an adult level of maturity in P21 rats.
The expression of IFN-γ and TGF-β1 mRNA did not significantly change after SE in either age group. In a recent study in adult rats, IFN-γ mRNA expression was increased in astrocytes of the hippocampus 24 h after lithium-pilocarpine-induced SE while the expression of its receptor was upregulated in neurons and, furthermore, neutralization of the IFN-γ receptor aggravated injury suggesting a protective role for IFN-γ [
34]. Also, the expression of TGF-β1 mRNA has been shown to be increased in response to KA-induced SE in microglial cells in the hippocampus of adult rats [
35]. However, TGF-β1 has recently been suggested to contribute to epileptogenesis induced by albumin after blood-brain barrier breakdown in adult rats [
36]. Our current findings suggest that these mechanisms, which are activated by SE in the adult hippocampus, are still underdeveloped in the juvenile P21 rat hippocampus.
Our present results together with earlier findings of other groups suggest that the more mature innate immunity system elicits a fulminant inflammatory response after SE, in which the pro-inflammatory reaction is marked. In contrast, the immature system evokes only a mild pro-inflammatory response, but a marked anti-inflammatory reaction. We propose that in the immature brain either the ability to induce acute inflammatory reaction is underdeveloped or its emphasis is on anti-inflammatory properties. Moreover, the inflammatory reaction in developing hippocampus is likely to at least partly contribute to age-dependent susceptibility to SE-induced neuronal injury and recurrence of seizures. However, other factors, such as molecular and functional differences in receptors and ion channels between the immature and adult rodent brain, may be of equal or even greater importance in determining changes in the seizure threshold [
37].
Glial cell activation after SE at the acute and the subacute phase
Glial cells, i.e. microglia and astrocytes, are age-dependently activated (in P21 and P15, but not in P9) in the postnatal rat hippocampus up to 24 h after KA-induced SE [
6]. Additionally, short-term microglial (a few days) and long-term (beyond 40 days) astrocytic activation have been observed after KA-induced seizures in P15 rats, and this has been proposed to contribute to neuronal injury, neurobehavioral impairment, and increased susceptibility to seizures [
38]. Our results further address this and show that also subacutely, 3 days after seizures, both types of glial cells remain activated in P21 rats and that glial cells remain inactivated in P9 rats also at this subacute phase. Furthermore, astrocyte activity seems to increase over time after SE in P21 rats, since GFAP expression was now significantly increased 3 d after SE, but not yet 1 d after SE. This further supports the notion that the more mature brain elicits a more prominent inflammatory response after SE, which may be in part due to a diminished ability of glial cells to self-limit their response. Additionally, it is of importance to note that the most pronounced subacute microglial activation was localized to the CA1 pyramidal cell layer in P21 rats after KA treatment, which is the region damaged by KA-induced SE in rats of this age [
8]. In contrast, an earlier study has shown that microglia are equally activated in all hippocampal regions 4 h and 24 h after onset of SE in P21 rats [
6]. Our results therefore suggest that the generalized microglial response in hippocampus seems to be gradually and specifically concentrated on the damaged CA1 region at the subacute phase in P21 rats, whereas the astrocytic response remains ubiquitous in the hippocampus. Furthermore, our recent transcriptomic study indicated pronounced upregulation of the GFAP gene in the CA1 region together with neuronal damage in P21 rats 7 days after SE [
39], which denotes long-term astrocyte activation specifically in the damaged area.
In contrast to prolonged glial cell reaction, increased expression of cytokine mRNAs is transient and, as our results show, at three days after SE in P21 rat hippocampi, cytokine mRNA expression has returned to control levels. This suggests that activated glial cells in juvenile hippocampus may have functions other than cytokine production during the subacute phase. Indeed, microglial response has been shown to be diverse after their activation, and can include both neurodegenerative and neuroprotective functions [
40]. The induction of COX-2 expression in adult and P21 rat brain has been shown to be transient as well, since its expression returns to the control level within three days after SE [
6,
41]. This further suggests that after the initial acute phase (which seems to last up to 24 h - 3 d), there is a change in the inflammatory response provoked by KA-induced SE in juvenile hippocampus, with attenuation of the pronounced, and perhaps exaggerated, expression of inflammatory mediators and modulation of microglial reactivity from widespread activation to a targeted response in the damaged area.
In our study, microglial activation was now detected in P21 rats 3 d after SE also in the basomedial amygdala, which together with the hippocampus is a part of the limbic system and suffers neuronal damage after KA-induced SE [
6]. Furthermore, the expression of COX-2 and c-Fos has been shown to be markedly increased in P21 rats after SE in amygdala, this again in tandem with hippocampus [
8,
42]. Thus, after the acute phase post-SE, the inflammatory response at the subacute phase seems to be more accurately targeted to damaged areas of limbic structures in juvenile rat brain.