The neuroinflammagenic potential of IL-1β is shown here through its induction of synthesis of itself and other proinflammatory cytokines including TNF, IL-1α, IL-1β, as well as the latter's maturation enzyme ICE. The additional impact of IL-1β on neuronal ApoE production shown here suggests that in neurological conditions where the expression of proinflammatory cytokines is elevated, the expression of IL-1-driven AD-related proteins such as ApoE would be elevated as well. Multiple MLKs--ERK, p38-MAPK, and JNK--were shown to be involved in elevated expression of ApoE in neurons exposed to IL-1β, Aβ, or sAPP. The increased expression of ApoE induced by glutamate was mediated by ERK and JNK, but not by MAPK-p38. Together, these findings have several implications for AD pathogenesis, particularly with respect to conditions in which neuroinflammation is prominent, especially those influenced by APOE genotype.
The actions of IL-1 and the other agents tested here--sAPP, Aβ, and glutamate--create the possibility for complex loops of influence analogous to the vicious circle of neuroinflammatory events we have termed the Cytokine Cycle . Glutamate can elevate neuronal expression of βAPP and its conversion to sAPP . βAPP is elevated in dystrophic neurites in and around plaques ; its breakdown into both sAPP and Aβ can result in induction of IL-1β in microglia [31, 50]. In addition to inducing IL-1β expression and release, sAPP and Aβ also stimulate microglia to release biologically relevant levels of glutamate and its cooperative excitatory amino acid D-serine [44, 51, 52]. Thus, future studies should address the potential for each of these agents to act indirectly through the elaboration of a key agent or agents that can directly stimulate ApoE expression.
Key to interpretation of our findings--and, indeed, to the role of APOE genotype in AD--is determining whether elevation of ApoE levels would be beneficial or harmful. Possession of the ε4 allele is associated with enhanced deposition of Aβ , consistent with in vitro studies wherein ApoE was shown to enhance Aβ fibrillogenesis . In this regard, ApoE4 has been shown to be more effective than ApoE3, fostering speculation that replacement of the ε3 allele by ε4 merely enhances an activity already present in ApoE3. This has been described as a toxic gain of function, implying that overabundance of any ApoE--even ApoE2 or ApoE3--would also create a gain in this function and thus be detrimental. Moreover, transient increases in cellular ApoE occur in response to injuries that promote AD, e.g., traumatic brain injury  and stroke . ApoE4 is generally reported to be present at higher steady-state levels than ApoE3 in CSF or brain parenchyma [57–61], though some studies have reported lower levels of total ApoE in ε4-positive individuals [62, 63].
In contrast to these connections to pathology, ApoE provides neuroprotection in many paradigms, and ApoE deficiency has proved detrimental in several respects . Therefore, inductions of ApoE by the stimuli we tested may represent a compensatory response, meaning that the distinction between ApoE3 and ApoE4 represents loss of a beneficial function. ApoE has anti-inflammatory effects, and even its interaction with Aβ can attenuate glial activation by the latter . However, ApoE3 is more effective than ApoE4 as an anti-inflammatory agent [31, 65, 66], so this putative compensatory response may be inadequate in ε4-positive individuals and thus allow more extensive propagation of the Cytokine Cycle. Such an allele-specific compensatory response may also extend to direct neuroprotective activity. We previously reported that ApoE3 induces βAPP expression but ApoE4 does not , confirming the findings of Ezra et al. . In this regard, elevations of ApoE by the process of neuroinflammation, or other stressors, would reflect a requisite role for the lipoprotein in enhancing the beneficial roles of βAPP and/or other acute-phase response proteins. Thus, it would be the inability of ApoE4 to participate in this chain of salutary events that makes it detrimental. We have previously shown that the increase in ApoE brain levels that occurs with aging continues to occur in AD, with a large fraction being deposited in plaques . This increase in ApoE levels is distinguishable from changes in βAPP, which rises with age but declines markedly in AD . This disease-associated severance of the coordinate regulation of ApoE and βAPP further strengthens the correlation of brain health with the coregulation of these two proteins; to wit, with ApoE expression itself, provided that the ApoE is not ApoE4.
Multi-lineage kinase pathways may be invoked in the regulation of ApoE expression, and can themselves be invoked by ApoE [68, 69], suggesting a feedback loop between MLK pathways and ApoE expression in neurons. Our findings of involvement of multiple MLKs--ERK, p38-MAPK, and JNK--in expression of ApoE in neurons exposed to IL-1β, Aβ, or sAPP, together with previous reports of ERK pathway invocation of ApoE expression and vice versa, are consistent with the existence of a complex feedback system that may be important in acute-phase responses to neuronal injury as well as potential exacerbation of neurodegenerative events. Our finding that glutamate regulates ApoE expression via ERK and JNK, but not by p38-MAPK pathways may be indicative of a correlation between glutamatergic induction of ApoE and neuronal survival. Excitotoxic effects of glutamate are largely dependent upon activation of extrasynaptic NMDA receptors, p38-MAPK, and the inhibition of ERK signaling; synaptic receptors, on the other hand, appear to activate ERK and promote survival [70–72].
In conclusion, the induction of neuronal ApoE by either neuroinflammatory or excitotoxic agents or neurotoxins, acting through MLK pathways suggests that alterations in these signaling pathways, together with other neuropathological entities in AD brain, may have consequences for ApoE expression. Differences in this expression may be critical, considering the role of APOE genotype in AD risk. The response of ApoE to IL-1β we show here in rodent brain suggests that elevation of IL-1 leads to the increases in ApoE that we and others have observed in the AD brain. This may have added significance with regard to the self-propagating nature of IL-1-driven cascades, especially when such cascades are instigated in the context of an ε4 allele of APOE. While induction of ApoE2 or ApoE3 may be anti-inflammatory or neuroprotective, and thereby act as a self-limiting influence on IL-1-driven cascades, ApoE4 may fail to participate and leave the brain vulnerable to prolonged activation of a maladaptive cycle.