Abstract
Zhao X, Liao Y, Morgan S, Mathur R, Feustel P, Mazurkiewicz J, Qian J, Chang J, Mathern GW, Adamo MA, Ritaccio AL, Gruenthal M, Zhu X, Huang Y. Cell Rep 2018;22:2080–2093.
Microglia are well known to play a critical role in maintaining brain homeostasis. However, their role in epileptogenesis has yet to be determined. Here, we demonstrate that elevated mTOR signaling in mouse microglia leads to phenotypic changes, including an amoeboid-like morphology, increased proliferation, and robust phagocytosis activity, but without a significant induction of pro-inflammatory cytokines. We further provide evidence that these noninflammatory changes in microglia disrupt homeostasis of the CNS, leading to reduced synapse density, marked microglial infiltration into hippocampal pyramidal layers, moderate neuronal degeneration, and massive proliferation of astrocytes. Moreover, the mice thus affected develop severe early-onset spontaneous recurrent seizures (SRSs). Therefore, we have revealed an epileptogenic mechanism that is independent of the microglial inflammatory response. Our data suggest that microglia could be an opportune target for epilepsy prevention.
Commentary
Changes in microglial morphology, physiology, and inflammatory phenotype have been observed both in models of epilepsy and in tissue from patients with epilepsy (1). However, it has been difficult to determine if alterations in microglial function are a cause of epilepsy or the effect of recurrent seizures. In a recent study, Zhao and colleagues explore noninflamma-tory roles of microglia and suggest that a noninflammatory, activated state may be sufficient to drive epileptogenesis.
Inflammation in epilepsy is a common finding, and data support a causative role for inflammation in the epileptogenic process (2–4). In many cases, microglia are mediators of the inflammatory response; however, other roles for microglia have been under considered. For example, in addition to the canonical pro-inflammatory reactive profile, microglia display a wide spectrum of two-way relationships with other cell types, including overtly anti-inflammatory activation and, in some cases, anti-neurodegenerative signaling.
Inspired by the increased pS6 immunoreactivity (indicative of enhanced mTOR signaling) observed in microglia from human patients with epilepsy, the authors selectively deleted the Tuberous Sclerosis Complex 1 gene (Tsc1, which encodes hamartin) in microglia through the use of a microglial selective Cx3cr1 Cre mouse line crossed with floxed Tsc1 mice. Defects in this gene resulting in enhanced mTOR activity are a common cause of TSC (5).
After deletion of Tsc1 (Tsc1cKO) in microglia, microglial morphology, physiology, and gene expression revealed the predominance of a noninflammatory, reactive-like state. Tsc1cKO microglia displayed larger cell bodies and shorter processes with less branching, reminiscent of the larger neurons observed in cortical dysplasia and also TSC. This reactive-like morphology has also been reported in other models of TSC (6), suggesting that overactive mTOR signaling in microglia per se drives these changes, rather than the altered morphology resulting from secondary effects of overactive mTOR signaling on other cell types. This reactive but noninflammatory state contrasts starkly with the ramified, resting-like state in control brains, or the activation state described after status epilep-ticus, which comprises a host of physiological changes and pro-inflammatory signaling. Strikingly, microglial Tsc1cKO mice developed severe spontaneous recurrent seizures and, as a result, suffered markedly increased mortality. A similar pathology and microglial response was observed after a temporally selective inducible deletion of Tsc1 in adult microglia. In addition to morphological changes, lysosome-related genes and phagocytotic phenotypes were upregulated in Tsc1cKO mice, a profile similar to that of a M2-polarized/anti-inflammatory state as compared to M1-polarized/pro-inflammatory state. As with the morphology, this phenotype stands in stark contrast to that observed following status epilepticus (7).
Although the canonical view places a microglial pro-inflammatory response as a contributor to epileptogenesis, deletion of Tsc1 in microglia was associated with only mild changes in inflammation in hippocampus. Interestingly, this response was not directly mediated by microglia. While TNF-α, IL-1β, IFN-β, and iNOS were moderately increased in total hippocampal homogenates, they were significantly downregulated in purified microg-lia. Thus, the inflammatory response was not due to enhanced microglial production of inflammatory mediators but was rather: 1) secondary to microglial reactive status or to seizures and/or 2) mediated by other cell types (e.g., neurons, astrocytes).
Elevation of mTOR signaling in microglia also resulted in a mild loss of both excitatory and inhibitory synapses in the cortex and hippocampus. Providing a potential mechanism for the synapse loss, astrocytes in these mice became hyperpro-liferative and adopted an activated-like morphology. Immu-noreactivity to GFAP and complement C3—both known to be upregulated when astrocytes encounter reactive microglia—was similarly enhanced. Consistent with the importance of the complement system to synaptic pruning and the observed synapse loss in the Tsc1cKO mice, mutant microglia displayed upregulation of complement and phagocytotic markers.
While compelling, findings in another recent study raise some caveats and contrasts. Zhang and colleagues tackled a similar question (8), and also used a Cx3cr1-Cre driver line to delete Tsc1 in microglia. This Cre driver line, while different than that used by Zhao, resulted in similar increases in microglial size, as well as the emergence of spontaneous seizures and a high degree of associated mortality. However, characterization of the Cre driver line revealed robust recombinase activity in neurons; similar characterization studies were not performed by Zhao and colleagues, raising the possibility that unintended deletion of Tsc1 in neurons may have contributed to the effects they observed. Importantly, using a similar inducible Cre strategy mitigated this off target activity, more selectively deleting Tsc1 in microglia; this resulted in increased microglial size, but was not associated with spontaneous seizures. The divergent findings between these two studies may be due to differences in the mouse lines employed, the age at postnatal Cre induction (8 weeks vs 2 weeks), or due to off-target deletion of Tsc1 in the Cre line used by Zhao. Supporting the latter concern, a prior report (9) found occasional leakiness of the same Cx3cr1 line used by Zhao, resulting in expression in neurons in ~10% of animals. A thorough characterization of the inducible Cre used by Zhao has not been reported. What impact do these unknowns have on the interpretations presented by Zhao? The non-inflammatory profile of Tsc1−/− microglia, which was extensively characterized in purified microglia, is unlikely to be impacted by off target recombination. However, the seizure phenotype may be—indeed, postnatal deletion of Tsc1 in even small numbers of neurons has been reported to result in an epilepsy phenotype (10).
Given these caveats, this study still raises an interesting set of questions. First, to what degree does a noninflamma-tory role for microglia contribute to epileptogenesis induced by other means (e.g., status epilepticus)? Moreover, selective deletion of Tsc1 in microglia, while interesting from a basic cell biology perspective, certainly does not mirror the clinical case in which Tsc1 would be disabled in neurons, astrocytes, or mi-croglia (actually, in all cells!). Microglia interact with other cell types in the brain, and it is unclear how microglial responses to Tsc1-deficient astrocytes or neurons would differ from their response to wild-type astrocytes or neurons.
Second, what is the sequence of events? Do mild inflammation, astrocytosis, and synapse loss result directly from microglial activation? Are they secondary to seizure activity or do they drive the seizures? Further studies with finer temporal resolution (i.e., direct comparisons of responses prior to and after the emergence of spontaneous seizures) will be needed to help elucidate these questions.
Third, the severe phenotype seen in these animals, which is similar to that reported in astrocytic or neuronal Tsc1 knockouts (11, 12), was reduced in the inducible adult cKO. While seizure occurrence and mortality were prevalent in both constitutive and inducible cKO mice, the question remains as to why deletion of Tsc1 in microglia in adult mice resulted in only 60% seizure occurrence and 50% mortality. Might this be due to a more selective deletion of Tsc1 in microglia in the inducible as compared to constitutive knockout lines? Alternatively, does microglial mTOR signaling play an especially important role during the critical stages of development, such as selective apoptosis or synaptic pruning?
Throughout the study, mortality was consistently observed shortly following prolonged seizure activity. Aside from the likely conclusion that the constitutive cKO mice reliably developed seizures severe enough to compromise breathing or other critical functions, a hypothesis that deserves further consideration is that the microglial cKO may directly compromise these underlying circuits, predisposing animals to die suddenly and unexpectedly from any given seizure. These findings may have particular relevance for the prevention and risk assessment of SUDEP, the most common form of death in refractory epilepsy (13).
Broadly, Zhao and colleagues have presented evidence suggesting that the role of microglia in epileptogenesis goes beyond the traditional view of damage causing inflammation and inflammation causing seizures. From this perspective, the major microglial contribution to epileptogenesis may be what microglia stop doing, and not what they start doing (maintaining homeostasis and responding to damage with inflammation, respectively). “Rounding up the unusual suspects,” or studying noninflammatory changes in microglia during epileptogenesis, may lead to novel and improved biomarkers and therapies for epilepsy.