In the present study, we investigated whether CCI traumatic brain injury results in a spatiotemporal phenotypic change in microglia and if these changes are linked to a neurologic and cognitive deficit in a mouse model. Microglia are the brain immune cells, and they are implicated in almost all physiological processes in the CNS and play an important role in several inflammatory and neurodegenerative diseases. Chronic changes in microglia phenotype and function, support the notion that chronic microglial dysfunction may contribute to the chronic progressive neurodegeneration that are observed years after head injury [19]. However, the precise role of microglia in neuroinflammation following TBI and leading to neurological, sensory-motor and cognitive, deficits at acute, subacute, and chronic time points after TBI remain to be defined.
Here, we provided insights into the dynamics of several microglial marker expression after injury using multi-color flow cytometry coupled with an unbiased bioinformatics approach to characterize microglia in the different brain compartment after injury. We used CX3CR1 and the classical CD11b and CD45 [31] markers to identify microglia although we have also verified the cell lineage with microglia-specific marker TMEM119, and macrophage-specific marker CCR2 [32] (data not shown).
At 48 hrs after CCI, both the mild and severe TBI groups showed statistically significant sensory-motor and cognitive deficits compared to the sham animals [33]. At this acute time point, the number of microglia increased in the ipsilateral cortex in the mild injury group, while the number doubled in the severe TBI group in both hemispheres suggesting diffuse activation. Similar studies done in rodents described an increase from 2 to 20-fold of the number of cells in the injured brain [34, 35]. An increase in the numbers of isolated cells is attributed to resident microglia proliferation and expansion and peripheral cell infiltration [34]. It is reported that infiltrating monocytes and neutrophils respond quickly by crossing the blood-brain barrier and penetrating the injured tissue [36]. We found that infiltrating macrophages were only observed in the ipsilateral hemisphere of severe TBI (Fig. 3) [37]. Surprisingly, infiltrating macrophages were not proinflammatory as they do not express TNFa and have a very high expression of CD206 the mannose receptor that is usually expressed on M2 macrophages, an observation in line with a previous report showing that macrophages in TBI have a mixed M1/M2 profile[38]. Interestingly, these cells were not found when the injury was mild, suggesting a possible threshold for the impact to solicit peripheral cells infiltration. Probably the mild injury produced here didn’t result in the disruption of the blood-brain barrier which leads to limited recruitment of circulating myeloid cells to the injured site [39]. Although neutrophils have an important role post-injury as they are the first responders, we elected to focus on microglia as the cells most likely responsible for the long-term damage [40]. When we looked at microglia markers at 48 hours post-injury, we found that almost all markers were upregulated mainly on the ipsilateral side suggesting activation of these cells. Microglia activation corresponds to new biological functions [34]. These changes are readily visualized in immunohistochemistry imaging, previously reported in rodents [41–44]. Early activation of microglia after TBI has been observed in animal models [45–47] and in humans [48]. Persistent inflammation is also known to occur for up to one year after injury in rodents [49] and for several years in humans [50].
While there is agreement about the persistent activation of microglia after TBI, the phenotype and functional aspects of these microglia remain less clear. Earlier studies have focused on the increased number of microglia at different time point after injury [51] and morphological evidence of activation [52]. Using the M1/M2 phenotype scheme adapted from peripheral macrophages, researchers reported dysregulation of the anti-inflammatory M2 phenotype in association with chronic microglial activation [49, 53, 54]. However, mass cytometry and single-cell transcriptomic analysis reveal that microglia often present mixed phenotypes [55, 56] and that the M1/M2 scheme may be an oversimplification[57].
An upregulation of Class II MHC and CD11b in chronically activated microglia [58] is consistent with our observations. High levels of CD11c and CD14 expression have been described in geriatric microglia, and chronic neuroinflammation may predispose to other neurologic disorders such as Alzheimer’s disease [59].
In normal brains, our results showed that microglia are predominantly of cluster 1 with a ratio close to 2 to 1 in comparison to cluster 2, this was seen in all investigated regions. At 48 hours after injury, all the identified microglial populations increased in the injured brains as an indication of active proliferation. We noted at this time point a shift in the predominance between Cluster 1 and 2, suggesting that the injury microenvironment at 48 hours is pro-inflammatory as cluster 2 express significantly higher TNFa. As for cluster 3, these cells were found highly upregulated only in injured brain, even in mice that underwent the sham procedure.
At seven days post injury, the animals with mTBI showed complete recovery of motor deficits but not the sensory functions. Both sensory-motor functions remained impaired after the severe TBI. Both animal groups showed persistent spontaneous object recognition deficits post injury[60, 61]. The highly activated microglia were nearly undetected at this time point with the persistence of the inflammatory cluster 2 in the cortex but not in the subcortical areas. Interestingly, at the chronic time point, sensory-motor and cognitive deficits were sustained post severe TBI and highly activated cells reappeared in injured brains along with persistence of cluster 2 cells suggesting an increase in the inflammatory background. An interesting observation in this study is the resolution of the highly activated microglia population at day 7 and its reappearance at the chronic time points. Previous studies have shown a bimodal change in microglia morphology [62] in the mouse model. This observation raises the possibility that after resolution of the acute injury, the second wave of activation may be triggered by late neuronal or axonal injury. This is in line with the delayed appearance of neurofibrillary tangles observed in humans after a single TBI [48].
In this study, we did not observe any patterns in the spatial distribution of the microglia populations in the severe TBI group suggesting that activation of microglia is already diffuse and includes ipsilateral, contralateral as well subcortical regions at 48 hours. In the mild TBI group, the proliferation of microglia was mostly observed on the ipsilateral cortex at 48hrs and ipsilateral cortex and hippocampus at 7 days.
Experimental studies in models of traumatic brain injury have shown that the majority of microglia and recruited macrophages at the site of injury have mixed pro and anti-inflammatory-like activation profiles, but that the anti-inflammatory-like response is short-lived and there is a phenotypic shift towards a proinflammatory-like dominant response within one week of injury [62]. Others have demonstrated prolonged activation of microglia in the subcortical areas at 28 days after injury [51, 63]. Here, the identified macrophages profile fits an anti-inflammatory response at 48h. Whether macrophages and their cross-talk with microglia have a role in downregulating inflammation at 7 days is difficult to judge in our experimental setting [64]. Since both pro-and anti-inflammatory paths associated with overexpression of the same markers, it may be necessary to add additional parameters, such as cytokine expression [65]. In-depth characterization of the phenotype is also important for therapy after TBI as the timing of attenuating therapy delivery is critical for success [66].
Sensory-motor and cognitive deficits are prevalent post traumatic brain injury of all intensities and based on severity may persist for years after the injury[60, 67]. Spontaneous object recognition and spatial learning and memory deficits found in this research are in line with our previous and other research studies[68, 69]. Yet, to our knowledge few researches characterized the cognitive deficits systematically over time, at acute, sub-acute and at chronic time points. The different tests used in this research study allowed for the early detection of the sensory motor and cognitive deficits and monitoring its persistence up to 35 days post injury while examining the course of microglial activation. The deficits seen after a TBI of mild intensity are often subtle and harder to detect in small animals. We have uncovered the sensory-motor and cognitive deficits at the acute and sub-acute time points in the animals with mild TBI. We have also found activation of the pro-inflammatory microglia at these same time points which warrant further investigation and analysis of these early intervals, as it may shed light into a possible therapeutic window for TBI management of any intensity.
Finally, to link microglial phenotypes to cognitive deficit we have investigated the microglia response within the subcortical areas, as they are critical in the formation and processing of spatial learning, memory and relaying sensory and motor signals. The deficits seen on the MWM were increasing in relation to the severity of the injury as previously reported [70]. We found that each 1 % increase in the frequency of highly activated microglia at day 35 corresponds to an approximately 6-second increase in the cumulative latency to platform on the MWM indicating a decrease in the learning curve. Whether the presence of highly activated microglia is the cause of the spatial learning deficit cannot be concluded from our studies. The microglia may become activated by injured neurons or axons and requires further investigation. We also need to explore the same microglial activation pattern while using both cognitive tests, the spontaneous object recognition and the MWM at the acute phases in both mild and severe TBI.
Our study has some limitations; first, the panel of antibodies and the numbers of markers included are limited by the capacity of the conventional flow cytometer. Future studies using high throughput mass cytometry will be of interest. Second, sex differences are also reported comparing activation of microglia following TBI in adult female and male mice [71]; in our study, we used only male mice.