Microglia induce neurogenesis by stimulating PI3K/AKT intracellular signaling in vitro

Background: Emerging evidence suggests that microglia can support neuronal survival, synapse development, and neurogenesis in classic neurogenic niches. Little is known about the ability of microglia to regulate the cortical environment and stimulate cortical neurogenesis outside classic neurogenic niches. We used an in vitro co-culture model system to investigate the hypothesis that microglia respond to soluble signals from cortical cells, particularly following injury, by altering the cortical environment to promote cortical cell proliferation, differentiation, and survival. Results: Analyses of cell proliferation, apoptosis, protein expression, and intracellular signaling were performed on uninjured and injured cortical cells in co-culture with an EOC2 microglial cell line. Microglia soluble cues enhanced cortical cell viability and proliferation of uninjured and injured cortical cells. Co-culture of injured cortical cells with microglial cells signicantly reduced cortical cell apoptosis. Microglial signicantly increased Nestin+ and a-internexin+ cells within and outside the injury site. NeuN+ cells increased in injured cortical cultures with microglia. Multiplex ELISA assays showed decreased levels of inammatory cytokines in conditioned media collected from injured cortical cell and microglial co-culture. RTPCR analysis of microglial mRNA was performed. AKT phosphorylation in uninjured, and particularly injured cortical cells, signicantly increased when co-cultured with EOC2 microglia. Inhibition of AKT phosphorylation in cortical cells blocked the microglial-enhanced cortical cell viability and expression of neurogenic markers in vitro . Conclusion: This in vitro model system allows for assessment of the effect of microglial-derived soluble signals on cortical cell viability, proliferation, and stages of differentiation during homeostasis or following injury. These data suggest that EOC2 microglia downregulate inammatory cytokine production following activation by acute cortical injury to enhance proliferation of new cells capable of neurogenesis. Inhibition of AKT signaling in cortical cells blocks the microglial-derived enhanced proliferation and expression of neurogenic markers in injured cortical cultures. In injured cortical cell and microglial co-cultures, signicantly increased expression of neurogenic markers Nestin and α-internexin was present within the site of injury where proliferating cells were observed. Expression of the mature neuronal marker NeuN increased in injured cortical cells outside the injury site when co-cultured with microglial cells. AKT phosphorylation was increased in cortical cells co-cultured with EOC2 microglial cells. Inhibition of AKT phosphorylation reduced the enhanced expression of neurogenic markers in cortical cell and microglial co-cultures. EOC2 microglial cells responding to acute injury downregulated their expression of pro-inammatory cytokines. These results show that this co-culture in vitro system provides a model to evaluate cortical cell responses to microglial derived soluble cues and to investigate the underlying mechanisms of the functional states of microglia in response to cortical signals during homeostasis or following injury. EOC2 cell lines provided an example of how microglial cells may inuence cortical cell differentiation following injury outside of neurogenic niches. This co-culture system provides a useful tool to further investigate the neuroimmune mechanisms important for primary microglial responses and cortical cell differentiation and survival in vivo. effector microglial-enhanced neurogenesis. together, these data suggest that microglial isolated primary cells or cell lines. These specic studies show that cortical cell injury activates neurogenic properties in EOC2 microglia. The specic EOC2 microglial cell line’s response to cortical injury results in at least the reduction of pro-inammatory cytokine gene expression and cytokine release. EOC2 microglial-derived soluble signals produced during homeostasis and, more so following activation by acute cortical injury, enhance neurogenesis by upregulating AKT signaling in cortical cells. Increasing our understanding of the mechanisms that drive cortical cell proliferation, differentiation, and survival as stimulated by microglia will provide insight into the neuroprotective role of immune activity in the CNS. 540-590nm on an ELISA plate reader. Three experiments were performed in triplicate.

intracellular signaling were performed on uninjured and injured cortical cells in co-culture with an EOC2 microglial cell line. Microglia soluble cues enhanced cortical cell viability and proliferation of uninjured and injured cortical cells. Co-culture of injured cortical cells with microglial cells signi cantly reduced cortical cell apoptosis. Microglial signi cantly increased Nestin+ and a-internexin+ cells within and outside the injury site.
NeuN+ cells increased in injured cortical cultures with microglia. Multiplex ELISA assays showed decreased levels of in ammatory cytokines in conditioned media collected from injured cortical cell and microglial coculture. RTPCR analysis of microglial mRNA was performed. AKT phosphorylation in uninjured, and particularly injured cortical cells, signi cantly increased when co-cultured with EOC2 microglia. Inhibition of AKT phosphorylation in cortical cells blocked the microglial-enhanced cortical cell viability and expression of neurogenic markers in vitro . Conclusion: This in vitro model system allows for assessment of the effect of microglial-derived soluble signals on cortical cell viability, proliferation, and stages of differentiation during homeostasis or following injury. These data suggest that EOC2 microglia downregulate in ammatory cytokine production following activation by acute cortical injury to enhance proliferation of new cells capable of neurogenesis. Inhibition of AKT signaling in cortical cells blocks the microglial-derived enhanced proliferation and expression of neurogenic markers in injured cortical cultures. This in vitro system is useful for continued studies with other microglial cell lines and primary microglial cells. Increasing our understanding of the mechanisms that drive cortical neurogenesis stimulated by microglial cells during homeostasis and following injury will provide insight into the potential mechanisms of the neuroprotective role of immune activity in the central nervous system (CNS).

Background
Microglia are resident immunocompetent and phagocytic cells of the central nervous system (CNS) and comprise anywhere from 5-12% of cortical cells [1,2]. During homeostasis, microglia survey the local CNS environment and communicate with neighboring glia and neurons through membrane bound and soluble signals [3,4]. Emerging evidence suggests that, given speci c activator(s), microglia function to support neuronal proliferation, differentiation, synaptic function, and survival [4,5]. Microglia contribute to synaptic development by re ning axonal branching and pruning synaptic connections through phagocytic activity [6][7][8][9]. Additionally, speci c microglial-derived cytokines, growth factors, and cell associated proteins play an important role in the modi cation and function of both excitatory and inhibitory synaptic connections in the CNS [10][11][12].
Microglia support neurogenesis in the classic neural stem cell niches of the subgranular zoneof the dentate gyrus in the hippocampus and the subventricular zone lining the lateral ventricles [13][14][15][16][17][18][19][20][21][22][23]. Further, in vivo studies have shown that speci cally stimulated microglia, macrophages, and in ltrating T cells protect neuronal axons from secondary degeneration following injury, degrade inhibitory proteins that restrict neuronal survival and regrowth, reduce pro-in ammatory cytokine production, and induce growth factor and neurotrophin production [24][25]. Production of anti-in ammatory mediators and neurotrophic factors by microglia are likely to be dependent on the nature and duration of the stimulus as well as the severity of injury to which microglia respond [26,27]. For example, microglia stimulated by damage signals from a human peripheral nerve increase BDNF and GDNF secretion; upregulate the expression of migratory cytoskeletal proteins; upregulate proteolytic and debris clearing enzymes; and, enhance both STAT protein expression and NFκB gene transcription [28].
Microglial proliferation and increased release of transforming growth factor -b (TGF-b) are correlated with neural stem cell proliferation in the adult dentate gyrus [13]. Secretion of insulin growth factor 1 (IGF-1) from microglia following status epilepticus in the adult dentate gyrus stimulates neurogenesis via activation of the p42/44 MAPK pathway [30]. Other work suggests that injury to adult CA1 neurons of the dentate gyrus stimulates IGF-1 release from microglia and astrocytes promoting neuronal survival via AKT phosphorylation and decreased MAPK phosphorylation [31] or via both AKT and MAPK phosphorylation [32]. Increased AKT phosphorylation by PI3K signaling is important for neurogenesis in classic stem cell niches as well as the cortex [31][32][33][34][35][36][37][38]. Recent studies have shown that microglial-derived FGF and EGF growth factor and IL-10/IL-13 cytokine secretion promote the proliferation and differentiation of adult neural stem cells in vitro [39]. While an evergrowing body of work supports the role of microglial soluble signals in proper neurogenesis and plasticity, neuroin ammation caused by microglia activity is also linked to neurodevelopmental and neurodegenerative diseases [40][41][42]. For example, the reduction in neuroin ammatory cytokines such as TNF-α, IFN-ϒ, MIP-1α and RANTES/CCL5, IL-1α, and IL-1β suppresses apoptosis and enhances neurogenesis [29,43]. Recent work demonstrates that during normal development and in ammatory states, primary microglia, even in the same region of the brain, express complex patterns of gene expression resulting in functionally diverse microglial phenotypes [43,44]. Taken together these data suggest that a complex milieu of microglial-derived soluble cues with neurogenic or neuroin ammatory properties work in combination to promote or restrict neuronal development, survival, and repair following injury [7,18,[39][40][41][42][43][44][45][46].
Outside the classic neural stem cell niches, CNS stem cells have the potential to generate neurons and glial cells following ischemia or traumatic injury [35,[47][48][49][50]. Microglial-derived soluble cues are potentially important for local endogenous neurogenesis, migration, and neuronal survival [51]. Interestingly, microglial invasion of the cortical plate overlaps with peak periods of cortical neurogenesis [1][2][3]52]. After invasion, microglia remain as morphologically and functionally dynamic cells within the environment of the cortex [2-3, 46, 49]. Microglialderived cytokines may promote neurogenesis by supporting progenitor cell survival and mitosis [53][54][55][56]. Given the diverse functional roles that microglia can play in the cortex, the development of an in vitro model system to evaluate the neurogenic potential of a uniform population of microglia is important. We hypothesize that microglia stimulate localized neurogenesis of cells throughout the cortex during homeostasis or following injury by activating speci c intracellular signaling pathways required for neuronal survival and differentiation. Little is known about the mechanisms by which homeostatic microglia or activated microglia responding to injury in uence neurogenesis and survival outside of hippocampal and lateral ventricular neurogenic niches.
In this study we present an in vitro model system for investigating the underlying mechanisms by which microglia respond to cortical cues to regulate neuronal differentiation and survival during homeostasis and following acute mechanical injury. Our in vitro system utilized an EOC2 microglial cell line suspended above primary rat cortical cells that were injured or left uninjured as control. Primary cortical cells were isolated from Sprague-Dawley rat embryos at embryonic day 16-18 because the differentiation of these cortical cells has been well characterized in vitro [57]. Neurogenesis or the differentiation of new neurons in primary cortical cell culture can be determined by evaluating the progressive expression of neurogenic proteins such as Nestin, GFAP, αinternexin, and NeuN [58,59]. Nestin expression has long been known to be upregulated in neural progenitor cells and cortical radial glia [60]. Several studies also report that Nestin protein is expressed in early differentiating neurons in rodents and humans [61][62][63][64][65][66]. Recently, Nestin has been shown to be co-expressed with doublecortin in immature neurons of the cortex [67,68]. GFAP expression is associated with early stages of neurogenesis and is co-expressed with Nestin in cells of the neurogenic niche, such as neural stem cells, and following hypoxia in the cortical parenchyma [69]. GFAP is also expressed in populations of mature astrocytes and can be used as a marker for these glial cells. Later in neurogenesis, cytoskeletal elements such as αinternexin and neuro lament are upregulated and highly expressed as neurons begin to mature [70][71][72]. Alphainternexin is expressed in postmitotic immature neurons as they commence differentiation and can be found colocalized with neuro lament triplets in mature neurons in the CNS [72]. The neuronal splicing regulator, NeuN, is expressed in both post-mitotic immature and mature cortical neurons [73]. We utilized the well characterized stages of neurogenesis in primary cortical cells to assess the effects of microglia during homeostasis and cortical injury on cortical cell viability, proliferation, differentiation, and intracellular signaling. While this coculture system investigates the interactions between cortical cells and EOC2 microglial cell lines, the co-culture system can be used with other microglial cells lines and primary microglia.
Our data suggest that EOC2 microglial-derived soluble cues promote cortical cell viability and enhance proliferation of cortical cells. Microglial-derived cues reduced apoptosis of primary cortical cells following acute mechanical injury in vitro. In injured cortical cell and microglial co-cultures, signi cantly increased expression of neurogenic markers Nestin and α-internexin was present within the site of injury where proliferating cells were observed. Expression of the mature neuronal marker NeuN increased in injured cortical cells outside the injury site when co-cultured with microglial cells. AKT phosphorylation was increased in cortical cells co-cultured with EOC2 microglial cells. Inhibition of AKT phosphorylation reduced the enhanced expression of neurogenic markers in cortical cell and microglial co-cultures. EOC2 microglial cells responding to acute injury downregulated their expression of pro-in ammatory cytokines. These results show that this co-culture in vitro system provides a model to evaluate cortical cell responses to microglial derived soluble cues and to investigate the underlying mechanisms of the functional states of microglia in response to cortical signals during homeostasis or following injury. EOC2 cell lines provided an example of how microglial cells may in uence cortical cell differentiation following injury outside of neurogenic niches. This co-culture system provides a useful tool to further investigate the neuroimmune mechanisms important for primary microglial responses and cortical cell differentiation and survival in vivo.

Results
The in vitro system utilized EOC2 microglial cells cultured on Transwell® inserts suspended directly above uninjured or injured primary cortical cells. These co-cultures were established to investigate the effect of microglial responses on cortical cell proliferation, survival, and differentiation during states of cortical cell homeostasis and following mechanical injury. Primary cortical cultures were established using methods previously described [57]. To characterize the cortical cell types at the time of microglial co-culture, immunocytochemical analysis of cortical cell protein expression was performed at two days in vitro (2DIV, Figure 1). In the control cortical culture, 56.3 ± 0.3% of cells expressed Nestin, 51.3 ± 2.0% expressed α-internexin, 41.7 ± 0.3% of cells expressed the mature neuronal marker TUJI, and 4.3 ± 0.3% expressed glial brillary acidic protein ( Figure 1D). Only 1.9 ± 0.6% of over 1000 cells counted were immunopositive for the microglial marker CD11b (CD11b+) demonstrating that the culture conditions did not support primary microglial cell proliferation and survival (Figure 1 E; over 300 cells were counted per experiment in three separate experiments; ± represents standard error of the mean (SEM)).
Immediately prior to co-culture with EOC2 microglial cells, cortical cells were injured using a sterile stylet to disrupt and remove cortical cells from the cell culture surface [74]. Uninjured cortical cells were used as controls. Uninjured and injured cortical cells were then cultured for two additional days with or without EOC2 microglial cells on Transwell® inserts. The site of injury (indicated by the dashed white line) was observable and few neuro lament immunopositive (NF+) cells or processes were found in the injury site in injured cortical cultures without microglia (Figure 2A Cortical cell viability following injury and co-culture with microglia was measured using 3-(4,5-Dimethylthiazol-2yl)-2,5-Diphenyltetrazolium Bromide (MTT) colorimetric assays that measure metabolic activity in living cells.
Injured cortical cultures without EOC2 microglial cells and uninjured cortical cultures with and without EOC2 microglial cells were also assessed using the MTT assay. Quanti cation of optical density (O.D.) of three separate MTT assays performed in triplicate shows that in the absence of cortical cell injury, microglial-derived soluble cues signi cantly enhanced cortical cell mitochondrial activity by a mean difference of 0.28 ± 0.03 O.D. units (*p<0.05, ± represents SEM, n=3) as compared to uninjured cortical cells cultured in control media alone ( Figure 2G). Co-culture of injured cortical cells with EOC2 microglial cells signi cantly increased mitochondrial activity by a mean difference of 0.39 ± 0.03 O.D. units (*p<0.05, ± represents SEM, n=3) when compared to uninjured conditions without EOC2 microglial cells and by 0.42 ± 0.02 O.D. units (**p<0.01, ± represents SEM, n=3) when compared to injured cortical cells co-cultured without EOC2 microglial cells ( Figure 2G). The metabolic activity of injured cortical cells alone was not signi cantly different from MTT activity in uninjured cortical cell control conditions ( Figure 2G, p>0.05, n=3).
Immunocytochemical assays speci c for measuring cell proliferation and survival were used to examine the response of primary cortical cells to co-culture with soluble signals from EOC2 microglial cells. Cell proliferation was measured by incorporation of a modi ed, uorescently labeled thymidine analogue EdU into newly synthesized DNA. Large eld confocal image analysis of uninjured cortical cells without EOC2 microglial cells showed the presence of EdU+ cells demonstrating that these cultures had at least a limited number of dividing cells upon isolation from the cortex ( Figure 3A). In the presence of EOC2 microglial cells, the number of EdU+ cells increased in uninjured cortical cell culture ( Figure 3B). Mechanical injury of cortical cells stripped away cortical cells from the culture surface as indicated by dashed white lines ( Figure 3C-D). Without EOC2 microglial cells, few EdU+ cells were observed in the damaged area ( Figure 3C). When injured cortical cells were cocultured with EOC2 microglial cells, an increase in proliferating EdU+ cells was seen throughout the culture and within the damaged area ( Figure 3D). Full magni cation of the boxed area within the injured site and EdU+ cells ( Figure 3D) is shown in Figure 3E. Quanti cation of the percent of proliferating cells in uninjured cortical culture without EOC2 microglial cells showed that 45.7 ± 5.0% of the cells were EdU+ ( Figure 3F). In the presence of EOC2 microglial cells, the average percent of EdU+ cells increased to 74.3 ± 5.6%. This 28.6 ± 7.5% increase in EdU+ cells in the presence of EOC2 microglial cells was signi cant ( Figure 3F, *p< 0.05, ± represents SEM, n=3). Following injury, the percent of EdU+ cells in cortical cultures without EOC2 microglial cells was 47.2 ± 9.3% and was not signi cantly different from the control, uninjured cortical cells cultured without EOC2 microglial cells ( Figure 3F, p>0.05, ± represents SEM, n=3). When cultured with EOC2 microglial cells, the percent of proliferating EdU+ cells in injured cortical cultures was 84.3 ± 3.3%. This increase was signi cantly different from the percent of EdU+ cells in uninjured control conditions ( Figure 3F, **p<0.01, ± represents SEM, n=3) and from the percent of EdU+ cells in injured conditions without microglia ( Figure 3F, *p<0.05, ± represents SEM, n=3) . The difference in the percent of proliferating cells between uninjured and injured cortical cells co-cultured with microglia did not reach signi cance ( Figure 3F, p>0.05, n=3).
To evaluate the effect of EOC2 microglial cells on cell survival, Click-iT® uorescent terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays were performed. TUNEL is a common method for detecting DNA fragmentation that results from apoptotic signaling cascades. Figure 4 illustrates TUNEL+ immunocytochemistry observed using large eld confocal imaging of uninjured and injured cortical cultures in the presence or absence of EOC2 microglial cells. In the absence of EOC2 microglial cells, very few TUNEL+ cells were present in uninjured cortical cell cultures ( Figure 4A). Following injury, cortical cells in the absence of EOC2 microglial cells, showed increased TUNEL expression particularly in the area of damage ( Figure 4C).
When injured cortical cells were co-cultured with EOC2 microglial cells, a decrease in TUNEL+ cells were observed both within the injured area and throughout the cell culture ( Figure 4D). Full magni cation of the boxed area within the injury site clearly revealed the presence of TUNEL+ cells ( Figure 4E). Quanti cation of TUNEL staining in uninjured cortical cultures without EOC2 microglial cells shows that 1.72 ± 0.2% of the cells were TUNEL+. Co-culture of EOC2 microglial cells with uninjured cortical cells did not signi cantly alter the percent of TUNEL+ cells (3.44 ± 0.6%, ± represents SEM, n=3, p>0.05, Figure 4F). Following injury, the percent of TUNEL+ cells in cortical cultures without EOC2 microglial cells signi cantly increased to 30.1 ± 4.9% (± represents SEM, n=3, *p<0.05, Figure 4F). When cultured with EOC2 microglial cells, the number of TUNEL+ cells in injured cortical cultures decreased to 5.6 ± 1.2% (± represents SEM, n=3, Figure 4F). The reduction of TUNEL staining by 24.6 ± 4.8% in injured cortical cells co-cultured with microglia was highly signi cant (**p<0.01, ± represents SEM, n=3, Figure 4F). The percent of TUNEL+ cells in uninjured cortical cultures in the presence of EOC2 microglial cells was not signi cantly different from the percent of TUNEL+ cells observed in uninjured cortical cultures used as control (p>0.05, n=3, Figure 4F).
In order to address whether microglial derived soluble signaling cues in uence the number of cells expressing Nestin, GFAP, α-internexin, and NeuN, co-culture experiments were established where cortical cells were injured or left uninjured and then cultured with EOC2 microglia as previously described. After 2DIV, cortical co-cultures were xed and immunostained for Nestin, GFAP, α-internexin, and NeuN. Nuclei were identi ed with DAPI (blue immuno uorescence). In all experiments at least 100 cells were counted per eld from three separate culture elds for each condition. The percent of immunopositive cells per eld was determined by dividing the number of immunopositive cells by the total number of DAPI stained cells per eld. The average percent of immunopositive cells was determined for three biological replicates and compared to controls. Two-way ANOVA with multiple comparisons followed by Tukey's multiple comparisons test was used to determine signi cance of these data. The presence of EOC2 microglial soluble cues signi cantly increased the percent of Nestin+ cells in uninjured cortical cultures by ~23% ( Figure 5A, ***p<0.001) as compared to control, uninjured cortical cultures without microglia. Microglial soluble cues increased the percent of Nestin+ cells within the site of injury by more than 45% and by approximately 38% outside the site of injury as compared to control, cortical cultures without EOC2 microglia. These differences were highly signi cant ( Figure 5A,****p<0.0001). Injury of con uent cortical cultures without EOC2 microglial soluble cues resulted in a signi cant loss of more than 20% of Nestin+ cells within the injury site while no signi cant change in the percent of Nestin+ cells outside the injury site was observed ( Figure 5A). Following injury, co-culture with EOC2 microglial soluble cues signi cantly increased the percent of Nestin+ cells at the site of injury by over 22% and outside the site of injury by approximately 15%  Figure 5C, ****p<0.0001). Co-culture of injured cortical cells with microglialderived soluble cues signi cantly increased the percent of α-internexin+ cells by nearly 73% within the site of injury to 88.8 ± 4.3% and over 74% outside the site of injury to 89.8 ± 4.1% ( Figure 5C, ***p<0.001 and ****p<0.0001, respectively). The percent of NeuN+ cells in uninjured cortical cultures in the presence of EOC2 microglia increased by approximately 5% as compared to uninjured cortical cultures without microglia ( Figure   5D, *p>0.05). Injury of cortical cells signi cantly decreased the percent of NeuN+ cells by nearly 20% at the site of injury as compared to uninjured controls in the absence or presence of microglia ( Figure 5D, ***p<0.001 and **p<0.01, respectively). Outside the site of injury, the percent of NeuN+ cells in cultures without microglia was not signi cantly different from uninjured control conditions ( Figure 5D, *p<0.05) and was approximately 20% higher than the percent of NeuN+ cells seen within the injury site ( Figure 5D, ***p<0.001). The presence of EOC2 microglia increased the percent of NeuN+ cells outside the site of injury by approximately 19% to 48.5 ± 1.1% as compared to sites outside of injury without EOC2 microglial in co-culture ( Figure 5D, **p<0.01). The increase in NeuN+ cells in the presence of microglia following injury is of interest since expression of neurogenic markers should lead to neuronal differentiation as indicated by NeuN expression. Higher power images of NeuN immunoreactivity show NeuN+ cells at the site of injury and outside the side of injury ( Figure 1S). An increase in the number of total cells was seen in cortical cultures in the presence of EOC2 microglial-derived soluble cues as compared to those cortical cultures alone ( Figure 1SA, B). Several cells within the injury site show low NeuN immunoreactivity and did not meet the criteria for positive immunoreactivity ( Figure 1SB).
Analysis of the relative immuno uorescence units (RFU) of immunocytochemical experiments was performed to assess the level of neurogenic protein expression. RFU for each labeled primary and secondary antibody conjugate directed against a neurogenic protein was calculated by measuring pixel intensity for each uorochrome in images acquired with the same exposure settings for all experimental conditions. For each protein, the immuno uorescence of at least 100 cells per eld from three separate culture elds was averaged and compared to controls. These experiments were performed using cortical cultures from three biological replicates ( Figure 6). Following injury, cortical cells were cultured in the presence of EOC2 microglial cells or in media alone. Immuno uorescent images for Nestin and GFAP, α-internexin and GFAP, or NeuN and GFAP showed increased expression of early neurogenic proteins in injured areas when co-cultured with EOC2 microglial cells as compared to media alone. Nestin and GFAP expression was signi cantly enhanced in injured neuronal cultures co-cultured with EOC2 microglial cells as compared to cortical cells cultured alone ( Figures   6A). Cells immunopositive for both Nestin (green) and GFAP (red) are indicated by yellow immuno uorescence.
No signi cant change in Nestin expression was observed in cortical cultures that were injured and not cocultured with EOC2 microglial cells (p>0.05, n=3, Figure 7B). GFAP, expression increased 1.6 ± 0.1fold in uninjured cortical cells co-cultured with EOC2 microglial cells and 1.9 ± 0.2fold in injured cortical cells cocultured with EOC2 microglial cells as compared to control uninjured cortical cells (± represents SEM, n=3, ***p <0.001, Figure 7B). Expression of GFAP also signi cantly increased ~1.6 fold in injured cortical cells without microglial co-culture (***p<0.001, n=3, Figure 7B). Injured cortical cells alone showed a 0.25 ± 0.1fold decrease in NeuN expression as compared to uninjured control cortical cultures (± represents SEM, n=3, *p <0.05, Figure   6B). NeuN protein expression in injured cortical cells co-cultured with EOC2 microglial cells was not signi cantly different than NeuN expression in control conditions as determined by western blot analysis of total protein (± represents SEM, n=3, p>0.05, Figure 7B).
In order to begin to examine changes in the cytokine environment of cortical cells in the presence of EOC2 microglia, multiplex ELISA assays were used to determine the presence of well-characterized cytokines in microglial-conditioned media. Media collected from co-cultures of injured cortical cells and EOC2 microglia was compared to media collected from co-cultures of uninjured cortical cells and EOC2 microglia. Analysis of three separate assays performed in triplicate showed that the concentration of several cytokines was signi cantly different from the levels observed in EOC2 microglial cells-conditioned media when suspended above uninjured cortical cells (Figure 8). The concentration of cytokines measured in media collected from uninjured cortical cell and microglial co-culture was used as the baseline, normalized, and set equal to one ( Figure 7A). When compared to control cytokine concentrations, MCP-1 concentration increased 22.0 ± 0.02% above control levels while IFN-g and TNF-a expression concentration decreased to 41.3 ± 0.07% and 73.5 ± 0.08% below control levels, respectively (± represents SEM, n=3, *p <0.05, Figure 8A). Concentrations of MIP-1a and RANTES decreased by ~20% in media from injured cortical and microglial co-cultures compared to media from uninjured cortical and microglial co-cultures but these decreases were not signi cant (p>0.05, Figure 8B). IL-1a, IL-1b, IL-2, IL-4, IL-6, and GM-CSF were either undetectable or not signi cantly different in conditioned media from uninjured and injured cortical and microglial co-cultures. IL-1a, IL-1b, IL-2, IL-4, IL-6, and GM-CSF were either undetectable or not signi cantly different in conditioned media from uninjured and injured cortical and microglial co-cultures ( Figure 2S). IL-1a, IL-1b, IL-2, IL-4 levels were undetectable in conditioned media from uninjured and injured cortical cultures with and without microglia. IL-6 was measured to be 3.5 ± 14.4 pg/ml (± is SD) in uninjured cortical cell and microglia conditioned media. The large standard deviation does not allow for accurate interpretation of IL-6 in this co-culture system. GMCSF was also detected in uninjured and injured cortical cell and microglia co-culture media. GMCSF concentration was 18.2 ± 13.2 pg/ml (± is SD) and 14.1 ± 16.3 pg/ml (± is SD) in uninjured cortical cell and injured cortical microglial co-culture media respectively ( Figure 2S). The large standard deviation does not allow for accurate interpretation of GMCSF in this co-culture system ( Figure  2S).
To begin to investigate whether the signi cant differences in cytokine concentrations detected in our microglialconditioned medias are due to changes in microglial cytokine expression, RT-PCR was used to compare mRNA levels of EOC2 microglial cells following co-culture with injured or uninjured cortical cells. Since EOC2 microglial cells are physically separated from cortical cells by TranswellÒ permeable supports, RNA can be speci cally isolated from EOC2 microglial cells upon removal of the TranswellÒ from the cortical culture. EOC2 microglial cells in co-culture with injured cortical neurons demonstrated decreased mRNA expression as compared to controls (indicated by the dashed line) for IFN-g (decreased by 22.2±10.2%), MCP-1 (decreased by 79.7 ± 2.9%), MIP-1α (decreased by 60.2 ± 6.7%) TNF-a (decreased by 97.6 ± 4.1%) and RANTES (decreased by 62.5 ±11.6%) (± is SEM, Figure 8B). Decreased expression of MCP-1 mRNA in EOC2 microglial cells suggests that the increase in MCP-1 protein levels was not microglial derived ( Figure 8B). Given that microglial cells, neurons, and astroglial cells can secrete a variety of soluble factors, studies directed at identifying neurogenic microglial, neuronal, or glial soluble signals are being addressed using this in vitro system. Several signaling pathways are activated by soluble signaling molecules during neurogenesis [29,[57][58] and may also underlie the microglial-enhanced neurogenesis observed in our co-culture system. To begin to investigate possible signaling pathways important for microglial-enhanced neurogenesis, injured and uninjured cortical cells in co-culture with and without EOC2 microglial cells were treated with inhibitors for intracellular signaling pathways. We then used the MTT viability assay to screen for those inhibitors that blocked microglialenhanced viability of cortical cells. Microglial co-culture increased viability of uninjured and injured cortical cultures as compared to cortical cultures alone as shown previously (***p <0.001, see supplemental data, Figure   3SA-D). Inhibitors for MEK (PD98059), p38 MAPK (SKF86002), PKCα/βI/βII/γ (GF109203X), and Janus Kinase 2 protein (AG490) did not block the increased metabolic activity and viability of cortical cells co-cultured with EOC2 microglial cells. AG490 and GF109203X at 40 µM did signi cantly in uence viability of cortical cultures but these affects were not speci c for microglial-enhanced viability ( Figure 3S). LY294002, an inhibitor of PI3K, and downstream AKT phosphorylation, speci cally reduced microglial-enhanced cortical cell viability by ~50% or 0.52 ± 0.02 and 0.51 ± 0.02 optical density (O.D.) units at 10 and 40µM respectively as compared to untreated, uninjured cortical cells co-cultured with EOC2 microglial cells (± represents SEM, n=3, ****p <0.0001, Figure 9A). Following injury, 10 and 40µM LY294002 treatment signi cantly reduced microglial-enhanced cortical cell viability by ~73% or 0.73 ± 0.02 and 0.74 ± 0.02 O.D. units (± represents SEM, n=3, ****p <0.0001, Figure 9A). Treatment of uninjured or injured cortical cells in the absence of EOC2 microglial cells with LY294002 did not signi cantly affect metabolic activity as measured by MTT (± represents SEM, n=3, p >0.05, Figure 9A). Western blot analysis of cortical cells con rmed that the presence of EOC2 microglial cells in suspension above cortical cultures increased phosphorylation of AKT, a PI3K target, in cortical cells ( Figure 9B).
Analysis of western blots showed that the presence of EOC2 microglial cells in uninjured cortical cultures increased AKT phosphorylation 3.6 ± 1.0fold as compared to uninjured cortical cells alone (± represents SEM, n=3, *p <0.05, Figure 9C). Injury alone did not signi cantly increase (~0.8fold) phosphorylation of AKT as compared to control levels ( Figure 9C, p>0.05). Following injury and co-culture with EOC2 microglial cells, AKT phosphorylation increased 5.0 ± 1.0fold as compared to injured cortical cells alone (± represents SEM, n=3, **p <0.01, Figure 9C). This increase was also signi cantly different from AKT phosphorylation levels measured in injured cortical cells without microglial co-culture (4.2 ± 1.0fold increase, *p<0.05, ± represents SEM, n=3, Figure   9C). These experiments suggest that phosphorylation of AKT may be necessary for microglial-enhanced expression of speci c neurogenic proteins.
To further investigate the necessity of the AKT phosphorylation for microglial-enhanced expression of neurogenic proteins, immunocytochemical analysis of neurogenic protein expression was assessed in injured and uninjured cortical cells co-cultured with EOC2 microglial cells in the presence of 40mM LY294002.
Uninjured cortical cells incubated with LY294002 but without EOC2 microglial cells served as the control and baseline for the normalization of protein expression. Incubation of cortical cultures with LY294002 completely blocked the increase in AKT phosphorylation and reduced neurogenic protein expression seen in cortical cells when cultured with EOC2 microglial cells ( Figure 9D). LY294002 reduced Nestin expression by 0.44 ± 0.3fold in uninjured cortical cells co-cultured with EOC2 microglial cells as compared to controls (**p<0.01, n=3, Figure  9E). LY294004 treatment reduced Nestin expression by 0.26 ± 0.0fold in injured cortical cells co-cultured with EOC2 microglial cells as compared to uninjured controls (***p<0.001, n=3, Figure 9E) and by 0.36 ± 0.1fold in injured cortical cells cultured without EOC2 microglial cells (**p<0.01, n=3, Figure 9E). Expression of the neuronal intermediate lament α-internexin was signi cantly reduced by 0.59 ± 0.1fold (****p<0.0001) in injured cortical cells co-cultured with EOC2 microglial cells when compared with uninjured controls and by 0.40 ± 0.1fold (**p<0.01) when compared with injured cortical cells not cultured with EOC2 microglial cells ( Figure 9E, n=3). Additionally, microglial-enhanced expression of GFAP in uninjured and injured cortical cells was signi cantly reduced following treatment with LY294002 ( Figure 9D and 9E). GFAP expression was reduced by 0.76 ± 0.0fold as compared to uninjured controls and 0.85 ± 0.1fold as compared to injured cortical cells without microglial co-culture (****p<0.0001, n=3, Figure 9E).
Immunocytochemical analysis was also used to evaluate the effect of blocking PI3K activity and AKT phosphorylation ( Figure 10). Co-cultures were established as previously described and cortical cells were incubated with LY294002. After co-culture, cells were xed and evaluated for the expression of the neurogenic markers-Nestin, a-internexin, and GFAP ( Figure 10A) since the expression of these markers was enhanced in our co-culture system. DAPI immuno uorescence was used to identify nuclei ( Figure 10A). Immunocytochemical analysis supported western blot data showing that LY294002 treatment and inhibition of AKT phosphorylation reduced Nestin, α-internexin, and GFAP expression in injured cortical cells co-cultured with EOC2 microglial cells ( Figure 10A). Quanti cation of neurogenic protein expression was determined by calculating RFU for each labeled primary and secondary antibody conjugate directed against neurogenic proteins in images acquired with the same exposure settings for all experimental conditions as described previously. Evaluation of RFU for Nestin expression showed that co-culture with EOC2 microglial cells signi cantly enhanced Nestin in uninjured or injured cortical cultures as compared to Nestin expression in cortical cultures alone (*p<0.05, Figure 10B).

Discussion
This study examines the ability of EOC2 microglial cells to support the viability, proliferation, neurogenesis, and survival of primary cortical cells outside of the classic neurogenic niche during homeostasis and following mechanical injury. Stem cell progenitors outside of the classic neurogenic niches have the capacity for neurogenesis during normal homeostasis and following injury or disease given exposure to the proper combination of neurogenic cues [47,[48][49][50]. Cortical microglial cells are a potential source of neurogenic signaling molecules [4,[39][40][41][42][51][52][53][54][55][56]. Microglia have been shown to contribute to neuronal synapse development, survival, and neurogenesis in neurogenic niches of the CNS during development and following injury [5-6, 8, 10-23, 51-56]. Neurogenic potential of microglial-derived cues may be controlled by the mechanism and duration of activation since infection, trauma, and disease also stimulate pro-in ammatory responses from microglia that can lead to neurotoxicity and neurodegeneration [26-27, 43, 75,76]. Several in vitro studies suggest that microglial-conditioned media is neurotoxic depending on the mechanism of microglial activation [43, 44-46, 75, 76]. Recent characterization of primary microglia using single cell sequencing and other methods shows that microglia are highly diverse and that subpopulations of functional microglia are present within speci c brain regions and in responses to particular environmental cues [45,46]. Because of the diversity of primary microglia, this study presents an in vitro model system using the EOC2 microglial cell line and primary cortical cells in co-culture to begin to examine the neurogenic potential of microglial soluble cues during cortical cell homeostasis and following cortical cell injury. Suspension of microglia on Transwell® cell culture inserts directly above cortical cells allows for evaluation of how the cortical environment is in uenced by microglialderived soluble cues. The methods of isolating primary cortical cells utilized in this in vitro system have been well characterized [52] and our data show that these cells express primarily neurogenic and neuronal cell markers with evidence of less than 5% astrocytic protein expression and less than 2% microglial protein expression ( Figure 1F). Using this system where EOC2 microglia and primary cortical cells are not in direct contact also allows for protein and RNA to be collected independently from either cell population. This reductionist in vitro system allows for preliminary questions to be asked about microglial neurogenic potential in the cortex and can be used for future studies with other microglial cell lines and primary microglial cells.
Results from experiments presented here show that soluble cues from EOC2 microglial cells responding to primary cortical cells enhanced proliferation, suppressed apoptosis, and promoted the expression of neurogenic and mature neuronal markers in the primary cortical cultures. Immunocytochemical analysis and MTT assays of cortical cells in co-culture with suspended EOC2 microglial cells following injury showed that EOC2 microglial cells enhanced viability of Neuro lament+ cells and increased extension of Neuro lament+ processes into injured areas of cortical cultures (Figure 2). Neuro lament is expressed in neuronal progenitor cells and mature neurons and is associated with structural maturation of neurons and axonal function [69,71] supporting these data showing that microglial-derived cues can promote neurogenic protein expression in cortical cells outside of neurogenic niches.
To more speci cally address how EOC2 microglial cells in uence cortical cell differentiation, the in vitro system was used to assess cell proliferation, apoptosis, and expression of neurogenic markers in uninjured and injured co-culture conditions. Immunocytochemical analyses revealed that co-culturing uninjured or injured cortical cells with EOC2 microglial cells signi cantly increased the percent of EdU+ cells in cortical cultures as compared to controls (Figure 3). EdU+ cells were observed within the injury site and throughout the culture (Figure 3). Particularly striking is the proliferation of cortical cells within the injury site. Enhanced proliferation is not seen within the injury site in injured cortical cultures without microglia (Figure 3). Microglial responses in the subgranular and subventricular zones can in uence the number of mature neurons that are generated by regulating neuronal stem cell and neuronal progenitor cell proliferation and differentiation throughout life [4-5, 10-11, 21, 39, 44]. In the subgranular zone and subventricular zones the microglial responses are dependent upon location and speci c developmental time points [4][5][51][52]. Enhancement of cortical cell proliferation outside of the neurogenic niche as shown using this in vitro system suggests that the microglial-derived soluble environment of cortical cells is able to signi cantly support cortical cell proliferation during homeostasis and following injury (Figure 3). Apoptosis as measured by TUNEL staining showed that co-culture with EOC2 microglial cells signi cantly reduced apoptosis in cortical cells following injury (Figure 4). Microglia can in uence apoptosis in neurogenic zones and this response is dependent upon the combination of proin ammatory cytokines, growth factors, and phagocytic activity of microglia [5, 22-23, 26-27, 29]. Recently it has been shown that neuroblasts can be recruited from the subventricular zone of hippocampus to sites of injury in the cortex by microglial-derived speci c cues suggesting that microglia, at least at speci c times following injury, function to attract differentiating neuronal progenitors into cortical tissue [77]. Further studies suggest microglia may also act locally to stimulate neurogenesis of responsive cortical cells at noncanonical neurogenic regions [47,[51][52][53][54][55][56]79].
We propose that this in vitro model system is useful to investigate whether microglial soluble signals following cortical injury have neurogenic potential outside of previously investigated subventricular and subgranular zones. Further this in vitro system allows for analysis of neurogenic protein expression and activation of intracellular signaling pathways mediated by microglia in cortical cells. . This effect was not limited to the injury site since an increase in the percentage of a-internexin+ cells was also seen outside of the injury site. Neuronal progenitor cells expressing a-internexin were clearly distinct from GFAP expressing cells. Immunocytochemical analysis of GFAP immunoreactivity in injured cortical cultures cocultured with EOC2 microglial cells showed an increase in GFAP+ cells with the morphology of mature astrocytes as compared to injured cortical neurons alone ( Figure 5-6). Western blot analysis con rmed a signi cant increase in GFAP protein expression in injured cortical cells cultured with EOC2 microglial cells ( Figure 7). It is possible that microglia responding to injury stimulate neurogenesis and astrogliogenesis. Microglial soluble signals have been shown to stimulate astrocyte differentiation from progenitor cells [55][56].
In order to determine whether EOC2 microglial cells were able to promote differentiation of mature neurons and enhance the survival of existing neurons following injury we used immuno uorescent and western blot analysis to visualize NeuN expression. NeuN is a marker for mature neurons [73]. The percent of NeuN+ cells increased in injured cortical cell and microglial co-cultures ( Figure 5D). Enhanced NeuN immuno uorescence was observed in injured cortical cultures exposed to microglial-conditioned media (Figures 6, 1S). Few NeuN+ cells are present in injured areas when neurons are cultured in control media. Western blot analysis showed an increase in NeuN expression when uninjured or injured neurons were exposed to microglial-conditioned media but this increase did not reach signi cance when compared with NeuN expression in cortical cultures without EOC2 microglial cells ( Figure 7A-B). There are several possible explanations for this result. For western blot analysis, protein was isolated from the entire neuronal culture which included injured as well as uninjured regions and would be less speci c for detecting proliferation and maturation in injured areas alone. Maturation of neuronal progenitors takes up to one week in culture [57][58]. The outcome of increased Nestin, GFAP, and ainternexin will require longer culture conditions. Our supplemental data showing increased NeuN immuno uorescence in the area of injury following cortical co-culture with EOC2 microglial cells, suggests that microglial-enhanced neurogenesis and differentiation of mature neurons is occurring ( Figure 1S, 2S). Additional, long-term experiments are underway to determine whether maturation of mature neurons results from effector microglial-enhanced neurogenesis. Taken together, these data suggest that microglial soluble signals released following co-culture with cortical cells during homeostasis and more so during activation by cortical injury promote the proliferation and survival of neurogenic cells.
Multiplex ELISA assays of microglial-conditioned media from co-culture experiments with uninjured and injured cortical cells revealed that the expression of several cytokines signi cantly changed following EOC2 microglia stimulation by neuronal injury. Speci cally, multiplex ELISA data showed signi cant upregulation of MCP-1/CCL2 and downregulation of IFN-g, MIP-1a, TNF-a and RANTES ( Figure 8A). Upregulation of MCP-1/CCL2 is interesting since MCP-1/CCL2 is associated with in ammation as well as subventricular zone and neocortical neurogenesis and neurogenic migration [85][86][87][88][89][90]. MCP-1/CCL2 is expressed by microglia, neurons, neural stem cells and astrocytes [85,89]. Our data suggest that the increase in MCP-1 protein is unlikely to be microglialderived since MCP-1 mRNA levels are lower in EOC2 microglia responding to injury than in control microglia ( Figure 8B following co-culture of microglia with injured neurons may also favor neurogenesis in cortical cultures [54,91]. In ammatory cytokines are known to act at speci c concentrations and in certain combinations to regulate neurogenesis and neuronal survival [29]. More speci cally, low levels of IFN-g have been shown to stimulate both neurogenesis and oligodendrogenesis [20,29,54,56]. TNF-a, while primarily associated with in ammatory responses associated with neurotoxicity [29], has varied effects on neurogenesis and can stimulate neurons to . However, these ligands may also play an important role in the development and migration of neurons [29,50,89]. Interestingly, secretion of pro-in ammatory cytokines by microglia such as IL-1b, IL-6, inducible nitric oxide synthase (iNOS), and TNF-a are suppressed by RANTES signaling [51,93]. These results suggest that RANTES may function to regulate microglia effector function and contribute to the neurotrophic and neurogenic properties of microglia.
Microglia can also secrete growth factors and neurotrophins [39,61]. Neurotrophin release may be enhanced when microglia respond to cortical injury [94,95]. Other soluble signals such as prokineticins may contribute to the microglial-enhanced neurogenesis presented in this in vitro system [77]. We suggest, as have others [18] that microglial effects on neuronal survival, proliferation, and differentiation are largely dependent upon the composition of soluble signals that are released by microglia in response to stimulation. Continued investigations are underway to better dissect the complex milieu of neurogenic soluble signals released by microglia or other secretory cells in the presence of microglia.
Taken together, our data provide an enticing view of the dynamic and multifunctional role of microglia in the cortex. Microglia responding to cortical cues may stimulate local neurogenesis and potential repair after injury. Downregulation of pro-in ammatory cytokine production by microglia could allow for increased proliferation, reduced apoptosis and increased neurogenesis [5,29]. Our results suggest that the microglia enhance neurogenesis and promote neuronal survival by stimulating the PI3K/AKT signaling pathway in cortical cells. While other intracellular signaling pathways are likely also stimulated, the inhibition of the PI3K/AKT pathway and not other pathways previously implicated in neurogenesis blocked EOC2 microglial-enhanced neurogenesis. Further elucidation of the intracellular mechanisms regulating neurogenic function of microglia is essential for understanding the intrinsic neuroprotective role of immune activity in the CNS and may aid in the development of methodologies to promote such activity during neurodegenerative disease or following traumatic injury. The in vitro model system presented here provides an experimental tool to investigate the mechanisms of primary microglial responses to cortical injury outside of the neurogenic niche.

Conclusions
Here we present an in vitro model system allowing for the assessment of microglial-derived soluble signals on cortical cell viability, proliferation, and differentiation during homeostasis or following cortical injury. Using this model system intracellular signaling pathways are readily investigated in isolated primary cells or cell lines.
These speci c studies show that cortical cell injury activates neurogenic properties in EOC2 microglia. The speci c EOC2 microglial cell line's response to cortical injury results in at least the reduction of pro-in ammatory cytokine gene expression and cytokine release. EOC2 microglial-derived soluble signals produced during homeostasis and, more so following activation by acute cortical injury, enhance neurogenesis by upregulating AKT signaling in cortical cells. Increasing our understanding of the mechanisms that drive cortical cell proliferation, differentiation, and survival as stimulated by microglia will provide insight into the neuroprotective role of immune activity in the CNS.

Isolation of rat cortical cells
Female timed-pregnant Sprague Dawley rats (200-250 g) were purchased from the Charles River Laboratories (USA). Use of animals was performed in strict accordance with the Institutional Animal Care and Use committee guidelines as approved by the IACUC committee at Creighton University (protocol #0793). Timed Sprague Dawley dams were housed for up to 3 days in Creighton's Animal Resource Facility that is AALAC accredited. Ad libitum food and water and normal 24h light dark schedules were followed. Rat cortical cultures were established as described by Meberg and Miller [57]. Brie y, Sprague-Dawley dams were euthanized by CO 2 asphyxiation. For CO 2 asphyxiation, dam were placed in a clear chamber with CO 2 delivery at 20% of the chamber volume per minute. After 1 min of cessation of all respiratory movement, toe pinch tests were performed to determine a lack of re exive responses and a thoracotomy was used to ensure death. E16-E18 rat embryos were removed from their placental sacs and immediately decapitated. The brains were removed and the cerebral cortices were dissected from day 15-16 embryonic Sprague-Dawley rats (Sasco, Wilmington, MA), mechanically dissociated in Ca 2+ /Mg 2+ -free Hank's balanced salt solution, with 0.035% sodium bicarbonate and 1mM pyruvate (pH 7.4) following 15 min digestion with 2.5% trypsin. Trypsin was neutralized with Dulbecco's Modi ed Eagles Media (DMEM: Hyclone, Thermo sher Scienti c, Waltham, MA) supplemented with 10% fetal bovine serum and the cell suspension was washed three times and resuspended with neurobasal media supplemented with B-27® and penicillin/streptomycin (Thermo sher Scienti c, Waltham, MA). Cells were then plated onto poly-D-lysine coated plates and coverslips (Sigma, St. Louis, MO) at a density of 1.5 x 10 6 cells/well in 6-well plates and 5 x 10 5 cells/well in 24-well plates and were maintained at 37 ºC in 5% CO 2 in neurobasal supplemented media. Each cortical culture from 1 pregnant dam was considered to be a biological replicate because embryonic brain tissue was used for co-culture, immunocytochemistry, multiplex ELISA, RT-PCR and western blot experiments all performed in triplicate. In total, 24 Sprague Dawley rats were used for these data.
Cultivation of microglia EOC 2 microglia isolated from brain tissue of Mus musculus were purchased from American Type Culture Collection (ATCC CRL-2467; Manassas, VA) and were maintained in DMEM (Hyclone, Thermo sher Scienti c, Waltham, MA) supplemented with 10% fetal bovine serum, 1% l-glutamine, 1% penicillin/streptomycin, and 20% LADMAC conditioned media. Cells were grown in 100-mm tissue culture dishes at 37 ºC in 5% CO 2 and allowed to reach 80% con uency before the cells were passed. LADMAC conditioned media was collected from Mus musculus bone marrow derived LADMAC cells (ATCC CRL-2420; Manassas, VA) 5-7 days after initial plating of cells at 1 x 10 5 in Eagle's Minimal Essential Media (MEM; Hyclone) supplemented with 10% FBS, 1% l-glutamine, 0.1mM nonessential amino acids, 1.0 mM sodium pyruvate, and 1% penicillin/streptomycin. LADMAC conditioned media was collected, lter sterilized, and frozen until needed as a media supplement for microglia. LADMAC conditioned media was used to provide colony-stimulating factor 1 (CSF-1) to the microglial cultures as outlined by the ATCC culture instructions for EOC microglial cells (ATCC CRL-2420; Manassas, VA).

Neuronal -Microglial Co-Cultures
Primary cortical cells were cultured for 48 hours then either injured by mechanical transection using a sterile stylet or left uninjured [74]. Brie y, mechanical transection using a sterile stylet involves the application of the sterile stylet tip directly to the cortical cell culture. Pressure is placed on the stylet while dragging the stylet tip across the cortical cell culture to form parallel sites of injury in the cortical culture. Microglia were pre-seeded directly onto 6-well permeable Transwells® at 5 x 10 5 cells/well or onto 24-well Transwells® at 4 x 10 4 cells/well (Corning,Tewksbury, MA) and cultured for 24 hours before being suspended above cortical cells using Transwells® in the co-culture model system. Microglia seeded onto Transwells® and injured or uninjured (control) primary cortical cells were co-cultured for an additional 48 hours in unsupplemented neurobasal media prior to cellular assays. For immunocytochemical analysis of EOC2 microglia only, EOC2 microglia were cultured on glass coverslips and placed into Transwells® and then removed for xing and immunostaining (see below). Qualitative and quantitative analysis of immunocytochemistry was performed by acquiring images with a Leica DMI4000B inverted microscope with a cooled CCD camera (Q Imaging, Surrey, BC) and uorescent capabilities. Quanti cation of the percent of cells expressing neurogenic markers was determined by counting the number of immunopositive cells for each marker and dividing that number by the total number of cells counted in the eld. Experiments were performed in triplicate with at least 300 cells counted manually per experiment for each condition. Quanti cation of relative uorescence intensity units (RFU) was calculated by subtracting pixel intensity from the background immuno uorescence of each uorochrome. Experiments were performed in triplicate with at least 300 cells counted manually per experiment for each condition. In all experiments, images were analyzed with Volocity (PerkinElmer,USA), and ImageQuant (GE Healthcare, USA) software were used for image analysis and presentation. For image data, 3 eld views of at least 100 cells from 3 separate experiments were analyzed for each condition.

Measurement of Cell Viability
Viability of cortical cells and microglia were measured by metabolism of thiazolyl blue, 3-[4,5-dimethylthiazol-2yl]-2,5-diphenyltetrazolium bromide (MTT, Sigma Aldrich). Injured and uninjured cortical cells cultured with and without microglia were incubated with 100 µl MTT in 1 ml of media for one hour. Media was removed and cells were dissolved in 300 µl dimethylsulfoxide (DMSO) and aliquoted to 100 µl/well in 96-well plates. Absorbance was read at 540-590nm on an ELISA plate reader. Three experiments were performed in triplicate.

Measurement of Cortical Cell Proliferation
Cell proliferation was measured using Click-iT® EdU Alexa Fluor 647 according to the manufacturer's instructions (C10340, Thermo Fisher Scienti c, Waltham, MA). Brie y, Click-iT® EdU Alexa Fluor 647 is a modi ed thymidine analogue EdU (5-ethynyl-2'-deoxyuridine, a nucleoside analog of thymidine) that is incorporated into newly synthesized DNA. The EdU is uorescently labeled with a photostable Alexa Fluor® dye during the click reaction. Uninjured and injured cortical cells co-cultured with and without microglia for 2 DIV (days in vitro) were xed in 3.7% formaldehyde in PBS for 15 min at RT. Fixed cells were washed twice with 1 ml of 3% BSA in PBS. Cells were permeabilized in 0.5% Triton®x-100 for 20 min at RT, washed and 1X Click-iT® EdU reaction cocktail was added for 30 min at RT. The reaction cocktail was removed, cells were washed in 3% BSA and PBS, counterstained with DAPI, mounted and imaged for analysis. Imaging was performed using IBIF Leica TCS SP8 MP Confocal Microscope at 20x magni cation. Experiments were performed in triplicate with at least 300 cells counted manually per experiment for each condition. Volocity (PerkinElmer,USA) and ImageQuant (GE Healthcare, USA) software were used for image analysis and presentation.

Measurement of Cortical Cell Apoptosis
Apoptosis of cortical cultures was measured following co-culture with microglia for 2 DIV (days in vitro) using a Click-iT® TUNEL Alexa Fluor 488 imaging assay (C10245, Thermo Fisher Scienti c, Waltham, MA). Manufacturer's instructions were followed. Brie y, injured and uninjured cortical cells cultured with or without microglia were xed with 4% PFA in PBS for 15 min then permeabilized with 0.25% Triton-X® 100 for 20 min.
Each condition was incubated with 100 µl of TdT reaction buffer for 10 min at RT then removed. Cells were incubated in 100 µl TdT reaction cocktail for one hr at 37 o C. Cells were washed twice in 3% BSA in PBS for 2 min then incubated with Click-iT® reaction buffer with additive for 30 min at RT protected from light. Cells were rinsed and counterstained with DAPI, mounted, and cover slipped for analysis. Imaging was performed using IBIF Leica TCS SP8 MP Confocal Microscope at 20x magni cation. Experiments were performed in triplicate with at least 300 cells counted manually per experiment for each condition. Volocity (PerkinElmer,USA) and ImageQuant (GE Healthcare, USA) software were used for image analysis and presentation.

ELISA Analysis
Conditioned media was collected from uninjured and injured neuronal and microglia co-cultures from three separate experiments and cytokine expression was determined by Q-Plex™ mouse cytokine -In ammation multiplex assay. Concentrations of mouse microglia-derived cytokines MCP-1, IFN-g, MIP-1a, TNFa, RANTES, IL-1a, IL-1b, IL-2, IL-4, IL-3, IL-6, IL-10, IL-12, 1L-17 and GM-CSF were evaluated by Quansys Biosciences (#110449MS, Logan, UT). Cytokine concentrations in media collected from uninjured cortical cells cultured with microglia were used as the reference and control for these experiments. In order to use cytokine concentrations from uninjured cortical and microglial co-cultures as our control condition, each cytokine concentration measured in the uninjured cortical cell and microglia co-culture condition was normalized and set equal to one.
Cytokine concentrations in media collected from injured cortical cells co-cultured with microglia were measured, normalized, and expressed as the percent change in cytokine concentration as compared to uninjured control concentrations for that cytokine. Multiplex ELISA assays were run in triplicate in three biological replicate experiments. Signi cance of the percent change from control was determined using student T test with Bonferroni correction. The percent change in cytokine concentration was considered signi cant if p<0.05, error bars represent the standard error of the mean of the percent change. Cytokine levels that were not consistently detected in either condition or were not signi cantly different in control and experimental conditions are not shown.

RT-PCR Analysis
For real-time PCR analysis of cytokines, total RNA of EOC2 microglial cultured on TranswellÒ permeable inserts that were physically separated from cortical cells was extracted using the mirVana miRNA Isolation kit (Ambion). An amount of 200 ng total RNA was reverse-transcribed using the Invitrogen™ NCode™ miRNA First-Strand cDNA Synthesis Kit (Thermo Fisher Scienti c). Comparative real-time PCR was performed using the Invitrogen™ SYBR GreenER™ qPCR SuperMix Universal (Thermo Fisher Scienti c) on the Bio-Rad CFX96 Touch™ Real-Time PCR Detection System. Primers were purchased from QIAGEN (Ccl3, Ccl5, Ifng, Mcpt1, Tnfα, Gapdh).
Normalization was performed using GAPDH. Relative expression was calculated using the comparative Ct (ΔΔCt) method.
Lysate supernatant were collected and heated at 95 °C for 5 minutes with 4X sample buffer plus 10mM DTT. Denatured protein samples were separated by SDS-PAGE gel electrophoresis on 10% TGS gels. Proteins were transferred to PVDF membranes in Tris-glycine transfer buffer. After transfer, membranes were blocked using Uninjured and injured cortical cells that were not cultured with microglia were used as controls. For control experiments, DMSO vehicle diluted in culture media was used in the experiments. After 48 hours, cells were analyzed using MTT assays (see above) or xed with 4% PFA in PBS to observe expression for neurogenic markers, Nestin, α-internexin, and GFAP using immunocytochemical methods as described above. To quantify imaging data, 3 eld views of at least 100 cells from 3 separate experiments were analyzed for each condition.
Immunoprecipitation for AKT/pAKT Analysis Immunoprecipitation for AKT and pAKT was used to increase speci city and detection of AKT protein in cellular lysates. Cellular cultures were lysed as described above and each condition was split into two aliquots (200 µl each). Primary antibodies AKT (pan) (C67E7) rabbit mAB (Cell Signaling Technology Cat# 4691, RRID:AB_915783) and Phospho-AKT (Thr308) rabbit mAB (Cell Signaling Technology Cat#9275, RRID:AB_329828) were added at 1:50 for each sample and rotated overnight at 4 °C. A 50% slurry of EZview Red Protein A A nity Gel Beads (Cat#P6486, EDM Millipore Sigma, Darmstadt, Germany) were added at 1:10 for each sample and rotated for 1 hour at 4 °C. Cells were centrifuged at 8,200g for 1 minute and washed with lysis buffer 3 times. Samples were heated at 95 °C for 5 minutes with 25 µl 3X sample buffer. Samples were run on 4-20% gradient SDS-polyacrylamide gels (Cat#4561096, BioRad, Hercules, CA) using SDS-PAGE and then transferred to PVDF membrane. After transfer, membranes were blocked using BSA Blocking Buffer™ in TBS for 1 hour and then gently rocked with the primary AKT or pAKT antibody at 1:1000 in BSA Blocking Buffer™ in TBS overnight at 4°C. Blots were washed and incubated in secondary anti-rabbit HRP conjugated antibody for 1 hour at RT. Membranes were developed using chemiluminescence as described above. Three separate experiments were performed. Statistical analysis involved analysis of densitometric images acquired with BioRad ChemiDoc QRS imaging system and software (BioRad,Hercules CA) which were performed as described above.

Statistical Analysis
Data are expressed as mean values and error bars represent standard error of the mean (SEM). Student T test with Bonferroni's correction or one-way ANOVA followed by Tukey-Kramer post hoc tests were performed where appropriate. For determination of signi cant differences between percents and for multiple comparisons between culture conditions, one-way or two-way ANOVA followed by Tukey-Kramer multiple analyses post hoc tests were used. Values of p<0.05 were considered to be signi cant. All statistical analyses were performed with Graphpad Prism 8 (La Jolla, CA). Publication costs are shared between these funding mechanisms and Creighton University.

List Of Abbreviations
Authors contributions: KC contributed to the design and implementation of the co-culture system, immunocytochemistry and intracellular signaling pathway analysis. KC was also a major contributor to the writing of this manuscript. NWM contributed to co-culture implementation, RT-PCR and analysis and was a signi cant contributor to the writing of this manuscript. ERW contributed to co-culture experimental design and immunocytochemistry analysis. AE contributed to ELISA experimental design, implementation and analysis. JC contributed to immunocytochemistry and western blot analysis. MB contributed to data analysis and to review and revision of this manuscript. XMC contributed to experimental design and to the writing of this manuscript. AS was responsible for conceptualization of the experiments, experimental design, data analysis, and writing of the manuscript. All authors read and approved the nal manuscript.  injured cortical cells alone or in co-culture with microglia in the presence of PKCα/βI/βII/γ inhibitor GF109203X. D) Quanti cation of viability of uninjured and injured cortical cells alone or in co-culture with microglia in the presence of Janus Kinase 2 inhibitor AG490. For each concentration, one-way ANOVA was used to determine signi cance of the inhibitor. *p,0.05, **p<0.01, ***p<0.001, ****p<0.0001, # indicates that OD values were signi cantly different from that of control, uninjured cortical cells alone, ns indicates not signi cant. Figure 1 Primary rat cortical cells in the in vitro system. A) Phase contrast image of cortical cells in primary culture. Cells have rounded cell bodies and extension of processes is visible. B) Fluorescent image of primary cortical cells stained for Nestin and GFAP protein expression. C) Fluorescent image of primary cortical cells showing α-internexin+ and GFAP+ cells. D) Fluorescent image of primary cortical cells showing TUJI+ and GFAP+ cells. E) Fluorescent image of primary cortical showing CD11b+ and GFAP+ cells. DAPI (Blue) was to identify nuclei of all cells within an imaged eld. Scale bar represents 50 µm and is the same for all images. All images were taken with the 20X Leica objective. F) Quanti cation of protein expression. The percent of cells immunopositive for each marker compared to the total number of cells was calculated for three separate elds with each eld having at least 100 cells or over 300 cells total. Error bars represent SEM.

Figure 2
Phase contrast and uorescent images showing the effect of microglial co-culture with primary cortical cells following injury in the in vitro system. A). Phase contrast image of primary cortical cells following injury mediated by stylet transection. The dashed white line indicated the site of injury. Two days in vitro (2DIV) following injury a few cells can be observed beyond the injury site. Black scale bar represents 100 µm. B) Fluorescent image of NF+ cortical cells 2DIV following injury. NF+ immunoreactivity indicated the location of neuronal progenitors and neurons in primary cortical cultures following injury and 2DIV without microglia. C) Phase contrast image of primary cortical cells following injury and co-culture with microglia for 2DIV. An increase in the number of cortical cells at and beyond the site of injury is visible (dashed white line represents the site of injury). D) Fluorescent image of NF+ cortical cells in microglia co-culture 2DIV following injury. Dashed white indicates the site of injury. A-D). Images were taken with the 20X Leica objective. E) Phase contrast images of cultured activated microglia cells co-cultured with primary neurons on Transwell® inserts. F) Microglia used in this co-culture system are immunopositive for the microglial marker CD11b-Alexa 488 (green immuno uorescence) antigen. E-F) Scale bar represents 50 µm. Images were taken with the 40X Leica objective. G) Viability of uninjured cortical cells or injured cortical cells with or without microglial co-culture for 2DIV was assessed using the MTT assay. MTT activity was read at OD 595 nm for three separate experiments. Data show average OD at 595nm with error bars representing SEM. One-way ANOVA with multiple comparisons were performed to determine the signi cance of viability data. Signi cance is *p<0.05, ** p<0.01. All images were taken with the Leica confocal using the 20 X objective. Scale bar represented 100 µm and applies to A-D) where E) shows full magni cation and partial view of the imaging eld. F). Quanti cation of EdU+ primary cortical neurons. The percent of EdU+ cells in each experiment was calculated from the total number of cells counted. Cells were identi ed by using Hoechst to immunostain nuclei (Blue). In three experiments, 300 or more cells in each eld were counted for each condition. Error bars represent SEM. One-way ANOVA followed by Tukey's multiple comparisons test was performed to determine the signi cance. Signi cance is *p<0.05, ** p<0.01. Microglial co-culture reduced apoptosis of injured cortical cells. A) Click-iT® TUNEL Alexa Fluor 488 (green) immuno uorescent staining of uninjured cortical cells cultured in the absence of microglia. B) Click-iT® TUNEL Alexa Fluor 488 immuno uorescent staining of uninjured cortical cells co-cultured with microglia. C) Click-iT® TUNEL Alexa Fluor 488 immuno uorescent staining of injured cortical cells cultured in the absence of microglia. At the site of injury and beyond, TUNEL+ cells were present. D) Click-iT® TUNEL Alexa Fluor 488 immuno uorescent staining of injured cortical cells co-cultured with microglia. A reduction of TUNEL+ cells at the site of injury and beyond was noticeable and signi cant (F). Hoechst immuno uorescence (blue) indicates nuclei A-D. All images were taken with the Leica confocal using the 20 X objective. Scale bar represented 100 µm and applies to A-D) where D) shows full magni cation and partial view of the imaging eld. F) Quanti cation of TUNEL+ primary cortical cells. The number of TUNEL+ cells was compared to the total number of Hoechst immunostained nuclei. In three experiments, 300 or more cells in each eld were counted for each condition for quanti cation. One-way ANOVA with multiple comparisons were performed to determine the signi cance of viability data. Signi cance is *p<0.05, ** p<0.01, ns is not signi cant.  determined by comparing the number of immunopositive cells to all cells that were counted in each culture condition. Cells were identi ed by using Hoechst to immunostain nuclei. Cortical cells cultured alone served as control. Uninjured and injured cortical cells were cultured alone or in the presence of EOC2 microglia suspended on Transwells®. Quanti cation of immunocytochemical data within the injury site and outside the injury site is shown. Error bars represent SEM. Two-way ANOVA followed by Tukey's multiple comparisons test was performed to determine the signi cance. Signi cance is *p<0.05, ** p<0.01,***p<0.001, *****p<0.0001, ns is not signi cant. EOC2 microglia stimulate increased expression of neurogenic markers in injured cortical cell co-cultures. A) Immuno uorescence of Nestin+ (green), α-internexin+ (green), NeuN+ (green) and GFAP+ (red) cells in injured neurons co-cultured with control or microglial conditioned media. DAPI (blue) was used to observe nuclei of all cultured cells. White line indicates the site of injury. Scale bar represents 500 µm. All images were acquired with a 20X Leica objective. B) Three separate elds within injured neuronal cultures were evaluated for protein expression using immuno uorescent measurement software to determine the uorescence intensity units for each protein marker. Averaged uorescent intensity data from injured cortical cell cultures were normalized and set equal to 1 to determine relative uorescent intensity units (RFU). Fold change in RFU in injured cortical cultures with microglia was determined and multiple Student T Tests were performed to determine signi cance. Error bars represent SEM. Signi cance is **p<0.01, *** p<0.001, **** p<0.0001.

Figure 7
Western blot analysis of Nestin, α-internexin, NeuN, and GFAP expression in uninjured and injured neuronal cultured exposed to EOC2 microglia or control media. A) Representative western blot images of protein from uninjured neuronal cultures in control media without microglia, uninjured neuronal cultures in microglialconditioned media, injured neuronal cultures in control media without microglia, and injured neuronal cultures in microglial-conditioned media. GAPDH was used as a total protein loading control. B) Quanti cation of relative protein expression in western blot experiments. Experiments were run in triplicate using primary cultures from three biological replicates. Error bars represent SEM. Two-way ANOVA followed by Tukey's multiple comparisons test was performed to determine the signi cance. Signi cance is *p<0.05, ** p<0.01,***p<0.001, *****p<0.0001.

Figure 8
Multiplex ELISA and RT-PCR analyses of in ammatory cytokine protein and mRNA following co-culture with uninjured or injured cortical cells. A) Relative cytokine levels in media collected from injured cortical cell and microglial co-culture as measured by multiplex ELISA assays are shown. The dashed line represents the normalized cytokine levels for IFN-γ, MCP-1, MIP-1α, TNF-α, and RANTES in media collected from co-cultures of EOC2 microglia and uninjured cortical cells. Normalized uninjured co-culture cytokine concentrations were set equal to one. The experimental data represent the average fold change in each cytokine as measured in the media collected from injured cortical cell and EOC2 microglial co-culture. Experiments were run in triplicate from three biological replicates. Error bars represent SEM. Students T Test was used to determine whether the fold change was signi cance as compared to normalized control, Signi cance is *p<0.05. B) qRT-PCR analysis of cytokine mRNA levels in EOC 2 microglia following stimulation with injured cortical cells. EOC2 mRNA was collected from microglia suspended above cortical cultures on Transwells®. Fold change in mRNA levels was normalized to Gapdh expression in EOC2 microglia following stimulation with injured cortical cells. Fold change is compared to mRNA in EOC2 microglia co-cultured with uninjured neurons. Control mRNA expression is indicated by the dashed line set at one. MIQE guidelines were followed. Mouse speci c primers were used for qRT-PCR analysis of mouse microglial cells. Experiments were run in triplicate for three biological replicates. Error bars represent SEM. Signi cance was determined using BioRad CFX Manager software, *p<0.05.

Figure 9
Effect of PI3K inhibition microglial-enhanced cortical cell viability and AKT phosphorylation in cortical cells following EOC2 microglial co-culture. A) Quanti cation of MTT viability following LY294002 treatment of uninjured and injured cortical cells alone and in microglial co-culture. For each concentration, one-way ANOVA was used to determine signi cance effect of the inhibitor. Signi cance is *p<0.05, **p<0.01, ***p<0.001, at p<0.05, ****p<0.0001, # indicates that OD values are signi cantly different from that of control, uninjured cortical cells alone, ns indicates not signi cant. Application of the MAPK inhibitor, PD98059 (10 µM and 40 µM) did not signi cantly reduce microglial enhanced mitochondrial activity in uninjured or injured co-culture experiments. Mitochondrial activity of neurons co-cultured with microglia remained signi cantly higher than that of neurons alone. B) Representative western blots illustrating pAKT phosphorylation in injured and injured cortical cultures with and without microglial co-culture. Culture conditions treated with 0 µm and 40 µm are shown. C) Quanti cation AKT phosphorylation as compared to total AKT protein levels normalized to GAPDH. Three separate western blot experiments were analyzed, data were averaged and error bars represent SEM. Oneway ANOVA was used to compare the signi cance the data for 0 µm and 40 µm LY294002 treatments.

Figure 10
Inhibition of PI3K blocks microglial-enhanced expression of neurogenic markers in injured cortical co-cultures.
A) Immuno uorescence of Nestin (green), α-internexin (green), or GFAP (red) and DAPI (blue to indicate nuclei) in injured neurons co-cultured with microglia in 0 µM or 40 µM LY294002. Application of 40 µM LY294002 signi cantly reduced Nestin, α-internexin, and GFAP immuno uorescence. DAPI (blue) was used to observe nuclei of all cultured cells. All images were acquired with a 40X Leica objective. Scale bar represents 50 µm. B) Quanti cation of immuno uorescence for neurogenic markers. Three separate elds within uninjured and injured cortical cultures that were treated with 0 µM or 40 µM LY294002 and stained for each neurogenic marker were evaluated using immuno uorescence measurement software. Fluorescence for each marker is shown as a relative uorescence intensity unit (RFU). Data represent the average RFU's for the three elds. Error bars represent SEM. Student's t-test were used to determine the signi cance of LY294002 treatment in each