In vitro neuronal and glial differentiation from embryonic or adult neural precursor cells are differently affected by chronic or acute activation of microglia


  • Emanuele Cacci,

    1. Department of Cell and Developmental Biology, “La Sapienza” University, Piazzale Aldo Moro 5, 00185, Rome, Italy
    Search for more papers by this author
    • E.C. and M.A.A-C contributed equally.

  • Maria Antonietta Ajmone-Cat,

    1. Department of Cell Biology and Neuroscience, Istituto Superiore di Sanità, Viale Regina Elena 299, 00161, Rome, Italy
    Search for more papers by this author
    • E.C. and M.A.A-C contributed equally.

  • Tonino Anelli,

    1. Department of Cell and Developmental Biology, “La Sapienza” University, Piazzale Aldo Moro 5, 00185, Rome, Italy
    Search for more papers by this author
  • Stefano Biagioni,

    1. Department of Cell and Developmental Biology, “La Sapienza” University, Piazzale Aldo Moro 5, 00185, Rome, Italy
    Search for more papers by this author
  • Luisa Minghetti

    Corresponding author
    1. Department of Cell Biology and Neuroscience, Istituto Superiore di Sanità, Viale Regina Elena 299, 00161, Rome, Italy
    • Department of Cell Biology and Neurosciences, Section of Degenerative and Inflammatory Neurological Diseases, Istituto Superiore di Sanità, Viale Regina Elena, 299, 00161 Rome, Italy
    Search for more papers by this author


The contribution of microglia to the modulation of neurogenesis under pathological conditions is unclear. Both pro- and anti-neurogenic effects have been reported, likely reflecting the complexity of microglial activation process. We previously demonstrated that prolonged (72 hr) in vitro exposure to lipopolysaccharide (LPS) endows microglia with a potentially neuroprotective phenotype, here referred as to “chronic”. In the present study we further characterized the chronic phenotype and investigated whether it might differently regulate the properties of embryonic and adult neural precursor cells (NPC) with respect to the “acute” phenotype acquired following a single (24 hr) LPS stimulation. We show that the LPS-dependent induction of the proinflammatory cytokines interleukin (IL)-1α, IL-1β, IL-6, and tumor necrosis factor (TNF)-α was strongly reduced after chronic stimulation of microglia, as compared with acute stimulation. Conversely, the synthesis of the anti-inflammatory cytokine IL-10 and the immunomodulatory prostaglandin E2 (PGE2) was still elevated or further increased, after chronic LPS exposure, as revealed by real time PCR and ELISA techniques. Acutely activated microglia, or their conditioned medium, reduced NPC survival, prevented neuronal differentiation and strongly increased glial differentiation, likely through the release of proinflammatory cytokines, whereas chronically activated microglia were permissive to neuronal differentiation and cell survival, and still supported glial differentiation. Our data suggest that, in a chronically altered environment, persistently activated microglia can display protective functions that favor rather than hinder brain repair processes. © 2008 Wiley-Liss, Inc.


Several studies have clearly demonstrated the persistence of neural stem/progenitor cells in discrete regions of the adult mammalian brain, that is, the hippocampal subgranular zone and the subventricular zone (SVZ) of the lateral ventricles (Doetsch et al., 1999; Gage et al., 1998). Neurons originate from neural stem cells (NSC) throughout the entire lifespan, and might continuously replace dying neurons or contribute to specific brain functions and plasticity (Merkle et al., 2007; Toni et al., 2007; van Praag et al., 2002). Multiple aspects of adult neurogenesis, such as stem cell self-renewal, differentiation, migration, survival, seem to be coordinated by complex instructive signals provided by several components of the so-called neurogenic “niche” where NSC reside (Alvarez-Buylla and Lim, 2004). In particular, astrocytes, vasculature, and basal lamina have been shown to provide both structural support and dynamic molecular cues in response to environmental changes (Campos, 2005; Shen et al., 2004; Song et al., 2002). Microglia, the macrophage population of the brain parenchyma, have also been recently recognized as a component of the neurogenic niche, positively contributing to hippocampal and SVZ neurogenesis (Walton et al., 2006; Ziv et al., 2006). The role played by microglia in the regulation of neurogenesis under specific pathological conditions is however a matter of hot debate.

The relevance of regenerative processes triggered by brain insults such as traumatic injury, cerebral ischemia or epilepsy, is limited by the poor survival of newly formed neurons, which is mainly attributed to the detrimental effect of the inflammatory events sustained by microglia and infiltrating immune cells, determining an unfavorable environment for mature and newborn neurons (Ekdahl et al., 2003; Liu et al., 2007; Monje et al., 2003). This hypothesis is supported by a number of in vitro studies, which demonstrate that the survival of hippocampal progenitors or NSC, is reduced when co-cultured with microglial cells activated by the bacterial endotoxin (lipopolysaccharide, LPS), or when exposed to their conditioned medium (CM) (Cacci et al., 2005; Liu et al., 2005; Monje et al., 2003). However, microglial cells activated by other agents, such as specific T-helper associated cytokines, or by a milieu in which massive apoptotic neuronal death occurs as a consequence of adrenalectomy, have been shown to favor neurogenesis (Battista et al., 2006; Butovsky et al., 2006). Indeed, microglia activation is emerging as a dynamic and complex process in which unique functional patterns can be recognized, depending on the nature, intensity and persistence of the activating agent and changes in the microenvironment (Perry et al., 2007; Schwartz et al., 2006).

In a previous study, we demonstrated that prolonged microglial exposure to LPS induces a sustained synthesis of immunomodulatory products whereas the release of proinflammatory molecules is progressively inhibited. This profound rearrangement of microglial functions reflects intrinsic changes in the activation process and, more specifically, in the signaling transduction machinery activated in response to chronic LPS exposure. Indeed, following chronic LPS, the ability to activate key signaling molecules in the process of microglial activation, such as the nuclear transcription factor NF-kB and the MAP kinase p38, is progressively reduced, whereas the ability to phosphorylate the transcription factor CREB (cAMP responsive element binding protein) is retained, indicating that chronic LPS-stimulation induces in microglia a molecular and functional state clearly distinct from that elicited by acute stimulation (Ajmone-Cat et al., 2003). A similar anti-inflammatory phenotype has been described for microglia phagocytosing apoptotic but not necrotic neurons and in few animal models of chronic neurodegenerative diseases (Minghetti et al., 2005; Perry et al., 2007). Interestingly, in experimental models of Prion and Parkinson's Disease, where microglia are exposed to the slow accumulation of pathological peptides and to apoptotic neurons, the proinflammatory cytokines expression profile [e.g. tumor necrosis factor (TNF)-α] is limited, whereas the synthesis of anti-inflammatory mediators such as IL-10, PGE2 and TGF-β1 is elevated (Depino et al., 2003; Minghetti et al., 2005; Perry et al., 2002). In the present study we further characterized the functional properties acquired by microglia upon prolonged and repeated exposure to LPS, mimicking a setting of chronic activation, and evaluated the effects of “chronically” activated microglia on neural precursor cell (NPC) properties. Such effects were compared with those exerted by microglia acutely activated by a single acute exposure to LPS.


Microglial Cultures

Microglial cultures were obtained from 9 to 12 days mixed primary glial cultures prepared from the cerebral cortex of 1-day-old rats, as previously described (Levi et al., 1993) in accordance with the European Communities Council Directive No. 86/609/EEC. Microglial cells, harvested from the mixed primary glial cultures by mild shaking, were resuspended in complete medium (BME-FBS) consisting in Basal Eagle's Medium (BME), supplemented with 10% heat-inactivated fetal bovine serum (FBS), 2 mM glutamine and 100 μg/ml gentamicin, and plated on uncoated plastic wells, or cell culture inserts (0.4 μm pore size, PET, Falcon), at a density of 1.25 × 105 cells/cm2. Cells were allowed to adhere for 20 min and then washed to remove nonadhering cells. All cell culture reagents were from Invitrogen (Grand Island, NY) and virtually endotoxin free (less than 0.3 E.U./ml as determined by the manufacturer).

Isolation and Expansion of Embryonic and Adult NPC

All the animal-related procedures were conducted in accordance with European Communities Council Directive No. 86/609/EEC. Embryonic cortical tissue from E13.5 mouse embryos (C57BL/10) was incubated in Dulbecco's modified Eagle's Medium (DMEM) for 20 min at 37°C and then mechanically dissociated by trituration to single cell suspension. The cells were transferred to a medium referred as to “embryonic expansion medium”. This medium consisted of basal medium (DMEM/F12, 1% penicillin/streptomycin, 0.1 M L-glutamine (Gibco), 23.8 mg/100 ml Hepes, 7.5% NaHCO3, 0.6% glucose), supplemented with 20 ng/ml of human recombinant epidermal growth factor (EGF; R&D), 10 ng/ml of basic fibroblast growth factor (bFGF; R&D), and N2 supplement (1%; Gibco). Living cells were counted by trypan blue exclusion method and plated at 5 × 104 cells/ml density in uncoated T25 plastic flasks (Nunc) in the embryonic expansion medium defined as above. After several days, spontaneously formed floating clusters were harvested, enzymatically dissociated (Accutase; Sigma) and single cells allowed attaching in flasks coated with 10 μg/ml poly-ornithine and 5 μg/ml laminin. Embryonic NPC (eNPC) were routinely dissociated twice a week by using accutase and plated at 0.4 or 0.6 × 104 cells/cm2 densities. In this study cells were used between passages 8 and 15.

Adult NPC (aNPC) were obtained from the SVZ of 2 months-old C57BL/10 male mice accordingly to a protocol previously described (Johansson et al., 1999). Briefly, coronal slices containing SVZ were collected in Leibovitz's L-15 medium and the lateral wall of the lateral ventricles, identified under light microscopy, was carefully removed from the surrounding brain tissue. SVZ explants were incubated in HBSS (Gibco) containing 5.4 mg/ml D-glucose, 18 mM Hepes, 0.74 mg/ml trypsin, 200 U/ml DNase, 0.7 mg/ml hyaluronidase, 2 mg/ml Kynurenic acid (all from Sigma) pH 7.5 for 20 min at 37°C, and mechanically dissociated by trituration. Living cells were counted by using trypan blue exclusion method and plated in uncoated plastic flasks at 5 × 104 cells/ml density. aNPC were routinely cultured on poly-ornithine and laminin coated flasks in a medium similar in composition to the embryonic expansion medium but containing 2% B27 instead of N2 supplement, and referred as to “adult expansion medium”. In this study cells were used between passages 8 and 15.

Both eNPC and aNPC were maintained under proliferative condition at 37°C in a 5% CO2 atmosphere in NPC expansion medium.

Acute and Chronic Stimulation of Microglial Cultures

Microgliae cells were cultured for 72 hr. All cultures were maintained in complete medium (BME-FBS) during the first 48 hr, then washed twice with serum free medium to remove any traces of FBS and further incubated for 24 hr in NPC basal medium (see below) with or without LPS (10 ng/ml). Unstimulated cultures received fresh medium without LPS every 24 hr, over a 72 hr period. Acutely stimulated cultures received fresh medium without LPS for the first 48 hr, and LPS-containing medium for the last 24 hr incubation. Chronically stimulated cultures received fresh medium containing LPS every 24 hr.

Culture media conditioned during the last 24 hr by unstimulated (U-CM), acutely (A-CM) or chronically stimulated cultures (C-CM), were collected, centrifuged, filtered and stored at −80°C until tested.

LPS (from Escherichia Coli, serotype 026:B6) was obtained from Sigma (St. Louis, MO).

Cytokines, PGE2, and NO Determination

The levels of IL-1α, IL-1β, IL-6, IL-10 and transforming growth factor (TGF)-β1 were assayed by specific ELISAs (from Endogen, Woburn, MA, for IL-1α, IL-1β, IL6, IL-10, and from DRG Diagnostic, Germany, for TGF-β1) following the manufacturer's instructions. The ranges of determination were: 51.2–2000 pg/ml for IL-1α, 15–1000 pg/ml for IL-1β, 31–2000 pg/ml for IL-6, 8–500 pg/ml for IL-10 and 10–600 pg/ml for TGF-β1.

Culture media PGE2 was measured by high sensitivity enzyme immunoassay (Assay Design, Ann Arbor, MI; detection limit: 7.8 pg/ml). All measurements were run at least in duplicate for each sample.

The production of nitric oxide (NO) was determined by measuring the content of nitrite, one of the end products of NO oxidation, as previously described (Minghetti et al., 1996). A standard nitrite curve (0.25–50 μM) was generated using a 10 mM solution of NaNO2. All chemical for the NO assay were from Sigma.

RNA Extraction and Real Time PCR

Total RNA was extracted from microglial cells, grown in the appropriate culture media for 7 or 24 hr after the last medium replacement, by Trizol reagent (Invitrogen), following the instructions provided by the manufacturer, and quantified by spectrophotometric analysis.

Two micrograms of RNA were reverse transcribed with M-MLV Reverse Transcriptase using random hexamers (Promega, Madison, WI) in a final volume of 25 μL.

Real-time PCR was performed on the reverse transcription (RT) products with the SYBR Green JumpStart Taq ReadyMix (Sigma) in a Lightcycler apparatus (Biorad), following the manufacturer's instructions. Annealing temperatures and primer sequences for all the cytokines, excepted for TGF-β1, were obtained from Peinnequin et al. (2004) and Gayle et al. (2004) (Gayle et al., 2004; Peinnequin et al., 2004). Primers for TGF-β1 were designed by using Beacon Designer software. The NCBI ( accession number for TGF-β1 is NM_021578. TGF-β1 for: AGATTCAAGTCAACTGTGGAG TGF-β1 rev: AAGCCCTGTATTCCGTCTC. Amplification efficiency for TGF-β1 and the housekeeping gene β-actin was higher than 1.9 as demonstrated by using different template dilutions. Thermal cycling conditions for TGF-β1 comprised an initial step at 95°C for 3 min, followed by 45 cycles at 95°C for 30 sec, 60°C annealing temperature for 30 sec and extension phase at 72°C for 30 sec. All samples were run in triplicate, and each well of PCR contained 25 μL as a final volume, including 2.5 μL of cDNA corresponding to 50 ng of total RNA, 0.2 μM forward primer, 0.2 μM reverse primer, and 12.5 μL SYBR Green JumpStart Taq ReadyMix and 0.2 μM internal reference dye.

The mRNA levels for each cytokine were compared between unstimulated, and acutely or chronically LPS-stimulated microglia using the ΔΔCT method. β-actin was used as internal control gene. Amplification specificity was checked using a melting curve, following the manufacturer's instructions.

NPC Differentiation

eNPC were plated on poly-ornithine/laminin-coated coverslips at 1.2 × 104 cells/cm2 or 0.75 × 104 cells/cm2 density, in expansion medium (defined as earlier). After 24 hr, the expansion medium was replaced by microglial CMs (U-CM, A-CM, or C-CM) supplemented with 10 ng/ml bFGF, 0.5% B27 and 1% N2S. In parallel cultures the expansion medium was replaced by the so called “differentiation medium” consisting of basal medium, supplemented with bFGF, B27 and N2S. All cultures were differentiated under these conditions for 3 or 6 days. In the latter case half of medium was replaced with fresh medium after 3 days. aNPC were differentiated according to the same protocol.

NPC/Microglia Co-Cultures

eNPC were co-cultured, without cell-to-cell contact, with microglial cells seeded into membrane inserts (0.4 μm pore size) that allow solute exchange. Before addition to eNPC, microglia were cultured for the first 48 hr on membrane inserts at a density of 1.2 × 105 cells/cm2 with culture medium (BME-FBS) containing or not LPS, and renewed every 24 hr. The inserts were then washed twice to remove serum and co-cultured with eNPC seeded on poly-ornithine/laminin-coated coverslips at 1.2 × 104 cells/cm2 density in NPC differentiation medium with or without LPS. After 24 hr of co-cultivation, the inserts were removed and eNPC allowed to differentiate for an additional 48 hr.

Cell Survival, Proliferation, and Total Cell Number

For cell survival, cell proliferation, and total cell number estimations, eNPC were plated at 0.75 × 104/cm2 on poly-ornithine/laminin-coated coverslips in NPC expansion medium. After 24 hr cells were exposed to U-CM, A-CM, or C-CM, supplemented with 10 ng/ml bFGF, 2% B27, and 1% N2S. Parallel eNPC cultures were exposed to differentiation medium or expansion medium.

Cell survival was evaluated after 24 and 48 hr of treatment by an in situ detection kit based on the incorporation of labeled nucleotides to DNA breaks by terminal deoxynucleotidyl transferase mediated fluorescein-dUTP nick-end labeling (TUNEL; Roche). The assay was performed according to manufacturer's instructions. After fixation with 4% paraformaldehyde (PFA) for 1 hr at RT, cells were rinsed three times with phosphate buffered saline (PBS) and then incubated in permeabilization solution (0.1% Triton X-100 in 0.1% sodium citrate) for 2 min on ice. TUNEL reaction mix, containing fluorescein-labeled dUTP (TUNEL Label), unlabeled dNTP nucleotides, and terminal deoxynucleotidyl transferase (TUNEL Enzyme), was added to each culture for 60 min at RT in humidified chamber. Cultures were then rinsed, and coverslips mounted on glass slides with glycerol based mounting medium. The total cell population was identified by incubating the cells for 15 min with 10 μg/ml of the nuclear dye Hoechst 33342 (Molecular Probes).

Cell proliferation was evaluated after 48 hr of treatment by quantifying the percentage of dividing cells positive for the proliferation antigen Ki67, with respect to total cell number (Hoechst 33342 labeled cells), taken as 100%. Number of cells in each culture condition was quantified by evaluating the average of Hoechst positive cells counted in 10 different and randomly chosen microscopic fields in each of three coverslips per condition, in at least three independent experiments. An average of 50 cells per field was counted.


Cultures were fixed with 4% PFA in PBS for 15 min. After fixation, cultures were washed with potassium-PBS (KPBS), pH 7.4, and then incubated in KPBS containing 5% of the appropriate normal serum and 0.025% Triton X-100 (preincubation solution) for 1 hr at RT. Subsequently, cultures were incubated at 4°C o.n. in preincubation solution containing primary antibodies listed below. Cells were washed thrice with KPBS and incubated for 2 hr at RT in the appropriate incubating solution containing 10 μg/ml Hoechst 33342 and the appropriate biotinylated (Vector) and/or Cy3-conjugated (Jackson Immunoresearch) secondary antibodies at dilution 1:200. After rinsing with KPBS, cells were incubated for 2 hr at RT with Alexa 488-conjugated streptavidin (Molecular Probes) at dilution 1:200, rinsed, and coverslipped with DAKO mounting medium (Dako). The following primary antibodies were used: monoclonal mouse anti nestin (Chemicon; 1:200); monoclonal mouse anti-βIII tubulin (Promega; 1:500); polyclonal rabbit anti-Tau (Sigma; 1:100); polyclonal rabbit anti-glial fibrillary acid protein (GFAP; Dako; 1:500); monoclonal mouse anti-Ki67 (Novocastra; 1:100); monoclonal mouse anti-CD68 (Serotec, 1:200), polyclonal rabbit anti-Iba1 (1:1000, Wako).

Immunostaining specificity was assessed by omission of primary antibodies during overnight incubations. The percentages of differentiated cells in each condition were evaluated by counting immunolabeled cells in 10 independent microscopy fields of three coverslips for each condition, in at least three independent experiments. An average of 100 cells per field was counted. An observer blind to treatment conditions conducted all analyses.

Statistical Analysis

Comparisons were performed using one way analysis of variance (ANOVA) followed by post hoc Bonferroni, or using Student's t-test where specified. Data are presented as mean ± SEM, with the number of independent experiments (run in duplicate or triplicate) indicated in parenthesis. Differences are considered significant at P < 0.05.


Microglial Cultures Purity and Morphology

We exposed microglial cultures to acute (single) or chronic (repeated) LPS challenges given in fresh LPS-containing medium every 24 hr over a 72 hr period. Upon acute LPS stimulation cell morphology was rounded and several cells showed vacuolated cytoplasm. Chronic LPS exposure did not substantially modify the morphology of the cells compared with the acute condition (Figs. 1A–C). The purity of the cultures was assessed by indirect immunofluorescence for the macrophage/microglia marker Ionized calcium binding adaptor molecule 1 (Iba1; Figs. 1D–F). An average of 96% of the cells were positive for the expression of Iba1 in all the culture conditions examined in this study. Similar results were obtained with another macrophage/microglia marker, the lysosomal membrane antigen CD68 (not shown).

Figure 1.

Microglial cultures and cytokine mRNA expression, following acute or chronic LPS stimulation. (AC) Phase contrast photomicrographs showing primary microglia cultured in different conditions: (A) unstimulated, (B) acutely LPS-stimulated, or (C) chronically LPS-stimulated. (DF) Iba1 staining (green) and Hoechst counterstaining (blue) in the corresponding microscopic fields represented in A–C. Scale bar in F is 50 μm. (G) Relative mRNA levels in acutely (A_m) or chronically (C_m) stimulated microglia at 7 hr after last medium replacement were assessed by real time PCR using the 2−ΔΔCT method. Data are expressed as the fold change in gene expression normalized to an endogenous gene (β actin) and relative to unstimulated microglia. Data are mean values ± S.E.M. (n = 3–4 in triplicate). Comparisons between A_m and C_m were made by two-tailed unpaired Student's t-test. *P < 0.05; **P < 0.005; ***P < 0.0005. [Color figure can be viewed in the online issue, which is available at]

Real Time PCR Analysis of Cytokine Transcripts in Microglial cells, Following Acute or Chronic LPS Stimulation

We evaluated the effects of acute or chronic LPS stimulation on microglial pro- and anti-inflammatory cytokine mRNA expression by means of real time PCR. Total RNAs were extracted after 7 or 24 hr following the third change of medium from unstimulated, acutely or chronically stimulated microglia, as detailed in Materials and Methods section. The mRNA levels of the proinflammatory cytokines IL-1α, IL-1β, IL-6, TNF-α, and the immunoregulatory cytokines IL-10 and TGF-β1 were determined and expressed as fold change in gene expression with respect to control samples from unstimulated microglia. The 2−ΔΔCt method was used to calculate the relative change of gene expression of target cytokine, normalized to β-actin RNA levels.

Unstimulated cultures constitutively expressed detectable mRNA levels of IL-1α, IL-1β, IL-6, TNF-α, IL-10, and TGF-β1. As shown in Fig. 1G, the relative levels of transcripts for IL-1α, IL-1β, IL-6, and TNF-α, measured in cells collected at 7 hr after the last medium replacement, were significantly higher in acutely stimulated cultures (A_m) than in chronically stimulated ones (C_m). Interestingly, induction of TNF-α mRNA expression was almost totally abrogated after chronic stimulation, being comparable to unstimulated cultures. On the contrary, IL-10 mRNA expression was upregulated by single LPS stimulation and further elevated in response to a chronic stimulation. TGF-β1 mRNA levels were not modified either by acute or chronic LPS exposure. Similar patterns of pro- and anti-inflammatory cytokine expression were observed in microglial cells harvested 24 hr after the last medium replacement, although relative mRNA levels were consistently lower with respect to those of cells collected 7 hr after medium replacement (data not shown).

These observations suggest that microglial cells respond to repeated LPS-challenges with a selective re-orientation of gene expression.

Analysis of Cytokine Levels in Culture Media Conditioned by Microglial Cells, Following Acute or Chronic LPS Stimulation

The different regulation of the cytokine mRNA levels observed after single or repeated LPS stimulation was mirrored by accumulation of the corresponding proteins in the culture media. As shown in Figs. 2A–C, the levels of IL-1α, IL-1β and IL-6 were strongly elevated in the conditioned media (CMs) of acutely stimulated microglia (A-CM) compared with media from unstimulated (U-CM) or chronically stimulated cultures (C-CM). In contrast with these proinflammatory cytokines, IL-10 levels remained elevated in C-CM (Fig. 2D). Despite the significant induction of IL-10 mRNA in chronically stimulated cultures, only a modest increase of IL-10 protein was found in C-CM compared with A-CM. This is probably due to posttranscriptional cytokine expression regulation (Powell et al., 2000). TGF-β1 levels were below the assay detection limit in all CMs.

Figure 2.

Levels of cytokines, NO and PGE2 in microglial CMs after acute or chronic LPS stimulation. CMs from unstimulated (U-CM), acutely (A-CM) or chronically (C-CM) stimulated microglia were collected at 24 hr from last medium replacement, and analyzed for IL-1α (A), IL-1β (B), IL-6 (C), IL-10 (D), PGE2 (F) content by means of specific ELISAs, and for nitrite levels (E) by means of Griess reaction. Data are mean values ± S.E.M. (n = 4–7 in duplicate). *P < 0.05; **P < 0.005; ***P < 0.0005.

In agreement with our previous findings (Ajmone-Cat et al., 2003), nitrite levels (as an index of NO production) were reduced whereas those of PGE2, an important immunomodulatory lipid mediator produced upon microglia activation (Minghetti, 2007), were enhanced in response to chronic LPS stimulation (Figs. 2E,F).

eNPC Properties

eNPC, obtained from E13.5 mouse cortical fetal tissue, were expanded as adherent cultures in embryonic expansion medium (containing bFGF/EGF; Fig. 3A). Under these conditions, virtually all eNPC were immunopositive for the cytoskeletric protein nestin, a marker for multipotential precursors (Fig. 3B) whereas they did not express the neuronal marker βIII tubulin or the astrocytic marker GFAP. This suggests that the majority of the cells were in an undifferentiated state. Moreover (41.8 ± 0.5)% of eNPC showed immunoreactivity for the proliferation marker Ki67 (see below and Figs. 5A, B, and K), and a doubling time of about 24 hr was estimated. When eNPC were cultured in the presence of differentiation medium (EGF−/bFGF+) for three days, βIII tubulin positive neurons began to appear [(13.1 ± 1.9)% in the total cell number] and their percentage increased up to (19.1 ± 0.7)% after 6 days (Fig. 3C). The neuronal identity was also confirmed by their expression of the microtubule-associated proteins Tau (data not shown). Parallel cultures, maintained for additional 6 days in basal medium containing only B27/N2 supplement, displayed a similar percentage of βIII tubulin+ neurons [(23.6 ± 3.2)%], although they were characterized by a more elaborated neuronal morphology with respect to cells differentiated for only six days (Figs. 3E,F). These data suggest that early differentiation of eNPC to neuronal precursors occurs after EGF removal.

Figure 3.

Expression of nestin, βIII tubulin or GFAP in eNPC. (A) Phase contrast photomicrograph showing eNPC propagated in expansion medium on poly-ornityne/laminin coated flasks. (B) eNPC propagated in expansion medium (bFGF+/EGF+) were virtually all immunopositives for the immature neural marker nestin. Scale bar = 30 μm. (C) Percentage of neuronal and (D) of astrocytic cells generated from eNPC cultured in NPC differentiation medium (bFGF+/EGF−) for 3 and 6 days. The percentages of βIII tubulin+ and GFAP+ cells significantly increased with time in culture. (n = 4 in triplicate). *P < 0.05. (E; F) βIII tubulin staining and (G; H) GFAP staining of eNPC cultures, after 6 days in differentiation medium (E; G), or after bFGF deprivation and additional six days in medium supplemented with B27/N2 only (F; H). Scale bar in F is 15 μm; in H is 30 μm. [Color figure can be viewed in the online issue, which is available at]

Poor glial differentiation was observed in eNPC cultures maintained for 3 or 6 days in the presence of differentiation medium (Fig. 3G). However, upon bFGF removal and additional 6 days of culture in medium supplemented with B27/N2, more than 50% of cells were immunostained for GFAP (Fig. 3H). Moreover, almost 100% of eNPC cultured for 6 days in medium containing 1% FBS differentiated to cells with typical astrocytic morphology and intense GFAP staining, according to previous observations (Conti et al., 2005).

Effects of Microglia CMs on Survival of eNPC

We investigated whether microglia could affect eNPC survival by using the TUNEL assay, together with Hoechst, to visualize fragmented nuclei for the identification of cell death at single cellular level. eNPC were cultured in expansion medium and after 24 hr the medium was replaced with U-CM, A-CM, C-CM, or differentiation medium. All media were supplemented, immediately before starting treatments, with bFGF and B27/N2 supplement. Cell survival was evaluated at 24 and 48 hr after treatment with the CMs or differentiation medium. As shown in Figs. 4A–H and I, A–CM, after 48 hr in cultures, increased the percentage of dead cells [(5.8 ± 0.2)%] compared with U-CM and C-CM [(2.3 ± 0.6)% and (3.2 ± 0.6)%, respectively]. No differences were found among the latter two conditions and NPC cultured without microglial CMs [(2.7 ± 0.8)%]. Similar results were obtained on NPC cultures analyzed after 24 hr.

Figure 4.

Analysis of apoptosis in eNPC cultures exposed to microglial CMs. (A–H) Representative photomicrographs of eNPC stained for the nuclear dye Hoechst (blue; A–G) and by TUNEL (green; B–H). The cells were maintained in the presence of basal medium (A; B), U-CM (C; D), A-CM (E; F), or C-CM (G; H), all supplemented with bFGF, B27, and N2S, for 48 hr. Note the higher number of TUNEL-positive cells in A-CM treated cultures (F) compared with U-CM (D) and C-CM (H) treated cultures and conversely, the lower number of Hoechst+ cells in A-CM treated cultures (E), if compared with U-CM (C) and C-CM (G). Scale bar is 50 μm. (I) The percentages of TUNEL positive cells in the total cell number, under the different culture conditions, are reported in the graph. (n = 3 in duplicate). *P < 0.005; **P < 0.0005. [Color figure can be viewed in the online issue, which is available at]

Effects of Microglia CMs on Proliferation of eNPC

To determine whether the percentage of proliferating NPC was altered following CM treatment, the expression of the proliferative marker Ki67 was investigated under different culture conditions. When eNPC were cultured for 48 hr in expansion medium, about 41% of the cells were immunopositive for Ki67 (Figs. 5A,B, and K). Withdrawal of EGF led to a drastic decrease of total cell number, possibly as a consequence of diminished cell proliferation as demonstrated by the percentage of Ki67 positive cells, which dropped to less than 10% after 48 hr (Figs. 5C, D, and K). These data are consistent with previous findings demonstrating that bFGF alone is insufficient to maintain established NS cell as a passagable cell line and to prevent their differentiation (Pollard et al., 2006). When eNPC were cultured in microglial CMs supplemented with bFGF, cell proliferation was strongly augmented, as indicated by the percentage of dividing Ki67 positive cells compared with cultures exposed to bFGF only-containing medium (differentiation medium) (Figs. 5E–L and K). The highest percentage of dividing cells was found in eNPC exposed to U-CM. No differences were found between A-CM and C-CM. In addition, the total cell number was higher in cultures exposed to CMs than in cultures exposed to differentiation medium (Fig. 5L). All together these data indicate that mitogenic factors released from microglia can increase eNPC proliferation.

Figure 5.

eNPC proliferation and total cell number in the presence of microglial CMs. eNPC were maintained in the different culture conditions for 48 hr and stained with the nuclear dye Hoechst (blue; A–I) and for the proliferative antigen Ki67 (red; B–L). (A–L) Photomicrographs of eNPC cultured for 48 hr in expansion medium (bFGF+/EGF+; A; B) or in bFGF+/EGF− medium (C; D) or in U-CM (E; F), A-CM (G; H), or C-CM (I; J). All CMs were supplemented with bFGF, N2S, and B27. (K) Percentages of proliferating Ki67+ cells in the total cell number and (L) total cell number per microscopic field, after 48 hr in each culture condition. All microglial CMs increased proliferation and total cell number if compared with bFGF+/EGF− medium. Data are mean values ± S.E.M. (n = 3–4 in triplicate). *P < 0.05; ***P < 0.0005. [Color figure can be viewed in the online issue, which is available at]

Effects of Microglial CMs on Glial and Neuronal Differentiation of eNPC

The effects of microglial CMs on eNPC differentiation were firstly evaluated by immunocytochemical detection of nestin expression. As shown in Figs. 6A,B, eNPC strongly down-regulated the expression of nestin when cultured for 3 days with CMs supplemented with bFGF and B27/N2, indicating a decrease in the number of immature cells. The effect was comparable with all the CMs (not shown). eNPC were then cultured for three or six days with microglial CMs or differentiation medium, and immunostained for the glial and neuronal markers GFAP and βIII tubulin, respectively. All CMs induced a dramatic increase in the percentage of GFAP positive cells compared with eNPC cultured in differentiation medium, at both times of culture (Figs. 6C,D and G-J). Interestingly, A-CM induced the highest percentage of GFAP positive cells, either after 3 or 6 days in culture. Since high cell density and extensive branching could confuse cell counting, quantification of GFAP labeled cells was also performed on eNPC plated at low density and exposed to the microglial CMs for only 24 hr. In these conditions, the percentage of GFAP positive cells was again higher with A-CM [(33.3 ± 1.5)%] than with U-CM [(20.7 ± 3.2)%] or C-CM [(10.9 ± 2.1)%].

Figure 6.

eNPC differentiation in the presence of microglial CMs. (A, B) Photomicrographs showing nestin down-regulation in NPC cultures exposed to U-CM for 3 days (B), with respect to cultures in expansion medium (bFGF+/EGF+; A). Similar effects were seen with A-CM and C-CM. Scale bar in B is 30 μm. (C; D) Percentages of immunolabeled GFAP+ cells or (E; F) of βIII tubulin+ cells in the total cell number after 3 (C; E) or 6 days (D; F), in the absence (–) or in the presence of CMs from unstimulated (U-CM), acutely (A-CM), or chronically (C-CM) stimulated microglia. Data are mean values ± S.E.M. (n = 4 in triplicate). *P < 0.05; **P < 0.005; ***P < 0.0005. (G-L) Photomicrographs showing GFAP staining in cultures differentiated for 6 days without CM (G) or with U-CM (H), A-CM (I) or C-CM (J). Note the higher number of GFAP+ cells in CMs treated cultures (H–J) than in cultures without CM (G). (K–N) Photomicrographs showing βIII tubulin staining in cultures differentiated without CM (K) or with U-CM (L), A-CM (M) or C-CM (N). The number of cells and the percentage of neurons were significantly lower in A-CM treated cultures. Scale bar in N is 15 μm. [Color figure can be viewed in the online issue, which is available at]

To examine whether CMs from microglia had effects on neuronal differentiation, we quantified the presence of neuronal cells, on the basis of βIII tubulin labeling and morphological criteria. Interestingly, eNPC cultured with U-CM or C-CM displayed a higher percentage of newly formed neurons than those cultured with A-CM, either after 3 or 6 days in culture (Figs. 6E,F and K–N). Moreover, a conspicuous number of these cells showed more elaborated processes than neuronal cells generated in the presence of A-CM. No morphological differences were found among newly formed neurons in U-CM and C-CM treated cultures. However, U-CM and C-CM did not increase the percentage of cells that differentiated to neurons, as eNPC exposed to these CMs for 3 or 6 days generated similar or even lower percentages of neurons than cultures differentiated in control conditions (in the presence of bFGF and B27/N2S only; Figs. 6E,F).

Effects of Microglial CMs on Glial and Neuronal Differentiation of aNPC

To investigate whether soluble mediators produced by unstimulated, acutely or chronically activated microglia could affect also aNPC properties we exposed NPC derived from adult mouse SVZ to the different CMs. As described earlier for the eNPC, treatment of aNPC with CMs for 6 days induced a strong increase in the number and intensity of GFAP positive cells (Fig. 7A). Neuronal generation was impaired by A-CM, as shown in Figs. 7B and C–F. The percentage of βIII tubulin+ cells in total cell number dropped to (3.2 ± 0.6)% with A-CM respect to (13.2 ± 3.4)% with U-CM and (13.5 ± 1.9)% with C-CM. These data demonstrated that microglia, activated by single or repeated exposure to LPS, affected the examined properties of both embryonic and adult NPC in a similar way.

Figure 7.

aNPC differentiation in the presence of microglial CMs. (A) Percentages of immunolabeled GFAP+ cells, or (B) of βIII tubulin+ cells in the total cell number after 6 days in the absence (–) or in the presence of CMs from unstimulated (U-CM), acutely (A-CM), or chronically (C-CM) stimulated microglia. Data are mean values ± S.E.M. (n = 3 in triplicate). *P < 0.05; **P < 0.005; ***P < 0.0005. (C-F): Photomicrographs showing β III tubulin staining of eNPC differentiated without CM (C), or with U-CM (D), A-CM (E), or C-CM (F) for six days. Note the poor neuronal differentiation and the lower number of Hoechst+ cells in A-CM treated aNPC (E), if compared with U-CM (D) or C-CM (F) treated aNPC. Scale bar = 30 μm. [Color figure can be viewed in the online issue, which is available at]

Microglia and eNPC Co-Cultures: Effects on Glial and Neuronal Differentiation

To take into account the possible contribution of labile, short-lived factors produced by microglia to the commitment of eNPC to neuronal and glial fates, we used a co-culture system that allowed medium sharing without direct cell-to-cell contact. Microglial cells were cultured in the absence of eNPC on culture inserts for 48 hr, in the appropriate fresh medium (with or without LPS), and then transferred to multiwell plates containing eNPC. Microglia and eNPC were then co-cultured for 24 hr in fresh medium containing or not LPS. At the end of the incubation period, the inserts were removed and eNPC were allowed to differentiate for further 48 hr in the same media, fixed, and stained for GFAP and βIII tubulin. Microglial cells strongly promoted eNPC glial differentiation, as assessed by the higher percentage of GFAP positive cells in the total cell number. Similarly to conditioned medium from single LPS exposure, acutely stimulated microglia (A_m) promoted a higher glial differentiation when compared with chronically stimulated microglia (Fig. 8A). Moreover, the percentage of cells immunoreactive for βIII tubulin was significantly lower in the presence of A_m, than in the presence of chronically stimulated microglia (C_m; Fig. 8B), confirming the detrimental effect of factors released following acute activation and the permissive environment for neuronal differentiation generated by chronically stimulated microglia. LPS did not affect glial or neuronal differentiation as assessed by culturing eNPC in the presence of LPS alone: with respect to the number observed in the absence of LPS, in the presence of LPS alone the percentages of GFAP and βIII tubulin positive cells were (111.1 ± 13.5)% and (116.1 ± 12.1)%, respectively.

Figure 8.

eNPC differentiation when co-cultured with microglia. Microglial cells, seeded on culture inserts, received fresh medium (with or without LPS) every 24 hr during the first 48 hr and were then co-cultured for 24 hr with eNPC, in the presence or in the absence LPS. The inserts were then removed and eNPC were allowed to differentiate for further 48 hr in the same media. (A) Percentages of immunolabeled GFAP+ cells, or (B) of βIII tubulin+ cells in the total cell number in the absence (–) or in the presence of unstimulated (U_m), acutely (A_m), or chronically (C_m) stimulated microglial cells. The highest percentage of GFAP+ cells and, conversely, the lowest percentage of βIII tubulin+ cells were observed in the presence of A_m. Data are mean values ± S.E.M. (n = 3 in triplicate). *P < 0.05; **P < 0.005; ***P < 0.0005.


The present study explored the possibility that distinct microglial phenotypes, acquired upon acute or chronic stimulation, have different repercussions on NPC properties. Our data, consistently with findings from other groups (Butovsky et al., 2006; Cacci et al., 2005; Liu et al., 2005; Monje et al., 2003), suggest that acutely activated microglia are detrimental for cell survival and neurogenesis. However, we show that during chronic stimulation microglia turn to a clearly distinct phenotype that, similarly to unstimulated microglia, does not impair cell survival and allows the generation of new neurons from either embryonic or adult NPC. We also show that microglia have a strong progliogenic effect.

Extending our previous observations (Ajmone-Cat et al., 2003), we found that chronically stimulated microglia progressively reduce their synthesis of proinflammatory cytokines, such as TNF-α, IL-1α, IL-1β, IL-6, and free radicals, such as NO, as compared with acutely stimulated microglia. On the contrary, the synthesis of the anti-inflammatory cytokine IL-10 and of the immunomodulatory PGE2 is fully retained or enhanced, respectively, during chronic stimulation. Both IL-10 and PGE2 are known as potent suppressors of the effector functions of microglia and other immune cells, including the synthesis of proinflammatory products and co-stimulatory molecules (Levi et al., 1998; Menendez Iglesias et al., 1997; O'Garra and Vieira, 2007), and have been shown to be neuroprotective in several models of brain injury (Akaike et al., 1994; Brewer et al., 1999; McCullough et al., 2004; Zhang and Rivest, 2001). The relevance of microglial functions in the outcome of neurogenesis is now beginning to be investigated, although a clear consistent picture is still lacking. It has been shown that microglia exposed to a specific range of T-helper associated cytokines express MHCII but do not release TNF-α, and promote neurogenesis and oligodendrogenesis from neural progenitor cells (Butovsky et al., 2006). Recently, Battista et al. (2006) proposed that the pro- or anti-neurogenic niche would depend on the degree of microglia activation and the balance between the pro- and anti-inflammatory cytokines produced (Battista et al., 2006). In addition, it has been reported that the production of newborn neurons generated in the dentate gyrus following electrically induced status epilepticus, though initially compromised by the acute inflammatory response accompanying the damage (Ekdahl et al., 2003), is not further prevented by chronic inflammation after 6 months. This observation raises the possibility that chronically activated microglia may turn into a beneficial or at least nondetrimental phenotype for newborn neurons survival and differentiation (Bonde et al., 2006), a hypothesis consitent with the observation of the present study.

To study the effects of distinct microglial functional profiles on NPC fate, we have generated NPC from embryonic mouse cortex and adult SVZ, and expanded them as adherent cultures according to the protocol previously described by Conti et al. (2005). This culture procedure is particularly suitable for avoiding spontaneous cell differentiation and for the maintenance of symmetrically expandable, apparently homogeneous adherent cultures, and represents a directly accessible system to study the effects of different experimental conditions on the stem cell state (Conti et al., 2005; Pollard et al., 2006). In our study we isolated neural precursors from embryonic cortex or adult SVZ and expanded them for several passages in the presence of growth factors without subsequent clonal selection, which implicates the existence of a mixed population of neural stem and progenitor cells in our cultures. Even though the presence of committed progenitors cannot be ruled out, we found that, in the presence of both EGF and bFGF, virtually all the cells expressed the neural stem cell marker nestin and their differentiation to neurons and glia was minimal, indicating a high number of cells in undifferentiated state. In addition, it is worth mentioning that neurogenic progenitors (transit-amplifyng C cells) isolated from adult mouse SVZ and exposed to EGF in vitro have been shown to convert into cells retaining stem cell competence (Doetsch et al., 2002).

In the presence of bFGF alone, eNPC proliferation was reduced. Cells showing neuronal morphology and βIII tubulin immunoreactivity appeared at 3 days. These data are in accordance with previous observations indicating that after EGF removal, established NS cell lines are no longer expandable and start to differentiate (Pollard et al., 2006). Interestingly, upon EGF removal, astrocytic differentiation was low and only after complete growth factors deprivation GFAP+ cells appeared in quantity (more than 50% in the total cell number). These observations suggest that eNPC can rapidly differentiate into neurons in the presence of bFGF alone but astrocytic differentiation is poor in the absence of other signals (see below). We have taken advantage of the plasticity of this cellular model to analyze the effects of microglia on the properties of undifferentiated NPC.

Neuronal differentiation, assessed as percentage of βIII tubulin+ cells, in cultures exposed to A-CM or co-cultured in the presence of acutely stimulated microglia, was strongly reduced. The percentage of newly generated neurons was extremely low after 3 or 6 days of differentiation when compared with all the other culture conditions. We also found that A-CM increased the fraction of TUNEL positive dead cells. These observations are consistent with previous data reporting negative effects of activated microglia on neurogenesis and neural cell survival (Cacci et al., 2005; Ekdahl et al., 2003; Nakanishi et al., 2007). The adverse effects on neuronal generation could result from the selective cell death of newborn neurons as suggested by the higher number of TUNEL+ cells in this condition, although data from Monje et al. on the effect of CM from activated BV2 microglial cells on NPC, do not support a selective vulnerability of neuroblasts (Monje et al., 2003). Alternatively, the decreased neuronal differentiation could be related to mechanisms favoring glial differentiation. In line with this interpretation, we found that the percentage of GFAP+ cells was significantly higher when NPC were cultured with acutely stimulated microglia than with unstimulated or chronically stimulated microglia, or their CMs. Nakanishi et al. (2007) have recently reported that microglia augmented astrocyte generation from NSC (Nakanishi et al., 2007), and this effect was prevented by neutralizing antibodies against IL-6 and LIF, two cytokines known to promote astrocytic differentiation (Chang et al., 2004; Taga and Fukuda, 2005). IL-6 binds to the heterodimeric gp130/IL-6 receptor and, through induction of the mammalian homologue of hairy-enhancer-of-split (Hes) 1, it can exert inhibitory effects on neuron generation from NSC by antagonizing proneural gene expression, and promote gliogenesis (Ishibashi et al., 1995; Monje et al., 2003; Nakamura et al., 2000; Sakamoto et al., 2003). In line with these findings, the strong tendency of NPC to generate astrocytes and, on the opposite, their poor differentiation into neurons when exposed to acutely stimulated microglia reported in our study, could partially be explained by the high levels of IL-6 released in their culture medium (see Fig. 2C). In addition to IL-6, it is likely that several other cytokines or soluble mediators contribute to the observed effects. Several studies have indicated a detrimental role for TNF-α in neural survival/differentiation (Cacci et al., 2005; Liu et al., 2005; Monje et al., 2003). In addition, preliminary observations from our group suggest that another cytokine highly induced by acute but not chronic LPS, IL-1α, contributes to the anti-neurogenic effect of acutely activated microglia. The dramatic decrease in the levels of these molecules, in chronically stimulated microglia, might create a permissive microenvironment for generation and survival of newborn neurons from NPC. In addition, we can not rule out a possible contribution of the increased and sustained synthesis of immunomodulatory products, such as PGE2 or IL-10, to the final outcome of NPC differentiation/survival.

Another important finding of the study is that all microglial CMs increased NPC proliferation and total cell number compared with cultures maintained in only bFGF. Previous works have failed to identify mitogenic effects of microglia on NSC (Butovsky et al., 2006; Monje et al., 2003; Nakanishi et al., 2007). One possible explanation is that, on the contrary to the other studies, in our culture conditions cell proliferation was evaluated in serum free medium and in the presence of bFGF. It is possible that bFGF cooperates with other factors released from microglia, in promoting cell proliferation. Furthermore, a mitogenic effect of microglia released neurotrophic factors has been reported for granule neuron precursor cells (Morgan et al., 2004).

In conclusion our data demonstrate that chronically activated microglia acquire an anti-inflammatory phenotype that is compatible with the generation and survival of newborn neurons from NPC. With respect to acute activation, chronic activation of microglia could enhance neurogenesis through increased neuronal differentiation and/or increased cell survival. The role that microglial cells with different states of activation play in the modulation of the neurogenic and gliogenic process needs to be further investigated in vivo. Recent findings suggest that newborn neurons generated from endogenous adult could partially replace dead cells following brain injury (Nakatomi et al., 2002; Thored et al., 2006). Therefore, the identification of suitable tools to direct microglial functions towards the proneurogenic chronic phenotype could represent a new strategy to enforce endogenous brain regenerative processes. In addition, the ability of microglia to induce NPC astrocytic differentiation is worth mentioning. The progliogenic effect has been proposed as a possible mechanism leading to the formation of the glial scar in lesioned areas after brain injury (Nakanishi et al., 2007). In contrast, given the role played by astrocytes in synapsis formation and stabilization, the enhancement of astrocytes generation could be regarded as a potentially beneficial mechanism in neurodegenerative diseases. These pathologies are often associated with severe synaptic and neuronal loss, and it has been proposed that symptomatic relief could be achieved, at least partially, by upregulation of neuronal plasticity events (Fischer et al., 2007).


We thank Dr. Zaal Kokaia for critical reading of the manuscript.