Characteristics of primary rat microglia isolated from mixed cultures using two different methods

DMEM/F12, 0.25% trypsin-EDTA, and fetal bovine serum (FBS) for cell culture were from Gibco.

Primary antibody against Iba-1 and Alexa Fluor 555-conjugated secondary antibody were from Abcam and Molecular Probe, respectively. The antibodies of CD11b, CD45, and CD68 for flow cytometry detection were purchased from BD, eBioscience, and Bio-Rad, respectively, and were diluted according to the manufacturer’s instructions. Recombinant rat IL-4 was from Amsbio. LPS was from Sigma-Aldrich. Recombinant rat IFNγ was from R&D. The stock solutions of IL-4, LPS, and IFNγ were prepared using sterile ddH₂O, and ddH₂O was used as control.

A PubMed search was conducted using the terms “microglia AND (rat OR mouse) AND primary AND culture NOT review [publication type]” from 2006 to 2016. This search resulted in 392 articles. One hundred twenty-five articles were excluded (81 did not use microglial culture, 25 microglia used cell lines, 19 were not available for full-text access), resulting in 267 articles that used primary cultures of rodent microglia. This quick survey of the literature suggested that two most commonly used methods to purify microglia involved shaking and mild trypsinization (204 used shaking and 26 of them used mild trypsinization) (Fig. 1). Based on this initial analysis, we compared microglia obtained with shaking versus trypsinization. Primary rat microglia cultures were prepared from cerebral cortices of 1–2-day-old neonatal Sprague-Dawley rats. After removing the meninges, the cortical tissues were digested with 0.25% trypsin-EDTA for 30 min at 37 °C, followed by mechanical triturating in DMEM/F12 with 10% fetal bovine serum. The mixed cortical cells were passed through a 70-μm nylon mesh cell strainer and plated on non-coated plastic dishes or plates in DMEM/F12 with 10% FBS, and the medium was completely replaced every 3–4 days. After achieving confluency at about 14 days in vitro, the microglia were isolated from mixed glial cultures via either mild trypsinization (enzyme, E) or shaking (S). The mild trypsinization was performed according to previously described methods [19, 20]. Incubation of mixed glial cultures with a trypsin solution (0.25% trypsin-EDTA diluted 1:4 in DMEM/F12) for 15–25 min resulted in the detachment of an intact layer of cells in one piece. Microglial cells remained attached to the bottom of the well. For the shaking method [21, 22], confluent mixed glial cultures were placed on an orbital shaker at 220 rpm for 1 h. The supernatant containing the detached microglial cells was collected and re-seeded for 1 h to allow microglial attachment. After 1 h, the nonadherent cells were removed. Microglia isolated from both methods were allowed to rest overnight prior to treatments. To compare the yield, we plated the cells from one brain of neonatal pups into one 6-well plate. Yield was calculated as cell numbers per field (six random fields of ×200 magnification per culture, n = 5 cultures). The cells were fixed in 4% paraformaldehyde for 30 min, blocked with 5% normal horse serum for 1 h, and incubated with primary antibody against Iba-1 (1:100) at 4 °C overnight. After washing, the cells were incubated with Alexa Fluor 555-conjugated secondary antibodies (1:200) for 1 h at room temperature. Negative controls were incubated without primary antibodies, and no immunoreactivity was observed in these controls.

The purity of microglia obtained from mild trypsinization or shaking was estimated by flow cytometry. The cells were collected and labeled with fluorochrome-conjugated monoclonal antibodies recognizing antigens (CD11b, CD45, and CD68) at 4 °C for 15 min. After labeling, the cells were washed twice in PBS and resuspended at a final volume of 400 μl. Flow cytometry was performed on a BD LSDII, and data were analyzed using FlowJo software.

Quantitative real-time PCR was used to measure mRNA expression. Cultured microglia were treated with IL-4 (20 ng/ml), LPS (50 ng/ml), or IFNγ (20 ng/ml) for 8 h. Total RNAs were extracted using miRNeasy kit (Qiagen) from primary cultured microglia with or without treatment. One hundred nanograms of total RNAs were reverse transcribed into cDNA using M-MLV reverse transcriptase (Invitrogen). Quantitative expression of inducible nitric oxide synthase (iNOS), CD86, CD206, arginase 1, CX3C chemokine receptor 1 (CX3CR1), toll-like receptor 2 (TLR2), and C-C chemokine receptor 2 (CCR2) were measured using gene-specific TaqMan Gene Expression Assays (ABI 7500HT, Applied Biosystems). Relative baseline gene levels were calculated by subtracting Ct value of β2-microglobulin (B2M) from Ct value of detected genes. Changes in gene expression (fold change) after various treatments were determined using the 2^−ΔΔCt method with normalization to B2M. All real-time PCRs were performed in triplicates. All experiments were repeated three to six times independently. Activated microglia include a spectrum of various states. The representative genes we selected, including iNOS, CD86, CD206, arginase 1, CX3CR1, TLR2, and CCR2, are closely relevant to different activating states of microglia. To compare the “average fingerprint” of cultured microglia, quantitative data were analyzed using two-way ANOVA (p < 0.05 for significance, SPSS 21).

Cultured microglia were treated with IL-4 (20 ng/ml), LPS (50 ng/ml), or IFNγ (20 ng/ml) for 24 h. Tumor necrosis factor α (TNFα) (eBioscience), IL-1β (R&D), IL-10 (eBioscience), and insulin-like growth factor-1 (IGF-1) (R&D) in cell culture supernatant were measured using ELISA kits, according to the manufacturer’s instructions.

To assess phagocytosis, microglial cells cultured in 6-well plates were digested using 0.25% trypsin-EDTA and re-seeded to a 96-well plate at a concentration of 3.0 × 10⁴ cells/well. Then, the cells were incubated with fluorescein-labeled Escherichia coli K-12 BioParticles (Invitrogen) for 2 h at 37 °C. The cells were rinsed with 0.25 mg/ml trypan blue to quench extracellular fluorescence. Intracellular fluorescence was read using a fluorescence microplate reader setup with excitation at 480 nm and emission at 520 nm. The experiments were performed with five replicates per condition and repeated four times.

Data were expressed as mean ± SE. Three to six separate experiments were performed. Data of real-time PCR were analyzed using two-way ANOVA. Data of ELISA that measures the cytokine release after various treatments were analyzed using one-way ANOVA. Other data were analyzed using t test between two isolation methods. Statistical significance was set at p < 0.05.

Phase contrast microscopy revealed distinct morphology of microglia isolated using two different methods. Microglia isolated using mild trypsinization showed uniform ramified morphology with short processes and small cell body (Fig. 2a). The morphology of microglia isolated using shaking was more heterogeneous. Most of them showed enlarged round cell body with reduced processes (Fig. 2b). Cells from both methods were Iba-1 positive (Fig. 2c, d). The yield of microglia cultures was higher by using mild trypsinization (76.18 ± 13.08 cells/field) than by using shaking (54.27 ± 9.19 cells/field) (p = 0.015, t test) (Fig. 2e).

Besides morphology, the size of microglial cells obtained from the two methods was also different. Measured by flow cytometry with forward scatter (FSC), the average size of microglia isolated using shaking was 1.17-fold larger than those obtained using mild trypsinization (p = 0.030, t test) (Fig. 2f–h). The purity of cultured microglia was determined by flow cytometry with CD11b and CD45 staining. The proportion of CD11b-positive cells (Fig. 2i–k) and CD45-positive cells (Fig. 2l–n) were significantly higher in the mild trypsinization group (93.68 ± 2.54% and 94.92 ± 3.64%) compared to the shaking group (82.9 ± 7.61% and 88.08 ± 3.32%) (p = 0.028 for CD11b, p = 0.024 for CD45, t test).

Because the morphology of microglia appeared to be different in the two preparation methods, we next asked whether these cells might also demonstrate different phenotypes. A representative panel of genes was examined to assess various activation states—iNOS, CD86, CD206, arginase 1, CX3CR1, TLR2, and CCR2. Baseline expression of these selected genes were significantly different in microglia obtained by shaking versus mild trypsinization (p = 0.003, 0.000, and 0.036 for methods, genes, and methods*genes, respectively, two-way ANOVA) (Fig. 3a).

Microglia are known to influence adjacent cells via the release of extracellular signals. Therefore, we used ELISA to assess key cytokines (TNFα, IL-1β, and IL-10) and the major microglial growth factor IGF-1 (Fig. 3b). Baseline levels of IL-1β (p = 0.122, t test) and IGF-1 (p = 0.032, t test) were about 2-fold higher in conditioned media from cultured microglia isolated by shaking compared with mild trypsinization. Even larger differences were observed for IL-10 (9-fold, p = 0.048, t test) and TNFα (24-fold, p = 0.020, t test), again with significantly higher levels from microglia isolated using shaking than using mild trypsinization.

In general, these differences in gene expression and cytokine release suggested that microglia isolated with shaking may be “more activated” compared with those obtained via mild trypsinization. CD68 is considered to be a general marker of activated microglia. Flow cytometry showed that the percentage of CD68-positive cells was slightly higher in microglia isolated using shaking (39.14 ± 6.94%) than using mild trypsinization (31.58 ± 7.80%), but there was no significant difference between the two methods (p = 0.185, t test) (Fig. 3c–e). However, an in vitro assay demonstrated that phagocytic capacity was significantly enhanced by about 2.5-fold in microglia by shaking versus mild trypsinization (p = 0.018, t test) (Fig. 3f–h).

Next, we tested the response of microglia isolated by these two methods to three typical stimuli that were commonly used to activate microglia into different phenotypes, i.e., IL-4, LPS, and IFNγ. After treatment with IL-4 (Fig. 4a) or LPS (Fig. 4b) for 8 h, the pattern of gene expression response (iNOS, CD86, CD206, arginase 1, CX3CR1, TLR2, and CCR2) appeared to be generally similar (for IL-4 treatment, p = 0.881, 0.000, and 0.408 for methods, genes, and methods*genes, respectively; for LPS treatment, p = 0.962, 0.000, and 0.430 for methods, genes, and methods*genes, respectively, two-way ANOVA). For example, both IL-4 and LPS treatment decreased CX3CR1 expression. IL-4 treatment upregulated the expression of M2-like phenotype marker CD206, whereas LPS treatment upregulated the expression of M1-like phenotype marker iNOS and downregulated CD206. The responses to 8-h IFNγ treatments were slightly more variable but once again, overall patterns were similar (p = 0.333, 0.023, and 0.916 for methods, genes, and methods*genes, respectively, two-way ANOVA). IFNγ increased the expression level of iNOS and decreased the level of CD206 in microglia from both shaking and trypsinization groups (Fig. 4c). Consistent with these gene expression findings, morphological changes also appeared to be mostly the same in both groups after various treatments for 24 h (20 ng/ml of IL-4, 50 ng/ml of LPS, or 20 ng/ml of IFNγ) (Fig. 5a–h).

For further comparisons, the conditioned media after 24 h treatment with the IL-4, LPS, or IFNγ stimuli were collected to measure the levels of secreted cytokines and growth factors. Consistent with the baseline data described above, pre-stimulation levels of TNFα, IL-1β, IL-10, and IGF-1 were significantly higher in conditioned media of microglia isolated using shaking compared with trypsinization. Stimulation with IL-4, LPS, or IFNγ induced distinct responses (Fig. 5i–l, one-way ANOVA). LPS significantly affected TNFα, IL-1β, and IL-10. IL-4 significantly increased the production of IGF-1, but it had no effect on IL-10, TNFα, and IL-1β. IFNγ had little effects. The direction of the response (increase or decrease) was mostly similar in microglia isolated with both methods (Fig. 5i–l). For example, in microglia isolated from either mild trypsinization or shaking, IL-4 treatment increased IGF-1 release, and LPS treatment increased the generation of TNFα and IL-1β. However, there were some key differences in the extent of response in some cases. To assess differences of the extent of response, we recalculated the data as fold change versus control in Fig. 5m–o. Compared with the baseline release, LPS treatment induced significantly higher levels of TNFα and IL-10 in microglia obtained by mild trypsinization than by shaking (Fig. 5n). For IFNγ stimulation, TNFα release was significantly elevated in microglia isolated using shaking, but not in microglia isolated using mild trypsinization (Fig. 5o).