P7 POMC-EGFP and 3-month Thy1-EGFP transgenic mice were obtained from the animal facility of Hamburg University, UKE, Germany. Animals were housed under standard laboratory conditions in the animal facility at the Center for Molecular Neurobiology Hamburg. All experiments were performed in accordance with the institutional guide for animal care (license number ORG 850). Genotyping was performed by PCR analysis of genomic DNA, as described previously (Overstreet et al., 2004; Vuksic et al., 2008). P0, P3, P7, P14, P30, and 3-month Kunming mice were obtained from the Experimental Animal Center of Xi’an Jiaotong University. All procedures were performed in accordance with the protocol approved by the Institutional Animal Care of Northwest A&F University and Xi’an Medical University.

The following primary antibodies were used for immunofluorescence studies and Western blot analyses: rabbit polyclonal anti-MIF (1:500, ABclonal, United States), mouse monoclonal anti-Reelin G10 (dilution 1:1,000, Millipore, United States), mouse monoclonal anti-Parvalbumin (1:500, Abcam, United States), mouse polyclonal anti-GFAP (1:500, DAKO, United States), mouse monoclonal anti-NeuN (1:500, Millipore, United States), mouse monoclonal anti–neurofilament 200 (1:1,000, Millipore, United States), mouse monoclonal anti–β-tubulin (1:500; Sigma-Aldrich, United States), goat polyclonal anti-DCX (1:1,000; Santa Cruz, United States), mouse monoclonal anti-BLBP (1:300, Raybiotech, Inc., United States), and rat polyclonal anti-BrdU (1:500, Bio-Rad Laboratories Inc., United States). The secondary antibodies used for Western blotting were horseradish peroxidase (HRP)–conjugated goat anti–rabbit IgG and goat anti–mouse IgG (1:2,000; Cell Signaling Technology, United States). The secondary antibodies for immunostaining were as follows: Alexa Fluor 488-conjugated goat anti–mouse, Alexa Fluor 568–conjugated goat anti–mouse, Alexa Fluor 488-conjugated goat anti–rabbit IgG, Alexa Fluor 568–conjugated goat anti–rabbit IgG, Alexa Fluor 488-conjugated donkey anti–goat, Alexa Fluor 488–conjugated goat anti-rat. Dyes used were DAPI (1:10,000; Invitrogen, United States) and TRITC-labeled phalloidin (1:4,000; Sigma, United States).

S, R-3-(4- hydroxyphenyl)-4, 5-dihydro-5-isoxazole acetic acid methyl ester (ISO-1), an antagonist of MIF, was purchased from Calbiochem, Germany. BrdU was purchased from Sigma-Aldrich, United States.

The procedures of pCAG-EGFP and pCAG-MIF-MYC construction were identical as previously described (Zhang et al., 2014). In brief, the MYC tag was inserted into the vector pCAG-MCS (Biowit Technologies, China) by using In-Fusion^® HD Cloning Kit (Clontech, United States) according to the manufacturer’s protocol, and then the full-length MIF was cloned into pCAG-MCS-MYC by the traditional cloning method. pCAG-EGFP was obtained by the insertion of EGFP fragment into linearized pCAG-MCS using the traditional cloning method.

HEK293T cells were grown in DMEM (Dulbecco modified eagle medium, low glucose; Invitrogen) with 10% fetal calf serum (FCS; Invitrogen) and 1% penicillin-streptomycin (P/S, 10,000 U/mL; Invitrogen) at 37°C in a humidified atmosphere of 5% CO₂. When reaching 60–70% confluence in 60 mm culture dishes, HEK293T cells were transfected with pCAG-MIF-MYC using Lipofectamine 2000 Kit (Thermo Fisher Scientific, United States). Forty-eight hours after transfection, the cells were collected, frozen in liquid nitrogen, and stored at -80°C for Western blotting.

In utero electroporation was performed according to the procedure described as previously (Chai et al., 2015, 2016). Timed-pregnant females were deeply anesthetized by intraperitoneal injection of chloral hydrate (4.3 mg per 10 g of body weight). Then, the uterine horns of the pregnant mice were carefully exposed via a midline abdominal incision by pinching gaps between embryos with ring-forceps. Approximately 1–1.5 μL of pCAG-MIF-MYC and pCAG-EGFP (1 mg/mL for each) plasmid solution was injected into the lateral ventricle of mouse embryos on embryonic day 15.5 (E15.5) using a mouth-controlled pipette system and a pulled-glass micropipette. Fast green solution (0.1%) was added to the plasmid solution in a ratio of 1:10 to monitor the injection. The heads of injected embryos in the uterus were placed between the tweezers-type electrodes. Electronic pulses (30–50 V; 50 ms) were charged five times at intervals of 900 ms with an electroporator (Gene Pulser Xcell Electroporation System, United States). The uterine horns were placed back into the abdominal cavity to allow the embryos to continue normal development until sacrificed at P0. Brains (n = 9) were collected and fixed with 4% paraformaldehyde (PFA) at 4°C overnight. After rinsing in 0.1 M PB, brains were embedded in 5% agar and cut into 40-μm-thick slices on a Leica VT 1000S vibratome (Leica Microsystems, Germany). The sections were placed in a 24-well plate containing 0.1 M PB for immunostaining.

P0 and P3 Kunming mice were sacrificed under frozen anesthesia. P7, P14, P30, and 3-month Kunming mice were anesthetized with an overdose of Narkodorm (500 μL/kg) first with 0.9% NaCl and decapitated (n = 4 for each). Hippocampi were prepared and frozen with liquid nitrogen for immediate protein extraction or stored at -80°C. Tissue pellets and cell cultures were taken out from -80°C and incubated in ice-cold hypertonic lysis buffer [pH 7.6, 50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA-Na₂, 1% (vol/vol) Non-idet P-40, 1% Triton X-100, 0.5% (wt/vol) sodium dodecyl sulfate (SDS), 0.25% (wt/vol) sodium deoxycholate] with 1% protease inhibitor (Sigma, United States), and phosphatase inhibitor cocktails (Sigma). The tissue pellets and cells were lysed by repeated freezing in liquid nitrogen and thawing at 37°C (five times). After the tissue was triturated with a pipette tip to homogenize larger pieces of tissue and sonicated for 5 min, the suspension was centrifuged at 20,000 × g at 4°C for 30 min. The resulting crude supernatants were taken, and protein concentration was measured by using BCA Protein Assay Kit (Pierce, United States). Aliquots of the supernatants were immediately used for Western blotting or were stored at -80°C until use.

Five to 10 μg of each sample was diluted in ultrapure water containing 1 × NuPAGE sample buffer (Invitrogen) and 1 × NuPAGE reducing agent (Invitrogen) to final volumes of 24 μL, boiled at 95°C for 5 min, and then immediately placed on ice. The samples were separated by NuPAGE bis-tris, pH 7.0, 4–12% PAGE (Invitrogen) with 1 × NuPAGE MES-SDS running buffer (Invitrogen), and transferred electrophoretically to polyvinylidene fluoride membranes with 1 × NuPAGE transfer buffer (Invitrogen). For blotting, membranes were blocked with 5% (wt/vol) non-fat milk or 5% bovine serum albumin (BSA) in tris-buffered saline at room temperature (RT) for 1 h and finally incubated with primary antibodies in blocking solution at 4°C overnight. After three washes, the blots were subsequently incubated with a HRP-conjugated goat anti–rabbit or mouse IgG secondary antibody (1:2,000; Cell Signaling, United States) for 2 h at RT. The immunoreaction was visualized by enhanced chemiluminescence and scanned with a chemiluminescent and fluorescent imaging system (BioRad, United States). Bands were normalized to β-tubulin level to provide a control for equal loading. At least three Western blots were analyzed for each experiment. Blots were quantified by densitometry using ImageJ (NIH, United States). Significance was determined by one-way analysis of variance across all samples followed by Tukey test (^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001) (GraphPad Prism 5, United States).

Quantitative real-time PCR was performed according to the procedure described as previously (Zhang et al., 2014). In brief, P0, P3, P7, P14, P30, and 3-month Kunming mice (n = 4 for each) were anesthetized, and decapitated; hippocampi were prepared and frozen with liquid nitrogen for immediate total RNA extraction or stored at -80°C. Total RNA was extracted using Trizol reagent (Invitrogen, United States). The integrity of RNA was verified by 1% agarose gel, and cDNA was synthesized from 1 μg of total RNA primed with oligo (dT) using RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, United States). Quantitative PCR was performed using the Bio-Rad CFX96 (Bio-Rad, CA, United States) with Actb as an internal control. All PCR products span an intron in the genomic DNA. The cDNA was amplified with SYBR Premix Ex TaqTM (Takara, Japan). Relative quantification of mRNA levels was performed using the comparative Ct method.

Dissociated cultures of hippocampus were prepared as described previously (Chai et al., 2009). For primary cell cultures, newborn (P0) Kunming mice (n = 9) were used. The hippocampi were dissected and put into ice-cold Hanks balanced salt solution (HBSS). After removal of the meninges, the hippocampi of both hemispheres were collected and maintained in 15 mL Falcon tubes (BD Biosciences, United States) containing 3 mL of 0.5% Trypsin and 0.53 mM EDTAx4Na (Invitrogen, United States) and incubated at 37°C for 10 min. The hippocampi were then washed twice in ice-cold HBSS and manually triturated with a polished glass Pasteur pipette for several times. Then, the cells were centrifuged at low speed (800 × g) for 3 min to discard dead cells. The sediment was collected in fresh Falcon tubes and centrifuged again. The supernatant was discarded, and cell pellets were resuspended in 1 mL of neurobasal-A medium (Invitrogen, United States), supplemented with 2% B27 (Invitrogen, United States), 1 mM Glutamax (Invitrogen, United States), 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen, United States). Hippocampal cells were counted using a hemocytometer and suspended at a density of 5 × 10⁵ cells in 0.5 mL aliquots on glass coverslips inserted in 24-well plates. The coverslips had been coated overnight with a solution of 20 μg/mL poly-L-ornithine (Sigma, United States) prepared in borate buffer, pH 8.1. After 2 h of incubation in a humidified incubator at 37°C and 5% CO₂ to allow for cell attachment, the medium was changed to remove dead cells. After 12 h of incubation, cultures were divided into three groups (n = 12 wells for each group): one group was allowed to grow about 5 days. The second group was used for ISO-1 treatment for 48 h. The third group was used for control treatment for 48 h. Then the cultures were fixed with 4% PFA in 0.1 M PB, pH 7.4, for 15 min at RT and rinsed in 0.1 M PB for 1 h followed by incubation in blocking solution for immunostaining.

ISO-1 (Calbiochem, Germany) was freshly dissolved in dimethyl sulfoxide (DMSO). Twelve hours after seeding of neurons in 24-well dishes, medium was replaced by serum-free medium containing 10 or 100 μM ISO-1, and neurons were incubated for 48 h before fixing with 4% PFA in 0.1 M PB for immunostaining.

P5, P14, P30, and 3-month Kunming mice (n = 4 for each) were injected with BrdU [intraperitoneal (i.p.), 50 mg/kg; Sigma-Aldrich, United States] dissolved in physiological saline 2 h (evidence proliferating cells only, Conboy et al., 2011) before being anesthetized with an overdose of Narkodorm (500 μL/kg) with 0.9% NaCl and then transcardially perfused with 4% PFA in 0.1 M PB (pH 7.4). For MIF inhibition experiment, P6 Kunming mice (n = 8) were injected with ISO-1 (i.p., 7 mg/kg; Calbiochem, Germany) or vehicle (DMSO, i.p.) daily for 9 consecutive days. On P15, mice were injected with BrdU 4 h (to evidence proliferating cells only, due to the blocking effect of MIF by ISO-1, the incubation time was elongated) before being anesthetized and perfused with 4% PFA. Brains were postfixed with 4% PFA in 0.1 M PB overnight at RT, washed 24 h, and cut into 50 μm thick sections on a Leica VT 1000S vibratome (Leica Microsystems, Germany).

BrdU-labeled sections were denatured by incubation with 2 M HCl for 30 min at 37°C and washed with 0.1 M borate buffer (pH 8.0) for several times. After rinsing with 0.1M PB three times, sections were placed in a 24-well plate for immunohistochemistry.

P0 mice were decapitated under hypothermic anesthesia; brains were removed and immersion fixed in 4% PFA. P7, P14, P30, and 3-month mice (including POMC-EGFP and Thy1-EGFP transgenic mice) were anesthetized with an overdose of Narkodorm (500 μL/kg) first with 0.9% NaCl and then transcardially perfused with 4% PFA in 0.1 M PB (pH 7.4). Brains were removed and postfixed in 0.1 M PB (PH7.4) containing 4% PFA overnight at RT. After washing in 0.1 M PB overnight, fixed brains were embedded in 5% agar and cut coronally into 40-μm-thick slices to contain the entire hippocampus on a Leica VT 1000S vibratome (Leica Microsystems, Germany). After washing for 2 h at RT in 0.1 M PB, sections were placed in a 24-well plate containing 0.1 M PB for immunostaining. Sections together with sections from in utero electroporation, BrdU experiment, and neuronal cultures treated with ISO-1 and control medium were preincubated with blocking solution (5% normal goat serum and 1% BSA in 0.1 M PB containing 0.2% Triton X-100) at RT for 1 h and then incubated with the primary antibodies diluted in blocking solution overnight at 4°C. After washing in 0.1 M PB for 15 min three times, sections or cultured cells were incubated in secondary antibodies diluted in 0.1 M PB for 2 h at RT. After washing in 0.1 M PB for 15 min three times, sections or cultured cells were counterstained with DAPI (1:10,000; Invitrogen). Some cell cultures were stained for F-actin with TRITC-labeled phalloidin (1:4,000; Sigma). After rinsing in 0.1 M PB 15 min three times, sections and cells were mounted with fluorescent mounting medium (Dako North America, Inc., United States) and photographed by a structured-illumination microscope (Zeiss observer Z1) or using a confocal laser scanning microscope (Olympus FV1000).

Quantifications of double-labeled cells with MIF and NeuN, MIF and GFAP, MIF and parvalbumin (Pv), and MIF and Reelin were performed using ImageJ software.

Morphometric measurements of cultured hippocampal neurons treated with vehicle, 10 μM ISO-1, and 100 μM ISO-1 were performed by ZEN 2011 software. Data were from three independent neuronal cultures. At least three fields were randomly photographed in each condition of every experiment; 2.5 DIV neurons have differentiated with a longest neurite, which later will develop into axons. Neurites were traced along the processes. The longest neurite, number of terminal tips, and total length of neurites were analyzed by GraphPad Prism 5. One-way analysis of variance with Tukey post test and Kruskal–Wallis test with Dunn post test were performed for parametric and non-parametric data, respectively. Graphs were represented as the mean ± SEM. Differences were considered significant when P < 0.05, and represented as ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001.

Quantification of newly generated cells was calculated by counting BrdU-positive cells in 1,000 μm sectional length in the SGZ and granular cell layer (GCL) of the hippocampus from P6 mice injected with either DMSO or ISO-1 for consecutive 9 days. Totally, 12 coronal sections for each group were photographed at 20× magnification with a Zeiss microscope. Values were analyzed for significance using the Student two-tailed t-test (GraphPad Prism 5).

In order to identify the effectiveness of MIF antibody purchased commercially, plasmid pCAG-MIF-MYC expressing MIF and MYC tag and plasmid pCAG-EGFP expressing EGFP were constructed and electroporated into embryonic cortical neurons. After birth of mice, brains were collected and cut into sections for MIF immunostaining. We found that neurons transfected with EGFP and MIF plasmids on E15.5 have reached their final destinations in the superficial layer of the cortex. They were perfectly recognized by MIF immunostaining (Figure 1A). Plasmid pCAG-MIF-MYC was also transfected into cultured HEK293T cells. Cells were harvested, and MIF protein expression was analyzed by Western blotting. As shown in Figure 1B, MIF antibody detected not only endogenous MIF-MYC, but also exogenous MIF (Figure 1B). Our observations indicate that MIF antibody purchased commercially is specific and sufficient to examine MIF expression by immunostaining.

To investigate the expression profile of MIF during development of hippocampus, we first checked mif mRNA by using quantitative real-time PCR at different developmental time points: P0, P3, P7, P14, P30, and 3 months. The results showed that expression level of mif mRNA gradually increased from P0 to P7, reached the peak at P7, and stably expressed until P30, and then declined dramatically to adulthood (Figure 2A). Compared with that at P0, mif mRNA increased by 0.6-fold at P7 and decreased by 0.4-fold at adult (Figure 2A). Western blotting result demonstrated that MIF was expressed at all postnatal ages (Figure 2B). Densitometric analysis from three independent experiments showed that the expression profile of MIF protein during hippocampal development is like an inverted bell with the lowest level at P7 and higher level before or after (Figure 2C). Compared to that at P7, the MIF expression was increased by 2.8-fold at P3 and 0.9-fold at P14 (Figure 2C). Statistical analysis revealed significant differences between P7 and other time points during development of hippocampus (Figure 2C).

Immunostaining of MIF in P0, P7, P14, and P30 hippocampus further proved that MIF expression at P7 was the lowest compared to those at P0, P14, and P30 ages (Figure 3A). In addition, we noticed strong MIF expression in stratum lucidum of CA3 from P14 (Figure 3A). To make sure that MIF was located in mossy fiber projections, the axons of granule cells, we performed double-immunolabeling of MIF and tubulin in sections of P14 hippocampus, and results revealed that MIF was perfectly overlapped with tubulin in mossy fibers (Figures 3B,C).

In order to explore the relationship between MIF and the nerve fiber structure, immunostaining of MIF and cytoskeletal molecules was performed in cultured neurons. Phalloidin recognizes F-actin (filamentous actin), which forms actin cytoskeleton, and tubulin recognizes microtubule cytoskeleton. Our results clearly showed that MIF colocalized with phalloidin in F-actin and tubulin in microtubules, indicating that MIF interacted with cytoskeletal structures (Figures 4A,B).

To examine whether the cells expressing MIF are neurons during development of hippocampus, we first performed double-immunostaining for MIF and NeuN in hippocampal sections from 3-month mice. NeuN is a specific marker for mature neurons. Our results showed that many MIF-positive cells were scattered throughout the whole hippocampus, and a group of MIF-positive cells was accumulated in subgranular zone (SGZ) of dentate gyrus (DG) and formed a densely packed cell layer with long processes extending into GCL (Figure 5A). Those MIF-expressing cells scattered in hilus of DG (Figure 5A) and in stratum oriens (so), stratum radiatum (sr), and stratum lacunosum molecular (slm) of hippocampus proper were found NeuN-positive (Figures 5B,C). However, neither NeuN-positive granule cells in the granule layer (GCL) nor NeuN-positive pyramidal neurons in the pyramidal cell layer (PCL) were seen MIF-positive (Figures 5A–C). Notably, the densely packed cell layer with MIF expression in SGZ showed NeuN-negative (Figure 5A). These results indicated that MIF was expressed in both NeuN-positive and NeuN-negative cells (Figures 5A–C). Here again a strong expression of MIF was observed in mossy fibers in sl attached to the pyramidal neurons in CA3 region (Figure 5C). We took an area of 0.1 mm² in the DG, CA1, and CA3 areas of each hippocampal slice; counted 10 MIF-positive cells; and calculated the percentage of MIF and NeuN double-labeled cells in MIF-positive cells. Five hippocampal slices in total were used. As shown, the percentages of MIF and NeuN double-labeled cells in MIF-positive cells in CA1,CA3 and DG (h + m) were much higher than that of MIF and NeuN double-positive cells in MIF-positive cells in SGZ (Figure 5D).

As the most of MIF and NeuN double-labeled cells in hippocampus were scattered throughout the whole hippocampus, not in the GCL and PCL, and morphologically very similar to interneurons (Figure 5), we speculated that these MIF and NeuN double-positive cells were interneurons. To confirm our supposition, we carried out double-immunolabeling of MIF with parvalbumin (Pv) and Reelin antibodies, respectively, in 3-month-old hippocampus. The results revealed that the majority of MIF-positive cells in hippocampus except the cells in SGZ were colabeled with parvalbumin (Figures 6A–C) or Reelin (Figure 6D). Given that parvalbumin and Reelin are markers for GABAergic interneurons, these results indicated that the scattered MIF-positive cells in hippocampus were GABAergic interneurons. Finally, we counted MIF-positive cells and parvalbumin-positive cells in 0.03 mm² area of CA1 and CA3 region, MIF-positive cells, and Reelin-positive cells in 0.03 mm² area of hippocampal fissure (HF) region. In Figure 6E, the rectangular boxes indicate the regions as shown in Figures 6A–D. Six hippocampal slices in total were used. We found that parvalbumin- or Reelin-positive cells were all MIF-positive, but not all MIF-positive cells were parvalbumin- or Reelin- positive. The percentage of parvalbumin and MIF double-labeled cells in MIF-positive cells was higher than that of Reelin and MIF double-labeled cells in MIF-positive cells (Figure 6F).

As MIF was strongly expressed in mossy fibers (Figures 3B,C) and previous study has suggested that NeuroD-positive immature granule cells express MIF (Conboy et al., 2011), wherefore, to find out whether the granule cells express MIF, we used P7 POMC-EGFP and 3-month-old Thy1-EGFP transgenic mice, in which most immature and mature granule cells are labeled by EGFP, respectively. In the transgenic mice, POMC promoter-driven EGFP expression in hippocampal granule cells remains only 1 week after generation. One week after, granule cells began to express EGFP under Thy1 promoter. Therefore, they can represent immature granule cells and mature granule cells. MIF immunostaining showed that neither POMC-EGFP–labeled immature granule cells nor Thy1-EGFP–positive mature granule cells express MIF (Figure 7). POMC-EGFP–positive immature granule cells in GCL were not overlapped with any MIF-positive cells distributed in DG (Figure 7A). Thy1-EGFP–positive mature granule cells in GCL were also MIF-negative (Figures 7B,C). Moreover, MIF-positive cells scattered in SGZ and hilus of DG (DH) displayed giant morphology prompting them as interneurons (Figures 7B,C). In Figure 7D, the rectangular boxes indicate the regions shown in Figures 7A–C.

We have mentioned above that many MIF-positive cells accumulated in SGZ of DG and formed an MIF-positive cell layer. Interestingly, in contrast to MIF-positive cells in other regions, the MIF-positive cells in SGZ were NeuN-negative and gave rise to MIF-positive processes pointing to molecular layer of DG (Figure 5A). It is well known that SGZ is the place where the neural stem cells and newly generated neurons are located, and neurogenesis occurs for whole life (Zhao et al., 2007). Thus, we wondered whether the neural stem cells or newly generated neurons express MIF. To test these possibilities, we performed double-labeling of MIF with GFAP, BLBP, and DCX, respectively, in 3-month-old hippocampus. Our results showed that most of MIF-positive cells in SGZ were colabeled by GFAP (Figure 8A) and BLBP (Figure 8B), markers for neural stem cells. We also found some MIF-positive astrocytes around the small blood vessels in the hippocampus proper (data not shown). In contrast, no MIF-positive cells in SGZ were found to be colabeled with DCX (Figure 8C), a marker for newly generated immature neurons. It is well known that GFAP is expressed by radial glial cells (neural stem cells), which are located in SGZ of the DG and act as neural precursor cells in developing hippocampus. In adulthood, most radial glial cells are transformed into astrocytes, whereas GFAP is retained by astrocytes. However, there are still some GFAP-positive radial glial cells in the SGZ acting as neural precursor cells (Zhao et al., 2004; Brunne et al., 2010). BLBP is another marker of neural precursor cells, which recognizes late differentiated radial glial cells (Hartfuss et al., 2003; Brunne et al., 2010). DCX is a microtubule-associated protein and expressed by newly generated neurons in SGZ of DG (Zhao et al., 2007). Our findings suggested that MIF is involved in maintenance of the feature of neural stem cells and cell proliferation. To quantify our results, we counted the number of GFAP-positive cells, BLBP-positive cells, and DCX-positive cells in 0.03 mm² area in DG respectively. Six hippocampal slices were counted. We found that the percentage of GFAP and MIF double-labeled cells in GFAP-positive cells was higher than that of BLBP and MIF double-positive cells in BLBP-expressing cells; no double-labeled cells with DCX and MIF were found in DG (Figure 8D). These results further indicated that MIF is maintained in GFAP- or BLBP-positive neural stem cells during hippocampal development.

Our data suggested that MIF is involved in cell proliferation. Conboy et al. (2011) have also shown that lack of MIF causes impairment of cell proliferation of adult hippocampus. So we were wondering whether MIF controls cell proliferation of neural stem cells in SGZ of DG during hippocampal development. Then we injected BrdU into P5, P14, P30, and 3-month-old adult mice 2 h before the animals were sacrificed and conducted double-labeling of MIF and BrdU in hippocampus. As revealed in Figure 9, BrdU emerged in MIF staining partially in SGZ of DG at all stages. By injection of ISO-1, an antagonist of MIF, into P5 mice for 9 consecutive days, we found a remarkable reduction of BrdU-positive cells in SGZ of DG in treated mice compared to that in control mice (Figure 10A). Statistical analysis was shown in Figure 10B. Our results indicate that inhibition of MIF impairs hippocampal cell proliferation at an early stage.

During hippocampal development, MIF was expressed in mossy fibers projected from granule cells to pyramidal neurons in CA3 region suggesting that MIF might participate in the regulation of axonal outgrowth (Figure 3). To test this possibility, we treated primary neuronal cultures with ISO-1 using different concentrations (10 and 100 μM) for 2 days. After fixation with 4%PFA in 0.1 M PB, cultures were visualized with neuronal marker neurofilament 200 and photographed. Three independent fields on a coverslip were imaged. At least three coverslips for each experiment were analyzed. Compared with control condition, treatment at both concentrations of ISO-1 seemed to affect neurite outgrowth (Figure 11A). To verify the extent of the effect, we measured the length of processes. The statistical analysis and cumulative probability results showed that 10 and 100 μM ISO-1 strongly decreased the length of the longest neurites of cultured neurons (Figures 11B,C). To identify the number of total neurites, we counted the terminal tips of neurites. The result revealed that 10 μM ISO-1 treatment led to a reduction in number of neurites, with significant difference compared with that in control and 100 μM ISO-1–treated neurons (Figure 11D). These data implied that appropriate concentration of ISO-1 is critical to block MIF function. In addition, the percentage of neurons with different number of neurites in control and ISO-1–treated cultures was evaluated, and the total length of processes except the longest neurite was measured (Figure 11E). No significant difference between control and ISO-1–treated neurons was observed (Figure 11F). It is well known that the longest protrusion from a neuron is axon, and the other protrusions differentiate to dendrites. Therefore, our results indicate that MIF might participate in the regulation of mossy fiber projection and dendritic bifurcation. Blocking of MIF with ISO-1 inhibited axonal elongation and dendritic formation.