Granulocyte-colony stimulating factor promotes brain repair following traumatic brain injury by recruitment of microglia and increasing neurotrophic factor expression

Abstract

Purpose: The overall objective was to elucidate cellular mechanisms by which G-CSF enhances recovery from traumatic brain injury in a hippocampal-dependent learning task.

Methods: Chimeric mice were prepared by transplanting bone marrow cells that express green fluorescent protein (GFP+) from a transgenic “green” mice into C57BL/6 mice. Two months later, the animals sustained mild controlled cortical impact (CCI) to the right frontal-parietal cortex, followed by G-CSF (100 μg/kg) treatment for 3 consecutive days. The primary behavioral end-point was performance on the radial arm water maze (RAWM) assessed before and after CCI (days 7 and 14). Secondary endpoints included a), motor performance on a rotating cylinder (rotarod), b) measurement of microglial and astroglial response, c) hippocampal neurogenesis, and d) measures of neurotrophic factors (BDNF, GDNF) in brain homogenates.

Results: G-CSF treatment resulted in significantly better performance on the rotorod at one week, and in the RAWM after one and two weeks. The cellular changes found 2 wks after CCI in the G-CSF group included increased numbers of hippocampal newborn neurons as well as astrocytosis and microgliosis in striatum and frontal cortex on both sides of brain. GFP+ cells that co-labeled with Iba1 (microglial marker) comprised a significant proportion of striatal microglia in G-CSF treated animals, indicating the capacity of G-CSF to increase microglial recruitment to the site of injury. Neurotrophic factors GDNF and BDNF, elaborated by activated microglia and astrocytes, were increased in G-CSF treated mice.

Conclusions: G-CSF serves as a neurotrophic factor that increases hippocampal neurogenesis (or enhances survival of new-born neurons), and activates astrocytes and microglia. In turn, these activated glia release a plethora of cytokines and neurotrophic factors that contribute, in a poorly understood cascade, to the brain’s repair response. G-CSF also acts directly on bone marrow-derived cells to enhance recruitment of microglia to the site of CCI from circulating monocytes to the site of CCI.

Keywords

Granulocyte-colony stimulating factor traumatic brain injury neuro-inflammation neurogenesis doublecortin astrocytosis microgliosis chimeric mice green fluorescent protein

1 Introduction

Traumatic brain injury (TBI) is an insult to the brain caused by an external physical force that may produce a diminished or altered state of consciousness and results in impairment of cognitive abilities or physical functioning. The neuropathogenesis of TBI is a complex process encompassing three overlapping phases: a) primary injury to brain tissue and/or the cerebral vasculature, b) the secondary injury which includes physiological, neuro-inflammatory and biochemical processes triggered by the primary insult and c) regenerative responses including enhanced proliferation of neural progenitor cells and endothelial cells (Cernak, 2005). The secondary injury evolves over hours to days and may significantly contribute to chronic post-traumatic neurologic disability. Closely overlapping and following the secondary injury phase is the complex regenerative response that ultimately determines the extent of functional recovery from the injury.

Treatment of TBI immediately or within hours of the insult with granulocyte-colony stimulating factor (G-CSF) has been reported to promote recovery from neurologic deficits in rodents (Yang et al., 2010). The mechanisms responsible for the reported beneficial effects are complex because of the pleiotropic actions of this hematopoietic cytokine. G-CSF is one of several colony stimulating factors (CSFs) that control the production of circulating blood cells by the bone marrow. Although G-CSF has been used primarily to treat leukopenia, the agent has been studied in animal models of stroke where it has been reported to reduce brain damage and improve outcome (Schabitz et al., 2003; Shyu et al., 2006; Six et al., 2003; Solaroglu, Cahill, Jadhav, & Zhang, 2006). Ongoing clinical trials are evaluating the effectiveness and safety of G-CSF for treatment of ischemic stroke (Schabitz et al., 2003; Sprigg et al., 2006). One recently completed large multi-center Phase II study of G-CSF for acute stroke failed to meet primary endpoints (changes in stroke score). However, a trend towards reduced infarct growth was found in the G-CSF treatment group (Ringelstein et al., 2013).

Several reports have documented benefical effects of G-CSF given for treatment of TBI. Intraperitoneal administration of G-CSF (via osmotic minipump) resulted in significantly better motor function recovery than the control group (Yang et al., 2010). The G-CSF group exhibited a greater increase in proliferative cells (BrdU+) and a significantly higher number of doublecortin expressing (DCX+) cells in the ipsilateral subventricular zone (SVZ) than the control group (Yang et al., 2010). In another recent study, intravenous (iv) infusion of bone marrow stromal cells (BMSC) was found to be as effective as G-CSF administration in a rat model of TBI in promoting recovery of motor function (Bakhtiary et al., 2010). Either mode of treatment (BMSC vs G-CSF) was significantly better than control (vehicle) treatment. Mechanisms responsible for the recovery were not addressed other than to state that both treatments resulted in increased BrdU+ cells at the border of injury.

Treatment as long as a week after the injury has also been beneficial. A single dose of G-CSF (300 μg/kg) administered one week after moderate TBI in rats resulted in reduction of brain lesion volume, decreased cell death, improved motor behavior, dampened neuro-inflammation and enhanced hippocampal neurogenesis, when measured 8 weeks after the injury (Acosta et al., 2014).

The mechanisms responsible for mediating the beneficial effects of G-CSF in TBI are incompletely understood. G-CSF treatment has direct effects on neural cells of brain (Jung et al., 2006; Sanchez-Ramos et al., 2009; Schneider, Kruger, et al., 2005), and is known for its capacity to increase circulating blood stem/progenitor cells and leukocytes. There is insufficient information on the extent to which blood marrow-derived cells (BMDCs) infiltrate the injured brain and differentiate into microglia and other cells. Moreover, it is not clear how the infiltrating blood-derived cells participate in the repair response. Hence, there is a need to analyze cellular and molecular mechanisms of action of G-CSF in the TBI model.

The overall goal of the present project is to understand how G-CSF mitigates the secondary injury that occurs hours to days in a mouse model of TBI. A significant component of this study utilizes a chimeric mouse model in which bone marrow-derived green fluorescent protein expressing (GFP+) cells can be tracked into brain to determine the role of microglial recruitment from blood monocytes to the site of injury. Specific objectives for the present report are 1) to study the impact of G-CSF treatment on water maze performance, 2) to measure the extent of infiltration of BMDCs from the blood to brain, 3) to determine the cellular fate of the BMDCs at the site of injury, and 4) to measure neurotrophic factors within regions of brain affected by controlled cortical impact (CCI). Alow dose of G-CSF (100 μg/kg) was chosen for administration in the 3 days after injury because earlier studies revealed that a single (300 μg/kg) administered intravenously 1 wk after CCI improved motor performance when tested 8 weeks later (Acosta et al., 2014). The dose of G-CSF (100 μg/kg×3 doses) administered in the present study was lower than that reported to be effective in mobilizing bone marrow in a rat model of stroke (300 μg/kg×10 days) but higher than that utilized by others in rodent models of stroke (Schabitz et al., 2003; Six et al., 2003; Solaroglu, Cahill, et al., 2006) and in a recently published report on the treatment of memory impairment in a mouse model of AD (50 microg/kg s.c. for 5 days) (Tsai, Tsai, & Shen, 2007).

2 Methods

This study was carried out in strict accordance with the recommendations from the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committee at the University of South Florida.

2.1 Animals

C57BL/6 mice, 8– 10 weeks old, were purchased from Harlan Laboratories, and transgenic GFP mice (C57BL/6-Tg [ACTB-EGFP] 1Osb/J, 003291) were obtained from Jackson Laboratory (Bar Harbor, ME). All the experiments, with the exception of 12 control C57BL/6 mice that did not undergo transplantation or CCI, utilized “chimeric mice”. These were prepared from C57BL/6 mice transplanted with green fluorescent protein expressing (GFP+) bone marrow. The procedure for bone marrow harvesting from transgenic (tg) GFP+ mice has been previously published (Sanchez-Ramos et al., 2009; Song & Sanchez-Ramos, 2002). Briefly, bone marrow cells are collected from femurs and tibias of adult male GFP transgenic mice by flushing the bone shaft with PBS + 0.5% bovine serum albumin (BSA) + 2 mM ethylenediaminetetraacetic acid (EDTA) (Sigma). To generate chimeric mice, C57BL/6J mice were lethally irradiated with 8 Gy total body irradiation (delivered in two fractions of 4 Gy, an interval of 4 hours) at dose rate of 1.03 Gy/min in a Gammacell 40 Exactor Following irradiation, the mice were given a bone marrow transplant (10×10⁶ mononuclear cells) from transgenic GFP mice infused via tail vein. Bone marrow-derived cells in the rescued mice were readily tracked by virtue of their green fluorescence. Examination of blood smears from tail clippings for the presence of green monocytes confirmed successful engraftment.

A total of 60 chimeric mice, divided into 3 groups of 20, were utilized in this study. Group 1 (n = 20) did not undergo RAWM testing and was euthanized 3 days after CCI. Group 2 (n = 20) had RAWM at baseline before CCI and on days 5 to 7 after CCI followed by euthanasia on day 7. Group 3 (n = 20) had RAWM at baseline and on days 12– 14 followed by euthanasia on day 14. In each group of 20 mice, 10 were randomly assigned to receive vehicle injections and 10 received G-CSF. In addition, 12 non-chimeric control mice that did not undergo CCI were utilized in the rotarod study and for assessment of expression of GFAP (astrocytes) and Iba1 (microglia).

A sub-set of 4 mice from each group of vehicle- or GCSF-treated mice were euthanatized, brains dissected and brain regions frozen for determination of brain levels of trophic factors (BDNF, GDNF). The remaining sub-set of 6 mice from each group of chimeric mice underwent histological analyses to determine extent of microgliosis and astrocytosis as well as alterations in hippocampal neurogenesis.

2.2 Surgery and CCI

Animals underwent an experimental TBI using a controlled cortical impactor (Pittsburgh Precision Instruments, Inc, USA) as described previously (Yu et al., 2009). Animals initially received Buprenorphine (0.05 mg/kg, s.c.) at the time of anesthesia induction (with 125 mg/kg Ketamine, 12.5 mg/kg Xylazine). Once deep anesthesia was achieved (by checking for pain reflexes), individual animals were fixed in a stereotaxic frame (David Kopf Instruments, Tujunga, CA, USA). After exposing the skull, craniotomy (size = 3 mm) to accommodate the impactor tip was performed over the right frontoparietal cortex (– 0.5 mm anteroposterior and +0.5 mm mediolateral to bregma). The pneumatically operated TBI device (with a convex tip diameter = 2 mm) impacts the brain at a velocity of 6.0 m/s reaching a depth of 0.5 mm, 1.0 mm or 2.0 mm for mild, moderate and severe TBI respectively, below the dura mater layer and remains in the brain for 150 ms. For the purposes of the present study, a mild TBI was induced. The impactor rod was angled 15° to the vertical to maintain a perpendicular position in reference to the tangential plane of the brain curvature at the impact surface. A linear variable displacement transducer (Macrosensors, Pennsauken, NJ), connected to the impactor, measured velocity and duration to verify consistency. Bone wax was used to cover the craniectomized region and the skin incision sutured thereafter. A computer operated thermal blanket pad and a rectal thermometer allowed maintenance of body temperature within normal limits. All animals were closely monitored until recovery from anesthesia and over the next 3 consecutive days.

2.3 Drugs

Human recombinant G-CSF (Neupogen) was purchased from Amgen, Inc. (Thousand Oaks, CA). The Neupogen was received in preservative-free vials containing 300 μg/mL. It was diluted to the appropriate concentrations in sterile 5% dextrose solution and injected at a dose of 100 μg/kg sub-cutaneously (s.c.). Alow dose of G-CSF (100 μg/kg) was chosen for administration in the 3 days after injury because earlier studies revealed that a single (300 μg/kg) administered intravenously 1 wk after CCI improved motor performance when tested 8 weeks later (Acosta et al., 2014). The dose of G-CSF (100 μg/kg×3 doses) administered in the present study was lower than that reported to be effective in mobilizing bone marrow in a rat model of stroke (300 μg/kg×10 days) but higher than that utilized by others in rodent models of stroke (Schabitz et al., 2003; Six et al., 2003; Solaroglu, Cahill, et al., 2006) and in a recently published report on the treatment of memory impairment in a mouse model of AD (50 microg/kg s.c. for 5 days) (Tsai et al., 2007).

2.4 Radial arm water maze (RAWM)

To study the cognitive effects of G-CSF in mice that had undergone mild to moderate CCI, a radial arm water maze (RAWM) task was employed. RAWM is a hippocampal- dependent, spatial learning task that does not rely on locomotor ability or swimming speed (Vorhees & Williams, 2006). RAWM was conducted at baseline and again on days 5 to 7 and days 12 to 14 post CCI. A six- arm radial arm maze was placed into a water tank of approximately 100 cm- diameter and a 25-cm-height, 5-cm-diameter platform was used. The platform was submerged 0.5 cm below the water surface. The temperature of the water was kept at 26°C. Mice were placed in the start arm at the beginning of every trial, and the platform was located in the goal arm. Every animal had an assigned platform/arm location throughout acquisition of learning, yet the starting zone was randomly changed per trial. A spatial-training protocol was followed. Mice were given 2 sets of 5 trials each separated by 30 minutes rest period per day, for a total of 10 trials a day for 2 days of acquisition of learning for both baseline and post-TBI training. Trials were only 60 seconds long, once animals found their goal arm/platform; they were allowed to remain on the platform for 30 sec between trials. If mice were unable to find their goal arm/platform within 60 seconds, mice were guided to their goal arm and allowed to rest on the platform for 30 sec. On day 3, a probe trial was given; this was reversal training in which the mice were placed 180 degrees from the goal arm. Mice were giving 5 trials to train for the new position (reversal training). RAMW performance analysis was done by averaging the 2 trials per block, for a total of 5 blocks per day. Reversal training was analyzed by counting the total of errors in each trial.

2.5 Motor, balance and coordination measured with rotarod performance

The rotarod (Ugo Basile S.R.L; 21025 Comerio VA Italy, Madel 47600, Rotarod for Mice) provided a motor balance and coordination assessment. Data were generated by averaging the scores (total time spent on treadmill divided by 3 trials) for each animal during training and testing days. Each animal was placed in a neutral position on a cylinder, then the rod was rotated with the speed accelerated linearly from 4 rpm to 40 rpm within 3 minutes, and the time spent on the rotarod was recorded automatically. For training, animals were given 1 trial before testing. For testing, animals were given 3 trials and the average score on these 3 trials was used as the individual rotarod score.

2.6 Immunohistochemistry

Mice were anesthetized with 150 mg/kg Ketamine, 15 mg/kg Xylazine and then transcardially perfused with 0.9% saline followed by 4% paraformaldehyde. Brains were stored in 4% paraformaldehyde, transferred to 25% sucrose solution in 4% paraformaldehyde, until the brains sank to the bottom. Then brains were slowly immersed into isopentane (cooled on dry-ice), left in isopentane for 20 seconds, removed, placed on a small piece of aluminum foil sitting on powdered dry-ice for 1-2 minutes (to let isopentane evaporate) and finally wrapped in the foil and stored at – 80°C until sectioning. Brains slices were cut 30 μm thick, in a cryostat (Leica, Germany) set to – 25°C. Every 6th coronal section was taken from the corpus striatum (caudate/putamen) spanning 1.2 mm in the anterior-posterior direction (from Bregma +1.32 mm to Bregma = 0 which corresponds to the beginning of the lateral ventricles to the anterior commissure). Serial sections were also cut from hippocampus, starting from Bregma – 1.28 to Bregma – 2.92. Every 6th section was kept for immunostaining.

Selective immunostaining of astrocytes and microglia was performed with antibodies to glial fibrillary acidic protein (GFAP) and (ionized calcium-binding adapter molecule-1 (Iba-1), respectively. Iba-1 is protein that is specifically expressed in macrophages/microglia and is upregulated during the activation of these cells. Antibodies to doublecortin (DCX) were used to label immature neurons in the dentate gyrus of the hippocampus. Brain sections were preincubated in PBS containing 10% normal serum (goat or donkey; Vector) and 0.3% Triton X-100 (Sigma) for 30 min. The sections were then transferred to a solution containing primary antibodies in 1% normal serum, 0.3% triton X-100/PBS and incubated overnight at 4°C. The specific antibodies used in each experiment were: rabbit anti-DCX (Abcam Inc.), 1:1000; rabbit anti-Iba1 (Wako Chemicals USA, Inc.), 1:500; rabbit anti-GFAP (BioGenex), 1:50 in PBS containing 1:100 normal serum without Triton X-100. After incubation with primary antibody, the sections were washed and incubated for 1 hour with Alexa Fluor 568 goat anti-rabbit IgG diluted 1:400 in PBS (Invitrogen) at room temperature. The sections were then rinsed in PBS three times and covered with a cover-glass. Green fluorescence signal (expressed by bone marrow-derived GFP+ cells) was visualized with fluorescence microscopy directly with the green band filter, with no need to stain with antibodies directed against mouse GFP.

2.7 Quantitative assessment of astrocytic, microglia responses and hippocampal neurogenesis

Quantitation of microgliosis and astrogliosis was made by computerized image analysis. The method for quantitative image analyses has been previously described and used by members of our lab (Boyd et al., 2010; Sanchez-Ramos et al., 2009). A total of 12 mice (6 vehicle and 6 GCSF-treated mice) were analyzed at 14 days after CCI. Images were acquired at a magnification of 200x as digitized tagged-image format files (TIFF) to retain maximum resolution using an Olympus BX60 microscope with an attached digital camera system (DP-70, Olympus, Tokyo Japan). Images of 8 sections (each 30 microns thick and 180 microns apart) were captured from serially sectioned striatum (approximately 1.2 mm in the AP dimension, starting at beginning of the lateral ventricles to the anterior commissure) on both left and right side from each animal. Using Image J software (NIH), the green channel was selected and converted into a monochrome signal to isolate the GFP+ signal of bone-marrow derived cells. Similarly, the red channel was selected and converted into monochrome to isolate the red fluorescence signal of the immunostained cells tagged with Alexa Fluor 568 anti-rabbit IgG. Then, a threshold optical density was obtained that discriminated staining from background. Each anatomic region of interest was manually edited to eliminate artifacts. The thresholded signal was automatically generated by Image J and expressed as the area of the threholded pixels divided by the total number of pixels encompassed by the 20x objective microscopic visual field (% area). Bias was eliminated by having the image analysis done by a blinded researcher.

In addition to estimating DCX expression in the dentate gyrus using the method above, an unbiased estimate of the number of DCX cells in DG was performed on coded sections as previously described (Catlow, Song, Paredes, Kirstein, & Sanchez-Ramos, 2013; Shors et al., 2001). Briefly, positively labeled cells were counted in every 6th section (each section separated by 180 μm) using a modification to the optical dissector method; cells on the upper and lower planes were not counted to avoid counting partial cells. Resulting numbers were tallied and multiplied by tissue thickness (30 μm) and number of intervening sections (n = 6).

2.8 BDNF and GDNF ELISA

Sub-sets of chimeric mice were euthanized at 3, 7 and 14 days after CCI for analysis of two neurotrophic factors, BDNF and GDNF. At each time point, the cohort included 4 vehicle and 4 GCSF-treated mice. Each brain was bisected in the sagittal plane and 3 brain regions (cortex, striatum and hippocampus) from each side were dissected for analysis. The brain tissue samples were homogenized in T-PER Tissue Protein Reagent (PI-#78510) with Protease and Phosphatase inhibitor cocktail (PI-78443) and each brain sample’s protein concentration was measured by a BCA kit (Fisher Scientific #23225). Levels of Mouse BDNF and GDNF were measured using a Mouse BDNF and GDNF ELISA Kit from Boster Biological Technology (Fremont, CA. Cat # EK0309 and # EK0935) and followed the Kit instruction and Preparation. The ELISA 96-well plate was read at a Plate Reader (BioTek Synergy) at O.D absorbance at 450 nm. The Cytokine concentration was expressed at Unit pg/ml with sample total protein concentration at 1 mg/ml.

2.9 Data analysis and statistics

Neurohistologic measures, as well as measured of neurotrophic factors and cytokines were expressed as mean ± SEM and statistically evaluated using. Neurohistologic measures, as well as measures of neurotrophic factors were expressed as mean ± SEM and statistically evaluated using 2- way ANOVA and multiple t-tests (comparing vehicle vs G-CSF) with the Holm-Sidak correction for multiple comparisons (GraphPad version 5.01). Analysis of the RAWM data utilized the repeated measures ANOVA, followed by post-hoc analysis Bonferroni test. In order for the reader to appreciate the significant differences during the probe trial test, they histogram was plotted as bar graphs keeping the same statistical results from the previous ANOVA analysis. All comparisons were considered significant at p < 0.05.

3 Results

The effects of CCI, followed by 3 days of G-CSF (100 μg/kg s.c. daily) treatment on motor performance on the rotarod, is shown in Table 1. Performance of control animals (without bone marrow transplant and without CCI) revealed no significant difference in latency between groups treated with vehicle or G-CSF for 3 days. Baseline performance of chimeric mice was less (shorter latency to fall) than in baseline performance of non-chimeric mice. The performance of G-CSF treated mice was improved 1 week after CCI compared to vehicle treated mice. At two weeks, the G-CSF treated group continued to improve compared to vehicle controls, but the difference did not reach statistical significance.

G-CSF-treated mice also performed better in the in a hippocampal-dependent learning task, the RAWM, compared to vehicle-treated mice at both one and two weeks after CCI (Fig. 1). During “probe trials”, in which the location of the platform was reversed, the G-CSF treated mice committed significantly less errors than vehicle-treated mice, both at 7 and 14 days after CCI.

Since TBI triggers cellular and inflammatory responses in the sub-acute period, it was important to determine the extent to which G-CSF treatment modulated these cellular repair responses. Both vehicle-treated and G-CSF treated animals experienced significant astrocytosis and microgliosis in multiple brain regions (cortex, corpus callosum, striatum, hippocampus) on both sides of brain. Quantitative analysis of the cellular response was focused on striatum and hippocampus on both right and left sides. The striatum was chosen because of its important role in the automatic execution of learned motor programs (such as required for automatic running on a rotating cylinder) and the hippocampus for its role in spatial learning and memory, as in the RAWM.

Euthanasia was performed 2 wks after CCI for immunohistologic analysis. Coronal brain sections were cut through cortex and striatum and were immunostained for visualization of microglial cells (Iba1+) astrocytes (glial fibrillary acidic protein–GFAP+) and bone marrow-derived cells (GFP+). Total microglial, astroglial and BM-derived GFP+ cell burdens were determined utilizing quantitative image analysis (see General Methods Section for details).

G-CSF treatment increased astrocytosis to a greater extent than vehicle treatment (Fig. 2). The GFAP+ cells were greatest at the cortical site of injury, and the signal extended along the corpus callosum to the contralateral side. GFP+ signal (indicative of blood-derived cells) was also present at the cortical site of injury. The extent of astrocytosis (GFAP+ signal) was much greater than the infiltration of BM-derived GFP+ cells. Quantitation of astrocytosis in the striatum revealed that both sides of brain exhibited an increased GFAP signal, a marker of astrocytosis (Fig. 3) but the extent of infiltration of GFP+ cells was much less in the contraleral striatum.

Microglia, indicated by Iba1 immunoreactivity, and GFP+ cells in striatum were also increased on both sides of brain (See Fig. 4). Quantitative analysis revealed that G-CSF treatment increased both GFP+ cell signal and total microglial (Iba1+) signal to a much greater extent on the side of CCI. In addition, it was clear that the total microglial burden was increased on both sides of brain after CCI even when G-CSF was not given.

Many of the Iba1+ cells co-expressed GFP (see Fig. 5). The insert boxes (Fig. 5A– C) show several double-labeled Iba1/GFP+ cells. A mean of 3.46% of cells in striatum from G-CSF treated mice exhibited overlap of GFP signal with the Iba1 signal on the side of the injury (Fig. 5H). Of the total bone marrow-derived GFP+ signal (9.6%), approximately 36% (3.46 ÷ 9.6) co-localized with Iba1 signal suggesting that a significant fraction of BM-derived cells differentiate into microglia. (See Methods, for details on measurements using image analysis). On the contralateral striatum, the overlapping of Iba-1+ and GFP+ signals was less than 0.5%.

To determine if G-CSF treatment increased the generation of new neurons in the SGZ of the dentate gyrus, DCX immunostaining was performed. Although DCX+ cells don’t indicate rate of neurogenesis, DCX is a marker of immature neurons, and has been used as an index of neurogenesis (Kempermann, 2006). The extent of DCX expression was measured by image analysis of the DCX signal, which includes both cell body and immunoreactive neurites. To complement this analysis, an unbiased estimate of numbers of DCX cells was determined (See Fig. 6). G-CSF treatment resulted in significantly increased numbers of DCX cells in the neurogenic niche on the right side only. This finding was replicated by image analysis of DCX signal area.

The effects of G-CSF treatment following CCI on the expression of two neurotrophic factors, BDNF and GDNF, is shown in Tables 2 and 3. G-CSF treatment resulted in signficantly increased levels of expression of BDNF in cortex at 3, 7 and 14 days after CCI on the right side of brain (Table 2). On side contralateral to the CCI, BDNF levels were increased in cortex on days 7 and 14. The other two brain regions did not exhibit this increase. In contrast, GDNF was significantly increased in striatum on days 3 and 7 on the right side only. In hippocampus, G-CSF treatment resulted in increased GDNF levels on day 14 on the right side. Interestingly, hippocampal GDNF was also increased significantly by GCSF treatment on the contralateral side on day 3.

4 Discussion

The findings presented here confirm and extend prior observations on the beneficial effects of G-CSF treatment in rodent models of TBI. Earlier studies had shown that acute intraperitoneal treatment with G-CSF (immediately after the injury) or sub-acute intravenous treatment one week after injury resulted in beneficial effects in rat models of TBI (Acosta et al., 2014; Yang et al., 2010), including improvement in performance in the Radial Arm Water Maze (RAWM) within a week of closed traumatic brain injury (Sikoglu et al., 2014). The salient findings reported in the present study include 1) the promotion of recovery in a hippocampal-dependent learning task (RAWM), 2) activation, mobilization of glial cells, 3) increased production of neurotrophic factors BDNF and GDNF, and 4) increased area of DCX+ signal (marker of new neurons) in the dentate gyrus of the hippocampus. G-CSF treatment increased astrocytosis on both sides of brain, with the side of injury showing the greatest increase. G-CSF treatment also increased the extent of infiltration of GFP+ bone marrow-derived cells (BMDC). Approximately one third of the microglial signal (Iba1) overlapped with the GFP+ signal in the striatum on the side of the lesion by day 14 after CCI.

Although increasing infiltration of BMDC into brain may theoretically worsen the injury, experimental evidence suggests that increasing the abundance of BMDCs by direct intracerebral grafting (or by intravenous administration) enhances recovery of neurologic deficits in rodent models of brain trauma (Mahmood, Lu, & Chopp, 2004; Mahmood, Lu, Qu, Goussev, & Chopp, 2006). The mechanisms for enhanced recovery is complex. Both humeral and cellular factors appear to be involved. The BMSC-secreted BDNF and VEGF may promote neurogenesis and angiogenesis, respectively. Brain natriuretic peptide (BNP) secreted by BMDC is similar to atrial natriuretic peptide (ANP) which has been reported to reduce vasogenic edema after sub-arachnoid hemorrhage (Doczi, Joo, & Balas, 1995). Moreover, BMDC may have other beneficial effects that includes rescue of injured neurons by fusion, though this is an infrequent phenomenon (Alvarez-Dolado et al., 2003; Terada et al., 2002; Weimann, Charlton, Brazelton, Hackman, & Blau, 2003). In addition, G-CSF appears to be a powerful immunodulator which increases expression of anti-inflammatory cytokines while dampening expression of pro-inflammatory cytokines (Hartung, 1998).

Another important property of G-CSF relates to its direct actions on neural cells to influence the repair response to injury. Direct interaction of G-CSF with its cognate receptor expressed on neural stem/progenitor cells results in proliferation and differentiation of neural progenitor cells (Schneider, Kruger, et al., 2005; Schneider, Kuhn, & Schabitz, 2005). Increased neurogenesis triggered by G-CSF has been previously reported in other disease models and was replicated in the present study of TBI (Acosta et al., 2014; Sanchez-Ramos et al., 2009; Schneider, Kruger, et al., 2005; Song et al., 2013; Yang et al., 2010). In the present study, G-CSF treatment increased the CCI-triggered increase in DCX signal in dentate gyrus on the side of injury but not the contralateral side. This finding was replicated by non-biased counting of DCX+ cells of the dentate gyrus. Although DCX+ cells don’t indicate rate of neurogenesis, DCX is a marker of immature neurons, and has been used as an index of neurogenesis (Kempermann, 2006). G-CSF treatment resulted in significantly increased numbers of DCX cells in the neurogenic niche on the right side consistent with the DCX signal area analysis. It should be noted that increased number of DCX cells does not necessarily indicate increased proliferation of neural stem/progenitor cells. Anti-apoptotic effects of G-CSF (Solaroglu, Tsubokawa, Cahill, & Zhang, 2006) may enhance survival of new-born neurons.

Others have explored the functional significance of alterations in hippocampal neurogenesis. For example, it is known that chemical ablation of neurogenesis impairs acquisition of hippocampal-dependent tasks (Shors et al., 2001; Shors, Townsend, Zhao, Kozorovitskiy, & Gould, 2002), and so it is possible that increased neurogenesis triggered by direct actions of G-CSF on neural progenitor cells was responsible for enhanced recovery of performance in the RAWM. However, the improvement in RAWM performance was evident as early as 1 week after CCI (4 days after the last dose of G-CSF administration), much sooner than the amount of time required for maturation of neural progenitors and integration of new neurons into hippocampal neural circuitry. Indeed, it takes two to four weeks for new neurons to mature and become fully integrated into hippocampal circuity (Cheng et al., 2011). The majority (71%) of birth-dated neurons express DCX by 14 days and action potentials are easily induced in these birth-dated cells, with a lower threshold current at 14 days compared to neurons at age 28 days (Cheng et al., 2011).

Another possible explanation for the benefits of G-CSF may be attributed to functional effects on hippocampal physiological activity. In other words, G-CSF can impact electrophysiological correlates of hippocampal neuroplasticity by altering synaptic activity. For example, a course of G-CSF treatment was reported to restore long-term depression (LTD) in ex vivo hippocampal slices in a mouse model of Alzheimer’s disease in which LTD was impaired by the disease process (Song, Wang, Sava, Weeber, & Sanchez-Ramos, 2014). In addition, G-CSF has been reported to increase synaptic efficacy in neonatal rat hippocampal slices, evidenced by reversal of perinatal hypoxia-induced decrease in magnitude of LTP (long-term potentiation) (Chen et al., 2011). Finally, it is known that absence of G-CSF (in a transgenic G-CSF knockout) results in disruption of memory formation and deficits in development of motor skills (Diederich et al., 2009). These alterations were accompanied by impaired induction of LTP in the CA1 region of hippocampus. Moreover, G-CSF deficiency led to decreased dendritic complexity in hippocampal neurons in the dentate gyrus and the CA1 region. So it is reasonable to infer that G-CSF treatment following CCI may result in the opposite effect, namely, increased dendritic complexity and connectivity.

Concentrations of the neurotrophic factors BDNF and GDNF were found to be increased in frontal cortex and striatum, respectively by G-CSF treatment. GDNF was also found to be increased in hippocampus on day 14 at the side of injury. This response is to be expected because these neurotrophic factors are elaborated by astrocytes and activated microglia (Batchelor et al., 1999). It is known that GDNF and BDNF stimulate hippocampal neurogenesis and possess anti-apoptotic properties (Boku et al., 2013; Chen, Ai, Slevin, Maley, & Gash, 2005). So the direct actions of G-CSF to stimulate neurogenesis, and its indirect actions to enhance BDNF and GDNF production by glial cells may have resulted in an augmented neurogenic response to injury which translated into significant recovery from the TBI-induced deficits in the RAWM.

Going forward, it will be important to further investigate the role of blood- borne monocytes in mediating the beneficial effects of G-CSF. One could test this hypothesis by using transgenic mice that have no receptor for monocyte chemotactic factor-1(MCP-1) and thereby do not respond to homing signals from the MCP-1 released at the site of injury. As an example, a recent study in a mouse model of TBI demonstrated that transgenic mice without the MCP-1 responded no differently than wild-type mice in terms of lesion size and cell death within the first week of injury (Semple, Bye, Rancan, Ziebell, & Morganti-Kossmann, 2010). In contrast, by 2 and 4 weeks, a delayed reduction in lesion volume, macrophage accumulation, and astrogliosis were observed in the injured cortex and ipsilateral thalamus of the tg MCP-1 deficient mice, corresponding to improved functional recovery as compared with wild-type mice (Semple et al., 2010). This study lends credence to the counter hypothesis that diminishment of recruitment of bone-marrow derived cells is likely to improve recovery from TBI. Even if the increased microgliosis stimulated by G-CSF treatment may slow recovery, the G-CSF effect on RAWM was definitely beneficial. The analysis remains complicated because there are direct actions of G-CSF on the hippocampal neural circuitry that mediates learning, at least as evidenced in an electrophysiological model of learning. For example, G-CSF treatment of a transgenic mouse model of Alzheimer’s disease (AD), followed by electrophysiological study of hippocampal slices revealed restoration of long term depression that was absent in the tg AD mice (Song et al., 2014).

In summary, G-CSF treated animals recovered faster than vehicle-treated mice from a spatial learning deficit that was produced by mild CCI. Enhanced recovery in the hippocampal-dependent learning task may be attributed to modulation by G-CSF of the sub-acute neuro-inflammatory response and direct actions on neural cells that result in increased hippocampal neurogenesis. Other direct effects of G-CSF on hippocampal neural circuitry, described in an earlier report (Song et al., 2014), may also play a role in the enhancement of spatial learning provided by G-CSF treatment immediately after TBI.

Conflict of interest

None of the authors have financial, personal or other relationships with organizations or other individuals that could inappropriately influence, or be perceived to influence, this submitted work.

Footnotes

Acknowledgments

Supported by Veterans Administration Merit Review Grant to S. Song and by the Helen Ellis Research Endowment Funds to J. Sanchez-Ramos.

The contents of this publication do not represent the views of the Department of Veterans Affairs or the United States Government.

References

Acosta

S.A.

, Tajiri

, Shinozuka

, Ishikawa

, Sanberg

P.R.

, Sanchez-Ramos

, Song

, Kaneko

, & Borlongan

C.V.

(2014). Combination therapy of human umbilical cord blood cells and granulocyte colony stimulating factor reduces histopathological and motor impairments in an experimental model of chronic traumatic brain injury. Public Library of Science One (PLoS One), 9(3), e90953.

Alvarez-Dolado

, Pardal

, Garcia-Verdugo

J.M.

, Fike

J.R.

, Lee

H.O.

, Pfeffer

, & Alvarez-Buylla

(2003). Fusion of bone-marrow-derived cells with Purkinje neurons, cardiomyocytes and hepatocytes. Nature, 425(6961), 968–973.

Bakhtiary

, Marzban

, Mehdizadeh

, Joghataei

M.T.

, Khoei

, Pirhajati Mahabadi

, Laribi

, & Moshkforoush

(2010). Comparison of transplantation of bone marrow stromal cells (BMSC) and stem cell mobilization by granulocyte colony stimulating factor after traumatic brain injury in rat. Iran Biomed Journal, 14(4), 142–149.

Batchelor

Peter E.

, Liberatore

Gabriel T.

, Wong

John Y.F.

, Porritt

Michelle J.

, Frerichs

Fenneke

, Donnan

Geoffrey A.

, & Howells

David W.

(1999). Activated macrophages and microglia induce dopaminergic sprouting in the injured striatum and express brain-derived neurotrophic factor and glial cell line-derived neurotrophic factor. The Journal of Neuroscience, 19(5), 1708–1716.

Boku

, Nakagawa

, Takamura

, Kato

, Takebayashi

, Hisaoka-Nakashima

, Omiya

, Inoue

, & Kusumi

(2013). GDNF facilitates differentiation of the adult dentate gyrus-derived neural precursor cells into astrocytes via STAT3. Biochem Biophysical Research Communications, 434(4), 779–784.

Boyd

T.D.

, Bennett

S.P.

, Mori

, Governatori

, Runfeldt

, Norden

, Padmanabhan

, Neame

, Wefes

, Sanchez-Ramos

, Arendash

G.W.

, & Potter

(2010). GM-CSF upregulated in rheumatoid arthritis reverses cognitive impairment and amyloidosis in Alzheimer mice. Journal of Alzheimers Disease, 21(2), 507–518.

Catlow

B.J.

, Song

, Paredes

D.A.

, Kirstein

C.L.

, & Sanchez-Ramos

(2013). Effects of psilocybin on hippocampal neurogenesis and extinction of trace fear conditioning. Experimental Brain Research, 228(4), 481–491.

Cernak

(2005). Animal models of head trauma. NeuroRx, 2(3), 410–422.

Chen

W.-F.

, Hsu

J.-H.

, Lin

C.-S.

, Jong

Y.-J.

, Yang

C.-H.

, Huang

L.-T.

, & Yang

S.-N.

(2011). Granulocyte-colony stimulating factor alleviates perinatal hypoxia-induced decreases in hippocampal synaptic efficacy and neurogenesis in the neonatal rat brain. Pediatric Research, 70(6), 589–595.

10.

Chen

, Ai

, Slevin

J.R.

, Maley

B.E.

, & Gash

D.M.

(2005). Progenitor proliferation in the adult hippocampus and substantia nigra induced by glial cell line-derived neurotrophic factor. Experimental Neurology, 196(1), 87–95.

11.

Cheng

, Li

, Huang

, Feng

, & Xiong

Z.-Q.

(2011). Pulse labeling and long-term tracing of newborn neurons in the adult subgranular zone. Cell Research, 21(2), 338–349.

12.

Diederich

, Sevimli

, Dörr

, Kösters

, Hoppen

, Lewejohann

, Klocke

, Minnerup

, Knecht

, Nikol

, Sachser

, Schneider

, Gorji

, Sommer

, & Schäbitz

(2009). The role of granulocyte-colony stimulating factor (G-CSF) in the healthy brain: A characterization of G-CSF-deficient mice. The Journal of Neuroscience, 29(37), 11572–11581.

13.

Doczi

T.P.

, Joo

, & Balas

(1995). Atrial natriuretic peptide (ANP) attenuates brain oedema accompanying experimental subarachnoid haemorrhage. Acta Neurochirurgica (Wien), 132(1-3), 87–91.

14.

Hartung

(1998). Anti-inflammatory effects of granulocyte colony-stimulating factor. Current Opinions in Hematology, 5(3), 221–225.

15.

Jung

K.H.

, Chu

, Lee

S.T.

, Kim

S.J.

, Sinn

D.I.

, Kim

S.U.

, Sinn

D.I.

, Kim

S.U.

, Kim

, & Roh

J.K.

(2006). Granulocyte colony-stimulating factor stimulates neurogenesis via vascular endothelial growth factor with STAT activation, 190–201.

16.

Kempermann

(2006). Adult Neurogenesis: Stem Cells and Neuronal Development in the Adult Brain: Oxford University Press, Inc..

17.

Mahmood

, Lu

, & Chopp

(2004). Marrow stromal cell transplantation after traumatic brain injury promotes cellular proliferation within the brain. Neurosurgery, 55(5), 1185–1193.

18.

Mahmood

, Lu

, Qu

, Goussev

, & Chopp

(2006). Long-term recovery after bone marrow stromal cell treatment of traumatic brain injury in rats. Journal of Neurosurgery, 104(2), 272–277.

19.

Ringelstein

E.B.

, Thijs

, Norrving

, Chamorro

, Aichner

, Grond

, Saver

, Laage

, Schneider

, Rathgeb

, Vogt

, Charisse

, Riebach

J.B.

, Schwab

, Schabitz

, Kollmar

, Fisher

, Brozman

, Skoloudik

, Gruber

, Serena

J.L.

, Veltkamp

, Kohrmann

, & Berrouschot

(2013). Granulocyte colony– stimulating factor in patients with acute ischemic stroke: Results of the AX200 for ischemic stroke trial. Stroke, 44(10), 2681–2687.

20.

Sanchez-Ramos

, Song

, Sava

, Catlow

, Lin

, Mori

, Cao

, & Arendash

G.W.

(2009). Granulocyte colony stimulating factor decreases brain amyloid burden and reverses cognitive impairment in Alzheimer’s mice. Neuroscience, 163(1), 55–72.

21.

Schabitz

W.R.

, Kollmar

, Schwaninger

, Juettler

, Bardutzky

, Scholzke

M.N.

, Sommer

, & Schwab

(2003). Neuroprotective effect of granulocyte colony-stimulating factor after focal cerebral ischemia. Stroke, 34(3), 745–751.

22.

Schneider

, Kruger

, Steigleder

, Weber

, Pitzer

, Laage

, Aronowski

, Maurer

M.H.

, Gassler

, Mier

, Hasselblatt

, Kollmar

, Schwab

, Sommer

, Bach

, Kuhn

G.G.

, & Schabitz

W.R.

(2005). The hematopoietic factor G-CSF is a neuronal ligand that counteracts programmed cell death and drives neurogenesis. Journal of Clinical Investigation, 115(8), 2083–2098.

23.

Schneider

, Kuhn

H.G.

, & Schabitz

W.R.

(2005). A role for G-CSF (granulocyte-colony stimulating factor) in the central nervous system. Cell Cycle, 4(12), 1753–1757.

24.

Semple

Bridgette D.

, Bye

Nicole

, Rancan

Mario

, Ziebell

Jenna M.

, Morganti-Kossmann

, Cristina (2010). Role of CCL2 (MCP-1) in traumatic brain injury (TBI): Evidence from severe TBI patients and CCL2–/– mice. Journal of Cerebral Blood Flow and Metabolism: Official Journal of the International Society of Cerebral Blood Flow and Metabolism, 30(4), 769–782.

25.

Shors

T.J.

, Miesegaes

, Beylin

, Zhao

, Rydel

, & Gould

(2001). Neurogenesis in the adult is involved in the formation of trace memories. Nature, 410(6826), 372–376.

26.

Shors

T.J.

, Townsend

D.A.

, Zhao

, Kozorovitskiy

, & Gould

(2002). Neurogenesis may relate to some but not all types of hippocampal-dependent learning. Hippocampus, 12(5), 578–584.

27.

Shyu

W.C.

, Lin

S.Z.

, Lee

C.C.

, Liu

D.D.

, & Li

(2006). Granulocyte colony-stimulating factor for acute ischemic stroke: A randomized controlled trial. Canadian Medical Association Journal, 174(7), 927–933.

28.

Sikoglu

E.M.

, Heffernan

M.E.

, Tam

, Sicard

K.M.

, Bratane

B.T.

, Quan

, Fisher

, & King

J.A.

(2014). Enhancement in cognitive function recovery by granulocyte-colony stimulating factor in a rodent model of traumatic brain injury. Behavioral Brain Research, 259, 354–356.

29.

Six

, Gasan

, Mura

, & Bordet

(2003). Beneficial effect of pharmacological mobilization of bone marrow in experimental cerebral ischemia. European Journal of Pharmacology, 458(3), 327–328.

30.

Solaroglu

, Cahill

, Jadhav

, & Zhang

J.H.

(2006). A novel neuroprotectant granulocyte-colony stimulating factor. Stroke, 37(4), 1123–1128.

31.

Solaroglu

, Tsubokawa

, Cahill

, & Zhang

J.H.

(2006). Anti-apoptotic effect of granulocyte-colony stimulating factor after focal cerebral ischemia in the rat. Neuroscience, 143(4), 965–974.

32.

Song

, & Sanchez-Ramos

(2002). Preparation of Neural Progenitors from Bone Marrow and Umbilical Cord Blood Human Press, In Zigova

Sanberg

P.R.

& Sanchez-Ramos

, Protocols for Neural Stem Cell Methods, 198, Totowa, NJ, pp. 79–88.

33.

Song

, Song

, Cao

, Lin

, Li

, Sava

, & Sanchez-Ramos

(2013). Hippocampal neurogenesis and the brain repair response to brief stereotaxic insertion of a microneedle. Stem Cells International, 2013, 205878.

34.

Song

, Wang

, Sava

, Weeber

E.J.

, & Sanchez-Ramos

(2014). In vivo administration of granulocyte colony-stimulating factor restores long-term depression in hippocampal slices prepared from transgenic APP/PS1 mice. Journal of Neuroscience Research, 92(8), 975–980.

35.

Sprigg

, Bath

P.M.

, Zhao

, Willmot

M.R.

, Gray

L.J.

, Walker

M.F.

, Dennis

M.S.

, & Russell

(2006). Granulocyte-colony - Stimulating factor mobilizes bone marrow stem cells in patients with subacute ischemic stroke: The Stem cell Trial of recovery EnhanceMent after Stroke (STEMS) pilot randomized, controlled trial (ISRCTN. Stroke, 37(12), 2979–2983.

36.

Terada

, Hamazaki

, Oka

, Hoki

, Mastalerz

D.M.

, Nakano

, Meyer

E.M.

, Morel

, Petersen

B.E.

, & Scott

E.W.

(2002). Bone marrow cells adopt the phenotype of other cells by spontaneous cell fusion. Nature, 416, 542–545.

37.

Tsai

K.J.

, Tsai

Y.C.

, & Shen

C.K.

(2007). G-CSF rescues the memory impairment of animal models of Alzheimer’s disease. Journal of Experimental Medicine, 204(6), 1273–1280.

38.

Vorhees

C.V.

, & Williams

M.T.

(2006). Morris water maze: Procedures for assessing spatial and related forms of learning and memory. Nat Protoc, 1(2), 848–858.

39.

Weimann

J.M.

, Charlton

C.A.

, Brazelton

T.R.

, Hackman

R.C.

, & Blau

H.M.

(2003). Contribution of transplanted bone marrow cells to Purkinje neurons in human adult brains. Proceedings of the National Academy of Sciences U S A, 100(4), 2088–2093.

40.

Yang

D.Y.

, Chen

Y.J.

, Wang

M.F.

, Pan

H.C.

, Chen

S.Y.

, & Cheng

F.C.

(2010). Granulocyte colony-stimulating factor enhances cellular proliferation and motor function recovery on rats subjected to traumatic brain injury. Neurological Research, 32(10), 1041–1049.

41.

, Kaneko

, Bae

, Stahl

C.E.

, Wang

, van Loveren

, Sanberg

P.R.

, & Borlongan

C.V.

(2009). Severity of controlled cortical impact traumatic brain injury in rats and mice dictates degree of behavioral deficits. Brain Research, 1287, 157–163.