Increased expression of the chemokines CXCL1 and MIP-1α by resident brain cells precedes neutrophil infiltration in the brain following prolonged soman-induced status epilepticus in rats

This model has been described previously [11, 25]. Briefly, adult male Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA; CRL: CD[SD]-BR, 250 - 350 g) were treated with HI-6 dichloride (1-(((4-(aminocarbonyl)pyridinio) methoxy)methyl)-2-((hydroxyimino)methyl)pyridinium dichloride)(BN44621, Starks Associates, Buffalo, NY; 125 mg/kg, i.p.) 30 minutes prior to GD administration and with atropine methyl nitrate (AMN, Sigma-Aldrich, St. Louis, MO; 2.0 mg/kg, i.m.) 1 minute after GD administration. Vehicle control animals received HI-6, AMN and saline, while naïve animals received no injections. GD (GD-U-2323-CTF-N, purity 98.8 wt%) was diluted in saline at the United States Army Medical Research Institute of Chemical Defense (USAMRICD) and administered subcutaneously (1.6 LD₅₀ = 180 μg/kg). The experimental protocol was approved by the Animal Care and Use Committee at USAMRICD, and all procedures were conducted in accordance with the principles stated in the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996), and the Animal Welfare Act of 1966 (P.L. 89-544), as amended. The animal care program at this institute is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International.

As previously described [11], piriform cortex, hippocampus and thalamus brain tissue samples were procured from experimental and vehicle control animals at 0.5, 1, 3, 6, 12, 24, 48 or 72 hours after onset of convulsions. Tissue lysates were produced by first rinsing the excised tissue with cold PBS followed by snap freezing in liquid nitrogen. A ratio of 1 ml ice-cold triple detergent lysis buffer containing a Complete™ protease inhibitor cocktail (Roche Biochemicals, Indianapolis, IN) to 50 mg of frozen tissue was used for homogenization. Two 30 sec pulses on a mini Beadbeater (Biospec Products Inc., Bartlesville, OK) using 3.2 mm stainless steel beads were used to homogenize the tissue. Samples were centrifuged at 8000 × g for 5 minutes to separate the lysate from the tissue pellet. Rat cytokine multiplex bead immunoassay kits were used to quantify the concentrations of CXCL1 (GRO KC), MIP-1α, G-CSF and GM-CSF (LINCO Research, St. Charles, MO). Individual standard curves were generated in duplicate using the supplied reference chemokine concentrations according to the manufacturer's instructions. A volume of 25 μl of sample (94 ± 8 μg protein) per well, assayed in duplicate, was used for data generation. The plate was read on a Bioplex™ 100 instrument (Bio-Rad Laboratories, Hercules, CA) and analyzed with either BioRad or STaRStation software (Applied Cytometry, Sacramento, CA). Values that were calculated by the assay to be below the minimum detectable concentration (MinDC) for that particular analyte were conservatively estimated to be the MinDC value minus 0.01 pg/ml for statistical analysis. The number of replicates for the experimental samples are as follows: piriform cortex, n = 6 for each time point and naïve; hippocampus, n = 6 for each time point except for naïve (n = 5), 6 hr (n = 5) and 24 hr (n = 7); and thalamus, n = 5 for each time point and naïve except for 0.5 hr (n = 6), 6 hr (n = 4), 12 hr (n = 3), 24 hr (n = 6) and 48 hr (n = 6). Time matched vehicle controls (n = 3 per time point) were analyzed individually and condensed into a single vehicle control comparison group when no significant statistical difference was found between these samples over time by analyte or brain region.

Separate from the animals used in the multiplex bead array immunoassay, experimental, vehicle control and naïve animals were deeply anesthetized and perfused with isotonic saline followed by 4% paraformaldehyde via cardiac puncture. Brains were processed and sectioned at 40 microns as previously described [11]. The 12 hour time point was selected based on the peak expression times of the analytes from the multiplex assays. Free float fluorescent IHC labeling was conducted as previously described [26]. The antibodies used were as follows: rabbit anti-Gro α (CXCL1) (1:100; ab9772), rabbit anti-MIP-1α (1:500, ab9781) and mouse anti-rat endothelial cell antigen (RECA,1:1000; ab9774) from Abcam (Cambridge, MA), mouse anti-NeuN to label neurons (1:1000; MAB377) and mouse anti-CD11b to label microglia and macrophages (1:1000; CBL1512) from Chemicon (Temecula, CA), and mouse anti-GFAP to label astrocytes (1:1000; MS-280-P) from NeoMarkers (Fremont, CA). Alexafluor™ fluorescent-tagged secondary and tertiary antibodies (Molecular Probes, Eugene, OR) were used for visualization. Tissue sections labeled with only secondary and tertiary antibodies were used as controls. Sections were viewed and digitally captured with an Olympus BX51 microscope equipped with an Olympus DP-70 high-resolution color CCD digital camera (Opelco, Dulles, VA). An Olympus BX61 equipped with a DSU spinning disk confocal system and DP-70 CCD camera and a Zeiss LSM 700 confocal microscope were used for subsequent IHC micrographs to confirm same cell co-localization. Images of 40-μm tissues were acquired using a z step interval of 1 μm and analyzed using Slidebook™ (Olympus) or Zen 2009 (Zeiss) software. Publication images were compiled using Adobe Photoshop CS digital image software. Color levels and background labeling were reduced and evened using the "levels" tool. All input levels (0-255) were normalized in the RGB channel as follows: highlight input levels were set at the peak of the image histogram, midtone levels were set at 0.8 and shadow levels were set either at the edge of the histogram closest to 255 or at 180, whichever was greater. This technique was successful at reducing background while not oversaturating specific labeling. For all time points, n = 3.

Sections were labeled with Mayer's hematoxylin/eosin-phloxine stain. Infiltrating neutrophils were visually identified by the user and quantified within the piriform cortex, hippocampus and thalamus using Stereologer 2000 software (Stereology Resource Center, Chester, MD) on an Olympus BX51 microscope equipped with an IK-C44H CCD Toshiba camera (Imaging Planet, Goleta, CA). Four to eight sections from tissue slabs of approximately 1440 to 3300 μm in length were used for counting in each case. Estimates used the optical fractionator method. For each tissue section analyzed, section thickness was assessed empirically and guard zones 2 μm thick were used at the top and bottom of each section. The tissue regions were outlined using 10 × magnification, and cells were counted using 40 × magnification. Approximately 50% of the outlined region was analyzed using a systematic random sampling design with a counting frame size of 175 μm and a disector height of 7 μm. The coefficients of error (CE) were calculated by the software and maximum CE was set at 0.1600.

Immunoassay data were evaluated by one-way ANOVA with a post-hoc Dunnett's analysis and expressed in pg/ml. Neutrophil stereology data were evaluated by one-way ANOVA with a post-hoc Newman-Kuel analysis and expressed as cells/mm³. A Pearson's correlation coefficient using a one measurement time lag between CXCL1 or MIP-1α concentration and neutrophil infiltration was also calculated. The one measurement time lag was used due to the many downstream molecular events that occur following neutrophil exposure to these chemokines that allow the neutrophil to traverse the vasculature into the injured tissue. Values are expressed as mean ± SEM. Differences were considered significant at the level of p ≤ 0.05.

Temporal and regional changes in CXCL1, MIP-1α, G-CSF and GM-CSF protein concentrations were determined using a bead-based multiplex immunoassay on tissue lysates from the piriform cortex, hippocampus and thalamus. CXCL1 and MIP-1α significantly increased in all brain regions investigated. G-CSF and GM-CSF concentrations did not significantly change and were not analyzed further (data not shown).

CXCL1 concentrations significantly increased in all three brain regions (Figure 1). The highest concentrations were in the hippocampus, where concentrations significantly increased by 6 hours (4674 ± 1504 pg/ml) and peaked by 12 hours (8441 ± 2152 pg/ml vs. 58 ± 6 pg/ml in vehicle controls) following GD-induced SE. In the piriform cortex, CXCL1 levels peaked at 6 hours (1164 ± 195 pg/ml) and remained significantly elevated up to 24 hours (501 ± 176 pg/ml) compared to vehicle controls (49 ± 9 pg/ml). In the thalamus, CXCL1 concentration became significant at 12 hours compared to controls (4571 ± 643 pg/ml vs. 50 ± 3 pg/ml).

Twelve hours following GD-induced SE, CXCL1 immunolabeling was present in the piriform cortex (Figure 2A, left), hippocampus (dentate gyrus shown; Figure 2B, left) and thalamus (Figure 2C, left), while CXCL1 labeling was absent in vehicle controls in the same regions (Figure 2A, B, & 2C, right). Specific labeling was also absent in secondary only controls at the 12-hour time point (Figure 2D) and in vehicle controls (Figure 2E) as exemplified by the piriform cortex. In the piriform cortex, CXCL1-positive cells were found predominantly in layer II but also in layer III. In the hippocampus, CXCL1-positive cells were found primarily in the granular layer of the dentate gyrus (GrDG) and the CA3 pyramidal layer closest to the dentate gyrus. CXCL1-positive cells were also found in the laterodorsal and lateral posterior nuclei of the thalamus. To identify these cells, sections were co-labeled with antibodies specific for neurons, astrocytes, microglia and endothelial cells and for CXCL1. CXCL1 immunoreactivity was found in neuronal populations in the regions mentioned above (Figure 2F). CXCL1 diffusely labeled the cytoplasm in these cells with interspersed fine punctate labeling. No co-localization was observed between CXCL1 and astrocytes (Figure 2G) or microglia (Figure 2H) regardless of state of activation. Co-localization was limited in endothelial cells, though a prevalent, but not exclusive, association between CXCL1 and the vasculature was often observed (Figure 2I).

Significant concentration increases were observed for MIP-1α protein in all three brain regions following GD-induced seizure (Figure 3). MIP-1α concentrations significantly increased in the hippocampus at 6 hours (152 ± 42 pg/ml), peaked at 24 hours (247 ± 90 pg/ml) and then rapidly decreased by 48 hours after SE onset compared to vehicle controls (2.06 ± 0.05 pg/ml). In the piriform cortex, MIP-1α concentrations significantly increased at 3 hours (149 ± 14 pg/ml), peaked at 24 hours (200 ± 34 pg/ml) and remained elevated through the 72-hour endpoint (139 ± 22 pg/ml) compared to vehicle controls (<1.94 pg/ml, MinDC). The pattern in the thalamus was different still, where a double peak was observed at 12 hours (245 ± 28 pg/ml) and 48 hours (248 ± 22 pg/ml) compared to controls (18 ± 11 pg/ml).

Twelve hours following seizure onset, MIP-1α immunolabeling was present in piriform cortex (Figure 4A, left), hippocampus (dentate gyrus shown; Figure 4B, left) and thalamus (Figure 4C, left) but not in vehicle controls (Figure 4A, B & 4C, right). Weak to moderate diffuse co-localization with neurons (Figure 4D) was observed in layers II and III of the piriform cortex, the CA1 and CA3 pyramidal layers of the hippocampus and the polymorphic (PoDG) but not the granular layer (GrDG) of the dentate gyrus. Astrocytes were not found to express MIP-1α in any region observed (Figure 4E). Activated microglia strongly expressed MIP-1α in all regions investigated. These cells had a myriad of morphological features including hypertrophy, spheroid shape, blebbing and dystrophy. Specifically, MIP-1α shows a high degree of cellular localization with the dystrophic microglial morphology (Figure 4F). Lastly, MIP-1α-positive cells were closely associated with large blood vessel endothelial cells in the piriform cortex and thalamus but co-localization was not observed (Figure 4G).

To determine whether neutrophil recruitment correlates to increases in CXCL1or MIP-1α, neutrophil counts in the piriform cortex, hippocampus and thalamus were quantified using stereological techniques and correlated to CXCL1 and MIP-1α concentration data using a one measurement time lag with Pearson's correlation analysis. Neutrophil infiltration significantly increases in all three observed brain regions following GD-induced SE (Figure 5). No neutrophils were found in vehicle controls in any brain region (0 ± 0 cells/mm³). In the piriform cortex, neutrophil infiltration significantly increased at 12 (1,117 ± 485 cells/mm³) and 24 hours (1,565 ± 618 cells/mm³) but not 6 hours (3 ± 5.8 cells/mm³) compared to vehicle. Significant, though less robust, neutrophil infiltration was also observed in the hippocampus at 12 (128 ± 85 cells/mm³) and 24 hours (589 ± 10 cells/mm³) but not 6 hours (0 ± 0 cells/mm³) compared to vehicle. In contrast, neutrophils in the thalamus significantly increased only at 24 hours (2,098 ± 824 cells/mm³) and not at 12 hours (158 ± 90 cells/mm³). Pearson's correlation analysis revealed a positive correlation between CXCL1 concentration and neutrophil infiltration (offset by one time point) in the piriform cortex, hippocampus and thalamus. For MIP-1α, a less robust positive correlation existed in the hippocampus and thalamus compared to CXCL1. No significant correlation was observed in the piriform cortex (Table 1).