Animals
A total of 152 male C57BL/6 and RAG1 deficient mice were studied in four different cohorts. In studies A and C, adult male, 8 weeks old, C57BL/6 mice (wild-type (WT) mice; Charles River Deutschland GmbH; Sulzfeld, Germany) were examined. In Study B and D, we investigated male knockout mice lacking RAG1 (RAG1-/-; Translational Animal Research Center of the University Medical Center of the Johannes Gutenberg-University). Mice were randomly assigned to experimental groups (www.pubmed.de/tools/zufallsgenerator). Experiments and analyses were performed by investigators blind towards group allocation and treatment. The studies were performed with the approval of the Animal Care and Ethics Committee of Rhineland-Palatinate, Germany in accordance with the institutional guidelines of the Johannes Gutenberg University, Mainz (protocol number: 23177-07/G13-1-046) and in compliance with the ARRIVE guidelines. The animals were kept under controlled light and environmental conditions (12-h dark/light cycle, 23 ± 1 °C, 55% ± 5% relative humidity), and had free access to food (Altromin, Germany) and water at all times before and after the experiments.
Experimental TBI and anesthesia
Animals were anesthetized by intraperitoneal (i.p.) application of midazolam (Hameln pharmaceuticals GmbH; Hameln, Germany), fentanyl (CuraMed, Karlsruhe, Germany) and medetomidine (Dorbene vet; Wien, Austria). An air mixture (40% O2 and 60% N2) was supplied via facemask in spontaneously breathing mice 20. Depth of anesthesia was verified by respiration rate and pedal withdrawal reflexes. Rectal temperature was maintained constant at 37°C by feedback-controlled heating pad (Hugo Sachs, Germany). TBI was performed by controlled cortical impact (CCI) as previously described in detail 8,21. Briefly, the animal´s head was fixed in a stereotactic frame (Kopf Instruments, USA) and a large craniotomy (4x4 mm) was drilled above the right parietal cortex between sagittal, lambdoid, coronal sutures and insertion of the temporal muscle. A custom-fabricated controlled pneumatic impactor (L. Kopacz, Mainz, Germany) was placed perpendicularly to the brain surface and the impactor tip (diameter: 3 mm) centered in the middle of the craniotomy. The impact parameters were as follows: velocity, 8 m/s; duration, 150 ms; brain penetration, 1 mm. Immediately after CCI, the craniotomy was closed with conventional tissue glue (Histoacryl, Braun-Melsungen, Germany) and filament sutures. After the procedure animals were placed in their individual cages and allowed to recover for 6 hours in an incubator heated to 33°C, at a humidity of 35% (IC8000, Draeger, Lübeck, Germany).
Treatment
Application of antibodies for neutrophil granulocyte depletion and control antibodies
For the depletion of neutrophils in WT mice (studies A and C) the Ly6G-specific antibody (anti-mouse, clone 1A8) was used. In the control antibody group, we used the isotype control antibody immunoglobulin IgG2a (rat, clone: 2A3). Both antibodies, anti-Ly6G (1A8) and IgG2a (2A3) (BXCell; West Lebanon, USA) were diluted in PBS with a final concentration of 2.5 mg / mL. We injected 0.2 mL (0.5 mg) of anti-Ly6G antibody (ND) and the same volume of the control IgG2a antibody (Ctrl) intraperitoneally (i.p.) 24 h before (studies A and C) and 24 h after experimental TBI (study C).
Application of the AT1 inhibitor candesartan or vehicle solution
In both studies (C, D) candesartan, a specific AT1 inhibitor was applied. Candesartan (CV-11974; Tocris bioscience; Bristol, UK) and the vehicle solution were prepared and applied as previously described 8: The crystalline form of the active drug candesartan was dissolved prior each set of experiments in 0.037 M Na2CO3 (vehicle solution) in a concentration of 10 µg / mL. The animals received 0.1 mg/kg candesartan (Cand) or vehicle solution (Veh) by subcutaneous (s.c.) injection 30 min after experimental TBI, followed by a daily injection, 24 and 48 h after TBI.
Experimental protocols (Figure 1)
Study A: Effect of neutrophil granulocyte depletion 4 and 24 hours after TBI
WT mice were randomized to i.p. injection with specific anti-Ly6G (1A8) for neutrophil depletion (ND) or control antibody (Ctrl; IgG2a (2A3)) 24 h before TBI. Lesion volume and cerebral inflammation were determined 4 h (ND-4h, Ctrl-4h; n = 6 / group) and 24 h after TBI (ND-24h, Ctrl-24h; n = 8 / group). Additionally, neurological outcome was assessed 24 h after TBI. In mice without TBI, neutrophil depletion was controlled by differential blood cell count 24 h after application (ND-0, Ctrl-0; n = 2 / group).
Study B: Effect of RAG1 deficiency mediated lymphopenia 24 hours and 5 days after TBI
In RAG1-deficient mice (RAG1-/-) and their RAG1+/+ wild type litter mates lesion volume and neurology were assessed 24 h (n = 10 / group) and 5 days (n = 8 / group) after CCI.
Study C: Effect of AT1 inhibition in neutrophil depleted mice 3 days after TBI
Mice were randomized to treatment (24 h before and repeated 24 h after TBI) with either anti-Ly6G (ND) or IgG2a control antibody (Ctrl). They were subjected to CCI and then randomly assigned to additional treatment with candesartan (Cand) or vehicle solution (Veh), performed 30 min after TBI and then repeated daily, 24 and 48 h after TBI. Therefore, the animals were randomly allocated to four treatment groups: Ctrl-Cand, Ctrl-Veh, ND-Cand and ND-Veh (n = 12 / group). After the 72-hours observation period, brains were removed for quantification of lesion volume, cytokine expression and activated microglia. Blood samples were withdrawn for hematological quantification of white blood cells (WBC), lymphocytes and neutrophils. For comparison we used naïve (non-operated) WT mice (n = 6; Fig. 1).
Study D: Effect of AT1 inhibition in lymphopenic RAG1-deficient mice 3 days after TBI
RAG1-deficient mice were randomly assigned to candesartan or vehicle solution treatment (RAG1-/--Cand, RAG1-/--Veh; n = 12 / group) at 30 min, 24 and 48 h after TBI. As in study C, 72 h after TBI, lesion volume, cytokine expression and activated microglia were quantified and hematologic assessment was performed. Additionally, we used naïve RAG1-/- mice (n = 6; Fig. 1).
Measurement of physiological parameters
Before, and after experimental TBI body weight of each mouse was controlled. Blood pressure was measured 5 minutes before and after CCI under general anesthesia at the tail using a modified NIBP system (RTBP 2000, Kent Scientific, Torrington, USA; A/D converter: PCI 9112, Adlink Technology, Taiwan; software: Dasylab 5.0, measX, Germany; Flexpro 6.0, Weisang, Germany) as previously described 20. Additionally, blood pressure values were determined in awake animals daily for 8 days before (training phase) and for 2 days after CCI. Perioperative body temperature was measured by a rectal temperature probe (Physitemp; Clifton, NJ, USA).
Assessment of functional outcome
In studies A, C and D neurological outcome was assessed using the rotarod performance test (Heidolph Instruments GmbH &Co.; Schwabach, Germany) as previously described 22-25. After a pre-training phase (mice remained on a rotating rod for 20 s at 4 rpm) two days before TBI, the time to fall from the accelerating rod in the 2-min test period was registered. This test assesses coordination and motoric function and was performed 1 day before, 24 and 72 h after CCI. In study B functional outcome was determined by Neurological Severity Score 26. In addition to the rotarod test, in studies C and D, functional outcome was also determined by modified neurological severity score (mNSS; modified after Tsenter et al., 2008 26) 1 day before and 24 and 72 hours after CCI 4. To calculate mNSS, general behavior, alertness, motor ability and balance were rated with 6 different tasks. Each task was scored from 0 (normal) up to 3 (failed task). The mNSS ranges from 0 (healthy) to 16 (severely impaired) points 27 (Table 1). All neurological tests were performed by investigators blinded towards experimental group allocations.
Flow cytometry and blood cell count
At the end of observation period, in deep anesthesia, EDTA anti-coagulated blood samples were taken from the retro-orbital veins as previously described 28. The differential blood cell count was obtained via flow cytometry with a full automated veterinary analyzer, validated for murine blood analyses (ADVIA 2120i Hematology System; Siemens Healthcare, Erlangen, Germany) by an experimenter blinded to experimental group allocation.
Histologic and immunohistochemical evaluation
According to our previous protocol 4 brains were removed in deep anesthesia. For tissue evaluation, the brains were frozen in powdered dry ice and stored at -20 °C. They were then cut in coronal plane with a cryostat (HM 560 Cryo-Star, Thermo Fisher Scientific, Walldorf, Germany) as previously described in detail 8. The first slide was defined according to the first section corresponding to bregma + 3.14 mm in the Mouse Brain Library (www.mbl.org). Sixteen sections (12 and 20 µm) were collected at 500 µm-intervals, placed on Superfrost+TM slides (Thermo Fisher Scientific, Germany). In cresyl violet (Merck, Darmstadt, Germany) stained sections (12 µm), the total area of both hemispheres and the injured brain tissue area were determined for each section and animal using a computerized image analysis system (Delta Pix Insight; Maalov, Denmark) by an investigator blind to the group allocation. The total hemispheric brain volumes and the lesion volumes were calculated by following formula: 0.5 [mm] x (area of slide 1 [mm2] + area of slide 2 [mm2] + … + area of slide 16 [mm2]) 4,8. Immunohistochemistry was performed as described before 27. For immunohistochemical staining, cryosections (20 µm) were fixed in 4 % paraformaldehyde in phosphate buffered saline (PBS), incubated with blocking solution containing serum (5 % goat serum, 2 % bovine serum albumin, Gibco) and 0.1 % TX-100 (Sigma) in PBS for 1 h at room temperature. Primary antibodies specific for anti-ionized calcium-binding adapter molecule-1 (Iba-1; rabbit anti-mouse, anti-Iba-1 antibody; Wako Chemicals GmbH, Neuss, Germany) were applied in blocking solution overnight at 4°C. The sections were washed, incubated with secondary biotin-conjugated antibodies (goat anti-rabbit IgG; Merck; Darmstadt, Germany) and processed according to the manufacturer’s instructions using Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, USA). Images were taken at ×20 magnification (Axiovert, Zeiss, Germany). The total number of positive cells were counted at bregma -1.28 mm in a region of interest (ROI) of 0.52 x 0.65 mm² in the cortical tissue adjacent to the lesion by an investigator blind to randomization, using ImageJ software (National Institutes of Health, USA). Iba-1-immunolabeled cells with appropriate morphology and appearance 29 were identified as activated microglia/macrophages and assessed in a region of interest (ROI) of 0.52 × 0.65 mm² in the cortical tissue adjacent to the lesion. The rationale for counting Iba-1-positive cells in an area adjacent to the lesion rather than within the lesioned area was, that inside the lesion, where the tissue is essentially destroyed, microglia/macrophages are almost absent. In the perilesional area, however, there is a robust activation of microglia/macrophages. Results are presented in number of activated Iba-1 positive cells / mm2.
Gene expression analysis
Brain tissue samples from the lesion and perilesional area of 500 µm coronal cryostat sections between histologic slice intervals were collected, snap frozen in liquid nitrogen, stored at -80°C. As described previously in detail 4,30, after tissue sampling, extraction of mRNA and cDNA synthesis qPCR were performed (lysis: Qiazol-reagent, Qiagen, Hilden, Germany; homogenization: MM300 mill mixer, Retsch, Haan, Germany; RNA isolation: RNeasy Lipid Tissue Kit, Qiagen, Hilden, Germany; RNA concentration determined by spectrometer: NanoVue System, GE Healthcare Europe, Munich, Germany; RNA to cDNA reverse transcription Verso cDNA Kit, ABgene, Hamburg, Germany; cDNA amplification: real-time Ligthcycler 480 PCR System, Roche). PCR fragments of all applied genes were generated by PCR on an Eppendorf Thermocycler gradient (Eppendorf, Hamburg, Germany). The PCR products were purified with QIA quick PCR Purification Kit (Qiagen) according to the manufacturer´s instructions, and the DNA concentration was determined using NanoVue. A standard curve for absolute quantification was generated with PCR DNA for each PCR product (101-107 DNA copies/µl), showing similar and good efficiency (90-110 %; LightCycler Software, Roche) and linearity. Equal amounts of cDNA (1 µl) of each sample were analyzed in duplicates and amplified by real-time Lightcycler 480 PCR System (Roche). Real-time RT PCR kits were used according to the manufacturer´s instructions. All assays were conducted by an investigator blinded to group allocation. Using mouse-specific primers and probes (Tab. 2) and optimized temperature conditions for qPCR, absolute copy numbers of the target genes, tumor necrosis factor α (TNFα), transforming growth factor β (TGFβ), interleukin 1β (IL1β), interleukin 6 (IL6) and inducible nitric oxide synthase (iNOS) were calculated, and were then normalized against the absolute copy numbers of cyclophilin A (PPIA) 4,6,8,31. The reference gene PPIA was chosen as single normalizer 32 based on recent findings in our housekeeping gene study 31. In order to improve comparability of the mRNA expression data between different treatment groups and to eliminate qPCR kit dependent differences and limitations, qPCR data was normalized with PPIA and then related to normalized naïve target gene expression from naïve tissue samples from the corresponding brain region 33. Therefore, normalized target gene expression values are expressed as % naïve expression 4.
Statistical analysis
All experiments were randomized and performed by investigators blinded toward the treatment groups (computer-based randomization: www.pubmed.de/tools/zufallsgenerator). In order to determine the required sample size, the a priori power analysis using G∗Power 34 was performed with the main variable, the primary endpoint, lesion volume data from previously published studies 4,8. Therefore, based upon the data of these studies the present a priori power analysis was performed to determine an effect size of d 1.75, with an actual standard statistical power (1 − β) of 0.95, and a significance level (α) of 0.05 and a sample size per group of n = 7. In order to have a sufficient power, we decided to have larger sample sizes (n = 8 – 12, per group) 35. Statistical analysis was performed using the GraphPad Prism 8 Statistical Software (GraphPad Software Inc., La Jolla, CA, USA). Data distribution was tested by Shapiro-Wilks test. The comparisons of parametric and non-parametric data between two independent groups were done using the Welch-t test and the Wilcoxon rank sum test, respectively. For the statistical analysis of mNSS we performed ANOVA on ranks with the Kruskal-Wallis test, corrected for multiple comparisons using the Dunn’s test. In this multi-arm parallel group randomized trial, for comparison of multiple independent groups, if the Shapiro-Wilk normality test was passed, one-way analysis of variance (one- way ANOVA) with post-hoc Holm-Šidák comparisons test (comparisons between all groups) was employed. In experimental groups where two separate treatment factors (neutrophil depletion and AT1 inhibition) are present, a two-way analysis of variance (two-way ANOVA) was performed. Physiologic data, blood cell count, lesion volumes, number of activated microglia and mRNA expression data were compared between experimental groups with two-way ANOVA and post hoc with all-pairwise multiple comparison procedures (Holm-Šidák method). To evaluate group differences in repeated measurements from the same animals (body weight, systolic blood pressure), repeated measures (RM) two-way ANOVA (two-factor repetition) was applied (factors: treatment and time), followed by Šidák’s multiple comparisons test. Whenever there were missing values in the repeated measures dataset and a two-way ANOVA was not possible, repeated measures (mNSS, rotarod) data were analyzed with the mixed effect model using the restricted maximum likelihood (REML) method with Holm Šidák’s multiple comparison test. The p values were adjusted for multiple comparisons. Values of p < 0.05 were considered significant. Data sets were tested for statistically significant outliers using the Grubbs’ test. Data are presented as mean and standard deviation (mean ± SD).