Animals

All mouse experiments were performed in the C57BL/6 genetic background using protocols approved by the Committee of Animal Research at the University of California, San Francisco, and in accordance with the National Institutes of Health guidelines. The program is accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care. C57BL/6 mice, B6.129P2-Cx3cr1tm1Litt/J (“Cx3cr1GFP/GFP”) mice [17], and B6.129-Ccr2tm2.1lfc/J (“Ccr2RFP/RFP”) mice [18] were purchased from the Jackson Laboratory for controlled cortical impact (CCI) experiments. Cx3cr1GFP/GFP and Ccr2RFP/RFP mice were crossed to yield Cx3cr1GFP/+Ccr2RFP/+ progeny. Fggγ390–396A mice were obtained from Dr. Jay Degen (University of Cincinnati, OH, USA) [7]. Males aged 17–24 weeks were used for all experiments. Mice were housed in groups of a maximum of five mice per cage under standard vivarium conditions with access to standard laboratory chow and water ad libitum and a 12-h light: dark cycle. All animal protocols were approved by the Committee of Animal Research at the University of California, San Francisco, and in accordance with the National Institutes of Health Guidelines.

Human tissue and immunohistochemistry

Human brain samples were obtained via the NIH NeuroBioBank. The “acute TBI” sample came from a 32-year-old male who had suffered a gunshot wound to the head within 24 h before autopsy. The sample arrived flash frozen, was processed and sectioned by the Gladstone Histology and Light Microscopy Core, fixed for 5 min in ice-cold acetone, and then subjected to immunohistochemistry. The “chronic TBI” sample came from a 21-year-old female who had survived for one year after blunt head trauma. The sample arrived formalin fixed, was processed and sectioned by the UCSF Biorepository and Tissue Biomarker Technology Core, deparafinized, rehydrated, and subjected to heat-mediated antigen retrieval (Target Retrieval Solution, Low pH, DAKO) for 1 h at 95 °C before proceeding with immunohistochemistry. Immunohistochemistry was performed as previously described [19] with the following antibodies: 1:500 mouse monoclonal anti-fibrinogen Ab [Abcam], 1:500 rabbit monoclonal anti-CD68 Ab [Abcam], 1:400 rabbit polyclonal anti-Iba1 [WAKO], 1:400 mouse monoclonal anti-MRP14 Ab [DAKO]. Images were acquired with an Axioplan II epifluorescence microscope (Carl Zeiss) equipped with dry Plan-Neofluar objectives (10 × 0.3 NA, 20 × 0.5 NA, or 40 × 0.75 NA), an Axiocam HrcCCD camera, and the Axiovision image analysis software. For each sample, high-powered fields were taken from a single slide with immune markers and fibrinogen immunoreactivity as histological landmarks.

Murine controlled cortical impact (CCI) model of TBI

CCI was performed as described previously [16], with modifications. Mice were anesthetized with buprenorphine (∼ 0.05 mg/kg) and ∼ 2% inhaled isoflurane through a non-rebreathing nose cone on a heating pad at 37 °C. After fixation with non-traumatic ear bars to a stereotactic frame (Kopf), eyes were provided with lubricant and the scalp was shaved and sterilized with betadine solution. Using sterile tip technique, a midline incision was made, and a ∼ 3.5 mm craniectomy was performed using an electric drill, with center point at 2.0 mm posterior and 2.0 mm lateral to bregma, taking care to not breach the dura. An electromagnetic cortical impactor (Hatteras, NC) delivered the injury with the following specifications: depth 0.9 mm, velocity 2.5 m/s, duration 500 ms. Afterwards, the scalp was stapled and the mouse was left to recover in a warmed cage (37 °C) until resumption of normal behaviors when it could rejoin its littermates. Sham controls underwent anesthetic induction, scalp incision, and craniotomy, but did not undergo cortical impact.

Mouse tissue immunohistochemistry

Dissection of brains and immunohistochemistry were performed as previously described [9], with following primary antibodies: rabbit polyclonal anti-fibrinogen (J. L. Degen, Cincinnati Children’s Hospital Medical Center), mouse monoclonal anti-NeuN (Millipore), and mouse monoclonal anti-4-Hydroxy-2-nonenal (HNE, Alpha Diagnostic International). Images were acquired with a BIOREVO BZ-9000 inverted fluorescence microscope (Keyence) equipped with a Nikon CFI 60 series infinite optical system and Keyence imaging software or an Olympus FluoView 1000 confocal microscope with water-immersed objectives (20 × 0.5 NA or 40 × 0.64 NA). ImageJ (NIH) was used for image analysis. Depending on the staining, quantification was performed on thresholded, binary images or counting of cells by researchers blind to the mouse genotype. Homozygous Fggγ390–396A mice were compared to C57BL/6 mice as non-transgenic wildtype colony controls for the CCI histology experiments.

For fibrin(ogen) staining, a region of interest (ROI) was established for each slice consisting of all cortex within 250 μm of the impacted edge (for TBI group) or craniotomy (sham). Relative intensity of fibrinogen within this region was compared to a 250 μm by 250 μm region of the contralateral, uninjured cortex. Three slices were averaged per mouse, with n = 4–7 mice per timepoint. Statistics performed with one-way ANOVA, with Dunnett’s multiple comparisons test to compare each timepoint to sham mice.

For quantification of fluorescent cells in Cx3cr1GFP/+Ccr2RFP/+, 150 × 150 μm were captured from a larger image of cortex containing the lesioned edge: ROI were obtained either adjacent to the lesion (i.e. an ROI with a complete border at the lesion edge and extending into cortex) or remote from the lesion (i.e. an ROI located at the furthest extent ∼ 300 μm away from injury). For each ROI, the fibrinogen intensity was quantified separately from the cell counting, where the scorer was blinded to location or fibrinogen content. Upon cell counting, CX3CR1 positive cells were further classified as to whether they were morphologically “amoeboid” (or “activated”): cells that lacked an obvious ramified/stellate morphology, instead demonstrating prominent, rounded cell somata with very few, shortened processes. Three adjacent and three remote ROIs were selected per mouse. These data were analyzed via linear regression or unpaired student’s t-test, as detailed.

For quantification of SMI-32 staining, images were acquired from an ROI consisting of cortex within 250 μm of the affected edge with an Axioplan II epifluorescence microscope (Zeiss) equipped with dry Plan-Neofluar objectives (10 × 0.3 NA, 20 × 0.5 NA, or 40 × 0.75 NA). Quantification of SMI-32 immunoreactive area was performed on thresholded images using ImageJ. Three sections were selected per mouse, with n = 4 mice. The relationship between fibrinogen and SMI-32 was determined by linear regression analysis. To measure oxidative damage levels, the CA1 region in the ipsilateral side was immunostained for 4-hydroxynoneal (4-HNE) protein adducts, a marker of lipid peroxidation. The 4-HNE immunoreactive area was quantified on thresholded images utilizing ImageJ software.

For quantification of cortical loss, immunostaining for NeuN was used to identify cortex in coronal slices at identical positions from bregma (i.e. -1.9 to -2.7 mm). The contralateral “cortical fraction” was calculated as a percent area of the depicted hemisphere that was designated as cortex. For ipsilateral cortex, an outline of a non-deformed hemisphere was produced (using the contralateral cortex as a guide); cortical fraction was calculated as percent area of the of the outlined cortex that was designated as remaining cortex.

Cerebral vessel casting and uDISCO

At the time of CCI, wild-type mice (n = 3) were injected with fluorescently tagged Alexa647-fibrinogen (Invitrogen). At 2 days following injury, mice underwent vessel casting and perfusion. Briefly, mice were injected with 2.5% tribromoethanol followed by transcardial perfusion with 20 mL of PBS and 20 mL of 4% PFA in PBS. This was followed by perfusion with 10 mL of fluorescent gel perfusate (2% w/v PBS porcine skin gelatin type A with 100 mg fluorescein 2MDa dextran) at 40 °C, followed by rapid body cooling, as previously described [20]. Brain and skull were kept intact as they underwent post-fixation in 4% PFA overnight at 4 °C. Mouse brain tissue was then processed using the uDISCO brain clearing protocol as previously described [21] with BABB-D15 refractive index matching solution. Samples were imaged using a Nikon A-100 light-sheet microscope equipped with a Vortran 4-line laser launch, an Andor Zyla 5.5 camera, and a 2x/0.2 AZ Plan Apo objective lens to obtain large-volume images. Image processing and analysis were performed with Imaris and ImageJ.

Fluorescence-activated cell sorting (FACS)

Following TBI experiments, animals were perfused with 4 °C DPBS and a 1 mm tissue block spanning the lesion sites of each mouse was isolated. To increase cell yield for RNA-seq analysis, two mouse brain tissue samples per genotype and timepoint were randomly pooled together for each biological replicate. Single cell suspensions were prepared using the adult brain dissociation kit (Miltenyi Biotec) as described [22]. Samples were then treated with Fc-block in DPBS supplemented with 0.2% BSA (Thermo Fischer Scientific) for 5 min at 4 °C, then incubated with primary antibodies for 30 min at 4 °C. The following primary antibodies were used at 1:200 dilution and purchased from Biolegend: CD11b (M1/70) conjugated to APC-Cy7 and CD45 (30-F11) conjugated to 488. Samples were incubated with Sytox blue live/dead stain for 10 min prior to sorting as previously described [13]. FACS of 40,000 live microglia with CD45lowCD11b+ expression or 25,000 live infiltrating myeloid cells with CD45hiCD11b+ expression were sorted directly into tubes containing RLT plus lysis buffer (Qiagen) supplemented with 1% 2-mercaptoethanol and 0.25% reagent DX (Qiagen) using FACSAria II with BD FACSDiva v8 software. Based on our prior bulk RNA-seq studies [13, 22], samples with > 20,000 sorted cells were used for RNA-seq library preparation. Microglia were sorted 1d and 7d after injury. Myeloid cells were only sorted at 1d after injury as fewer than 20,000 cells were detected at day 7. Gating strategy: first gate, cells based on SSC-A and FSC-A size; second gate, doublet discrimination using FSC-H and FSC-W parameters; third gate, sytox blue negative live cells; final gate, either CD45loCD11b+ or CD45hiCD11b+. Lysates were frozen and stored at -80 °C until processed for RNA-seq.

Bulk RNA sequencing (RNA-seq) and data analysis

RNA was isolated from microglia or macrophage lysates using the RNAeasy micro kit without modification (Qiagen). RNA quality control (QC) was determined by Bioanalyzer pico chip analysis (Agilent) and all samples with RNA integrity number > 8.0 were used for library preparation. cDNA libraries were generated using the Ovation RNA-seq System V2 low input kit following manufacturer’s instructions (NuGEN). Libraries were QC checked by KAPA qPCR (Roche) and Bioanalyzer DNA chip analysis (Agilent). All 21 libraries were equimolar pooled and sequenced across three lanes on Hiseq4000 with paired-end 100 base pairs (Illumnia). The median sequencing depth was ∼ 63 million reads per sample. FASTQ files were processed using the open-access Nextflow RNA-seq pipeline [23] using nextflow v20.12.0-edge in Singularity with default nf-core/rnaseq v3.0 parameters for paired-end FASTQ files. The following packages were used: Bioconductor-summarized experiment v1.20.0; bedtools v2.29.2; deseq2 v 1.28.0; dupradar v1.18.0; fastqc v0.11.9; picard v2.23.9; preseq v2.0.3; rseqc v3.0.1; salmon v.1.4.0; samtools v1.10; star v2.6.1d; stringtie v2.1.4; subread 2.0.1, trimgalore v0.6.6 and ucsc v377. FASTQ files were mapped to GRCm38 genome. Gene analyses was performed in R v4.2.0 using DeSeq2 v1.36.0 on the salmon.merged.gene_counts_scaled file produced by Nextflow. Reads with fewer than 5 counts per gene across all replicates were filtered out. For differential gene analysis the results function in DeSeq2 was used with contrast to test between two genotypes and timepoints of interest. The data were then filtered for significance using abs(log2FC) > 1 and adjusted P value < 0.05 unless otherwise stated. Unbiased KEGG analysis was performed using clusterProfiler and enrichplot with default parameters and pvalueCutoff set to 0.1.

Statistical analysis

Data were analyzed with Graphpad Prism with statistical testing as described in the text. Data are expressed as the mean ± s.e.m. All mice survived until the end of the study, and all of the data were analyzed. Mice were assigned and coded in a blinded manner for group allocation and subsequent data collection after the analysis. Statistical significance was determined using a non-parametric, two-sided Mann-Whitney test, or one-way or two-way analysis of variance, followed by Dunnett’s or Bonferroni’s multiple comparisons test. All attempts at replication following the protocols described in the methods were successful. For IHC experiments, image acquisition and quantification were performed by observers blinded to experimental conditions. Images were quantified independently by blinded observers. Sample sizes were determined by prior studies rather than statistical approaches.