Abstract

   Spinal cord injury (SCI) triggers microglial/monocytes activation with
   distinct pro-inflammatory or inflammation-resolving phenotypes, which
   potentiate tissue damage or facilitate functional repair, respectively.
   The major integrin Mac-1 (CD11b/CD18), a heterodimer consisting of
   CD11b and CD18 chains, is expressed in multiple immune cells of the
   myeloid lineage. Here, we examined the effects of CD11b gene ablation
   in neuroinflammation and functional outcomes after SCI. qPCR analysis
   of C57BL/6 female mice showed upregulation of CD11b mRNA starting from
   1 d after injury, which persisted up to 28 d. CD11b knockout (KO) mice
   and their wildtype littermates were subjected to moderate SCI. At 1 d
   post-injury, qPCR showed increased expression of genes involved with
   inflammation-resolving processes in CD11b KO mice. Flow cytometry
   analysis of CD45^intLy6C^−CX3CR1^+ microglia, CD45^hiLy6C^+Ly6G^−
   monocytes, and CD45^hiLy6C^+Ly6G^+ neutrophils revealed significantly
   reduced cell counts as well as reactive oxygen species (ROS) production
   in CD11b KO mice at d3 post-injury. Further examination with NanoString
   and RNA-seq showed upregulation of pro-inflammatory genes, but
   downregulation of the ROS pathway. Importantly, CD11b KO mice exhibited
   significantly improved locomotor function, reduced cutaneous
   mechanical/thermal hypersensitivity, and limited tissue damage at 8
   weeks post-injury. Collectively, our data suggest an important role for
   CD11b in regulating tissue inflammation and functional outcome
   following SCI.

   Keywords: spinal cord injury, Mac-1/CD11b, microglia/monocytes,
   neuroinflammation

1. Introduction

   Spinal cord injury (SCI) is a devastating event that leads to long-term
   disability and even life-threatening consequences. After the initial
   mechanical damage of the spinal cord, secondary injury processes
   produce a cascade of biochemical changes within the lesion site and
   surrounding tissue. These delayed and often long-lasting processes
   include activation of microglia and astrocytes, infiltration of
   macrophages and other peripheral immune cells, overproduction of
   reactive oxygen species (ROS), and neurotoxicity [[37]1,[38]2,[39]3],
   leading to extended tissue loss and neurological dysfunction. Following
   activation, microglia/macrophages present distinct disease-associated
   phenotypes shortly after injury: pro-inflammatory and
   inflammation-resolving [[40]4]. The former can exacerbate tissue
   damage, while the latter can potentiate functional recovery. This
   process is driven by the complex interaction between the central
   nervous system (CNS)’s resident microglia and infiltrated immune cells.
   However, the underlying mechanisms for phenotype determination are
   still not fully understood.

   Many studies have demonstrated that integrins, a large family of
   transmembrane proteins, regulate the interaction between immune cells
   and their extracellular environment, thus playing a major role in
   driving the polarization of macrophages and microglia
   [[41]5,[42]6,[43]7,[44]8]. Examples include microglia/macrophage beta2
   integrins binding to vascular adhesion molecules in the vascular lumen
   and promoting migration toward the injury site [[45]9]. Using
   monoclonal antibodies, studies showed that blocking integrin αDβ2
   (CD11d/CD18) activity reduced intraspinal inflammation, improved
   neurological functions, and was able to attenuate systemic inflammatory
   syndrome after SCI [[46]9,[47]10,[48]11,[49]12]. Moreover, β1-integrin
   has been reported to regulate astrogliosis and glial scar formation
   following SCI [[50]13], suggesting that immune modulation through the
   integrin family is a potential route for future strategies. More
   recently, studies have shown that inhibition of integrin β1 and its
   downstream signaling pathway can improve blood spinal cord barrier
   repair and promote neurological recovery after trauma [[51]14,[52]15].

   The major integrin CD11b/CD18 (also known as Mac-1, αMβ2, and CR3) is a
   heterodimer of CD11b (αM) and CD18 (β2) subunits [[53]16,[54]17]. The
   CD11b/CD18 integrin is highly expressed in all myeloid lineage immune
   cells, including microglia, macrophages, and neutrophils. As early as
   1999, a study reported on the expression levels of intercellular
   adhesion molecule 1 (ICAM-1) and CD11b in the acute stages of CNS after
   mechanical compression, showing significant upregulation in both the
   lesion center and perifocal zones within the first week of injury
   [[55]18]. More recently, studies using mouse models of optic nerve
   injury, hypertensive cardiac remodeling, liver fibrosis, pneumococcal
   pneumonia, and others have shown that CD11b knockout (KO) can promote
   debris clearance, attenuate neural degradation, and reduce macrophage
   infiltration [[56]19,[57]20,[58]21,[59]22,[60]23]. The extensive
   reporting on the role that CD11b/CD18 integrin plays in regulating
   myeloid cells has led us to explore the possibility of using CD11b as a
   therapeutic target in SCI. Prior work has shown that CD11b ablation can
   have neuroprotective effects in CNS diseases
   [[61]24,[62]25,[63]26,[64]27]. However, CD11b also mediated
   anti-inflammatory activity following activated protein C (APC)
   treatment of sepsis [[65]28,[66]29,[67]30,[68]31]. Thus far, only a few
   studies have examined the role of Mac-1 in SCI, with one recent paper
   reporting that Mac-1/CR3 is an important component in the development
   of injury-induced neuropathic pain [[69]32], while another earlier
   study found that CD11b contributed to myelin-induced inflammatory
   response after injury [[70]33]. Overall, the specific role of CD11b in
   SCI pathophysiology is still understudied and has yet to be fully
   addressed.

   Here, we utilize a contusion mouse SCI model to examine cellular and
   molecular changes as well as behavior and histopathology in the absence
   and presence of the CD11b gene. We demonstrate that SCI-induced acute
   inflammatory response is increased, whereas production of free oxygen
   species in microglia/macrophages/neutrophils is reduced in the CD11b KO
   mice, as measured by qPCR, flow cytometry, NanoString, and bulk RNA-seq
   analysis. Importantly, we report that functional deficits in locomotion
   and hypersensitivity to mechanical and thermal stimuli are
   significantly limited in CD11b KO mice and were associated with reduced
   tissue damage. Together, our studies demonstrate that CD11b regulates
   the status of beneficial neuroinflammation in the early-phase response
   to SCI, potentially contributing to SCI pathophysiology.

2. Materials and Methods

2.1. Animals and Mouse Spinal Cord Contusion Model

   All surgical and experimental procedures in this study were performed
   under protocols approved by the Institutional Animal Care and Use
   Committee (IACUC) at the University of Maryland School of Medicine. To
   reduce the risk of bladder infection [[71]34,[72]35], we only used
   female mice in this study. Young adult (10–12 weeks) female CD11b KO
   mice were obtained from Dr. Li Zhang’s lab and maintained in the UMB
   animal facility. Following the induction of anesthesia by isoflurane, a
   laminectomy was carried out at approximately the T9 to T11 levels, and
   the spinal column was secured with metal clamps to the lateral
   processes. The mice were subjected to a midline contusion injury of the
   spinal cord at the T10 level using the Infinite Horizon Spinal Cord
   Impactor (Precision Systems and Instrumentation) with a force of 60
   kilodyne, classified as moderate injury [[73]36]. Manual bladder
   expression was carried out thrice daily for the first 7–14 days after
   SCI until reflex bladder emptying was re-established. For mice used as
   control animals, sham surgery was performed after anesthesia,
   consisting of only a laminectomy without contusion. Individuals who
   performed functional assessment and data analysis were blinded to group
   designations and genotypes throughout all experiment stages.

2.2. Flow Cytometry

   Following euthanization, mice were perfused with 40 mL of ice-cold
   phosphate buffered saline (PBS). A segment (5 mm length) of spinal cord
   tissue surrounding the lesion area was isolated, weighed, and placed in
   complete Roswell Park Memorial Institute (RPMI) 1640 (Cat# 22400105,
   Invitrogen, Waltham, MA, USA) medium with 10% fetal bovine serum (FBS).
   The tissue samples were mechanically and enzymatically digested in
   collagenase/dispase (Cat# 10269638001, 1 mg/mL; Roche Diagnostics,
   Indianapolis, IN, USA), papain (Cat# [74]LS003119, 5 U/mL; Worthington
   Biochemical, Lakewood, NJ, USA), 0.5 M EDTA (Cat# 15575020, 1:1000;
   Invitrogen), and DNAse I (Cat# 10104159001, 10 mg/mL; Roche
   Diagnostics) for 1 h at 37 °C on a shaking incubator (200 rpm). The
   cell suspension was washed with RPMI and filtered through a 70-µm cell
   strainer before adding RPMI to a final volume of 3 mL and kept on ice.
   The samples were then transferred to FACS tubes and washed with FACS
   buffer, followed by a 10 min incubation period with Fc Block (Cat#
   101320, Clone: 93; Biolegend, San Diego, CA, USA) on ice. The spinal
   cord cells were stained with the following surface antigens:
   [CD45-eF450 (Cat# 48-0451-82, Clone: 30-F11; eBioscience, Waltham, MA,
   USA), CD11b-APC/Fire™750 (Cat# 101262, Clone: M1/70; Biolegend),
   Ly6C-APC (Cat# 128016, Clone: HK1.4; Biolegend), Ly6G-AF700 (Cat#
   128024, Clone: 1A8; Biolegend), Cx3CR1-PE (Cat# 149006, Clone:
   SA011F11, Biolegend)] and Zombie Aqua fixable viability dye (Cat#
   423102, Biolegend). The spinal cord cells were washed in FACS buffer,
   fixed in 2% paraformaldehyde (PFA) for 8 min, and washed once more
   before adding 500 µL FACS buffer. We used H2DCFDA (DCF, Cat# C6827, 5
   µm; Thermo Fisher Scientific, Waltham, MA, USA) to measure ROS
   production according to the manufacturer’s instructions. As described
   in previous publications [[75]37,[76]38], DCF was added to RPMI media,
   vortexed, and incubated for 30 min in a 37 °C water bath. All data were
   acquired on a BD LSRFortessa cytometer FACSDiva 6.0 (BD Biosciences,
   Franklin Lakes, NJ, USA) and analyzed using FlowJo v10.10 (Treestar
   Inc., San Carlos, CA, USA), with a minimum of 5 million events
   collected for each sample. Countbright^TM Absolute Counting Beads (Cat#
   [77]C36950; Invitrogen, Waltham, MA, USA) were used to estimate cell
   counts, expressed as total counts/mg spinal cord tissue weight.
   Leukocytes were first gated using a splenocyte reference (SSC-A vs.
   FSC-A), followed by singlets (FSC-H vs. FSC-W) and live cell gating
   based upon exclusion of Zombie Aqua (SSC-A vs. Zombie Aqua-Bv501).
   Resident microglia were identified as the CD45^intLy6C^−CX3CR1^+
   population, whereas peripheral myeloid cells were identified as the
   CD45^hiLy6C^+CX3CR1^+ population. Within the myeloid subset, further
   separation was achieved by identifying monocytes as Ly6C^hiLy6G^− and
   neutrophils as Ly6C^+Ly6G^+. Cell type-matched fluorescence minus one
   (FMO) control was used to determine the positivity of each antibody,
   and the DCF dye [[78]39].

2.3. Total RNA Extraction and Real-Time PCR (qPCR)

   RNA samples were obtained after processing with the RNeasy mini kit
   (Cat# 74104, Qiagen, Tegelen, The Netherlands) from a
   5-millimeter-length spinal cord tissue at various post-injury time
   points. Complementary DNA (cDNA) was synthesized by a Verso cDNA RT kit
   (Cat# AB1453B, Thermo Scientific, Waltham, MA, USA) per the
   manufacturer’s protocol. Real-time PCR for target mRNAs was performed
   with TaqMan gene expression assays for Itgam (Mm00434455_m1), CD83
   (Mm01350412_m1), P2yr12 (Mm01950543_s1), Tmem119 (Mm00525305_m1), Trem2
   (Mm04209424_g1), Gfap (Mm01253033_m1), Gdf15 (Mm00442228_m1), Opalin
   (Mm00463365_m1), Fcrls (Mm01219428_m1), Cxcl10 (Mm00445235_m1), Fkbp5
   (Mm00487406_m1), H2-T23 (Mm00439246_g1), Csf1r (Mm01266652_m1), Pllp
   (Mm00452740_m1), Ennp6 (Mm00624107_m1), and Plekhm1 (Mm00805590_m1).

2.4. Sample Processing and Western Blot Analysis

   At 3 d post-injury, mice spinal cord tissue at a length of 5 mm was
   extracted from the epicenter of injured WT or CD11b KO mice. For sham
   animals, an equal length of 5 mm was extracted at the same time from
   the approximate area of T10. For the processing of tissue samples, RIPA
   lysis buffer (Cat# R0278, Sigma-Aldrich, Burlington, MA, USA)
   supplemented with 1 × protease inhibitor cocktail (Cat# P8340,
   Sigma-Aldrich), Phosphatase Inhibitor Cocktail II (Cat# P5726,
   Sigma-Aldrich), and Phosphatase Inhibitor Cocktail III (Cat# P0044,
   Sigma-Aldrich, USA) was used for homogenization on ice. Following
   sonication and centrifugation at 20,000× g for 20 min, protein
   concentration was determined with the Pierce BCA method (Cat# 23227,
   Thermo-Fisher Scientific). For SDS-PAGE, samples were run on a
   criterion TGX 4–20% gradient gel (Bio-Rad, Hercules, CA, USA) and
   transferred to a 0.2 μm nitrocellulose membrane (Bio-Rad). After
   blocking with 5% non-fat skim milk in PBST, membranes were incubated
   overnight in primary antibodies diluted with blocking buffer. The next
   day, membranes were incubated in HRP-conjugated secondary antibodies
   and visualized with Supersignal West Pico PLUS Chemiluminescent
   Substrate (Cat# 34577, Thermo-Fisher Scientific). Images were acquired
   with the Chemi-Doc MP Imaging system (Bio-Rad) and optical density was
   quantified with the Image Lab software 6.1 (Bio-Rad). Primary
   antibodies and their respective dilutions are as follows: NLRP3 (1:500;
   Cat# 15101S, Cell Signaling Technologies, Danvers, MA, USA), Iba1
   (1:1000, Cat# 17198S, Cell Signaling Technologies, Danvers, MA, USA),
   and β-actin (1:10,000; Cat# SAB1305567A1978, Sigma-Aldrich, Burlington,
   MA, USA).

2.5. NanoString Analysis

   RNA samples were acquired from 5 mm of spinal cord tissue one day after
   SCI. We used the NanoString nCounter^® system to analyze total RNA (20
   ng/L) and obtain the transcript counts for 757 genes and 13
   housekeeping genes from the Mouse Neuroinflammation panel (NanoString
   Technologies, Seattle, WA, USA). The gene transcript counts were
   normalized before further analysis and paired differential expression
   analysis was carried out using NanoString’s nSolver software Version
   4.0. The statistical analysis of NanoString data was conducted with R
   version 4.2.2 in RStudio version 2023.6.0, build 421. Partial least
   square discriminant analysis (PLS-DA) was performed with the Mixomics
   package [[79]40], while the online Enrichr toolkit from Maayan Lab was
   used for pathway enrichment analysis [[80]41,[81]42]. All plots were
   drawn with the ggplot2 package v3.5.1 [[82]43], with volcano plots and
   heatmaps being rendered by the EnhancedVolcano v1.22.0 and
   ComplexHeatmap v2.20.0 [[83]44] packages, respectively.

2.6. Bulk RNA Sequencing and Transcriptomic Analysis

   Total RNA was extracted using RNeasy mini kit (Qiagen) and sent to
   Novogen (Sacramento, CA, USA) for RNA quality test and RNA-seq. The RNA
   quality was initially analyzed using Nanodrop and Agarose Gel
   Electrophoresis, and the quality of cDNA library is analyzed using an
   Agilent 2100 Bioanalyzer. RNA-seq transcriptomic analysis of all
   libraries was performed on Illumina, with data being converted into
   FastQ format for downstream analysis. The RNA-seq results were
   validated for selected genes using reverse-transcription real-time
   quantitative PCR (RT-qPCR). Bioinformatics analyses were performed by
   Novogen in RStudio version 1.2 with R v3.6.1. The FastQC v0.11.8
   package [[84]45] was used for data quality assessment and ribosomal RNA
   were filtered with SortMeRNA [[85]46]. For quantification of
   transcript-level abundance, the quasi-mapping-based mode was used in
   Salmon v1.0.0 [[86]47], with mapping to GRCm38 (mm10) mouse reference
   genome. Transcript-level abundance was aggregated with gen-level
   abundance with tximport [[87]48], while the package DESeq2 v1.44.0
   [[88]49] was used for differential expression analysis. Volcano plots
   and heatmaps were rendered with the EnhancedVolcano and ComplexHeatmap
   [[89]44] packages in Rstudio. Pathway enrichment analysis was performed
   with the online Enrichr toolkit from Maayan Lab [[90]41,[91]42].

2.7. Tissue Processing and Histologic, Immunohistochemistry Analysis

   The mouse spinal cord was dissected following intracardial perfusion
   with ice-cold normal saline and 4% paraformaldehyde. The segments
   containing the lesion area, or an equal length of spinal cord for sham
   mice, were embedded in an optimal cutting temperature (O.C.T) compound
   and cut into 20-µm-thick serial sections. In the assessment of lesion
   volume, sections spaced 1 mm apart within a segment of 5 mm on each
   side from rostral to caudal of the injury epicenter were stained with
   GFAP (1:1000; Cat# Z0334, Dako, Carpinteria, CA, USA) and DAB (Cat#
   PK-6100, Vector Labs, Newark, CA, USA) as the chromogen. The Cavalieri
   method was used with a grid spacing of 200 µm in the Stereo
   Investigator Software version 2022.2.1 (MBF Biosciences, Williston, VT,
   USA) for quantification as described in previous publications [[92]50].
   Areas absent of GFAP are considered part of the lesion cavity, from
   which the total volume was calculated. For visualizing and
   quantification of spared white matter (SWM), Luxol fast blue (LFB, Cat#
   S3382, Sigma-Aldrich) staining was performed as described in previous
   publications [[93]50,[94]51]. The section with the least amount of SWM
   was defined as the lesion epicenter, which was imaged at 4×
   magnification. Using ImageJ 1.54f (RRID: SCR_003070), each section’s
   total LFB positive area was calculated for SWM assessment and presented
   as a percentage of the total area.

   For immunohistochemistry (IHC) analysis, five coronal spinal cord
   sections around the lesion site with a 200 μm interval were applied for
   IHC staining as described in previous publications [[95]52]. Primary
   antibodies of rabbit Iba1 (1:200; Cat# 17198S, Cell Signaling), rabbit
   GFAP (1:200; Cat# PA5-16291, Invitrogen), rat CD86 (1:50; Cat# ab955,
   Abcam, Cambridge, MA, USA), and rat PDGFrB (1:100; Cat# 14-1402-82,
   Thermo Fisher) were used along with secondary antibodies of Alexa Fluor
   488 goat anti-rabbit IgG (1:800, Cat# A11008, Invitrogen) and Alexa
   Fluor 546 goat anti-rat IgG (1:800, Cat# A21247, Invitrogen).
   Fluorescent images were obtained through tile scanning with a Nikon
   Ti-E inverted microscope at 20× magnification (CFI Plan APO VC 20X NA
   0.75 WD 1 mm). The images were then analyzed using the NIH ImageJ
   software (1.54), wherein a counting window of 165.46 um × 165.46 um was
   randomly selected in the specified region for signal sampling. Two
   sections per mouse at 400–600 um distance rostrally from the epicenter
   were used for manual calculation of Iba1/CD86 positive cells. For
   GFAP/PDGFrB distribution, two sections per mouse close to the epicenter
   were used, in which the counting windows were applied in the injury
   core for PDGFrB or in the peri-injury area for GFAP. Identical imaging
   parameters and data analysis were adopted for all the samples per
   outcome.

2.8. Neurological Behavioral Tests

   Basso mouse scale (BMS) for locomotion: For evaluation of hindlimb
   locomotor function with BMS [[96]53], mice were placed within a level,
   enclosed arena 100 cm in diameter. Two trained researchers, unaware of
   the genotype of each mouse, observed them for at least 4 min. The
   animals were evaluated on a scale of 0 to 9, where 0 indicated total
   paralysis and 9 was normal activity. The grading criteria are derived
   from evaluating joint mobility, weight distribution, plantar stepping,
   and coordination. The mice underwent BMS score testing on day 1, day 3,
   and at weekly intervals for up to 6 weeks after SCI.

   Catwalk XT automated gait analysis: At 7 weeks after SCI, CatwalkXT
   automated system (Noldus; RRID: SCR_004074) was used for gait analysis
   as described previously [[97]54,[98]55]. Each mouse underwent a single
   testing session to maintain situational novelty and encourage
   exploration of the CatWalk. The CatWalk apparatus features a glass
   walkway with a light source emitting green LED light, which would be
   refracted upon contact with the paws. The areas of contact are then
   captured with a high-speed color camera and recorded in the CatWalk XT
   v10.6. software. Data acquisition was performed in a darkened room by a
   single researcher blinded to the mice genotype and handling each
   subject. Animals were placed in the open end of the CatWalk and allowed
   to walk across the walkway to the darkened escape enclosure. A minimum
   of three valid runs, or complete walkway crossings, were obtained for
   each subject. Trials where the animal stopped partway across or turned
   around during a run were excluded from the analysis.

   Hot plate test: To test SCI-induced cutaneous hypersensitivity of the
   hind paws, mice were placed on the contact probe of a computerized
   thermal stimulator on an Incremental Hot/Cold Plate Analgesia Meter
   (PE34, IITC Life Science, Woodland Hills, CA, USA). The temperature was
   increased from 30 to 50 °C with the incremental rate at 10 °C per
   minute. The test was stopped when the tested mouse licked either of its
   hind paws, and the threshold temperature was recorded. The test was
   conducted twice with the interval of 3 h, and the average stop
   temperature and latency time was recorded.

   Von Frey test: The von Frey filament method was used to detect hind-paw
   withdrawal from a mechanical stimulus according to the “up-down” Von
   Frey method outlined by Caplan et al. [[99]56] and simplified by Bonin
   et al. [[100]57]. At the start of the experiment, each mouse was
   individually placed in Plexiglass cubicles on a wire mesh platform and
   allowed to acclimate for 30 min. The von Frey filaments (MyNeuroLab,
   St. Louis, MO, USA), with incremental stiffness ranging from 0.04 g to
   2.0 g, were applied to the plantar surface of each hind paw. The
   experiment tests the response of the hind paw, with the first stimulus
   being set at a force estimated to be close to the 50% withdrawal
   threshold (0.4 g). A positive response is a brisk paw withdrawal (at
   least 3 times out of 5 applications) in response to the filament, while
   no paw withdrawal was considered a negative response. If there is no
   response, the next filament with a higher force is tested; if there is
   a response, the next lower-force filament is tested. This continues
   until 20 readings are obtained for each paw, and the sequence of
   outcomes (−for no response or +for response) is recorded. The
   statistical formula published by Dixon et al. was used to determine the
   mechanical force required to elicit a paw withdrawal response in 50% of
   animals [[101]58]. The average 50% pain threshold across three trials
   (one trial per day) was used to estimate the mechanical force threshold
   accurately.

2.9. Statistical Analysis

   Quantitative data are all plotted as mean ± standard error of mean, and
   individual data point is shown. The animal numbers required for each
   experiment were derived from variability estimated from published data
   and power calculations based on effect sizes defined by Cohen
   [[102]59]. All statistical analyses were conducted with GraphPad Prism
   (Version 9.00) for Windows (RRID: SCR_002798, GraphPad Software,
   Boston, MA, USA). Two-way ANOVA with repeated measures was used to
   analyze BMS scores, followed by Sidak’s multiple comparisons post hoc
   test. One-way or two-way ANOVA was used for multiple comparisons
   between groups, followed by Tukey’s multiple comparisons post hoc test
   parametric data (normality and equal variance passed). For lesion
   volume, stereological data were analyzed using Student’s t-test.
   Statistical analysis in each assay was detailed in figure legends. A
   p-value of ≤0.05 was considered statistically significant. Using a
   compilation of the Catwalk data, PLS-DA was used to calculate the
   overall percentage of variation that mice genotype would contribute to
   gait. For pairwise Spearman correlation analysis, all Catwalk
   parameters were included with the added rank variable of injury (0 =
   sham, 1 = SCI), genotype (0 = WT, 1 = CD11b KO), and the final BMS
   scores at 6w post-injury.

3. Results

3.1. Ablation of CD11b Gene Limits Acute Proinflammatory Response to SCI

   As a first step in exploring the role that CD11b plays in acute
   neuroinflammation, we used qPCR analysis to examine the changes to its
   expression level at multiple timepoints after SCI. We observed a
   significant increase in CD11b mRNA expression starting at 1 d
   post-injury, reaching its peak at 7 d and remaining persistently high
   for up to 28 d ([103]Figure 1A). Next, we examined the effects of CD11b
   ablation on microglia and monocyte markers, along with genes that
   reflect their phenotypic functions. At 1d after SCI, qPCR analysis
   showed near-zero levels of CD11b (Itgam) mRNA expression in both sham
   and SCI of CD11b KO mice, as expected ([104]Figure 1B). Furthermore,
   the gene Cxcl10, a chemokine well-known for inducing microglia
   migration and initiating microglial activation [[105]60], showed
   significant increases in the injured spinal cord of both genotype
   groups, with CD11b KO mice showing even higher levels ([106]Figure 1C).
   In contrast, the gene Trem2, which is expressed in microglia and other
   myeloid cells of the CNS, showed a marked increase in SCI/WT mice, but
   no injury-induced changes were observed in CD11b KO mice ([107]Figure
   1D). As a member of the transforming growth factor beta superfamily and
   known as macrophage inhibitory cytokine-1 (MIC-1), the expression
   levels of Gdf15 were very low in both sham groups but showed
   significant increases after injury. Pairwise comparison of SCI/WT with
   SCI/CD11b KO mice showed significantly lower expression levels in the
   latter group ([108]Figure 1E). Other microglia and myeloid cell markers
   that we tested included P2ry12, Tmem119, CD83, and Csf1r. The markers
   P2ry12 ([109]Figure 1F) and Tmem119 ([110]Figure 1G), both of which are
   abundantly expressed in ramified microglia [[111]61,[112]62], showed a
   significant decrease after 1d SCI without genotype effects. The gene
   CD83 plays a critical role in controlling and resolving immune
   responses, showed significantly lower mRNA levels in injury groups
   ([113]Figure 1H), but no genotypic differences were observed. Finally,
   the gene that encodes microglial receptor Csf1r showed neither injury
   nor genotype effects ([114]Figure 1I). In addition, both WT and CD11b
   KO mice showed marked upregulation of Gfap at 1 d post-injury, but no
   genotype effects were observed, suggesting that genetic depletion of
   CD11b did not affect astrocyte function ([115]Figure 1J). Collectively,
   acute SCI leads to increased levels of genes that initiate
   inflammation-resolving processes and drive myeloid cells towards
   clearance of damaged tissue and debris, which is partially affected by
   CD11b ablation.

Figure 1.

   [116]Figure 1
   [117]Open in a new tab

   CD11b mediates acute inflammatory response to SCI. (A) qPCR analysis
   was performed to examine the mRNA expression of CD11b in the injured
   spinal cord of C57BL/6 WT female mice. n = 7 (sham) and 5 (SCI). ** p <
   0.01, *** p < 0.001 vs. sham group. One-way ANOVA following Dunnett’s
   multiple comparisons test. (B–J) qPCR analysis was used to examine
   inflammatory responses in the injured spinal cord of young adult CD11b
   KO and WT female mice at 1d SCI. * p < 0.05, ** p < 0.01, *** p < 0.001
   vs. sham/WT. # p < 0.05, ## p < 0.01 vs. SCI/WT. n = 5–6/group. Two-way
   ANOVA followed by Tukey’s post hoc test.

3.2. Depletion of CD11b Reduces the Number of Microglia and Infiltration of
Neutrophils in the Injured Spinal Cord

   At d3 after SCI, when macrophage infiltration peaks in the spinal cord,
   flow cytometry was used to examine the cellular inflammatory response
   in the lesion site. Due to the absence of CD11b protein in the CD11b KO
   mice, we used other known microglia and monocyte markers for gating
   myeloid cell populations. The cell surface marker CX3CR1 is expressed
   in microglia and macrophages [[118]63]. As indicated in [119]Figure 2A,
   microglia were gated on the criteria of CD45 intermediate (int), Ly6C
   low, and CX3CR1 positive (+), which was validated in WT groups with the
   traditional gating strategy of CD45^intCD11b^+. Although there is a
   question of whether CD11b would affect microglia survival in sham mice,
   we observed no differences in the microglia cell counts of WT and CD11b
   KO mice. For infiltrating myeloid cells, the gating strategy of
   CD45^hiLy6C^+Ly6G^− was used to identify monocytes, while Ly6G^+
   indicated neutrophils. Using this new strategy, we were able to observe
   a significant increase in microglia cell counts at the lesion site of
   both WT and CD11b KO mice compared to sham groups of the same genotype;
   however, the number of microglia were markedly lower in SCI/CD11b KO
   mice ([120]Figure 2B), suggesting lower levels of injury-induced
   proliferation. Although the number of infiltrating monocytes remained
   the same between the two genotypes, CD11b KO mice had significantly
   lower number of neutrophils ([121]Figure 2C). Furthermore, ROS
   production was significantly attenuated in both microglia ([122]Figure
   2D,E) and infiltrating myeloid cells of SCI/CD11b KO ([123]Figure 2F,G)
   mice compared to the injured control group, as determined by DCF mean
   fluorescence intensity. To confirm the anti-inflammatory effects of
   CD11b KO, we used a separate batch of mice for Western blot analysis at
   3d post-injury. After SCI, Western blot analysis showed a dramatic
   increase in inflammasome NLRP3 ([124]Figure 3A,C) and
   microglia/macrophages marker Iba1 ([125]Figure 3B,D) in the injured
   spinal cord. As expected, the expression levels of both molecules were
   significantly lower in CD11b KO mice compared to their WT counterparts,
   indicating that CD11b plays a key role in the inflammatory response and
   microglia/macrophages activation process. Together, these results
   demonstrate that CD11b is critical to promoting microglia
   proliferation, infiltration of peripheral immune cells, and ROS
   production after SCI.

Figure 2.

   [126]Figure 2
   [127]Open in a new tab

   Genetic ablation of CD11b attenuates pro-inflammatory activation of
   microglia and monocytes at 3 days post-injury. (A) Gating strategy for
   microglia (CD45^intLy6C^−CX3CR1^+), monocytes (CD45^hiLy6C^+Ly6G^−),
   and neutrophils (CD45^hiLy6C^+Ly6G^+) in CD11b KO mice. (B,C)
   Quantification of microglia (B) and myeloid (C) cell counts are shown.
   (D–G) Reactive oxygen species (ROS) and oxidative stress were measured
   using DCF. Main effects of genotype and injury are seen in the ROS
   production levels of microglia (D,E) and myeloid cells (F,G). *** p <
   0.001 vs. sham groups. # p < 0.05, ## p < 0.01, ### p < 0.001 vs.
   SCI/WT. n = 5–6 mice/group. Two-way ANOVA followed by Tukey’s post hoc
   test.

Figure 3.

   [128]Figure 3
   [129]Open in a new tab

   Genetic ablation of CD11b suppresses NLRP3 expression and Iba1
   microglia/macrophages activation at 3 days post-injury. (A,B) Western
   blotting (WB) images show expression of inflammasome marker NLRP3 and
   microglia/macrophages marker Iba1. (C,D) WB quantification. **** p <
   0.0001 vs. sham groups. ## p < 0.01, ### p < 0.001 vs. SCI/WT. n =
   6/group. Two-way ANOVA followed by Tukey’s post hoc test.

3.3. CD11b KO Mice Show Robust Changes in Neuroinflammation-Related Genes
after SCI

   To determine the consequences of CD11b deletion on acute inflammatory
   response following SCI, we evaluated spinal cord tissue at the injury
   site with NanoString Neuroinflammation panel. Partial least square
   discrimination analysis (PLS-DA) was used on all normalized
   transcription count data to reveal a distinct separation between
   samples of each group, which was clustered into four quadrants
   ([130]Figure 4A). The two main variants of the PLS-DA model separated
   samples by injury (variate 1) and genotype (variate 2), accounting for
   52% and 10% of the total variation between samples, respectively.
   Furthermore, the gene Itgam regulating expression of CD11b showed
   significant reduction in the spinal cord of KO mice compared to their
   WT littermates, which further confirms the validity of the global KO
   model ([131]Figure 4B). Pairwise comparison between groups yielded a
   total of 95 (20 downregulated, 75 upregulated) in sham/CD11b KO vs.
   sham/WT and 123 (54 downregulated, 69 upregulated) genes in SCI/CD11b
   KO vs. SCI/WT, which indicated robust genotype effects in both baseline
   and after SCI conditions ([132]Figure 3C,D). Interestingly, more than
   double the number of genes showed genotype-induced downregulation in
   SCI mice than sham groups. In terms of injury effects, we were able to
   observe a total of 390 (175 downregulated, 215 upregulated) when
   comparing 1d SCI vs. sham in WT ([133]Figure S1A), as well as 385 (184
   downregulated, 201 upregulated) genes in SCI/CD11b KO vs. sham/CD11b KO
   groups ([134]Figure S1B). As expected, the majority of genes related to
   neuroinflammation showed dramatic increase at 1d post-injury. Next, we
   sought to find differentially expressed genes (DEGs) between groups,
   which yielded the top 20 DEGs in the SCI/CD11b KO vs. SCI/WT comparison
   set ranked by p-value ([135]Figure 4E). Amongst the top DEGs with the
   lowest p-value, Itgam and Mapk14 are involved in the regulation of
   innate immune response, both of which showed significant reduction in
   CD11b KO mice after SCI. For cellular function, genes involved with
   microglia (Stmn1, Dst) are also downregulated in the injured spinal
   cord of CD11b KO mice, while the gene Dlg1 (neurons and
   neurotransmission) and Opalin (oligodendrocyte function) followed
   similar trends. In addition, four genes (Apc, Agt, Fkbp5, and Cp) are
   involved with the regulation of astrocyte function. Moreover, the genes
   Apoe and Ep300 are regulators of lipid metabolism, which show
   significant upregulation in the spinal cord of CD11b KO mice. There are
   also several genes that are part of the epigenetic regulation process:
   Smarca5, Kdm4a, Eif1, Brd4, and Smarca4. Finally, the genes Cd47 and
   Ms4a4a, both of which show significant upregulation in the KO mice, are
   enriched on the surface of neurons and microglia, respectively.

Figure 4.

   [136]Figure 4
   [137]Open in a new tab

   NanoString analysis of spinal cord tissue is performed at 1 day after
   SCI. (A) Partial least square differentiation analysis (PLS-DA) was
   performed with all normalized gene counts from the NanoString
   neuroinflammation panel. The two main components of variation were
   captured on the x- and y-axis and show a clear separation of clusters
   between the injury and genotype groups. (B) Log10 abundance of the
   Itgam (CD11b) gene was obtained from NanoString transcriptional data.
   **** p < 0.0001 vs. sham groups. #### p < 0.0001 vs. SCI/WT. n =
   6/group. Two-way ANOVA followed by Tukey’s post hoc test. (C,D) Volcano
   plot of differentially expressed genes (DEGs) in the spinal cord of
   sham (C) and SCI (D) spinal cord samples with log2(fold-change) larger
   than 0.4 and log10(P) higher than 1.3 at the cutoff points depicted by
   dashed lines. (E) Heatmap of the top 20 DEGs in pairwise comparison
   between SCI/CD11b KO and SCI/WT. n = 6/group.

   To confirm the changes observed in NanoString, we performed validation
   on several DEGs with qPCR. We first examined the mRNA expression levels
   of four genes that had significant downregulation in NanoString
   analysis. Opalin, a marker for oligodendrocytes [[138]64,[139]65],
   showed significant injury-induced downregulation in WT mice
   ([140]Figure S2A) and even lower levels in CD11b KO mice after SCI. The
   gene Fcrls, which modulates Fc receptor-like protein 2, only showed a
   significant injury-induced upregulation in WT mice ([141]Figure S2B)
   but not in CD11b KO mice, which was markedly lower at the same
   timepoint after SCI. The genes Pllp and Ennp6 showed significant
   downregulation after SCI in both WT and CD11b KO, with no significant
   differentiation between the two genotypes after injury ([142]Figure
   S2C,D). We next examined three DEGs that were upregulated in NanoString
   analysis; the first was Plekhp1, which showed significant upregulations
   after injury but no differences between genotypes ([143]Figure S2E). In
   the case of the genes H2-T23 and Fkbp5, SCI/CD11b KO mice showed
   significant increases at 1 d SCI compared to sham groups ([144]Figure
   S2F,G), which was also significantly higher than the SCI/WT group.

   NanoString pathway enrichment analysis of DEGs showed interferon alpha
   response as the top pathway upregulated in the SCI/CD11b KO vs. SCI/WT
   comparison set, with interferon gamma response, inflammatory response,
   apoptosis, and TNF-alpha signaling via NF-kB as part of the top
   enriched pathways ([145]Figure 5A). Within the top enriched pathway of
   interferon alpha response, genes included Ifitm2, Rsad2, Irf1, Ifih1,
   Cd47, and Gbp2 ([146]Figure 5B). Pathway enrichment analysis of
   downregulated DEGs in SCI/CD11b KO vs. SCI/WT comparison yielded E2F
   targets as the top pathway ([147]Figure 5C). Other downregulated
   pathways include UV response Dn, hypoxia, reactive oxygen species
   pathway, and apoptosis. Specific DEGs within the top downregulated
   pathway include Xrcc6, Prkdc, Stmn1, Rpa1, and Rad1 ([148]Figure 5D).
   Due to the critical role that CD11b plays in the acute immune response
   to SCI, it is reasonable to assume that the pathways activated by
   injury in WT mice may be different than those activated CD11b KO mice.
   Based on this hypothesis, we next examined the injury-induced DEGs in
   WT and CD11b KO groups through pathway enrichment analysis. For WT
   mice, SCI led to the upregulation of genes in the pathways of
   interferon gamma response, inflammatory response, IL-6/JAK/STAT3
   signaling, TNF-alpha signaling via NF-kB, and apoptosis ([149]Figure
   S1C). Moreover, SCI also led to downregulation of reactive oxygen
   species pathway, E2F targets, apoptosis, TGF-beta signaling, and UV
   response Dn in WT mice ([150]Figure S1D). On the other hand, in CD11b
   KO mice, the top upregulated pathway is IL-6/JAK/STAT3 signaling, while
   the top downregulated pathway is UV response Dn ([151]Figure S1E,F).
   For analysis of classic pro- and anti-inflammatory genes, we examined
   eight pro-inflammatory genes (CD68, CD86, Cxcl10, Il1b, Il6ra,
   Tnfrsf10b, Tnfrsf1a, Tnfrsf1b) and seven anti-inflammatory genes (Ccl5,
   Cx3cr1, Il10rb, Il2rg, Tgfb1, Tgm2, Vegfa). These genes showed
   significant upregulation after injury, but no genotype differences were
   observed ([152]Figure S3A–C).

Figure 5.

   [153]Figure 5
   [154]Open in a new tab

   Enrichment analysis reveals potential pathways involved with
   CD11b-mediated neuroinflammation. (A,B) Pathway enrichment analysis of
   upregulated DEGs (A) with MSigDB Hallmark 2020 and genes involved with
   interferon alpha response (B). (C,D) Pathway enrichment analysis of
   downregulated DEGs (A) and genes involved with E2F targets (D). n = 6
   mice/group. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001 vs.
   sham groups. # p < 0.05, ## p < 0.01, vs. SCI/WT. n = 6 mice/group,
   two-way ANOVA followed by Tukey’s post hoc test.

   To further investigate the effects of CD11b ablation on the
   transcriptomic profile of injured spinal cords, we used bulk RNA
   sequencing to examine spinal cord tissue at 1d post-injury. Using a
   cutoff point of FDR < 0.05, we were able to screen a total of 137 DEGs
   (59 downregulated and 78 upregulated) from SCI/CD11b KO vs. SCI/WT and
   depict them in a volcano plot ([155]Figure 6A). In addition, our
   analysis also yielded the top 10 DEGs ([156]Figure 6B), with CD11b,
   Cfh, Gpr182, Ube2a, and Cyp51 being the top 5 genes with the lowest FDR
   value ([157]Figure 6C). Moreover, pathway enrichment analysis with the
   Bioplanet

Figure 6.

   [158]Figure 6
   [159]Open in a new tab

   RNA sequencing of injured spinal cord samples at 1d SCI. (A) Volcano
   plot of DEGs obtained from bulk RNA-seq. (B) Heatmap of top 10 DEGs in
   SCI/CD11b KO vs. SCI/WT. (C) Log10 abundance of top 5 DEGs obtained
   from pairwise comparison. (D,E) Pathway enrichment analysis of up- and
   downregulated genes. * p < 0.05, ** p < 0.01, **** p < 0.0001 vs.
   SCI/WT. n = 6 mice/group, two-way ANOVA followed by Tukey’s post hoc
   test.

   2019 database shows “Interleukin-1 regulation of extracellular matrix”
   as the top upregulated pathway ([160]Figure 6D), while “Cholesterol
   biosynthesis” is the top downregulated one in the KO mice ([161]Figure
   6E).

   In summary, our NanoString and bulk RNA-seq results show upregulation
   of proinflammatory genes involved with interferon alpha, interferon
   gamma, and other pathways that participate in the development of innate
   and adaptive immune responses, suggesting that CD11b KO mice have
   upregulated acute neuroinflammation in response to spinal cord injury,
   which is a necessary mechanism of defense for the CNS protection.

3.4. Ablation of CD11b Improves Functional Recovery and Reduces Tissue Damage
after SCI

   Finally, we assessed behavior and histopathology to determine the
   chronic impact of CD11b ablation on recovery. Weekly assessment of BMS
   scores and sub-scores showed that SCI/CD11b KO mice were recovering at
   a faster rate than their WT littermates ([162]Figure 7A,B). Starting
   from as early as 7 days post-injury, the average score for the SCI/WT
   mice was 1.11 ± 0.423, indicating that most WT mice within the group
   only showed slight ankle movement. At the same time point, SCI/CD11b KO
   mice had an average score of 1.875 ± 0.524, which is indicative of
   extensive ankle movement but no plantar placement. This dramatic
   genotype difference at 7 d post-injury (n = 9 for WT, n = 8 for CD11b
   KO, p = 0.025) was the start of a persistent trend that would continue
   for 6 weeks (p < 0.001), after which both injury groups appeared to
   have reached a plateau for motor function recovery. The average BMS
   scores at the plateau period was 4.6 for WT mice, indicating that mice
   had occasional or frequent plantar stepping. In contrast, CD11b KO mice
   had an average BMS of 5.7, which suggests that animals had frequent or
   consistent plantar stepping. Moreover, BMS sub-scores also reflected
   consistent plantar stepping and better coordination. We detected a
   significant main effect of genotypes [F(1, 15) = 5.549, p = 0.033 for
   BMS scores and F(1, 15) = 5.525, p = 0.033 for BMS sub-scores].

Figure 7.

   [163]Figure 7
   [164]Open in a new tab

   CD11b ablation improves motor and sensory functional recovery following
   SCI. (A,B) BMS scores and sub-scores were recorded weekly to quantify
   hindlimb locomotor recovery after SCI. * p < 0.05, ** p < 0.01, *** p <
   0.001 vs. SCI/WT. n = 8–9 mice/group. Repeated two-way ANOVA followed
   by Holm–Sidak post hoc analysis. (C,D) von Frey monofilament
   stimulation and hot plate tests were utilized at 7 weeks post-injury
   for assessment of mechanical and thermal allodynia. * p < 0.05, **** p
   < 0.0001 vs. sham/WT. ## p < 0.01 vs. SCI/WT. n = 5 for sham/CD11b KO
   and 8–10 for other groups. Two-way ANOVA followed by Tukey’s post hoc
   test. (E,F) Representative images and quantification of lesion volume
   (LV) at 8 w SCI via GFAP-DAB staining. (G,H) Representative images and
   quantification of spared white matter (SWM) at 8 w SCI. ** p < 0.01 vs.
   SCI/WT. n = 6–8 mice/group. Unpaired t-test.

   Next, we tested whether CD11b is involved in the development of
   post-injury allodynia evoked by mechanical and thermal stimuli. At 6 w
   following SCI, mice that regained adequate locomotor function to be
   able to withdraw a hind paw from a stimulus were selected for further
   nocifensive behavioral testing. In both tests of nociceptive behavior,
   there was no difference in mechanical/thermal threshold between the
   sham groups of either genotype ([165]Figure 7C,D). After SCI, the 50%
   mechanical pain threshold of SCI/WT was considerably lower than their
   uninjured counterparts ([166]Figure 7C), whereas CD11b KO mice showed
   little difference between sham and injury groups. Further examination
   of results demonstrated that the 50% pain threshold of SCI/CD11b KO
   mice was significantly higher than SCI/WT mice at 6 w post-injury
   ([167]Figure 7C), which indicates that hyperesthesia was effectively
   alleviated by CD11b genetic ablation. In the hot plate test, the
   temperature threshold for WT mice was significantly decreased after SCI
   (p < 0.05 vs. SCI/WT, [168]Figure 7D), but no differences were observed
   between the sham and SCI groups of CD11b KO mice, which further
   confirms the attenuation of SCI-induced allodynia. After completion of
   the behavioral tests, we examined tissue damage by lesion volume (LV)
   and spared white matter (SWM). Quantification of lesion volume by
   unbiased stereology showed a much smaller area of glial scarring in
   CD11b KO compared to their WT littermates (p < 0.01 vs. SCI/WT,
   [169]Figure 7E,F). Similarly, histological staining with solvent blue
   showed less myelin damage in SCI/CD11b KO mice compared to WT mice,
   while quantification showed a significantly higher percentage of SWM (p
   < 0.01 vs. SCI/WT, [170]Figure 7G,H). Further investigation of the
   injured spinal cord with IHC showed a reduced density of
   microglia/macrophage Iba1+ and CD86+ cells in the dorsal and lateral WM
   ([171]Figure 8A–D). Although no genotype differences were observed in
   the Iba1+ cell numbers in the ventral WM ([172]Figure 8B), the number
   of CD86+ cells were significantly lower in CD11b KO mice. Examination
   of the spinal scar with the astrocyte marker GFAP and fibrotic scar
   marker PDGFrB showed a lower ratio of positive cell areas in the
   peri-injury region and injury core ([173]Figure 8E–G). Taken together,
   the results indicate that CD11b depletion improves recovery after SCI,
   which is associated with reduced tissue damage.

Figure 8.

   [174]Figure 8
   [175]Open in a new tab

   Chronic neuroinflammation and formation of fibrotic scars was
   alleviated in CD11b-KO mice after SCI. (A) IHC representative images of
   Iba1+(green)/Cd86+(red) cells at 0.4–0.6 mm rostral to the epicenter.
   Areas of dorsal white matter (DWM), lateral white matter (LWM), ventral
   white matter (VWM) are outlined for quantification. (B) KO mice exhibit
   significantly reduced Iba1 positive cells in DWM and LWM, not VWM after
   SCI. (C) Significantly decreased Cd86 positive cells are detected in
   all three regions in KO SCI mice. (D) The ratio of Cd86 and Iba1
   double-positive cells show significant difference only in LWM between
   the two SCI groups. (E) Representative images demonstrate that PDGFrB
   (red) is mainly distributed inside the injury core while GFAP in the
   peri-injury area. (F) Either the relative GFAP-positive area or the
   relative PDGFrB-positive area are significantly lowered in KO mice in
   the two regions, respectively. (G) Representative images display the
   visual difference from GFAP or PDGFrB imaging between the two SCI
   groups in the peri-injury area or injury core. * p < 0.05, ** p < 0.01,
   *** p < 0.001 vs. SCI/WT. n = 6–7/group. Unpaired t-test. Scale bars =
   500 µm (A,E) and 50 µm (G).

   For further assessment of motor coordination, we used Catwalk XT gait
   analysis to examine fine motor differences beyond that recognizable by
   BMS scores. Stride length is the distance between successive placements
   of the same paw ([176]Figure 9A), which showed a significant main
   injury effect between groups [F (1, 28) = 6.130, p = 0.0196], but no
   main genotype effect. Measurement of print length and width also found
   significant main injury effects [F (1, 28) = 18.03, p = 0.0002 for
   print length, F (1, 28) = 17.29, p = 0.0003 for print width], along
   with marked decrease in both parameters in WT mice following SCI (p <
   0.01, [177]Figure 9B,C), but deficits were not significant in CD11b KO
   mice. Representative footprint images are indicated in [178]Figure S4.
   Next, we evaluated motor coordination with regularity index, a
   parameter that tracks the order of paw placement in a step cycle
   ([179]Figure 9D). A single step cycle is defined as each of the four
   paws being placed on the walking surface in sequence, which was
   analyzed by attributing each set of steps into either a normal stepping
   pattern or abnormal gait. The result is a percentage of normal stepping
   out of all step cycles analyzed. As expected, the step sequence
   regularity index was significantly lower in SCI/WT mice compared to
   sham/WT (p < 0.01 vs. sham/WT, [180]Figure 9D and [181]Figure S4),
   indicating clear deficits in motor coordination, but this decrease was
   not significant in CD11b KO mice. Phase dispersions, a parameter that
   describes the temporal relationship between placement of two paws
   within a step cycle, was used to measure inter-paw coordination
   ([182]Figure 7E,F). Assessment of the phase dispersions between right
   forepaw (RF) and left hind paw (LH) yielded significant increase in WT
   mice following SCI (p < 0.001, [183]Figure 9E), but the deficits in
   CD11b KO were not significant. On the other side, which examines the
   diagonal dispersion of left forepaws (LF) and right hind paws (RH),
   deficits from SCI could be observed in both WT (p < 0.001) and CD11b KO
   mice (p < 0.05, [184]Figure 9F). Moreover, both parameters showed
   significant main injury effects [F (1, 28) = 18.58, p = 0.0002 for RF-
   > LH; F (1, 28) = 31.56, p < 0.0001 for LF- > RH]. Print position is
   defined as the distance between a pair of hind paw and forepaw of the
   same side. Ideally, healthy C57BL/6 mice should be able to place their
   hind paw next to the location of the forepaw that has just been lifted
   from the walkway. Following SCI, the print positions of both WT and
   CD11b KO mice were significantly increased (p < 0.001, [185]Figure 9G)
   while also showing an injury main effect [F (1, 28) = 44.44, p <
   0.0001]. However, pairwise comparison of SCI/WT and SCI/CD11b KO
   yielded no statistical significance. Hindlimb base-of-support is a
   parameter that measures the average width of the track (distance
   between RH and LH) made by the animal, in which the farther apart the
   feet are placed during locomotion, the less likely the animal is to
   fall and the larger the base-of-support (BOS). Following SCI, WT mice
   showed significant decrease in hindlimb BOS (p < 0.05, [186]Figure 9H),
   indicating a lack of coordination and trunk stability. We next examined
   print area, max contact area, and max contact max intensity, which
   could reflect spontaneous pain activity in mice ([187]Figure 9I,J). The
   SCI/WT group showed a significant decrease in all three parameters,
   which suggests reduced contact with the Catwalk surface and the
   potential presence of spontaneous pain. However, no significant
   decrease was observed in CD11b KO mice, which suggests a lack of
   spontaneous pain and coincides with our results in mechanical and
   thermal allodynia.

Figure 9.

   [188]Figure 9
   [189]Open in a new tab

   CD11b KO mice showed less deficits of motor coordination after chronic
   SCI. At 7 weeks post-injury, Catwalk automated gait analysis was used
   to assess stride length (A), print length (B), print width (C), step
   sequence regularity (D), phase dispersions (E,F), print positions (G),
   base of support (H), print area (I), max contact area (J), and max
   contact max intensity (K). * p < 0.05, ** p < 0.01, *** p < 0.001 vs.
   sham groups. n = 5 for Sham/CD11b KO, n = 8–10 for other groups.
   Two-way ANOVA followed by Tukey’s post hoc test.

   In order to obtain an overview of the effects that SCI and genotype may
   have on Catwalk gait analysis, we used PLS-DA to examine their
   respective contributions to clustering of the gait parameters. The
   cluster diagram shows that injury contributed 69% to the differential
   variation, whereas genotype contributed 14% ([190]Figure S6A).
   Furthermore, pairwise Spearman correlation analysis on all Catwalk
   parameters showed high levels of correlation with injury severity and
   BMS scores ([191]Figure S6B). Based on the correlation coefficients,
   the parameters of print length (PL), print width (PW), step sequence
   regularity index (SSRI), and base of support (BOS) have high levels of
   (coefficient higher than 0.7) positive correlation with endpoint BMS
   scores, indicating that mice with higher BMS scores would also perform
   better in these gait parameters. On the other hand, both sets of phase
   dispersions (PD) and print position (PP), along the categorial scores
   of injury (0 = sham, 1 = SCI) have a negative correlation with BMS,
   indicating that injury severity would impact these scores negatively.
   In addition, the categorical score of genotype (0 = WT, 1 = CD11b KO)
   shows a negative correlation with Catwalk parameters, suggesting that
   CD11b KO mice have better recovery of gait. These data suggest CD11b KO
   mice could recover locomotor function more fully than WT mice.

4. Discussion

   In the present study, we examined the role that the major integrin
   CD11b/CD18 (Mac-1, CR3, αMβ2) plays in the acute inflammatory response
   after SCI, along with its contribution to locomotor dysfunction and
   neuropathic pain. Our findings show that experimental SCI causes
   pro-inflammatory activation of microglia/macrophage, characterized by
   qPCR, flow cytometry, NanoString, and bulk RNA-seq analysis. Genetic
   deletion of CD11b gene further increases the neuroinflammation in the
   injured spinal cord tissue. Importantly, CD11b KO mice show
   significantly improved locomotor functional recovery and reduced
   cutaneous hypersensitivity to mechanical and thermal stimuli, which
   were associated with reduced tissue damage. Thus, our data suggest that
   the CD11b gene may play a deleterious role in the recovery of
   neurological functions following SCI by suppressing the beneficial
   acute inflammation.

   Using qPCR, we first observed a significant increase in CD11b mRNA in
   the acute and sub-acute phase of SCI. This is consistent with studies
   that examined the temporal activation of microglia and macrophages
   after SCI [[192]18,[193]66]. Studies showed that on d1 after SCI,
   post-traumatic neuroinflammation mainly involves neutrophils, while
   microglia/macrophages start to play a more prominent role between d3
   and 7 following injuries [[194]67,[195]68]. The ubiquitous nature of
   CD11b in all types of myeloid cells required examination of both stages
   following SCI. Thus, we utilized qPCR and other methods to assess the
   transcriptomic profile of the injured spinal cord at 1d post-injury.
   Furthermore, we used flow cytometry due to the heterogeneity of immune
   cells activated at d3 after SCI.

   Using CD11b KO mice, we observed higher expression levels of genes
   involved in the activation of myeloid cells and their associated
   inflammatory responses. One of the genes that we examined, Cxcl10,
   regulates a chemokine that actively participates in the initiation of
   microglial activation following injury [[196]60]. Studies have shown
   that microglial activation in the early stages of SCI has beneficial
   effects in debris clearance, working with reactive astrocytes to form a
   glial scar at the lesion site and promoting neuronal regrowth
   [[197]69,[198]70,[199]71]. We further observed lower expression levels
   of Trem2 and Gdf15 mRNA in the lesion area of CD11b KO mice. High
   levels of Trem2 have been reported to be a sign of poor prognosis in
   Alzheimer’s disease (AD) and a mediator of microglia hyperactivation in
   neuropathic pain [[200]72,[201]73,[202]74], whereas Gdf15, a mediator
   of macrophage inhibitory cytokine-1 (MIC-1), has been reported to an
   important marker of mortality in the aged population and stroke
   patients [[203]75,[204]76,[205]77]. Although reports have also
   identified myeloid Gdf15 as a potential mediator for regenerative
   inflammation [[206]78,[207]79], the acute timepoint of our data
   excludes regeneration from the scope of study. Taken together, these
   results suggest that microglia/macrophage in CD11b-deficient mice
   demonstrate an inflammation-resolving phenotype that facilitates debris
   clearance.

   The surface adhesion receptor CD11b is routinely used in flow cytometry
   and other techniques as a well-known marker of the microglia/macrophage
   population. In the current study, the genetic ablation of CD11b/CD18
   has created a unique challenge for characterizing the neuroimmune
   profile of the injured spinal cord. Using a combination of microglia
   and monocyte markers, we were able to devise a novel gating strategy
   for identifying the cell populations. In this gating strategy,
   antibodies against CX3CR1 were used to differentiate resident microglia
   from infiltrating monocytes. Furthermore, differentiation of monocytes
   and neutrophils was achieved using Ly6C, a surface marker expressed in
   most blood-borne cells of monocytic origin [[208]80]. Using Ly6C and
   Ly6G, Saiwai et al. were able to identify a population of infiltrating
   monocytes that promotion inflammatory resolution after SCI [[209]81].
   Similarly, we used Ly6C and Ly6G to separate monocytes from
   neutrophils. Following initial gating of Ly6C, we used Cx3CR1 to
   isolate resident microglia from macrophages. Early studies that
   utilized in situ hybridization demonstrated that the majority of
   Cx3CR1+ cells in the spinal cord of experimental autoimmune
   encephalomyelitis (EAE) rats were microglia [[210]63]. The high
   specificity of Cx3CR1 to microglia is further confirmed by later
   studies that developed Cx3CR1-cre lines as a tool for genetic
   manipulation of microglia [[211]82,[212]83].

   By using flow cytometry to examine the injured spinal cords’
   inflammatory profile, our data also demonstrate that CD11b KO mice have
   less microglia accumulation and reduced infiltration of macrophages and
   neutrophils from the blood circulation following SCI, along with lower
   ROS production. To this end, we first show that SCI/CD11b KO have
   significantly lower expression levels of Trem2 than their WT
   counterparts. The Trem2 gene modulates the expression of triggering
   receptor expressed on myeloid cells-2 (Trem2), which has been shown to
   be upregulated in the serum of SCI patients [[213]84], and controls
   microglia hyperactivation in a model of progranulin deficiency
   [[214]74], along with the transition of microglia towards a
   proinflammatory phenotype and exacerbation of neuropathic pain
   [[215]72]. Moreover, a study comparing the mechanisms of inflammatory
   responses between traumatic and neurodegeneration injuries found that
   Trem2 plays a major role in the determination of phenotype [[216]85].
   In our qPCR results, we also observed downregulation of the
   oligodendrocyte marker Opalin, which is required for promotion of
   oligodendrocyte growth and differentiation [[217]64,[218]65]. One
   possible explanation for decreased Opalin levels is enhanced
   phagocytosis and cleanup of damaged cells at the acute stage of SCI.

   Consistent with the results from flow cytometry, both NanoString and
   RNA-seq showed that CD11b KO mice exhibited significant upregulation of
   pro-inflammatory transcriptomic factors and genes involved in
   phagocytosis. Pathway enrichment analysis reveals interferon alpha as
   the top pathway involved in upregulated DEGs for SCI/WT vs. SCI/CD11b
   KO. A recent study on chronic constriction injury in rats demonstrated
   that interferon alpha can improve their mechanical pain threshold,
   possibly having an analgesic effect [[219]86]. Amongst the genes
   upregulated in SCI/CD11b KO mice compared to their WT littermates, Cd47
   has been to be instrumental to the phagocytosis engulfment of apoptotic
   neurons, as CD47 KO mice exhibit excessive synaptic pruning [[220]87].
   Moreover, the downregulation of E2F targets coincides with our lab’s
   prior findings on cell cycle-related genes and neuroinflammation after
   SCI [[221]88]. Combined with the attenuation of allodynia observed in
   the present study, these results provide a possible mechanism for
   neuroprotection in CD11b KO mice. Thus far, we are the first to study
   the consequences of CD11b genetic ablation in a rodent model of
   traumatic SCI. In other models of neuronal injury, studies have shown
   that CD11b is critical in the development of experimental autoimmune
   encephalomyelitis and myelin phagocytosis [[222]89,[223]90]. In
   neutrophils, studies show that CD11b regulates adhesion and recruitment
   in pathologic inflammation [[224]91,[225]92].

   Unfortunately, the ubiquitous expression of CD11b in CNS immunity cells
   has made it difficult for our study to focus on one specific cell type,
   highlighting an innate limitation. In general, both microglia and
   macrophages play critical roles in inflammatory responses following
   trauma. However, it is hard to distinguish the role that CD11b plays in
   different cell types with a global knockout model. Such differentiation
   can only be performed with CD11b-flox mice, which we plan to
   incorporate in future studies focusing primarily on cell-type-specific
   mechanisms.

   Despite advancements in emergency medicine and higher survival rates,
   no effective therapy has emerged for SCI patients, who often suffer
   from temporary or permanent motor and sensory deficits. Immunotherapy
   strategies have achieved exciting progress in treating many CNS
   degenerative disorders such as Alzheimer’s disease (AD)
   [[226]93,[227]94,[228]95], Parkinson’s disease (PD)
   [[229]96,[230]97,[231]98], and multiple sclerosis (MS)
   [[232]99,[233]100,[234]101], and have even been adapted for traumatic
   brain injury [[235]102]. Our study shows for the first time that
   genetic ablation of CD11b could suppress oxygen free radical-associated
   tissue injury by microglia/macrophages/neutrophils while promoting
   neuroprotection. Although blockade of CD11b activity through use of
   antibodies has been reported in other disease models, with one recent
   study by Wolf et al. showing that ligand-specific blockade of CD11b
   with CD40L could target regional inflammation [[236]103], such
   therapeutic methods has yet to be employed in the context of traumatic
   SCI. Thus, one possible future direction will be the intrathecal
   injection of ligand-specific antagonists, which has more therapeutic
   and translational value. Based on the positive results shown in this
   study, delayed administration of CD11b antagonists may attenuate
   allodynia following SCI. Ultimately, it will be intriguing to see
   whether pharmacological manipulation of CD11b/Mac-1 has the same
   beneficial effects as genetic ablation.

5. Conclusions

   Taken together, we showed that CD11b KO significantly altered the
   ability of microglia/monocytes to migrate and infiltrate into the
   lesion site following SCI. We further demonstrated that SCI/CD11b KO
   mice had a pro-inflammatory profile and enhanced phagocytic activity
   with lower ROS production, thus promoting functional recovery following
   SCI. This was further supported by results from the NanoString
   Neuroinflammation panel and bulk RNA-seq. In chronic SCI, behavior
   assessment showed reduced motor deficits in CD11b KO mice compared to
   their WT littermates, along with mechanical and thermal allodynia
   attenuation. These findings suggest that CD11b is a major
   pathophysiological factor in SCI-mediated inflammatory response and
   related neurological impairments.

Acknowledgments