Deletion or Inhibition of Astrocytic Transglutaminase 2 Promotes Functional Recovery after Spinal Cord Injury

Following CNS injury, astrocytes become “reactive” and exhibit pro-regenerative or harmful properties. However, the molecular mechanisms that cause astrocytes to adopt either phenotype are not well understood. Transglutaminase 2 (TG2) plays a key role in regulating the response of astrocytes to insults. Here, we used mice in which TG2 was specifically deleted in astrocytes (Gfap-Cre+/− TG2fl/fl, referred to here as TG2-A-cKO) in a spinal cord contusion injury (SCI) model. Deletion of TG2 from astrocytes resulted in a significant improvement in motor function following SCI. GFAP and NG2 immunoreactivity, as well as number of SOX9 positive cells, were significantly reduced in TG2-A-cKO mice. RNA-seq analysis of spinal cords from TG2-A-cKO and control mice 3 days post-injury identified thirty-seven differentially expressed genes, all of which were increased in TG2-A-cKO mice. Pathway analysis revealed a prevalence for fatty acid metabolism, lipid storage and energy pathways, which play essential roles in neuron–astrocyte metabolic coupling. Excitingly, treatment of wild type mice with the selective TG2 inhibitor VA4 significantly improved functional recovery after SCI, similar to what was observed using the genetic model. These findings indicate the use of TG2 inhibitors as a novel strategy for the treatment of SCI and other CNS injuries.


Introduction
Astrocytes are the most abundant cell type in the central nervous system (CNS) and play an essential role in supporting neuron homeostasis and function [1]. For example, as a key component of the tripartite synapse (which also includes presynaptic and postsynaptic nerve terminals), astrocytes play a pivotal role in modulating synaptic transmission and plasticity [1,2]. Astrocytes also play a central role mediating the response of the CNS to injury. Following an insult, astrocytes respond by becoming reactive and undergoing a rapid change in morphology, gene expression and molecular functions [3,4]. Depending on the type, severity and location of the injury, astrocytes exhibit different responses, taking on a phenotype that is either beneficial or detrimental to neuron survival and axonal regeneration [5]. Although it is clear that astrocytes can take on either supportive or harmful phenotypes after injury, the cell autonomous determinants that direct astrocytes towards either phenotype have not been well defined. Recently, we have identified transglutaminase 2 (TG2) as a key factor in determining the molecular response of astrocytes to injury [6][7][8][9].
TG2 is a multifunctional protein that is upregulated in astrocytes in response to cell stressors including ischemia and inflammatory signals. Several studies have demonstrated that TG2 plays a role in facilitating either cell death or survival in a range of cell types, depending on the cell type and stressor [8]. For example, TG2 supports neuronal health [10]. Overexpression of TG2 in neurons improves their survival following an ischemic insult, and, conversely, when TG2 is depleted from neurons it significantly decreases their survival [11,12]. In contrast, deletion or depletion of TG2 from astrocytes significantly increases their survival in response to oxygen and glucose deprivation (OGD) and significantly increases their ability to protect neurons from ischemic-induced cell death injury [6,7,13]. In addition, we have demonstrated that mice with astrocyte-specific TG2 deletion (glial fibrillary acidic protein (Gfap)-Cre+/−/TG2fl/fl, referred to here as TG2-A-cKO) exhibited significantly less astrocytic scarring in an in vivo model of a contusion spinal cord injury (SCI) when compared to TG2fl/fl mice expressing normal levels of astrocytic TG2 [7]. Given that the reduction of astrocytic scarring in the TG2-A-cKO mouse is indicative of a reduction in reactive astrogliosis in the mice with TG2-depleted astrocytes, it is important to understand how deletion of TG2 impacts functional recovery following SCI.
In this study, we demonstrate that astrocytic-specific TG2 deletion results in a significant improvement in functional recovery in a mouse T9 spinal cord contusion injury model. GFAP immunoreactivity was significantly decreased in the TG2-A-cKO compared to TG2fl/fl mice. We further extend these findings and demonstrate that expression of both the transcription factor SOX9 and the chondroitin sulfate proteoglycan (CSPG) NG2 (Cspg4) are significantly decreased in TG2-A-cKO spinal cord compared to TG2fl/fl mice at 3 and 7 days after injury. These are exciting findings as NG2 is part of the astroglial scar and may block successful CNS regeneration [14], while deletion of SOX9 improves outcomes after CNS injury [15][16][17].
To begin to understand the molecular mechanisms contributing to improved outcomes following SCI when TG2 is specifically deleted from astrocytes, we carried out RNA-seq on spinal cords from TG2-A-cKO and TG2fl/fl mice 3 days after SCI. Intriguingly, 37 genes were differentially expressed after injury in TG2-A-cKO spinal cords compared to TG2fl/fl mice and all were upregulated. Interestingly, the majority of genes that were upregulated were involved in fatty acid metabolism, energy pathways and lipid storage. Overall, these genes are involved in promoting neuron-astrocyte metabolic coupling, which enhances the ability of the astrocytes to protect and support the neurons [18][19][20].
Finally, to test whether acute inhibition of TG2 activity after SCI is sufficient to promote recovery, we used a pharmacological approach. In vitro, the highly specific and irreversible TG2 transglutaminase inhibitor VA4 [21] phenocopies the effects of TG2 deletion in that it protects astrocytes from OGD-induced cell death [13]. In addition, previous studies with the parent compound of VA4 showed that it was well tolerated by mice [22]. Interestingly, in mice expressing normal levels of TG2, treatment with VA4 significantly improved functional recovery following SCI compared to mice treated with vehicle only. Overall, the results of our studies demonstrate that removal of astrocytic TG2 or inhibition of TG2 significantly improves outcomes after CNS injury and strongly suggest that the use of TG2 inhibitors may be a therapeutic strategy for treating acute SCI.

Mice
All procedures with animals were in accordance with guidelines established by the University of Rochester Committee on Animal Resources, with approval from the Institutional Animal Care and Use Committee. TG2fl/fl mice were generously provided by Drs. R. Graham and S. Iismaa (Victor Chang Cardiac Research Institute). These mice contain loxP sites flanking Tg2 exons 6-8 [23]. B6·Cg-Tg(Gfap Cre)73.12Mvs/J (Gfap Cre+/−) mice were obtained from Jackson Laboratories and have been described and characterized previously [24]. See [7] for detailed information on the generation and genotyping of the Gfap Cre+/− TG2fl/fl (TG2-A-cKO) and TG2fl/fl mice.

Mouse Spinal Cord Contusion Injury
All mice were genotyped prior to surgery and after tissue harvest. For the locomotor evaluation, female mice were used for studies that continued for 42 days. For all other measures both male and female mice were used. TG2-A-cKO and TG2fl/fl mice, 12-16 weeks of age, were anesthetized using a combination of ketamine and xylazine given intraperitoneally. A laminectomy was performed at the thoracic 9 (T9) level and the dura mater was exposed. The spinal cord was moderately contused bilaterally using a force-defined injury device (65 kDyn, Infinite Horizon Impactor 400, Precision Systems and Instrumentation). Muscle and skin were closed in layers at the injury. A heating pad was used to maintain body temperature at 37 • C during surgery and the initial recovery period. Mice were given buprenorphine subcutaneously at time of surgery and three times daily for 3 days thereafter. Food and DietGel ® Recovery was placed in the bottom of the cages and lactated Ringer's solution was administrated i.p. for the first 5 days to ensure hydration. Following SCI, bladders were manually expressed every 12 h until spontaneous voiding recurred.
For the studies with the TG2 inhibitor VA4 [21], wild type mice expressing normal levels of TG2 were used. One hour after SCI, they were administered VA4 at 15 mg/kg i.p. in Cremophor or Cremophor only (as a control) and then again 24 and 48 h later. Mice were maintained after injury as described above.

Locomotor Evaluation
For gait analysis, mice were video recorded traversing an illuminated, 80 cm long glass walkway. A minimum of three complete crossings per session were recorded and two individuals blinded to the genotype of the mice viewed the videos and scored the motor function of the hind limbs using the Toyama Mouse Score (TMS) [25]. Movements of the left and right hind limbs were evaluated independently. The scores were averaged for each animal. Scoring of motor function was performed pre-injury, and then at: 1, 3, 7, 14, 21, 32 and 42 days post-injury (dpi).

Histology
Animals were anesthetized and transcardially perfused on the indicated days postinjury with 0.1 M phosphate-buffered saline (PBS) followed by 4% paraformaldehyde (PFA) in 0.1 M PBS. Dissected spinal cords were cryo-protected by sequential equilibration in 10%, 20% and 30% sucrose. Transverse sections (20 µm) of spinal cords were collected rostral, central and caudal of injury site. For immunofluorescence staining, sections were washed in PBS, post fixed with 4% PFA for 10 min, washed 3 × 5 min with PBS, permeabilized with 0.5% Triton X-100 in PBS for 15 min, blocked with 10% goat or donkey serum, 1% bovine serum albumin and 0.2% Triton X-100 in PBS for 1 h at room temperature, then stained with: rabbit polyclonal  Table S1). Tissue sections were counterstained with 1 µg/mL DAPI (Jackson ImmunoResearch, West Grove, PA, USA) in antibody buffer for 30 min. Sections were rinsed in Milli-Q water for 3 × 5 min and cover slipped with ProLong Gold Antifade (Invitrogen, P36930).
For immunofluorescence imaging and quantification, at least three sections per animal at 2 mm rostral (−2 mm), epicenter and 2 mm caudal (+2 mm) of injury site were collected, stained and imaged. All sections used for signal quantification were processed simultaneously, using the same batch of reagents. Images were acquired using a Leica TCS SP5 laser scanning confocal microscope with consistent settings across conditions. All analyses were performed in sequential scanning mode to prevent cross-bleeding between channels. Quantification of intensity and cell counts was performed using the Measurement and Analyze Particles functions in FIJI software (version 1.41, National Institutes of Health, Bethesda, MD, USA) (supplementary code 1 & 2.ijm; Supplemental Figure S1) [26].

RNA Isolation and cDNA Synthesis
Animals were anesthetized and transcardially perfused at 3 dpi with 0.1 M PBS. Spinal cords were removed and an 8 mm section (4 mm rostral to caudal of lesion center) was collected and snap frozen in liquid nitrogen. Total RNA was isolated following homogenization in Trizol as per manufacturer's instructions. Samples were then collected and spun through a HighBind ® RNA spin column (Omega Biotek) and processed according to the manufacturer's instructions (E.Z.N.A. ® Total RNA Kit I). RNA was eluted in 35 µL RNAse-free water.

RNA Sequencing and Analysis
RNA-seq was performed at the University of Rochester Genomics Research Center using RNA isolated from the spinal cords of 4 TG2-A-cKO and 4 TG2fl/fl mice collected 3 dpi. Analysis was also performed on 3 TG2-A-cKO and 3 TG2fl/fl mice without injury. RNA concentration was determined with the NanopDrop 1000 spectrophotometer (NanoDrop, Wilmington, DE, USA) and RNA quality was assessed with the Agilent Bioanalyzer (Agilent, Santa Clara, CA, USA). The TruSeq RNA Sample Preparation Kit V2 (Illumina, San Diego, CA, USA) was used for next generation sequencing library construction according to the manufacturer's protocols. In brief, mRNA was purified from 100 ng total RNA with oligo-dT magnetic beads and fragmented. First-strand cDNA synthesis was performed with random hexamer priming followed by second-strand cDNA synthesis. End repair and 3 adenylation was then performed on the double stranded cDNA. Illumina adaptors were ligated to both ends of the cDNA, purified by gel electrophoresis and amplified with PCR primers specific to the adaptor sequences to generate amplicons of approximately 200-500 bp in size. The amplified libraries were hybridized to the Illumina single end flow cell and amplified using the cBot (Illumina, San Diego, CA) at a concentration of 8 pM per lane. Single end reads of 100 nt were generated for each sample. Sequenced reads were cleaned according to a rigorous pre-processing workflow (Trimmomatic-0.32) before mapping them to mouse reference genome (GRCm38.p4) with STAR-2.4.2a (https://github.com/alexdobin/STAR, accessed on 26 August 2021). Cufflinks2.0.2 (cuffdiff2-Running Cuffdiff) was used with the gencode M6 mouse gene annotations to perform differential expression analysis with an FDR cutoff of 0.05 (95% confidence interval). Preliminary evaluation of the RNA-seq data was carried using Functional Enrichment Analysis Tool (http://www.funrich.org/download/) [27], STRING (https://string-db.org/) and gene ontology pathway analysis with Enrichr [28][29][30].

Statistical Analysis
Data were graphed and analyzed using GraphPad Prism software. Statistical significance was tested by two-way ANOVA followed by Bonferroni's post hoc comparisons of TG2-A-cKO vs. TG2fl/fl and VA4 versus control data sets. Data are expressed as mean ± SEM. The level of significance was set to less than 0.05.

Deletion of TG2 from Astrocytes Significantly Improves Functional Recovery after SCI
For these studies, mice were scored using the TMS scale before injury and then at the indicated days after injury up to 42 days [25]. Using the TMS scale, each blinded observer was able to isolate each attribute of gait, score each attribute and assign a score for functional recovery. TMS categories include frequency and mobile extent of ankle movement; movement in knee and hip; sole touch at rest and in step, coordination, parallel or rotative movement in step; and trunk stability. Each scorer totaled their score for the left hind limb and right hind limb. Both scores were averaged for the final value. The results of these measures are shown in Figure 1 and clearly demonstrate that deletion of TG2 from astrocytes significantly improves functional recovery. By 7 dpi, TG2-A-cKO mice demonstrate a significantly higher TMS score, with an average of 5 points higher than TG2fl/fl mice. While TG2fl/fl mice also demonstrate a gradual improvement over time, animals with astrocyte-specific deletion of TG2 achieve and maintain a significantly higher TMS score through day 42 post-injury.

Statistical Analysis
Data were graphed and analyzed using GraphPad Prism software. Stat icance was tested by two-way ANOVA followed by Bonferroni's post hoc co TG2-A-cKO vs. TG2fl/fl and VA4 versus control data sets. Data are express SEM. The level of significance was set to less than 0.05.

Deletion of TG2 from Astrocytes Significantly Improves Functional Recovery a
For these studies, mice were scored using the TMS scale before injury an indicated days after injury up to 42 days [25]. Using the TMS scale, each blin was able to isolate each attribute of gait, score each attribute and assign a sc tional recovery. TMS categories include frequency and mobile extent of ankl movement in knee and hip; sole touch at rest and in step, coordination, paral movement in step; and trunk stability. Each scorer totaled their score for the l and right hind limb. Both scores were averaged for the final value. The res measures are shown in Figure 1 and clearly demonstrate that deletion of TG cytes significantly improves functional recovery. By 7 dpi, TG2-A-cKO mice a significantly higher TMS score, with an average of 5 points higher than T While TG2fl/fl mice also demonstrate a gradual improvement over time, anim trocyte-specific deletion of TG2 achieve and maintain a significantly highe through day 42 post-injury.  TG2fl/fl and TG2-A-cKO mice at 12-16 weeks of age (12-15 mice in each group) were anesthetized and their spinal cords moderately contused bilaterally at T9. Motor function was assessed prior to injury and at the days indicated post-injury using the Toyama mouse score (TMS) [25]. Data are expressed as an average of the right hind limb and left hind limb. Data were analyzed by two-way ANOVA with a post hoc analysis using Bonferroni's multiple comparisons test. ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Deletion of TG2 from Astrocytes Results in a Significant Attenuation of Reactive Gliosis after SCI
As TG2 is rapidly upregulated in astrocytes in response to CNS injury [9], we next examined histological measures of CNS injury and reactive gliosis in spinal cord sections from mice at 3 and 7 dpi. GFAP immunoreactivity was significantly reduced both 2 mm proximal (−2 mm) and at the epicenter of the injury in TG2-A-cKO mice compared to TG2fl/fl mice that express normal levels of astrocytic TG2. This decrease in immunoreactivity was observed at both 3 and 7 dpi (Figure 2). We also found that immunoreactivity of another important component of the glial scar, the CSPG NG2 [14], was significantly reduced in the lesions of TG2-A-cKO mice as compared to TG2 expressing mice ( Figure 3). This reduction in glial scar CSPG was born out by a reduction in the number of SOX9 positive astrocytes. SOX9 is a transcription factor that upregulates genes associated with glial scar formation [16], including CSPGs such as NG2 (Cspg4) [31]. The number of SOX9 expressing cells, was reduced both at 3 and 7 dpi at the epicenter of injured TG2-A-cKO mice; numbers were also significantly lower at 3 dpi at 2 mm rostral and at 7 dpi at 2 mm caudal from the injury site ( Figure 4).
Lastly, to determine if there was increased neuron survival post SCI in the TG2-A-cKO mice, the number of NeuN positive cells were determined. At 3 dpi, there was a slight but significant increase in the number of NeuN positive cells in the TG2-A-cKO mice caudal to the lesion center. However, by 7 dpi, there was no significant difference in NeuN ( Figure 5).

Deletion of TG2 from Astrocytes Results in an Upregulation of Genes Following SCI
To begin to understand how deletion of TG2 from astrocytes impacts gene expression after SCI, we carried out whole tissue RNA-seq. At 3 dpi, RNA was extracted from 5 mm sections of the spinal cord centered on the site of injury from TG2-A-cKO and TG2fl/fl mice (n = 4). Intriguingly, the 37 genes that were differentially expressed in the TG2-A-cKO mice after injury were all upregulated compared to TG2fl/fl mice ( Figure 6, Supplemental Table S3). Unexpectedly, gene ontology-based pathway analysis using Enrichr revealed 22 out of 39 GO terms of biological processes with an adjusted p-value of less than 0.05, are involved in lipid and fatty acid uptake, storage and metabolism (Supplemental Table S4). A STRING network representation shows the interconnectedness and preponderance of lipid metabolism pathways (Figure 7). We then validated the expression of three key genes that are involved in lipid processing, lipoprotein lipase (Lpl), perilipin 1 (Plin1) and fatty acid binding protein 4 (Fabp4) via qPCR and found that all three were significantly upregulated in the spinal cords from the TG2-A-cKO mice at 3 dpi (Figure 8).   Lastly, to determine if there was increased neuron survival post SCI in the TG2-A-cKO mice, the number of NeuN positive cells were determined. At 3 dpi, there was a slight but significant increase in the number of NeuN positive cells in the TG2-A-cKO mice caudal to the lesion center. However, by 7 dpi, there was no significant difference in NeuN ( Figure 5).

Deletion of TG2 from Astrocytes Results in an Upregulation of Genes Following SCI
To begin to understand how deletion of TG2 from astrocytes impacts gene expression after SCI, we carried out whole tissue RNA-seq. At 3 dpi, RNA was extracted from 5 mm sections of the spinal cord centered on the site of injury from TG2-A-cKO and TG2fl/fl mice (n = 4). Intriguingly, the 37 genes that were differentially expressed in the TG2-A-cKO mice after injury were all upregulated compared to TG2fl/fl mice (Figure 6, Supplemental Table S3). Unexpectedly, gene ontology-based pathway analysis using Enrichr revealed 22 out of 39 GO terms of biological processes with an adjusted p-value of less than 0.05, are involved in lipid and fatty acid uptake, storage and metabolism (Supplemental Table S4). A STRING network representation shows the interconnectedness and preponderance of lipid metabolism pathways (Figure 7). We then validated the expression of three key genes that are involved in lipid processing, lipoprotein lipase (Lpl), perilipin 1 (Plin1) and fatty acid binding protein 4 (Fabp4) via qPCR and found that all three were    . Deletion of TG2 from astrocytes results in an upregulation of genes involved in lipid uptake and metabolism following SCI. Five mm of spinal cord surrounding the lesions were collected from TG2-A-cKO and TG2fl/fl mice at 12-16 weeks of age (4 mice in each group) 3 days post-injury, RNA isolated and used for RNA-seq. Thirty-seven genes were increased in injured spinal cord from TG2-A-cKO compared to TG2fl/fl mice. Graphic representation of log fold changes in gene expression in spinal cords from TG2-A-cKO is shown.

Treatment of Wild Type Mice with the Irreversible TG2 Inhibitor VA4 Significantly Improves Functional Recovery after Acute Spinal Cord Injury
In a previous study, we had found that treatment of astrocytes with VA4 phenocopied the effects of TG2 deletion, i.e., they were more resistant to OGD-induced cell death [13]. To determine if VA4 was also efficacious in vivo, we administered VA4 (or vehicle only as a control) to wild type mice starting at 1 h post SCI and then 24 and 48 h later. Remarkably, we found that by 7 dpi the VA4 treated mice showed significantly better motor function compared to mice treated with vehicle only (Figure 9). As with our TG2-A-cKO mouse model, wild-type control animals do not recover past an average TMS score of 18-19, while VA4-treated animals achieve a recovery of 28 points, very similar to TG2-A-cKO animals ( Figure 1). Cells 2021, 10, x FOR PEER REVIEW 11 of 16 Figure 8. Deletion of TG2 from astrocytes increases the expression of genes that facilitate neuronastrocyte metabolic coupling following SCI. TG2-A-cKO and TG2fl/fl mice at 12-16 weeks of age (4 mice in each group) were collected at 3 days post-injury, spinal cords removed and a 5 mm section of the spinal cord surrounding the lesion was used to prepare RNA followed by reverse transcription and for qPCR. Fatty acid binding protein 4 (Fabp4) is a carrier protein of fatty acids and likely facilitates the transport of fatty acids from neurons to astrocytes. Lipoprotein lipase (Lpl) degrades lipoprotein particles and thus upregulation would allow more fatty acids to enter the astrocytes and perilipin 1 (Plin1) increases the storage capacity of lipid droplets in astrocytes. Gapdh was used as a control. N = 3 for each group. ** p < 0.01, unpaired t-test.

Treatment of Wild Type Mice with the Irreversible TG2 Inhibitor VA4 Significantly Improves Functional Recovery after Acute Spinal Cord Injury
In a previous study, we had found that treatment of astrocytes with VA4 phenocopied the effects of TG2 deletion, i.e., they were more resistant to OGD-induced cell death [13]. To determine if VA4 was also efficacious in vivo, we administered VA4 (or vehicle only as a control) to wild type mice starting at 1 h post SCI and then 24 and 48 h later. Figure 8. Deletion of TG2 from astrocytes increases the expression of genes that facilitate neuronastrocyte metabolic coupling following SCI. TG2-A-cKO and TG2fl/fl mice at 12-16 weeks of age (4 mice in each group) were collected at 3 days post-injury, spinal cords removed and a 5 mm section of the spinal cord surrounding the lesion was used to prepare RNA followed by reverse transcription and for qPCR. Fatty acid binding protein 4 (Fabp4) is a carrier protein of fatty acids and likely facilitates the transport of fatty acids from neurons to astrocytes. Lipoprotein lipase (Lpl) degrades lipoprotein particles and thus upregulation would allow more fatty acids to enter the astrocytes and perilipin 1 (Plin1) increases the storage capacity of lipid droplets in astrocytes. Gapdh was used as a control. N = 3 for each group. ** p < 0.01, unpaired t-test. [13]. To determine if VA4 was also efficacious in vivo, we administered VA4 (or vehicle only as a control) to wild type mice starting at 1 h post SCI and then 24 and 48 h later. Remarkably, we found that by 7 dpi the VA4 treated mice showed significantly better motor function compared to mice treated with vehicle only (Figure 9). As with our TG2-A-cKO mouse model, wild-type control animals do not recover past an average TMS score of 18-19, while VA4-treated animals achieve a recovery of 28 points, very similar to TG2-A-cKO animals (Figure 1).  Motor function was assessed prior to injury and at the days indicated post-injury using the Toyama mouse score (TMS) to assess locomotor function [25]. Data are expressed as an average of the right hind limb and left hind limb. Data were analyzed by two-way ANOVA with Bonferroni's multiple comparisons test, *** p < 0.001, **** p < 0.0001.

Discussion
Astrocytes play an indispensable role in determining outcomes subsequent to SCI. When the CNS is injured, astrocytes undergo a process called reactive astrogliosis [4]. Acutely, this response is a defensive reaction aimed at limiting tissue damage. However, astrogliosis often results in processes that may inhibit functional recovery or fail to provide adequate support to promote axonal outgrowth [4,5,[32][33][34]. Conversely, transplantation of non-reactive astrocytes can promote recovery after SCI [35][36][37] and, in an animal model of Parkinson, neurodegeneration [38]. In the injured CNS, astroglial response is a highly regulated process that is characterized by changes in astrocyte gene expression, morphology, proliferative state and mobility, and is dependent upon numerous factors such as location, severity and time after insult [1]. Astrocytic modulatory proteins play key roles in determining how they respond to injury and the extent to which they inhibit or support functional recovery. Identifying and understanding the role different modulatory proteins play in the injury response is essential for the development of intervention strategies to promote the ability of reactive astrocytes to support and not hinder functional recovery. One such modulatory protein that plays an important role in regulating astrocyte function and response to injury is TG2.
Previously we demonstrated that deletion of TG2 from astrocytes or treatment with the selective TG2 inhibitor VA4, significantly increases their survival in OGD conditions [6,13]. Deletion of TG2 from astrocytes also significantly increases their ability to protect neurons from ischemic-induced death [6,39]. Given the fact that TG2 regulates transcriptional processes [40,41], we carried out RNA-seq on wild type and TG2-/-astrocytes and found that pro-survival/pro-regenerative genes were upregulated in the TG2-/-astrocytes [39]. Although these findings indicate that deletion of TG2 from astrocytes is beneficial in the context of CNS injury, this needs to be tested in an in vivo model. Therefore, we used TG2-A-cKO and TG2fl/fl mice [39] in a contusion SCI model followed by analysis of motor function over a 42-day period. Remarkably, motor function in the mice with astrocytic TG2 deleted was significantly better starting at 7 dpi and continuing through 42 dpi. In fact, by 35 dpi many of the TG2-A-cKO mice had recovered most of their motor function based on the TMS score ( Figure 1). These data clearly indicate that in vivo deletion of TG2 from astrocytes enhances their supportive role post traumatic injury.
Immunohistochemical analyses of the spinal cords from the TG2-A-cKO and TG2fl/fl mice at 3 and 7 dpi suggests a modulation of the glial scar in animals with astrocytic deletion of TG2. In agreement with what we had observed previously [39], GFAP immunoreactivity as a general indicator of astrogliosis was significantly reduced in the TG2-A-cKO mice. Likewise, we find a significant reduction in the expression of the neurite growth inhibitory CSPG NG2 and the upstream regulator of inhibitory CSPGs, SOX9. Sox9 is associated with negative outcomes following CNS injury as it upregulates inhibitory CSPGs as well as other genes associated with glial scar formation [15], including NG2 (Cspg4) [31]. Deletion of SOX9 reduces inhibitory CSPG levels, increases reactive sprouting and improves outcomes after a contusion SCI [16,17]. In line with these previous studies, and the functional recovery observed in TG2-A-cKO mice, the number of SOX9 positive cells and NG2 immunofluorescence were significantly reduced in the TG2-A-cKO mice at both 3 and 7 dpi compared to TG2fl/fl mice. It has been suggested that NG2 plays a role in preventing axon regeneration across newly formed glia scars [34]. An immunohistochemical analyses of post mortem samples from control and injured/lesioned human spinal cords revealed that NG2 was present in the evolving astroglial scar, and therefore it was suggested that it played a role in inhibiting successful CNS regeneration [14]. While the direct role of reactive astrocytes in the production of CSPGs remains unclear [32], our finding of reduced SOX9 and NG2 strongly indicate that deletion of TG2 from astrocytes results in an environment that is more permissive for successful regeneration following a CNS injury.
To begin to understand how deletion of TG2 from astrocytes impacts the genetic landscape following SCI, thereby promoting regeneration, we carried out RNA-seq on spinal cords from TG2-A-cKO and TG2fl/fl mice at 3 dpi. Thirty-seven genes were differentially expressed in the TG2-A-cKO spinal cords following SCI and all were upregulated. This is in line with the finding that, when TG2 is deleted from astrocytes, the majority of genes that are differentially expressed are upregulated [39]. Using the FunRich interaction network analysis tool [27] (see Supplemental Figure S2), it was found that the upregulated genes were primarily involved in energy pathways, fatty acid metabolism and lipid uptake and storage, which is also illustrated in the STRING analysis ( Figure 7) and gene ontology analysis with Enrichr (Supplemental Table S4). At first glance, these results were puzzling; however, upon further investigation, these results are quite exciting given that astrocyte-neuron metabolic coupling is crucial for maintaining healthy neurons. SCI induces neuronal hyperactivity along with glial activation creating a maladaptive environment [42] as hyperactive neurons release toxic fatty acids. These fatty acids need to be transferred to astrocytes in lipid particles for astrocytes to protect the neurons [19], and both Plin1 and Lpl, which are significantly increased in injured TG2-A-cKO spinal cords, may play a role in fatty acid uptake by astrocytes. Furthermore, Lpl contributes to the degradation of triglyceride particles and release of fatty acids, which plays a significant role in maintaining energy balance in astrocytes and neurons [18]. At the same time, increased Plin1 expression may increase the storage capacity of lipid droplets in astrocytes [43] which would protect neurons from oxidative stress, and promote a more regenerative environment [44]. Further, prior studies have shown that the ability to transport lipids to astrocytes for lipid droplet formation is essential to maintain healthy neurons [20]. The increased expression of the genes involved in fatty acid metabolism and lipid uptake and storage in injured TG2-A-cKO spinal cords could also promote remyelination [45]. Additional studies will be needed to determine whether the increased expression of genes involved in fatty acid uptake and metabolism contribute to improved recovery in TG2-cKO animals through re-uptake of lipids released post-injury, improved astrocyte-neuron metabolic coupling and/or increased remyelination. Nonetheless, our results indicate that TG2 ablation in astrocytes promotes a more regenerative phenotype.
In vitro, both deletion of TG2 from astrocytes and treatment with VA4 significantly attenuated OGD-induced cell death [13]. Therefore, the final aspect of this study was to determine if the TG2 inhibitor VA4 would improve motor function after SCI in a manner similar to what was observed with astrocytic TG2 deletion. For these studies wild type mice were administered either 15 mg/kg VA4 or vehicle 1 h after SCI and then again at 24 and 48 h. Remarkably, this simple three dose paradigm with VA4 significantly improved motor function. VA4 is an irreversible inhibitor that bears an acrylamide "warhead", which is quickly becoming the most popular warhead [46] in novel targeted covalent inhibitors on the market. What is critically important to the specificity of the drug is the "targeting domain", which in the case of VA4 is highly specific for TG2; VA4 does not inhibit any other transglutaminase tested [46]. Consistent with the low toxicity profile in vitro towards neurons, and the high potency specificity for inhibiting the transamidating activity of TG2 [21], we did not observe any apparent toxic side effects in the VA4 treated animals. Whether VA4 is acting locally at the site of injury or alters the injury response through effects in the periphery, or both, remains to be seen. However, both VA4 and our genetic model of TG2 ablation in astrocytes yielded similar results, i.e., improved functional outcomes. These findings suggest that selective, irreversible TG2 inhibitors have the potential to be considered for the treatment of SCI and other acute CNS injuries. Future studies will continue to explore the mechanisms of action and window of treatment for the use of VA4 in improving outcomes following SCI.