Pharmacological Inhibition of Astrocytic Transglutaminase 2 Facilitates the Expression of a Neurosupportive Astrocyte Reactive Phenotype in Association with Increased Histone Acetylation

Abstract

Astrocytes play critical roles in supporting structural and metabolic homeostasis in the central nervous system (CNS). CNS injury leads to the development of a range of reactive phenotypes in astrocytes whose molecular determinants are poorly understood. Finding ways to modulate astrocytic injury responses and leverage a pro-recovery phenotype holds promise in treating CNS injury. Recently, it has been demonstrated that ablation of astrocytic transglutaminase 2 (TG2) shifts reactive astrocytes towards a phenotype that improves neuronal injury outcomes both in vitro and in vivo. Additionally, in an in vivo mouse model, pharmacological inhibition of TG2 with the irreversible inhibitor VA4 phenocopied the neurosupportive effects of TG2 deletion in astrocytes. In this study, we extended our comparisons of VA4 treatment and TG2 deletion to provide insights into the mechanisms by which TG2 attenuates neurosupportive astrocytic function after injury. Using a neuron-astrocyte co-culture model, we found that VA4 treatment improves the ability of astrocytes to support neurite outgrowth on an injury-relevant matrix, as we previously showed for astrocytic TG2 deletion. We hypothesize that TG2 mediates its influence on astrocytic phenotype through transcriptional regulation, and our previous RNA sequencing suggests that TG2 is primarily transcriptionally repressive in astrocytes, although it can facilitate both up- and downregulation of gene expression. Therefore, we asked whether VA4 inhibition could alter TG2's interaction with Zbtb7a, a transcription factor that we previously identified as a functionally relevant TG2 nuclear interactor. We found that VA4 significantly decreased the interaction of TG2 and Zbtb7a. Additionally, we assessed the effect of TG2 deletion and VA4 treatment on transcriptionally permissive histone acetylation and found significantly greater acetylation in both experimental groups. Consistent with these findings, our present proteomic analysis further supports the predominant transcriptionally repressive role of TG2 in astrocytes. Our proteomic data additionally unveiled pronounced changes in lipid and antioxidant metabolism in astrocytes with TG2 deletion or inhibition, which likely contribute to the enhanced neurosupportive function of these astrocytes.

Keywords: astrocytes; lipid metabolism; neurite outgrowth; proteomics; transcriptional regulation.

Figure 1

VA4 inhibition of TG2 facilitates the ability of WT astrocytes to promote neurite outgrowth on a chondroitin sulfate proteoglycan (CSPG) growth-inhibitory matrix. (a) Schematic of neurite outgrowth experimental paradigm including VA4 treatment. Created with BioRender.com. (b) Representative images of MAP2 staining of neurons without paired astrocytes, neurons paired with vehicle (DMSO)-treated astrocytes, or neurons paired with VA4-treated astrocytes (scale bar = 20 µm). (c) Quantification of length of longest neurite for neuron experimental groups on CSPG matrix. Shown as mean and SEM (n = 86–111 neurons per group from two biological replicates, Kruskal Wallis test ** p < 0.01, **** p < 0.0001).

Figure 2

Irreversible inhibition of TG2 with the drug VA4 reduces interaction between TG2 and Zbtb7a. (a) Input controls of V5-TG2 and FLAG-Zbtb7a transfected into HEK 293TN cells treated with VA4 or vehicle control (DMSO). (b) Immunoprecipitation of V5-TG2 pulls down less FLAG-Zbtb7a in cell lysates that were treated with VA4. In (a,b) the position at which molecular weight markers (kDa) migrated is indicated at the left of the immunoblots. (c) Quantification of the amount of FLAG-Zbtb7a pulled down normalized to the amount of V5-TG2 immunoprecipitated. Treatment of cells with VA4 resulted in a significant reduction in the Zbtb7a co-immunoprecipitated with TG2 compared to DMSO control. Shown as mean and SEM (n = 5 samples per condition from 4 independent biological replicates, unpaired t-test *** p < 0.001). Original images can be found in Supplementary File S1.

Figure 3

TG2−/− astrocytes have significantly greater acetylation of histone H3 at lysine residue 9 (H3K9ac) compared to WT astrocytes. (a) Representative Western blot of WT and TG2−/− astrocyte lysates probed for Histone H3 acetylated at lysine residue 9 (H3K9ac). The position at which molecular weight markers (kDa) migrated is indicated at the left of the immunoblots. (b) Quantification of H3K9ac levels in WT and TG2−/− astrocytes. TG2−/− astrocytes showed significantly greater acetylation of H3K9 than WT astrocytes (~60%). Shown as mean and SEM (n = 1–4 samples per condition from 4 independent biological replicates, unpaired t-test ** p < 0.01). Original images can be found in Supplementary File S1.

Figure 4

WT astrocytes treated with VA4 show significantly greater H3K9 acetylation compared to DMSO vehicle control treated WT astrocytes. (a) Western blot of 10 μM VA4- and DMSO-treated WT astrocyte lysates probed for H3K9ac. The position at which molecular weight markers (kDa) migrated is indicated at the left of the immunoblots. (b) Quantification of H3K9ac levels showed significantly greater acetylation in VA4-treated WT astrocytes compared to DMSO-treated WT astrocytes (~65%). Shown as mean and SEM (n = 3 samples per condition from 2 independent biological replicates, unpaired t-test * p < 0.01). Original images can be found in Supplementary File S1.

Figure 5

TG2−/− and VA4-treated astrocytes share significant alterations in proteins associated with lipid metabolic and antioxidant pathways. (a) Volcano plot of significant differentially regulated proteins comparing TG2−/− to WT astrocyte cultures (n = 5 samples per condition from 1 biological replicate) with thresholds set for log2 fold changes at ±0.5 and for adjusted p value < 0.05, and (b) DAVID GO Biological Process analysis of up- and downregulated proteins. (c) Volcano plot of significant differentially regulated proteins comparing VA4-treated to DMSO-treated WT astrocyte cultures (n = 6 samples per condition from 2 independent biological replicates) with thresholds set as above, and (d) DAVID GO Biological Process analysis of up- and downregulated proteins. (e,f) Enrichr transcription factor enrichment of up- and downregulated proteins in (e) TG2−/− astrocytes (f) VA4-treated astrocytes (Fisher Exact Test, * p < 0.05). (g) Venn diagram showing overlap of significant differentially regulated proteins in both TG2−/− and VA4 datasets.

Figure 6

Proposed mechanism by which TG2 modulates the neuroprotective phenotype of reactive astrocytes. In stressed astrocytes, TG2 is able to move in and out of the nucleus. While in the nucleus, TG2 can interact with Zbtb7a and Sin3a, resulting in increased histone deacetylase (HDAC) activity and neurosupportive gene repression. If TG2 is deleted from the astrocytes, or the astrocytes are treated with VA4, the recruitment of Zbtb7a/Sin3a/HDAC to the DNA is diminished, resulting in de-repression of neurosupportive genes leading to a more neurosupportive phenotype. The “?” indicates that the effect of VA4 treatment on the nuclear/cytosolic localization of TG2 is unclear. Created with BioRender.com.