Transcriptomic signature of astrocytes isolated from the cortex of sham, PTE + and PTE- mice
Dysregulation of astrocytes has been implicated in the development and progression of epilepsy (9, 10). Understanding how the transcriptomic landscape is altered in astrocytes will uncover key molecular pathways and reveal potential therapeutic targets to prevent the development of post-traumatic epilepsy (PTE). We assessed transcriptomic changes by performing RNA-seq on astrocytes isolated from the ipsilateral and contralateral cortex of sham and severe CCI-injured mice at four months post-injury. Mice were stratified based on seizure development (PTE + vs PTE-), as previously described (11). Transcriptomic analysis was performed in conjunction with gene ontology pathway enrichment to detect differentially expressed genes (DEGs) and biological processes (BP) respectively. Ipsilateral cortical astrocytes showed 207 DEGs that were altered in CCI-injury compared to sham and 106 DEGs altered in CCI-injured PTE + compared to PTE- mice, while 11 genes were common between the comparison groups (Fig. 1A). The top five (Log2FC) upregulated genes in sham vs. CCI were Cntf, Ahcy, Tmem14c, Ucma, and C3, while the top five (Log2FC) downregulated genes were Mag, Tmem125, Hapln2, Opalin, and Mog (Fig. 1D and 1E). When comparing PTE + vs. PTE- genes, we found that the top five upregulated genes (Log2FC) were Pam, Pfas, Gm14296, Ttyh2, and Gli3, while the top five downregulated genes (Log2FC) were Hspa1a, Fos, Gm10197, Clca3a1, and Ccl4 (Fig. 1F and 1G). Gene ontology of the ipsilateral injured cortex vs. sham revealed changes in genes associated with gliogenesis, glial development, and differentiation (Fig. 1B), while myeloid cell differentiation and homeostasis are predominant biological processes associated with gene changes in PTE + mice (Fig. 1C). These data highlight key astrocytic genes altered in the ipsilateral cortex of PTE + mice including upregulation of peptidylglycine α-amidating monooxygenase (PAM), an essential enzyme for the synthesis of amidated neuropeptides.
Evidence suggests an important functional and structural basis of contralateral interictal activity in focal, unilateral epilepsy (12). We next addressed associated changes in the contralateral cortical astrocytic transcriptome in our focal CCI model (11). CCI injury showed changes in 421 genes compared to sham, and when comparing PTE + vs PTE- injured-mice we observed 457 genes that were altered in astrocytes from animals that developed generalized seizures (Fig. 2A). Only nine genes were commonly expressed between these groups. The top five (Log2FC) upregulated genes between sham and CCI were Tmem14c, Ceacam1, Nupr1, Slc35c1, and Rbm46, and the top five (Log2FC) downregulated genes were Gm10800, Timp1, Slc47a1, Prg4, and Lcn2 (Fig. 2D and 2E). The top five (Log2FC) upregulated genes in PTE + astrocytes compared to PTE- were Slc47a1, Pnpla3, Sgpp2, Penk, and Vwf, while the top five (Log2FC) downregulated genes were annotated genes without a canonical name including Gm10801, Gm10197, Gm20390, Gm10217, and Gm10320 (Fig. 2F and 2G). Gene ontology of the post-injury contralateral cortex vs sham predominantly showed top GO terms associated with RNA splicing and stability as well as catabolic processes (Fig. 2B), while GO terms in PTE + mice related to ameboidal-type cell migration, vasculogenesis, and endothelial processes (Fig. 2C). Interestingly, Tmem14c, a key player in mitochondrial heme metabolism (13), was a top gene upregulated in both ipsilateral and contralateral astrocytes in CCI-injured compared to sham mice. These findings show significant changes in astrocytic genes in the uninjured contralateral cortex of PTE + mice that are distinct from the ipsilateral cortex.
Ingenuity pathway analysis (IPA) was performed to comprehensively analyze differential gene expression and infer predicted downstream effects for potential mechanistic targets. A graphical summary of the major canonical pathways, upstream regulators, and biological processes shows the significant genes altered in PTE + cortical astrocytes reveal predicted inhibition of Il6 and other pathways with inferred relationship to Il6 including IFNG, IGF1, IL1A, IL5, IL17A (Fig. 3A) based on top ready analysis (Fig. 3B) and other significantly altered genes. The top 8 canonical pathways show a negative z-score for Il17A, p38 MAPK, NOD1/2, IL-6 and a positive z-score for PKR in interferon (IFN) signaling (Fig. 3C). The IPA summary of contralateral PTE + cortical astrocytes shows predicted activation of serine peptidase inhibitor Kazal type 1 (SPINK1) general cancer pathway, as well as IL-15 production, KLF4, SOX4 and migration of phagocytes/myeloid cells (Fig. 3D). Based on top ready analysis genes (Fig. 3E) and other statistically significant genes, top canonical pathways include SPINK1, IL-15 production with positive z-scores (Fig. 3F).
A broader IPA synopsis of PTE + cortical astrocytes yielded predicted activation of several pathways. Analysis of contralateral PTE + cortical astrocytes showed downregulation of genes such as MT1, SOX9, and PPARA, with upregulation of several genes including SOX17, and CD34 among others via predicted activation of AGT, EPAS1, SOX4, NFKB1, KLF4, and IL10RA (Fig. 4A). Analysis of ipsilateral PTE + cortical astrocytes demonstrated downregulation of Ccl4 via predicted activation of APOE, and other genes such as Fos were downregulated via predicted activation of SCD, SFTPA1, TSC2, and TWF1 (Fig. 4B). Overall, we found that 217 genes were downregulated in contralateral PTE + cortical astrocytes, 84 genes were downregulated in ipsilateral PTE + cortical astrocytes, and 12 genes were downregulated in both contralateral and ipsilateral samples (Fig. 4C). In contrast, 240 genes were upregulated in contralateral PTE + cortical astrocytes, 22 genes were upregulated in ipsilateral PTE + cortical astrocytes, and 5 genes were upregulated in both contralateral and ipsilateral samples (Fig. 4D). In total, 17 genes were differentially expressed in both contralateral and ipsilateral samples of PTE + cortical astrocytes. These included downregulation of GM10197, FOS, TTC38, NNAT, MT3, MACROD1, MGST1, NDUFA6, CST3, GNPTG, ID4, and TOR1AIP1, and upregulation of GDF10, NCKAP1, ASAP1, UBR2, and PAM (Fig. 4E). Pam was the most upregulated gene, and the most downregulated genes were GM10197, and Fos based on log2FoldChange in both hemispheres.
Finally, we previously identified Cystatin 3 (Cst3), a highly expressed gene in astrocytes (14), to be altered in the ipsilateral and contralateral hippocampus of PTE + astrocytes (11). We demonstrate here that this is the only gene influenced across all brain regions in isolated astrocytes from PTE + mice. Immunohistochemistry of Cst3 on GFAP-expressing peri-lesional cortical astrocytes is evident at 4 months post-CCI injury in both PTE- and PTE + mice (Fig. 4F-L), however, we observed its expression was more prominent on IBA1-positive microglia. The cell-type specific role of Cst3 may be a viable target for further exploration.