Targeting SRSF3 restores immune mRNA translation in microglia/macrophages following cerebral ischemia

Abstract

We recently described a novel ribosome-based regulatory mechanism/checkpoint that controls innate immune gene translation and microglial activation in non-sterile inflammation orchestrated by RNA binding protein SRSF3. Here we describe a role of SRSF3 in the regulation of microglia/macrophage activation phenotypes after experimental stroke. Using a model-system for analysis of the dynamic translational state of microglial ribosomes we show that 24 h after stroke highly upregulated immune mRNAs are not translated resulting in a marked dissociation of mRNA and protein networks in activated microglia/macrophages. Next, microglial activation after stroke was characterized by a robust increase in pSRSF3/SRSF3 expression levels. Targeted knockdown of SRSF3 using intranasal delivery of siRNA 24 h after stroke caused a marked knockdown of endogenous protein. Further analyses revealed that treatment with SRSF3-siRNA alleviated translational arrest of selected genes and induced a transient but significant increase in innate immune signaling and IBA1+ immunoreactivity peaking 5 days after initial injury. Importantly, delayed SRSF3-mediated increase in immune signaling markedly reduced the size of ischemic lesion measured 7 days after stroke. Together, our findings suggest that targeting SRSF3 and immune mRNA translation may open new avenues for molecular/therapeutic reprogramming of innate immune response after ischemic injury.

Graphical abstract

Rahimian and colleagues show that SRSF3 orchestrates a ribosome-based regulatory checkpoint controlling innate immune gene translation and microglial activation after stroke. Timely targeting of SRSF3 using siRNA approach may open new avenues for therapeutic/molecular reprogramming of immune response in brain ischemia.

Introduction

Ischemic stroke is one of the leading causes of mortality and morbidity worldwide.^1,2 To date, in spite of extensive research efforts, the clinical treatments remain poorly effective. Growing evidence suggests that inflammation is one of the key components of the pathobiology of stroke. Experimentally and clinically, stroke is followed by an acute and prolonged inflammatory response characterized by the activation of resident glial cells, production of inflammatory cytokines, and leukocyte infiltration in the brain.^1,2 A current view is that the acute and/or early inflammatory response after stroke may contribute to expansion of the ischemic lesion, while at the subacute and/or delayed stages post-stroke inflammation mediates infarct resolution, as well as remodeling and repair.^3,4 Thus, timely modulation/targeting of the ischemia-induced immune responses may represent a relevant therapeutic strategy for stroke.

Microglial cells are the principal innate immune cells of the brain implicated in the initiation as well as resolution of the brain immune response to ischemic injury.^5,6,7,8,9 A consensus today is that once activated, microglia can acquire a wide repertoire of the context-dependent immune phenotypes ranging from the classical pro-inflammatory to alternative, anti-inflammatory polarization phenotypes.^10,11,12,13 However, the molecular mechanisms that drive changes in the microglia activation phenotypes following ischemic injury are not yet fully elucidated.

We recently described a novel ribosome-based checkpoint mechanism involved in the control of innate immune response and microglia/macrophage activation. The mechanism is based on a selective 3′UTR-mediated translational repression of highly expressed, ribosome-bound immune mRNAs orchestrated by RNA binding protein SRSF3.¹⁴ The SRSF3-mediated translational arrest led to a marked dissociation of mRNA and protein networks in activated microglia.¹⁴ SRSF3 belongs to the Serine-arginine-rich (SR) proteins family that contains several RNA binding proteins with a functional implication in RNA metabolism.¹⁵ Like other SR proteins, SRSF3 is involved in alternative splicing events; however, recent reports emphasize its role in the mechanisms involved in post-transcriptional regulation, such as mRNA export, surveillance, stability, and translation.^16,17

We recently showed that SRSF3 is implicated in the control of innate immune gene translation following an acute lipopolysaccharide (LPS)-mediated challenge. In the current study, we investigate the role of SRSF3 in the control of innate immune response and microglia/macrophage activation in the context of sterile inflammation such as ischemic brain injury. Here, we took advantage of our well-characterized in vivo model system for analysis of the dynamic translational state of microglia/macrophage ribosomes and performed a high-affinity immunoprecipitation followed by a parallel analysis of ribosome-bound mRNAs and peptides. Using this strategy, we identified mRNA and protein signatures associated with the acute microglia/macrophage activation after stroke.

We report here that highly up-regulated innate immune genes are not translated resulting in a marked dissociation of mRNA and protein molecular signatures 24 h after transient middle cerebral artery occlusion (MCAO). Of note, the apparent translational arrest was limited to a cluster of previously identified innate immune genes including Lcn2 and Saa3, etc., known to be targeted for regulation by SRSF3 (it activated phosphorylated form-pSRSF3).¹⁴ Next, the post-stroke inflammatory response was characterized by a robust increase in levels of pSRSF3/SRSF3, its expression being restricted to CD11b/IBA1-positive microglia/macrophages. Targeting SRSF3 in the brain by intranasal delivery of small interfering RNA (siRNA) 24 h after stroke resulted in a marked knockdown of endogenous protein starting 4 days after initial injury. Our results revealed that treatment with SRSF3-siRNA (1) alleviates translational arrest of targeted genes; (2) induces de novo synthesis of selected immune proteins such as LIRB4, LCN2, SAA3, and TIMP1; and (3) induces transient increase/boost in the TLR2 response/signaling that was therapeutically beneficial after stroke. Together, our results suggest a role for SRSF3 as a regulator of innate immune response and microglia/macrophage activation in post-ischemic inflammation. Furthermore, our results revealed that a timely targeting of SRSF3 and immune mRNA translation may open new avenues for therapeutic modulation of innate immune response in brain ischemia.

Results

Characterization of the CD11brGFP transgenic mice following MCAO

To assess the mRNA and protein signatures of microglia activation in response to ischemic injury in vivo, we took advantage of the CD11brGFP transgenic mouse model generated in our laboratory.¹⁴ As previously described, this mouse line expresses FLAG-EGFP fused to the N terminus of the large subunit ribosomal protein L10a (FLAG-EGFP-RPL10a), under transcriptional control of the human CD11b promoter (Figure 1A), thus allowing isolation of ribosome-bound mRNAs and associated peptides by a simple high-affinity immunoprecipitation 24 h after stroke induction (Figure 1B). As revealed in Figures 1C and 1D, and as well as in our previous work, in the brain, the human CD11b promoter was shown to efficiently drive microglial transgene expression.^6,14 Indeed, the double immunofluorescence analysis revealed that the CD11b-driven transgene (GFP staining) co-localized with the endogenous CD11b immunostaining in control conditions and 24 h following MCAO (Figures 1C and 1D). In our experiments, the transcriptome/proteome analysis was performed 24 h following MCAO because at this time point the infiltration of the peripheral monocyte/macrophage is not significant in comparison to the control non-stroked brains.^18,19 Consistent with previous evidence, this was further evaluated by flow cytometry analysis. As shown in Figures 1E and 1F, we examined the dynamics of immune cells (microglia and monocytes) in the post-stroke brain by flow cytometry. The FACS analysis indicates that 24 h after MCAO the percentage of resident microglia (CD11b⁺ CD45^Low) is significantly increased in comparison with the control non-stroked brains (Figure 1E). Next, as indicated in the lower panels (Figure 1F) the CD45^high population was gated to compare the percentage of monocyte (LyC6⁺) cells in MCAO and control condition. Our analysis revealed that the monocyte population in the ipsilateral side of the brain-stroke lesions was a bit higher (an increase of 4%) but not significantly increased in comparison with the control non-stroked brains, thus further suggesting that 24 h after MCAO more than 90% of the CD11b+ cells in the brain are resident microglia.

Figure 1.

Characterization of the CD11brGFP transgenic mice following MCAO

(A) Schematic representation of FLAG/EGFP-tagged murine Rpl10a construction under control of the CD11b promoter. I, intervening sequence (IVS); II, SV40 polyA. (B) Schematic representation of the experimental timeline. (C and D) Representative confocal microscopy images of EGFP expression in brain sections in control mice (sham operated) (C) or in MCAO mice 24 h following stroke induction (D), magnification ×20; Scale bar, 50 μm. GFP immunostaining colocalized with CD11b. White squares locate the regions where the pictures have been taken in (C) and (D). Images were taken with ×10 and ×40 objectives. (E and F) Twenty-four hours following stroke induction, the FACS analysis demonstrated that the population of resident microglia (CD11b⁺ CD45^Low) is higher in comparison with control non-stroked brains (sham operated) (upper panels, ∗p = 0.0301 determined by unpaired t test with Welch’s correction, n = 5 for MCAO group and n = 4 for control group) (E). Also, as indicated in lower panels (F) the CD45^high population was gated to compare the percentage of monocyte (LyC6⁺) cells in MCAO and control condition (sham) (p = 0.0613 determined by unpaired t test with Welch’s correction, n = 5 for MCAO group and n = 4 for control group). All graphs are expressed as mean ± SEM.

mRNA profiling of microglia depicts clusters of highly upregulated immune transcripts

Next, we sought to establish the RNA and protein molecular signatures of the activated microglia population after stroke. Twenty-four hours following MCAO, the stroked cortical area homogenates were divided in two aliquots and immunoprecipitated using an anti-FLAG agarose affinity resin. The polyribosome complexes were used either for (1) mRNA extraction followed by Affymetrix Mouse Genome 430 analysis or (2) peptide extraction followed by a high-resolution label-free proteomic analysis. The different steps are schematically demonstrated in Figure 2A. Statistical analyses revealed that brain ischemia significantly regulated the expression of 1,824 transcripts (5.29% of total transcripts) compared with the control non-stroked mice (1.2 > Fold change > −1.2 and p < 0.05). Of those, 1,132 transcripts were upregulated (62.06%) whereas 692 transcripts were downregulated (37.94%) (Figure 2B). Hierarchical clustering of differentially expressed genes (1.2 > Fold change > −1.2 and p < 0.05) provides transcriptional patterns of the MCAO and control non-stroked samples and shows distinctive clusters between the two conditions (Figure 2C). The majority of the transcripts included in these clusters were highly upregulated in stroke condition when compared with controls and were involved in immune and cytokine responses (see Figure 2C, zoom 1, zoom 2, and zoom 3). Comparison with our previously described transcriptome profile of activated microglia following a standard LPS challenge¹⁴ revealed certain overlap among the highly upregulated immune-related transcripts found after stroke (23 common genes) (Figure 2D), suggesting that the similar gene activation patterns are associated with innate immune responses mediated by a sterile and non-sterile inflammation. The most highly upregulated transcripts after MCAO are as follows: leukocyte immunoglobulin-like receptor, subfamily B, member 4 (Lilrb4) (22.3-fold change), chemokine (C-C motif) ligand 3 (Ccl3) (15.6-fold change), and heme oxygenase 1 (Hmox1) (11.46-fold change) (Figure 2C, zoom 1). Lilrb4a belongs to a family of innate immune receptors with immunomodulatory functions. These receptors interact with MHC class I molecules on antigen-presenting cells and initiate a negative signal that inhibits the stimulation of an immune response.²⁰ The second most upregulated transcript was Ccl3. This cytokine belongs to the CC chemokine family that is involved in the acute inflammatory responses.²¹ It is well established that Ccl3 gene expression by microglia modulates disease severity in different models of neurodegenerative diseases.^22,23 The third most upregulated transcript was Hmox1. This inducible enzyme degrades heme to carbon monoxide, free iron, and biliverdin, is involved in the cell defense against oxidative stress, and it has been shown that it might be a new therapeutic target for neuroprotection.²⁴ Another highly upregulated transcript after stroke was tissue inhibitors of matrix metalloproteinases (Timp1) (8.95-fold change) (Figure 2B, zoom 2). It has been shown that Timp1 knockout mice display increased levels of MMP-9 activity and stroke volume following focal cerebral ischemia while adenoviral-mediated gene transfer of Timp1 to the ischemic brain elicits neuroprotection.^25,26 Figure 2E summarizes the top regulated genes. Interestingly, the top upregulated mRNAs were not detected as peptides (Figure 2E). Raw transcriptomic data are presented in Table S1.

Figure 2.

Upregulated immune transcripts following MCAO

(A) Schematic representation of the EDTA-TRAP protocol after stroke. (B) Pie chart 1 illustrates the percentage of regulated genes (purple) vs. non-regulated genes (gray) and pie chart 2 illustrates the percentage of the upregulated vs. dowregulated (blue) genes. (C) Hierarchical clustering of microglia analyzed with the Affymetrix 2.0 ST chip 24 h after stroke induction visualized by heatmap. Experiments are conducted in three biological replicates (n = 6 mice per group). (D) Venn diagram showing overlapping of highly upregulated mRNAs after MCAO or LPS challenge. (E) List of the top 10 upregulated immune-related mRNAs after stroke. Differences between transcriptomic and proteomic data analysis: Upregulated immune-related transcripts presented in cluster 1 (zoom 1), cluster 2 (zoom 2), and cluster 3 (zoom 3) were not observed in the sequenced peptide list.

Highly upregulated immune-related transcripts are not translated in acutely activated microglia after stroke

To compare the microglial molecular signature at mRNA level with an actual cell specific proteome, we performed a parallel transcriptome and proteome analysis 24 h after stroke. Briefly, we observed that brain ischemia significantly altered the expression levels of 472 proteins of which 42.8% were upregulated and 57.2% were downregulated by at least 1.2-fold. The top upregulated immune-related transcripts were not detected among sequenced peptides (Figure 2E). This was further validated by western blot analysis. Consistent with the results obtained by mass spectrometry, the most upregulated mRNAs including Lilrb4a, Ccl3, Hmox1, Timp1, Lcn2, and Saa3 were not regulated and/or significantly increased at the protein level (Figures 3A and 3B). As shown in Figures 3A and 3B, Western blot analysis from the whole brain tissue homogenates revealed no significant changes in the expression levels of LIRB4a, CCL3, HMOX1, TIMP1, LCN2, and SAA3, 24 h after MCAO when compared with the control non-stroked group. Next, we used volcano plots to compare expression patterns of two datasets, i.e., microglia transcriptome and proteome 24 h after stroke. As revealed in Figures 3C and 3D, with the exception of genes related to the Serpina cluster, there is a marked divergence in the top upregulated mRNAs (Figure 3C) and peptides (Figure 3D). The expression level of the top upregulated peptides detected by quantitative mass spectrometry was further validated and confirmed by western blot analysis (Figures 3D and 3E). In alignment with the results obtained by mass spectrometry, quantitative analysis revealed a significant increase in expression levels of SERPINA 3k, SERPINA 3n, S100a9, PFKFB3, and HSPA1 (Figures 3E and 3F). Taken together, our results suggest that translation of the selected upregulated transcripts is suppressed in microglia/macrophages 24 h after stroke. It is notable that the observed translational arrest was restricted to a cluster of highly upregulated immune-related genes whereas other upregulated and/or unregulated ribosome-bound mRNAs were normally translated and detected by quantitative mass spectrometry. Raw proteomic data are presented in Table S2.

Figure 3.

Highly upregulated immune-related genes are not translated in activated microglia after stroke

(A) Western blot analysis of highly upregulated mRNA expression after stroke. (B) Quantitative analyses of western blot indicate that the protein levels of highly upregulated mRNA did not change significantly between stroke and non-stroked (sham) group (n = 4; LILRB4, p = 0.58; CCL3, p = 0.11; Hmox1, p = 0.89; TIMP1, p = 0.53; LCN2, p = 0.83; and SAA3, p = 0.7). (C and D) Volcano plots of mRNA identified by transcriptomic analysis and proteins identified by proteomic analysis. Gray dots represent non-significantly regulated mRNA/proteins (−1.5 < fold change < 1.5 and/or p > 0.05); Blue dots represent significant downregulated mRNAs/proteins (fold change < −1.5 and p < 0.05) and red dots represent significantly upregulated mRNAs/proteins (fold change >1.5; p < 0.05). (E) Validation of top upregulated proteins detected by mass spectrometry by western blot. (F) Quantitative analysis of western blot showed a significant increase of top upregulated proteins after stroke (SERPINA3k [2.89 ± 0.054-fold change, n = 4, ∗∗∗p = 0.0001], S100a9 [3.74 ± 0.30-fold change, n = 4, ∗∗∗p = 0.0001], PFKFB3 [2.097 ± 0.013-fold change, n = 4, ∗∗∗p < 0.0001], SERPINA3n [12.32 ± 0.047-fold change, n = 4, ∗∗∗p = 0.0001], and HSPA1a [HSP70] [1.26 ± 0.026-fold change, n = 4, ∗∗p = 0.0035]). All samples were analyzed 24 h after MCAO. All graphs are expressed as mean ± SEM and p value was determined by Student’s two-tailed t test.

Divergent molecular RNA and protein signatures in MCAO-activated microglia

To further assess microglial transcriptome and proteome molecular signatures after MCAO, we used Cytoscape and ClueGO cluster analysis^27,28,29 and functionally characterized the top 100 significantly upregulated transcripts and peptides. As shown in Figures 4A and 4B, we identified the top 20 biological functions related to upregulated transcripts. The majority of these biological functions are related to processes associated with inflammation including terms such as the following: inflammatory response, defense response, response to cytokines, innate immune response, and TLR2 signaling pathway. In contrast to a highly specialized immune mRNA response, the top 20 biological functions related to upregulated proteins revealed more diversified biological response to ischemic injury. The top biological functions were associated with regulation of complement cascade, neurogenesis, housekeeping functions like spliceosome, vesicle recycling, endocytosis, regulation of proteolysis, and regulation of metabolic process. One term related to inflammation, named “immune system,” was also associated with peptide analysis (Figure 4B). Next, we compared the genes associated with the term-function of immune response at RNA and the protein level (Figures 4C and 4D). Surprisingly, further analyses of the immune networks at RNA protein levels revealed that they upregulate two sets of functionally distinct genes. At 24 h after stroke, the RNA network was highly enriched in genes/transcripts related to innate immune response including several cytokines and chemokines, whereas at the protein level, we observed an enrichment of genes related to the biology of the complement immune system. The complement system is an essential part of innate immune response and usually elicits protection via eliminating debris and pathogens. However, uncontrolled activation of the complement system can aggravate inflammatory and degenerative responses in different pathological contexts.³⁰ Finally, as revealed in Figure S1, analysis of the top downregulated function additionally further confirmed functional divergence of RNA and protein signatures in activated microglia. While top downregulated RNA functions were highly enriched in terms related to chemokine signaling, the top downregulated functions at the protein level reflected enrichment in genes related to neurodegeneration and phagosome functions. Together, our results suggest that during the first 24 h after stroke, activated microglial cells acquire functionally distinct mRNA and protein immune signatures.

Figure 4.

Distinct mRNA and proteomic signature in activated microglia after stroke

Top 20 upregulated biological transcriptomic (A)/proteomic (B) functions from Top 100 upregulated mRNAs/proteins and summarized in bar plot based on p value (−log₁₀[p value]). (C and D) Molecular signature of the immune system term at the transcriptomic (C) and proteomic (D) levels. All transcriptomic and proteomic analyses were performed in parallel, 24 h after MCAO.

Marked induction of pSRSF3/SRSF3 in ischemic brain is restricted to IBA+ cells

We previously reported that in the context of innate immune challenge, SRSF3 may act as a suppressor of innate immune transcript translation in acutely activated microglia/macrophages.¹⁴ Given that we observed a similar pattern of translational repression of the selected and highly regulated ribosome-bound immune mRNAs after stroke, we next asked whether activation of microglia after ischemic injury is associated with the changes in SRSF3 activation patterns. As revealed in Figures 5A and 5B, western blot analysis revealed a significant increase in pSRSF3 level in ipsilateral brain lysates 24 h following MCAO. The levels of pSRSF3 then gradually decreased reaching a baseline level at 72 h after stroke (see supplementary evidence Figure S2). The double immunofluorescence analysis further confirmed a marked increase of pSRSF3 in IBA1-positive cells. Analysis of the DAPI staining revealed a cytoplasmic localization of pSRSF3 (Figures 5D and 5E). As stroke incidence is only slightly higher in men,³¹ and given the known biological sex differences in male vs. female immune response after stroke, we also investigated the pSRSF3 expression patterns in female mice as well as in the female mice subjected to chronic estrogen deficiency (chronically overiectomized mice we developed as a model of menopause), reflecting the group with increased risk and vulnerability after stroke. In all experimental groups, pSRSF3 levels were analyzed 24 h after MCAO. As revealed in the supplementary evidence (see Figures S3A–S3F), western blot analysis 24 h after stroke revealed that the post-stroke pSRSF3 levels were significantly increased in male and OVX mice (the most robust increase) when compared with levels observed in the ischemic brains of young females (observed downregulation after stroke). This suggests that, early after stroke, the SRSF3-mediated mechanisms implicated in the control of immune gene translation may be preferentially activated in male and aging female brains, but not in the brains of young females, a group developing normally a more protective microglia phenotype after stroke.³²

Figure 5.

Expression of pospho-SRSF3 is increased in the ipsilateral side of ischemic brain

(A) One day after MCAO, the levels of pSRSF3 increased significantly in ipsilateral brain lysates in comparison with control non-stroked (sham operated) animals. (B) Quantitative analysis showed a significant increase of pSRSF3 24 h after stroke (n = 4, ∗∗∗p = 0.0001). (C) Schematic representation of coronal brain section. The black square shows the region where the images were taken. (D and E) The immunofluorescence studies also indicated the upregulation of pSRSF3 in CD11b-positive cells in ipsilateral cortex after stroke (E) when compared with control non-stroked mice (D). Magnification ×20; Scale bar, 50 μm. (F and G) Western blot (F) and its quantitative analysis (G) demonstrated that the expression of SRPK1 (n = 4, ∗∗∗p = 0.0008) and not CLK1(n = 4, p ˃ 0.05) increased 1 day following MCAO. (H) Potential SRSF3-binding sites on the 3′UTR of upregulated immune-related mRNAs after stroke. All graphs are expressed as mean ± SEM and p value was determined by Student’s two-tailed t test.

All SR proteins have one or two N-terminal RNA recognition motifs (RRMs) and one C-terminal RS domain. Peptide mapping and cellular analyses indicate that phosphorylation of the RS domain by SRPKs and/or CLKs controls the subcellular distribution of the SR family.³³ Interestingly, our results demonstrated that the expression of SRPK1 but not CLK1 increases following MCAO (Figures 5F and 5G), suggesting that upregulation of SRPK1 after ischemic insult might be responsible for the observed increase in pSRSF3 expression, which in turn may facilitate its binding to 3′UTR of highly regulated innate immune genes and consequently arrests translation.¹⁴ Of note, we previously reported that 3′UTRs of the innate immune genes regulated by SRSF3 are highly enriched in the SRSF3 putative binding sites. To address this, here we analyzed the 3′UTRs of immune genes that were highly upregulated at the mRNA level but whose expression was not regulated at the protein level (see Figures 2 and 4). We took advantage of the RBP map bioinformatics tool to map the 3′UTRs of Lilrb4a, Ccl3, Hmox1, and Timp1.³⁴ We observed that the 3′UTR of these mRNAs are indeed enriched in putative binding sites of SRSF3 (Figure 5H), suggesting that SRSF3 may act as a regulator of immune gene translation and microglia/macrophages in the context of post-stroke inflammation.

SRSF3-siRNA treatment transiently increases the TLR2 signaling and microglia activation after MCAO

We recently showed that SRSF3 controls innate immune cascade in activated microglia/macrophages in vivo following LPS challenge.¹⁴ Based on our results, we hypothesized that the stroke-induced increase in pSRSF3 levels is responsible for the observed translational repression of the selected immune transcripts. Therefore, targeting endogenous SRSF3 would alleviate translational arrest of the selected genes in vivo after MCAO and would induce de novo synthesis of immune proteins. To silence endogenous protein, 15 μg of siRNA directed against SRSF3 or Scrambled-siRNA were delivered intranasally starting 24 h after MCAO. A timeline of the experiment is schematically presented in Figure 6A. Based on our previous work, the efficiency of knockdown was examined by western blot, 4 days after intranasal delivery. As shown in Figures 6B and 6C, western blot analysis revealed a robust (80%) decrease in the expression levels of the endogenous SRSF3 in siRNA-treated mice when compared with controls. To visualize the effects of SRSF3 knockdown on innate immune response after stroke, we took advantage of the TLR2-luc-GFP reporter mice previously generated and validated in our in our laboratory.⁷ In this transgenic model, luciferase and GFP are co-expressed under transcriptional control of the murine TLR2 gene promoter, and thus the innate immune response/microglial activation can be visualized in real time, as a bioluminescence signal, from the brains of living mice using a high-resolution/high-sensitivity charge-coupled device (CCD) camera. In this context, the effects of SRSF3/Scrambled-siRNA treatment on the TLR2 signal induction were investigated over a 7-days period. Of note, our previous work demonstrated that after MCAO, over 90% of the TLR2 signal arises from the IBA1 positive microglial cell.^7,35 As expected, at baseline conditions the low intensity bioluminescence signal was restricted to the olfactory bulb area. After stroke, a robust signal arising from the ischemic part of the brain was detected in all animals. The signal peaked 24 h after stroke and started to decline between 3 and 5 days after initial ischemic injury for the MCAO (black line) and Scrambled-siRNA groups (green line). Interestingly, the quantitative analysis of photon emissions revealed that SRSF3-siRNA treatment, administrated 1 day following MCAO surgery via intranasal delivery, had a direct effect on innate immune signaling causing a transient boost in the TLR2 bioluminescence signal at day 5 after stroke; i.e., 4 days after intranasal delivery (Figures 6D and 6E). As indicated in Figures 6D and 6E, there were no statistical differences in the bioluminescence signal intensities between the Scrambled and MCAO-only groups, clearly suggesting that the selected Scrambled-siRNA does not affect the TLR2 signal at any time points. Importantly, the transient SRSF3-siRNA-mediated increase in TLR2 signaling was also associated with a marked increase in microglia activation. As revealed by immunofluorescence analysis 5 days following MCAO (Figures 6F and 6G), IBA1+ microglia cells surrounding cortical ischemic lesions display a more activated ameboid morphology following SRSF3-siRNA treatment as compared with the Scrambled-siRNA experimental group (Figures 6F and 6G). The observed increase in microglia activation was further confirmed by western blot analysis revealing a significant increase in expression of IBA1 in cortical brain lysate of MCAO-SRSF3-siRNA treated animals when compared with the corresponding control (Figures 6H and 6I). Next, we investigated whether the observed transient boost and/or increase in TLR2 signaling and microglia/macrophage activation following SRSF3-siRNA treatment is caused by the effects of SRSF3 on microglia/macrophage immune gene expression profiles. To assess the effects of SRSF3 on the selected immune protein expression profile directly on the CD11+ microglia/macrophages, we used CD11b magnetic beads and purified microglia/macrophages from the ipsilateral lesions. CD11b-positive cells were afterward lysed and analyzed by immunoblot (Figures 6J and 6K). A cell-specific western blot analysis revealed that the expression levels of LILRB4a, HMOX1, and TIMP1 were significantly increased in microglia/macrophages that were pulled-down from the ipsilateral lesions of MCAO-SRSF3-siRNA-treated mice when compared with control, the Scrambled siRNA-treated mice. This further confirms that the siRNA-mediated knockdown of endogenous SRSF3, post-MCAO, alleviates translational repression of the selected immune genes and thus may reprogram microglia protein expression profile in vivo (Figures 6J and 6K). Finally, we investigated whether targeting SRSF3 may have a therapeutic potential after stroke. Although increasing an acute inflammatory response after stroke may be harmful, increasing evidence suggests that a delayed phase of post-ischemic inflammatory processes is important for the initiation of repair mechanisms.^3,36 Our previous work suggests that targeting delayed inflammatory response after stroke may have some benefits.³⁷ As presented in the experimental diagram (Figure 6A), stroke area was assessed 7 days after MCAO. To our surprise, a single intranasal dose of SRSF3-siRNA delivered 24 h after stroke exerted marked beneficial effects on the size of ischemic lesions (Figures 6L and 6M). The quantitative analysis of the stroke area measured on Cresyl-Violet-stained brain sections revealed a significant decrease in the size of the ischemic lesion in SRSF3-siRNA-treated mice 7 days post MCAO (Figure 6M). Hence, targeting SRSF3 and immune mRNA translation may have a therapeutic potential after stroke.

Figure 6.

SRSF3 regulates innate immune cascade following brain ischemia

(A) Schematic representation of the experimental timeline. (B) Five days after MCAO, the levels of total SRS3 in ipsilateral brain lesions were measured and analyzed by western blot. (C) Quantitative analysis showed an 80% SRSF3 downregulation after SRSF3-siRNA delivery to the stroke area (n = 4, ∗∗∗p = 0.0001). (D) In vivo imaging of TLR2 induction after MCAO. Quantitative analysis of photon emission showed that SRSF3-siRNA treatment (single dose administration of 15 μg siRNA 1 day after stroke) has a notable effect on TLR2 signal in subacute phase of stroke. Indeed, TLR2 signal, in the SRSF3-siRNA group, increases with time to become significantly higher than the scrambled group at day 5 (n = 5–7, ∗p = 0.0022). Multiple t test was used to analyze in vivo imaging results over the time. (E) Representative photographs at 2, 3, and 5 days after stroke. The color calibrations at the right are photon counts. (F and G) Immunofluorescence studies show that 5 days after stroke microglia are more ameboid in cortical lesions of SRSF3-siRNA-treated mice (G) when compared with the control non-stroked group, magnification ×20; Scale bar, 50 μm (F). (H) Western blot and its quantitative analysis (I) indicate the increase of IBA1 expression in cortical brain lysate of SRSF3-siRNA animals (n = 4, ∗p = 0.04). Images were taken with ×40 objective. (J) Western blot analysis performed on the purified microglia/macrophage isolated from the fresh brain homogenates by magnetic CD11b microbeads. (K) Quantitative analysis showed that the SRSF3 knockdown following MCAO alleviated translational repression of the highly upregulated transcripts. Five days after MCAO LILRB4a (1.54 ± 0.13-fold change, n = 4, ∗p = 0.0065), HMOX1 (2.61 ± 0.13-fold change, n = 4, ∗∗∗p = 0.0001) and TIMP1 (3.68 ± 0.078-fold change, n = 4, ∗∗∗p = 0.0001) were increased significantly in microglia/macrophage that were pulled down from ipsilateral lesions of SRSF3-siRNA-treated mice. (L) Infarct volume was measured 7 days after MCAO by cresyl violet staining. (M) The size of ischemic lesions decreased after SRSF3-siRNA treatment (14.2%) when compared with the control (22.2%; Scrambled-siRNA) 7 days following MCAO (n = 8, ∗∗∗p = 0.0007). All graphs are expressed as mean ± SEM and p value was determined by Student’s two-tailed t test.

Discussion

A robust activation of microglial cells in response to ischemic injury is considered as a key event in the pathogenesis of stroke.^2,4,38 Over the years, we and others have shown that a timely activation of microglia and associated innate immune cascade are instrumental in the control of injury-induced CNS damage.^6,9,39,40,41 Normally essential for maintenance of the brain tissue homeostasis,^42,43 after ischemic injury, microglial cells are the first and the predominant cell population to be activated in the neuroanatomical area surrounding the brain infarction.^18,19 Given their capacity to develop a wide range of the context-dependent activation profiles, in this study we set out to decipher the molecular signatures of activated microglia 24 h after stroke. By using our well-characterized in vivo model-system for analysis of the dynamic translational state of microglial ribosomes with mRNAs as input and newly synthesized peptides as an output, we found a marked dissociation of mRNA and protein networks in acutely activated microglia after stroke. We report here that highly upregulated selected immune transcripts were not translated resulting in functionally diverging mRNA and protein molecular signatures in activated microglia. Importantly, as in the context of an acute LPS-mediated innate immune challenge, microglia activation after stroke was characterized by a robust increase in expression of RNA binding protein SRSF3. Based on our previous work¹⁴ and the current evidence, our data strongly suggest that the same ribosome-based checkpoint mechanism is implicated in the control of innate immune gene translation in activated microglia/macrophages in the context of sterile and non-sterile inflammation. This posttranscriptional regulation of immune response is orchestrated by RNA-binding protein SRSF3.

SRSF3 belongs to the Serine-arginine-rich (SR) protein family of RNA-binding proteins with the functional implication in RNA metabolism and regulation of multiple aspects of gene expression program.^15,16 The activities of SR proteins including SRSF3 are regulated by cycles of phosphorylation and dephosphorylation and changes in expression levels.⁴⁴ To date, a disruption of biological functions of SRSF3 has been mostly linked to the pathobiology of cancers. For example, increased levels of SRSF3 have been associated with increased cellular proliferation and cancer progression, while downregulation of SRSF3 in cancer cells has been associated with cellular maturation/senescence.^45,46,47,48 While to date, SRSF3 may represent a prime target in cancer,⁴⁸ little is known about its role in microglia biology and/or its implication in regulation of immunity in healthy and/or injured brain. We have recently shown that in the context of neuroinflammation and acute LPS-mediated innate immune challenge, SRSF3, by binding to 3′UTR of the selected highly upregulated innate immune genes acts as a regulator/suppressor of immune gene translation in activated microglia.¹⁴ Conversely, in our well-controlled experimental paradigm, targeted knockdown of SRSF3 alleviates translational arrest and induces de novo synthesis of immune proteins. As mentioned above, after stroke, as well as after LPS challenge, aberrant expression of SRSF3 is restricted to IBA+ immunoreactive cells suggesting a cell-type-specific regulatory and/or pathogenic function.

Following several disappointing attempts to translate anti-inflammatory therapeutics into the clinics, the question that arises here is: can we therapeutically target SRSF3 to restore more functional immunity? Does timely molecular reprogramming of immune cells, i.e., microglia/macrophages represent a valid strategy to limit neuronal damage after stroke? As brain revealed in our proof-of-concept experiment (see Figures 6A–6J), targeting SRSF3 in ischemic brain by using intranasal delivery of siRNA resulted in a marked knockdown of endogenous protein. As expected, decrease in SRSF3 levels consequently induced de novo synthesis of selected immune proteins. At a functional level, a knockdown of endogenous SRSF3 resulted in a transient and controlled boost in innate immune signaling, revealed as an increase in the TLR2 bioluminescence signal in reporter mice, and a significant increase in IBA1 immunoreactivity. It is noteworthy that transient boost of innate immunity peaked between days 4 and 5 after initial ischemic injury. Remarkably, transient silencing of SRSF3 by a single intranasal delivery of siRNAs 24 h post stroke resulted in a significant decrease in the size of the ischemic lesion 7 days after initial injury, further suggesting that a transient and delayed boost in innate immunity may be beneficial after stroke. Indeed, as revealed previously, upregulation of the TLR2 signaling in the subacute and chronic phase of stroke is neuroprotective following MCAO through induction of insulin-like growth factor (IGF-1) signaling³ and/or galectin-3-mediated microglia alternative activation.^9,49

Based on our findings, we propose that increased activity/expression of pSRSF3/SRSF3 in microglia/macrophages early after stroke suppresses translation of highly regulated immune genes, resulting in a marked divergence between cellular transcriptome and functional outcomes. Furthermore, our studies revealed that pSRSF3 expression levels are significantly upregulated in the ischemic brains of males and aged (OVX) female mice, two groups carrying an increased risk of stroke. Here it is worth noting that the additional cell-specific transcriptome/proteome analyses revealed that the activated microglia/macrophages from the brains of aged OVX mice after stroke upregulate a same cluster of the SRSF3-regulated transcripts including ccl3, lirb4, lcn2, and Hmox1 (R.G., R.R. H.B. and J.K. unpublished data). Although some additional experiments are needed, our results so far strongly suggest that SRSF3 may represent a valid target in both paradigms. Therefore, a timely therapeutic targeting of SRSF3 using siRNA and/or corresponding gene silencing approaches may open new avenues for molecular reprogramming of microglia/macrophage after stroke. Finally, in this study we used the intranasal route to deliver SRSF3-siRNA into the CNS. Intranasal delivery is a non-invasive and convenient method that rapidly targets therapeutics to the CNS bypassing the blood brain barrier and minimizing systemic exposure. Therefore, using this methodology together with the silencing of SRSF3 may lead to a development of novel therapeutic concepts in treatment of stroke.

Materials and methods

Ethical approval

All experimental procedures were approved (protocol no. 017–133/protocol no.2021-960) by the Université Laval Animal Care Ethics Committee and are in accordance with The Guide to the Care and Use of Experimental Animals of the Canadian Council on Animal Care.

Experimental animals

All experiments were done on two main transgenic lines: CD11brGFP transgenic mice (C57BL/6NCrl background) in which the transgene FLAG-EGFP-RPL10 is expressed under the transcriptional control of human CD11b promoter, and Toll-like receptor 2 (TLR2)-luc-GFP transgenic mice (B6 Albino background), in which luciferase and GFP reporters are driven under the transcriptional control of the murine TLR2 gene promoter. The TLR2-luc-GFP transgenic mice were detected by the amplification of the luciferase transgene as described previously.⁷ To avoid known effects of sex on immune response in stroke, the experiments were performed on adult 3- to 4-month-old male mice.⁹ Additional pathway validation experiments presented in the supplementary evidence were performed on 3- to 4-month-old female mice. As previously described,⁵⁰ ovariectomy (OVX) was performed in 2- to 3-month-old female mice. The MCAO procedure was performed 40 days after OVX.

Surgical procedures: Middle cerebral artery occlusion

Ischemic stroke was induced as described earlier,³ by occluding the middle cerebral artery (MCA) for 1 h followed by 1-day, 5-day, and 7-day reperfusion periods. Adult male mice (2–3 months old) of 20–25 g were used for stroke induction. The animals were anesthetized using 2% isoflurane. To avoid cooling, the body temperature was regularly checked and maintained at 37°C with a heating pad during and after surgery. Under an operating microscope, the left common carotid artery, external carotid artery (ECA), and the internal carotid artery were exposed through a midline neck incision and were carefully isolated from surrounding muscles. The proximal part of the ECA was isolated and carefully separated from the adjacent nerve plexus and veins, a 12-mm-long 6–0 silicon-coated monofilament with the tip diameter of 170–190 μm was inserted via the incision made on the proximal ECA into the ICA up to 9-mm length blocking the blood flow in the MCA, thus occluding the MCA.^3,6 The filament was left for 60 min in place without disturbing and after occlusion time, it was slowly removed from the artery and the incision on the ECA was sealed. The midline incision was closed by suturing the animal using biodegradable 6.0 suture filament and animals were injected with saline post-surgery. All animals were allowed free access to water and food before and after surgery. In the current study, all sham-operated animals (control groups) underwent surgery including all the manipulations except the incision on the ECA and occlusion of the MCA. The correct placement of the filament was confirmed by Laser Doppler measurements (PF5001, Perimed, Sweden). As previously described, to additionally confirm successful MCAO, at 6 and 24 h after surgery, the animals were examined for early neurological deficits. Based on these early screenings we excluded the animals that we had doubt about the efficacy of stroke induction.

Flow cytometry analysis

One day following MCAO induction, mice (n = 4–5) were transcardially perfused with ice-cold 1X DPBS to remove all blood from brain tissue. Lesion areas were enzymatically (Dispase II, 2U/ml, Sigma) and mechanically dissociated and filtered through a 70-μm cell strainer to have a single cell suspension (Becton Dickinson). To isolate mononuclear cell fraction, the single cell suspensions were loaded on a 30%-37%-70% Percoll tri-gradient (GE Healthcare) and centrifuged for 40 min at 300 × g. After centrifugation, mononuclear cells in-between the 37%–70% interface were collected and other two layers of supernatant were discarded. The collected cells were stained with anti-CD11b (PE), anti-CD45 (PerCP), anti-Ly6C (FITC), anti-Ly6G (APC-Cy7), and anti-CD3 (FITC) (BD Biosciences). Samples were analyzed on a flow cytometer FACStar Plus and further using FlowJo v10.6 the cells were gated using side and forward scatter to eliminate non-viable cells.⁵¹

TRAP protocol

We used the modified translating ribosome affinity purification (EDTA-TRAP) protocol.¹⁴ EDTA-TRAP is based on TRAP protocol described by Heiman and colleagues.^52,53 The transgenic line CD11brGFP developed in our laboratory expresses FLAG-EGFP-RPL10a transgene under the control of the human CD11b promoter. It allows us to use a cell-type-specific genetic targeting of transgene expression by the incorporation of the affinity tags, like FLAG and enhanced green fluorescent protein (EGFP), on the large ribosomal subunit protein L10a. FLAG tag was used for ribosome immunoprecipitation and EGFP was used for cell visualization in the tissue. Brain tissue from ipsilateral cortical and subcortical lesions and the corresponding penumbra were dissected and placed on ice-cold dissection buffer followed by a homogenization (10% w/v) in tissue lysis buffer. The homogenates were then centrifuged at 2,000 × g for 10 min at 4°C. Supernatants were transferred into new tubes to which, 1/9 sample volume of 10% NP-40 and 1/9 sample volume of 300 mM DHPC were added and mixed well, and it was incubated for 30 min at 4°C on an orbital shaker. The insoluble material was recovered by centrifugation at 20,000 × g for 10 min at 4°C. Each sample supernatant was equally divided into two aliquots (one aliquot to be used for mRNA extraction and the other for peptide elution). Each sample was mixed with anti-FLAG agarose affinity resin and incubated overnight at 4°C on an orbital shaker. The following day, the beads were recovered by centrifugation and washed three times with high-salt washing buffer (20 mM HEPES-KOH [pH 7.3], 200 mM KCl, 12 mM MgCl₂, 1% NP-40, 0.5 mM DTT, and 100 μg/mL cycloheximide). The beads pellet was used either for mRNA purification or peptide purification.¹⁴ Throughout the purification of the microglial ribosomal complex, cycloheximide and RNAse inhibitors were used to preserve the integrity of the complex.

Purification of mRNA from F/EGFP-RPL10a mice after TRAP protocol

After the last washing, the bead pellet was resuspended in 100 μL Nanoprep lysis buffer with beta-mercaptoethanol and incubated for 10 min at room temperature on an orbital shaker. The RNA isolation was done according to the kit manufacturer’s instructions (Absolutely RNA Nanoprep kit, Agilent Technologies). Three biological replicates were performed for each experiment (for each replicate, n = 6 per experimental group). Collected RNA was subjected to Affymetrix mouse gene chip.¹⁴

Affymetrix microarray

DNA microarray was performed by the Plateforme d’Expression Génique team in CHUL (Quebec City, Canada). Before the microarray, RNA quality and quantity were measured with an Agilent BioAnalyzer (Agilent Technologies) and an ND-1000 Spectrophotometer (NanoDrop Technologies), respectively. cDNA from the previously isolated microglial mRNAs using cRNA hybridization cocktail (Affmetrix) following Affymetrix Gene Chip WT cDNA Synthesis and Amplification Kit protocol. After washing, cDNA was stained and scanned with the Affymetrix GCS 3000 7G. Resulting data were analyzed on an Affymetrix Expression Console. To evaluate the regulation levels between ALS and control mice lumbar spinal cord microglial transcripts, we used a Transcriptome Analysis Console (Affymetrix). A transcript was considered to be regulated if the absolute fold change value between ALS and control samples was equal to or higher than 2 and its p value, lower than 0.05.

Purification of peptides from F/EGFP-RPL10a mice after TRAP protocol

After the last washing, all remaining wash buffer was removed and the bead pellets were resuspended in EDTA-elution buffer (10 mM HEPES-KOH [pH 7.3], 150 mM KCl, 5 mM MgCl₂, 20 mM EDTA, and protease inhibitors) and incubated for 30 min at room temperature on an orbital shaker. EDTA elution buffer was used to dissociate ribosomes and release nascent chain peptides. Eluate was recovered by centrifugation at 5,000 rpm for 15 min.¹⁴ Three technical replicates were performed for this experiment (n = 6 per experimental group).

Mass spectrometry analysis: Sample preparation

Samples were concentrated on a desalting column Amicon 3 kDa (Millipore) and washed three times with ammonium bicarbonate 50 mM. Protein concentration was determined by colorimetric Bradford assay. Equal amounts of protein were solubilized in the denaturation buffer. Then samples were heated to 95°C for 5 min in a solution of DTT and iodacetamide. Finally, 1 μg trypsin was added, and the mixture was incubated at 37°C, overnight. The precipitated sodium deoxycholate was eliminated by 10-min room temperature (RT) incubation and 5-min RT centrifugation at 16,000 × g. The supernatant was desalted on C18 Empore filter. Peptides were eluted in 80% ACN – 0.1% TFA, and dried in a speed vac.

Mass spectrometry analysis: Mass spectrometry

Samples were analyzed by nanoLC/MSMS as triplicates for statistical information. For each injection, 750ng of peptide samples were injected and separated by online reversed-phase (RP) nanoscale capillary liquid chromatography (nanoLC) and analyzed by electrospray mass spectrometry (ESI MS/MS). The experiments were performed with a Dionex UltiMate 3000 nanoRSLC chromatography system (Thermo Fisher Scientific/Dionex Softron GmbH, Germering, Germany) connected to an Orbitrap Fusion mass spectrometer (Thermo Fisher Scientific) equipped with a nanoelectrospray ion source. Mass spectra were acquired using a data-dependent acquisition mode using Thermo XCalibur software version 3.0.63. Full scan mass spectra (350–1,800 m/z) were acquired in the orbitrap using an AGC target of 4e5, a maximum injection time of 50 ms, and a resolution of 120,000. Each MS scan was followed by acquisition of fragmentation MS/MS spectra of the most intense ions for a total cycle time of 3 s (top speed mode). Dynamic exclusion of previously fragmented peptides was set for a period of 20 s and a tolerance of 10 ppm. Mass spectrometry analyses were performed by the Proteomics platform of the Eastern Quebec Genomic Center, CHU de Quebec, Canada. Database searching and Label Free Quantification Spectra were searched against a mouse proteins database (UniprotKB – taxonomy Mus musculus – 84,675 sequences) using the Andromeda module of MaxQuant software v. 1.5.0.25. Only unique and razor peptides were used for quantification. A protein was considered as quantifiable only if at least two replicate values in one of the two samples to compare were present. A protein was considered as variant if the fold change between the two compared samples was higher than 1.2 and the associated p value was lower than 0.05.

Cluego analysis

Data from gene chip Affymetrix or mass spectrometry were analyzed with ClueGo application (version 2.1.6) using the cytoscape environment (3.2.1). Differentially expressed genes/proteins (with corresponding fold changes and p values) were used to generate biological networks using different ontology sources like the Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, and WikiPathways. The GO interval was between 4 (Min level) and 11 (Max level). The Kappa score was 0.7. For the enrichment of biological terms and groups, we used the two-sided (Enrichment/Depletion) tests based on the hyper-geometric distribution. We set the statistical significance to 0.05 for transcriptomic result, and we used the Bonferroni adjustment to correct the p value for the terms and the groups created by ClueGO. The leading group term is based on %genes/term vs. cluster.

Administration of SRSF3-siRNA

Scramble-siRNA (Dharmacon; 15 μg/mouse) or SRSF3-siRNA treatment (Dharmacon; 15 μg/mouse) was administrated intranasally as a single dose 1 day following MCAO in anesthetized mice using in vivo jetPEI reagent (Polyplus) according to the manufacturer’s protocol. Briefly, in vivo jetPEI and siRNA were diluted separately in 10% glucose solution. Then, siRNA and PEI solutions were mixed and incubated for 15 min at RT for a total of 50 μL. Mice received 25 μL of solution in each nostril.¹⁴ Five days after stroke, mice were euthanized for biochemical and immunofluorescence experiments. Bioluminescence imaging was done in the time course of 7 days and stroke volume was measured 7 days following the induction of brain ischemia.

Brain tissue fixation

Mice after 24 h of stroke onset were treated with Scrambled-siRNA and SRSF3-siRNA, after they were anesthetized via an intraperitoneal (i.p.) injection of ketamine/xylazine (100–10 mg/kg). The mice thorax was opened after they underwent deep anesthesia, and a needle of 23G size connected to a drain tube was inserted into the left ventricle of the heart. After cutting the right atrium, mice were perfused with 20 mL of 1X PBS, followed by 20 mL of PBS-buffered 4% paraformaldehyde (PFA) (at pH 7.4) with a speed of 6 mL/min, using a syringe pump. After flushing with PBS and PFA, the brain was removed from the skull through dissection and the collected brains were fixed for 2 days in 4% PFA on shaker (for even PFA perfusion into the brain tissue), and later were transferred into phosphate-buffered 20% sucrose for 48 h. Brains were embedded in Tissue-Tek (OCT compound; Sakura), frozen at −80°C overnight, and cut into coronal sections (25 μm thick) with a cryostat and stored at −20°C.⁴⁹

Immunofluorescence (IF) analysis

Brains sections were blocked in 4% goat serum, 1% bovine albumin serum (BSA), 0.45% of Triton X-100 (Blocking solution) prepared in 1XPBS for 1 h at RT, and then incubated overnight at RT with primary antibodies diluted in blocking solution (refer to the primary antibodies listed in Table S1). After three washes with 1XPBS for 10 min each, the sections were then incubated in corresponding fluorescent goat secondary antiserum (as listed in Table S2) for 2 h (1:1,000), washed with 1XPBS three times, and were incubated with DAPI (1:1,000, Invitrogen) for 5 min. To remove the excess DAPI, they were washed again three times with 1XPBS and mounted using Fluromount G (ThermoFisher) and stored at 4°C until imaging. For the negative control, the tissue sections were incubated only with the secondary antibody solution. The immunohistochemical staining patterns obtained with all the antibodies used in our experiments are in agreement with the known expression patterns of the corresponding proteins, supporting the specificity of these antibodies.

Microscopy

Confocal microscope (Zeiss LSM T-PMT, 2012), Zeiss upright laser-scanning was used to acquire images. Images were further processed in ImageJ software v8.0. The image z stacks were collected with a 2 μm step using ×20 and ×63 objectives. The settings for PMT, laser power, gain were identical between experimental groups. The acquired images were processed in ZEN lite v2.6 software in all the figures.

Histological evaluation for the size of infarction

The size of infarction was evaluated at 7 days after MCAO in Scrambled-siRNA and SRSF3-siRNA treated animals. Brain sections (n = 8–9 mice/group) were pre-treated in ascending graded ethanol and stained with Cresyl-Violet for the development of violet coloration, then they were exposed to 95% Glacial acetic acid to differentiate the nucleus and the cytoplasm of the cells. The de-stained slides were then subjected to descending graded ethanol and Xylene and were cover-slipped with mounting media. The dried slides were digitized with TISSUEscope 4000 and the whole brain section was scanned at ×10 magnification. The area of infarction, i.e., the direct stroke area, was quantified with ImageJ (version 1.42q, NIH), and the infarct area was calculated as a percentage of stroke volume in both scrambled-siRNA and SRSF3-siRNA ipsilateral hemisphere.³⁵

In vivo bioluminescence imaging

The images were gathered with an IVIS Spectrum Imaging System (PerkinElmer, Hopkinton, MA). Before imaging sessions, the mice were intraperitoneally (i.p.) injected with D-luciferin (150 mg/kg IP, PerkinElmer), luciferase substrate, dissolved in 0.9% saline and sterile filtered through a 0.2-μm cellulose acetate filter. The mice were then anesthetized with 2% isoflurane in 100% O₂ at a flow rate of 2 L/min and placed in a heated, light-tight imaging chamber. Images were collected with a high-sensitivity CCD camera with wavelengths ranging from 300 to 600 nm. Exposure time for imaging was 1–2 min with different fields of view and an f/1 lens aperture. Images were analyzed as previously described, bioluminescence emission was normalized and displayed in physical units of surface radiance (photons · s⁻¹ cm⁻² · steradian⁻¹ [sr]). Light output was quantified by determining the total number of photons emitted per second with the use of Living Image 2.5 acquisition and imaging software (PerkinElmer). Region of interest measurements on the images were used to convert surface radiance (photons · s⁻¹ cm⁻² · sr⁻¹) to source flux or the total flux of photons expressed in photons/second. The data are represented as pseudocolor images indicating light intensity (red and yellow, most intense), which were superimposed over gray-scale reference photographs.^7,49,54 In imaging studies, we designed three different groups. The first group received SRSF3 siRNA through intranasal delivery in in vivo jetPEI reagent 1 day after MCAO. The second group received scrambled sequence through intranasal delivery in in vivo jetPEI reagent 1 day after MCAO, and the third group only received in vivo jetPEI reagent by intranasal delivery 1 day following MCAO. The comparison of second and third groups gives us information about possible nonselective effect of scrambled sequence like immunogenicity on TLR2 response following MCAO.

Western blotting

Total protein extracts were obtained from the ipsilateral hemisphere of the brain 1 day and 5 days after cerebral ischemia by homogenization in a 1× Cell Lysis Buffer (Ray Biotech kit) with Phosphatase inhibitor (Thermoscientific #1862495) and Protease inhibitor cocktail (Sigma, Cat log #P8340). The protein lysates from different samples were quantified by Bradford assay (detection at 595 nm). Samples containing 20 μg of protein were boiled in SDS-mercaptoethanol sample buffer, separated on 10% or 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and electrically transferred to nitrocellulose membranes. Non-specific binding was blocked by pre-incubation of the nitrocellulose membrane in PBS containing 0.1% Tween 20 (PBS-T) or TBS containing 0.1% Tween 20 (TBS-T) and 5% skimmed milk or BSA for 1 h. The nitrocellulose was then incubated overnight at 4°C with antibodies against the targeted proteins listed in Table S2.⁴⁹ Primary antibody was detected with horseradish peroxidase-conjugated anti-rabbit or anti-mouse antibody (1:1,000–1:5,000), and blots were developed using an enhanced chemiluminescence detection system (ECL kit; Thermo Fisher Scientific). The density of the specific bands was quantified using ImageJ software and normalized to β-actin as a housekeeping protein.

Isolation of brain CD11b-positive cells with magnetic beads

Mice treated with Scramble-siRNA or SRSF3-siRNA were transcardially perfused with ice-cold PBS, stroke areas were dissected and enzymatically digested by Dispase II (Sigma) 30 min at 37°C with a gentle trituration each 15 min. The undigested tissue clumps were removed by passing the cell suspension through a 70-μm cell strainer. The cells were centrifuged for 10 min at 300 × g, and the pellet containing the cells was resuspended in 30% Percoll (GE Healthcare) and centrifuged for 10 min at 700 × g. The supernatant containing the myelin was removed and the pelleted cells were washed with 1XHBSS and subjected to magnetic CD11b beads separation according to the kit manufacturer’s instructions (CD11b [Microglia], MicroBeads mouse; Miltenyi Biotec). Collected cells were subjected to western blot analysis.¹⁴

Statistical analysis

The Welch test was used to analyze proteomic findings and Student’s t test was employed for transcriptomic analysis. Student’s unpaired t test was employed to analyze western blot analysis of control and MCAO group or Scrambled-siRNA and SRSF3-siRNA treated groups. Multiple t test was used to analyze in vivo imaging results over the time. Two-way ANOVA was used to analyze FACS data. Statistical analyses were performed using the GraphPad Prism 8 software (GraphPad, La Jolla, CA). Data were expressed as the mean ± SEM.

Data and code availability

All data generated during this study are included in this article or are available on reasonable request from the corresponding authors.