Single-cell analysis identifies Ifi27l2a as a gene regulator of microglial inflammation in the context of aging and stroke in mice

Abstract

Inflammation is a significant driver of ischemic stroke pathology in the brain. To identify potential regulators of inflammation, we performed single-cell RNA sequencing (scRNA-seq) of young and aged mouse brains following stroke and found that interferon alpha-inducible protein 27 like 2 A (Ifi27l2a) was significantly up-regulated, particularly in microglia of aged brain. Ifi27l2a is induced by interferons for viral host defense and has been linked with pro-inflammatory cellular mechanisms. However, its potential role in neurodegeneration is unknown. Using a combination of cell culture, experimental stroke models in mice, and human autopsy brain samples, we demonstrated that induction of Ifi27l2a occurs in microglia in response to aging, ischemic stroke, and pro-inflammatory molecules. We further showed that induction of Ifi27l2a in microglia was sufficient to stimulate mitochondrial ROS production and promote a pro-inflammatory phenotype. Lastly, using an ischemic stroke model, we demonstrated that hemizygous deletion of Ifi27l2a (Ifi27l2a^+/- mice) reduced gliosis (microgliosis and astrogliosis), acute and chronic brain injury, and motor function deficits. Together, these findings identify Ifi27l2a as a critical neuroinflammatory mediator in ischemic stroke and provide support for the therapeutic strategy of disrupting Ifi27l2a to attenuate inflammation in the post-stroke brain.

The role of Ifi27l2a, an interferon-induced gene, remains poorly understood in diseased brains. Here, authors show age and stroke-dependent upregulation of Ifi27l2a in microglia, and that reduction of Ifi27l2a leads to reduced brain injury and functional deficits after ischemic stroke.

Introduction

Microglia (MG) are resident immune cells in the central nervous system (CNS). Critically, MG feature extensive heterogeneity, with subtypes evident across different regions of the brain, across different developmental stages and ages, and between sexes^1,2. These various MG subtypes play pivotal roles in both initiating and resolving inflammation, often acting in coordination with other glial cells, such as astrocytes and oligodendrocytes³. The MG response to inflammatory challenge and ischemic stroke is a crucial component for maintaining/restoring brain homeostasis, salvaging tissue, and minimizing brain damage⁴. However, aging results in altered MG responsiveness to inflammatory signaling, often exacerbating the inflammatory response⁵. This exaggerated response contributes to impaired brain function and contributes to the pathological aging processes in the aged brain^6,7. Selectively ‘dampening’ the MG response to ischemic injury so that they adopt a less pro-inflammatory and more anti-inflammatory-like state may represent an effective therapeutic interventional strategy to decrease ischemic damage and improve functional recovery after stroke, particularly in the context of the aged brain.

To better define specific differences in how the aged brain responds to ischemic stroke, we performed single cell transcriptional profiling (scRNA-seq) of young and aged brains of mice 2 weeks after animals underwent either sham surgery or stroke. For these studies we chose a permanent stroke model induced by permanent distal middle cerebral artery occlusion (pdMCAO), which results in primary infarction in the cortex and delayed secondary injury in the thalamus⁸. Analysis of differentially expressed transcripts showed that interferon alpha-inducible protein 27 like 2 A (Ifi27l2a) was upregulated in MG two weeks after stroke in young mice compared to sham-operated controls. Notably, Ifi27l2a was significantly further upregulated in MG of aged mice compared to young mice, suggesting a disproportionate effect of stroke on Ifi27l2a induction with aging. Recently, a distinct transcriptional analyses study found that ischemic-like damage upregulated multiple interferon-stimulated genes (ISGs) in brain, including Ifi27l2a^9,10. Both interferons (IFN α and β) and certain types of viral infection upregulate Ifi27l2a expression in the brain^11,12. However, to our knowledge, functional studies analyzing the role of Ifi27l2a in ischemic stroke have not been reported. Herein, we provide multiple lines of evidence supporting a role for Ifi27l2a in regulating neuroinflammation in the aged brain and in the pathologic setting of stroke, and demonstrate that reducing Ifi27l2a attenuates ischemic injury and improves functional outcomes following ischemic injury. Together, these studies provide compelling evidence that targeting Ifi27l2a expression or function may mitigate brain injury and inflammation following ischemic stroke.

Results

Ifi27l2a is highly upregulated after stroke in MG and aging significantly enhances this upregulation

To define the transcriptional signature across multiple cell types in the post-stroke brain, we performed scRNA-seq of young (3-month-old) and aged (20-month-old) male and female C57BL/6 J mice subjected to pdMCAO or sham surgery (Table 1). The pdMCAO stroke model was used as it produces both primary injury (cortical infarct) and delayed secondary injury in the thalamus (by 1,2 weeks post-stroke)⁸. Since the cortex and thalamus also feature clear increases in microgliosis and astrogliosis following stroke, we isolated cells for scRNA-seq at 14 post-stroke days (PSD) from a coronal brain slab containing both peri-infarct cortex and ipsilateral thalamus⁸. We performed QC of the scRNA-seq datasets prior to subsequent analyses (Supplementary Fig. 1). We evaluated differential gene expression (DGE) patterns as an initial approach for determining how the cell-specific molecular signatures are altered by stroke within the young and aged brain. To measure the effect of aging in stroke brains, we integrated young stroke (AGGR2) and aged stroke (AGGR4) data for analysis (using the Seurat package¹³). Then, a single integrated analysis (data integration, PCA, UMAP and clustering, and DGE) was performed. Visualization of this merged dataset of 20,514 cells from the aged and young stroke mouse brains through dimension reduction by uniform manifold approximation and projection (UMAP) identified eight clusters of unique cell types based on gene expression differences. Identities were assigned to each of the eight cell clusters based on the expression of conserved markers, including microglia (MG, n = 7489) (Trem2), oligodendrocytes (Oligo, n = 4498) (Mog), endothelial cells (EC, n = 3651) (Cldn5), astrocytes (Astro, n = 1666) (Aldoc), lymphocytes (Lym, n = 1623) (Plac8), epithelial cells (Epi, n = 1131) (1500015O10Rik), Vascular smooth muscle cells, arterial (VSMCA, n = 332) (Myl9), and Vascular leptomeningeal cells (VLMC, n = 124) (Dcn) (Fig. 1a-b). In the MG cluster assigned, microglia were the predominant cell type; however, other macrophage cell types, such as border-associated macrophages and infiltrated macrophages, were also present within this group. In the aged stroke brain, the proportion of MG and Lym clusters were highly increased, whereas oligodendrocytes were reduced (Fig. 1c). These findings are consistent with more extensive white matter injury in the aged stroke brain and correlate with increased MG-mediated neuroinflammation and lymphocyte infiltration. As we and others have demonstrated that MG are highly sensitive to inflammation and ischemic stress and act to regulate innate immunity in brains^4,14, we focused our subsequent analysis on transcriptional changes within MG.

Table 1.

scRNA-seq samples analyzed (total 12 samples)

	Young (6)		Aged (6)
Sample	Sham (1 male, 1 female)	Stroke (2 male, 2 female)	Sham (1 male, 1 female)	Stroke (2 male, 2 female)
Name	AGGR1	AGGR2	AGGR3	AGGR4
# of cells analyzed	5616	12723	5073	7791

Fig. 1. scRNA-seq identification of cell clusters and upregulation of Ifi27l2a in young and aged brains after stroke with distinct transcriptional signatures.

a UMAP plot shows the clusters in young and aged stroke (Seurat FindClusters resolution at 0.019). MG; microglia, Oligo; oligodendrocytes, EC; endothelial cells, Astro; astrocytes, Lym; lymphocytes, Epi; epithelial cells, VSMCA: vascular smooth muscle cells, arterial, VLMC: vascular leptomeningeal cells) b Feature plot verifying clustering assignments by representative cell-specific marker gene expression (Trem2; MG, Plp1; Oligo, Cldn5; EC, Aldoc; Astro, Plac8; Lym, 1500015l10Rik; Epi, Myl9; VSMCA, Dcn; VLMC). c Pie plot showing the percentage of the total for each cluster in young stroke and aged stroke. d Normalized Ifi27l2a expression from total cell population in young stroke (12,723 cells) and aged stroke (7791 cells). The overall expression of Ifi27l2a was greater in aged stroke. ****p < 0.0001, two-tailed unpaired t-test. e Lgals3bp, Lyz2, Ifitm3, Lgals3, and Ifi27l2a were identified as highly expressed genes in aged stroke (red dot, AS: aged stroke), compared to young stroke (blue dot, YS: young stroke). Dot size indicates the percent of cells that express the respective gene in the cluster. f Feature plots showing the distribution of Ifi27l2a, Ifitm3, and C1qa in MG of young and aged stroke brains. Violin plots showed the increased expression of g Ifi27l2a, h Lgals3, i Ifitm3, and j Lgals3bp on cell type-specific in young and aged stroke (showing increased Ifi27l2a expression in MG, Lym, and VLMC clusters in aged stroke samples).

Our unbiased scRNA-seq analyses of 20,514 combined cells from young and aged stroke brains revealed Ifi27l2a as a highly upregulated gene (aged stroke > young stroke), based on both fold change and statistical significance (Table 2). Total normalized expression of Ifi27l2a from all cells showed increased Ifi27l2a levels in cells of aged stroke brains compared to young stroke animals (Fig. 1d). Within MG, the top genes showing significant upregulation in the aged stroke brain included MG-related genes (Lgals3, Lyz2, Lgals3bp) and another interferon-stimulated gene (ISG), Ifitm3. The expression level and percent of cells expressing these upregulated genes are indicated by dot plots in Fig. 1e. Within the MG cluster, there was a notable increase in the percentage of cells expressing Ifi27l2a between young stroke and aged stroke (30.6% vs 74.4%). Ifi27l2a was more highly upregulated in MG of aged stroke brain, suggesting that aging or the context of an aged brain may act synergistically with ischemic stroke to enhance Ifi27l2a expression in MG (Fig. 1f–g). Other MG-associated genes, such as Lgals3, Ifitm3, and Lgals3bp, were also upregulated following stroke (Fig. 1h–j). As expected, a known marker of activated MG, Cst7, was increased in aged stroke compared to young stroke brain (Supplementary Fig. 2a), confirming that MG were more highly activated in aged stroke brains than in young stroke brains. In addition, we observed increased transcripts for Apoe and Lyz2 in aged stroke compared to young stroke samples, whereas Aif1 appeared unchanged (Supplementary Fig. 2b–d, Supplementary Data file 1). While the vast majority of Ifi27l2a-expressing cells belonged to the MG cluster, Ifi27l2a expression was also detected in lymphocyte (Lym) and vascular leptomeningeal cell (VLMC) populations (Fig. 1g). VLMC showed increased Ifi27l2a expression in aged stroke, whereas Ifi27l2a expression in Lym from the stroke brain was not markedly altered by aging.

Table 2.

Identification of Ifi27l2a as a top gene that is upregulated in MG (young stroke vs. aged stroke)

Top 5 Genes	Young Stroke	Aged Stroke	Ave_log FC	Ave_FC
Ifi27l2a	1.09	2.11	1.02	2.03
Lgals3	0.98	1.81	0.83	1.78
Ifitm3	1.15	2.05	0.89	1.89
Lgal3bp	1.42	2.14	0.72	1.64
Lyz2	3.24	3.93	0.68	1.61

Given the extensive heterogeneity evident within MG and macrophage (MΦ) populations, we subjected the aged scRNA-seq datasets to more granular analysis to determine if Ifi27l2a expression profiles correlated with different functional roles. With this deeper sub-clustering, we resolved a total of 26 populations in brain cells of aged sham and aged stroke mouse brains, with eight clusters assigned an MG or monocyte/MΦ identity based on the expression of conserved cell markers (Supplementary Fig. 3). Two MG homeostatic clusters were identified based on the expression of MG genes such as Siglech, Tmem119, Gpr34, P2ry12, and Selplg. These MG were annotated as Siglech homeostatic MG and P2ry12 homeostatic MG. We also identified two different MG populations that appeared to be in an activated state (Lag3⁺ activated MG and Tyrobp⁺ activated MG). Two monocyte-macrophage populations were identified using conserved marker genes (Mrc1, Ms4a7, Pf4, Flt3, and Clec12a). We also identified a disease-associated MG (i.e. DAM) like cluster with high expression of Lpl, Itgax, Cst7, and Spp1 (Supplementary Data file 2). These same genes are upregulated in DAMs in other neurodegenerative diseases, such as Alzheimer’s Disease (AD)^15,16. Comparison of Ifi27l2a gene induction (sham vs. stroke) among MG sub-clusters in aged brain revealed that Ifi27l2a expression shows a range of expression levels across the various MG/MΦ subtypes (Supplementary Fig. 4, Table 3, Supplementary Data file 2). Furthermore, Ifi27l2a transcripts were markedly elevated with aging among the MG subtypes (Supplementary Fig. 5). Expression of Ifi27l2a in the sham group was higher in the monocyte/MΦ population than MG, however, MΦ reflected a small percentage of total MG/MΦ in sham samples (3.5%, 81 of 2195 cells).

Table 3.

Differential expression of Ifi27l2a in each MG sub-clusters in sham and stroke in aged brains

MG sub-clusters	Aged sham (expression level)	Aged stroke (expression level)	Fold change
Homeostatic MG_1 Siglech+	1.16	2.33	2.0
Homeostatic MG_2 P2ry12+	1.30	3.42	2.62
DAM like sub-cluster	2.17	4.6	2.12
MG_Activated_1	2.27	5.25	2.31
MG_Activated_2	1.51	5.37	3.53
MG_Progenitor	2.88	6.70	2.32
Mono-Macro_1	5.23	6.34	1.21
Mono-Macro_2	1.99	6.02	3.02

scRNA-seq revealed that aging alone is sufficient to increase Ifi27l2a transcripts in MG

Following our finding that Ifi27l2a is upregulated following stroke in an age-dependent manner, we next sought to determine if aging alone impacts Ifi27l2a expression. We initially conducted an analysis integrating all groups (i.e. including young and aged sham brains; AGGR1 and AGGR3) within an integrated plane to observe the expression patterns across four groups: young sham, aged sham, young stroke, and aged stroke. Our findings revealed that aging alone led to a modest increase of Ifi27l2a expression in MG/macrophage clusters (young sham vs aged sham: 0.109 vs 0.327, p < 0.0001, Supplementary Fig 6). We then examined the effect of aging on Ifi27l2a expression within specific sub-clusters. Eight clusters were identified (Supplementary Fig. 7a and b), including oligodendrocytes (Oligo) (Plp1), MG (C1qa), ECs (Cldn5), astrocytes (Astro) (Gpr37l1), epithelial cells (Epi) (Ttr), lymphocytes (Lym) (Nkg7), vascular smooth muscle cells, arterial (VSMCA) (Des), and B cells (CD79a). To determine how aging affects the transcriptional landscape in MG, we compared the levels and percent of cells expressing the previously identified top five MG genes (Ifi27l2a, Lgals3, Ifitm3, Lyz2, and Lgals3bp) in young and aged sham brains. All 5 genes that were upregulated in MG from aged stroke brain were also increased by aging alone (Supplementary Fig. 7c and 8). Notably, Ifi27l2a transcript levels significantly increased with aging (MG cluster, young sham vs aged sham, 0.35 vs 0.76, p < 0.0001), as did Rps27rt. On the other hand, C1qa was not dramatically altered between young and aged sham animals (Supplementary Fig. 7d). We also confirmed the increased expression of Ifi27l2a in B cell clusters in aged brains (Supplementary Fig. 8a). Interestingly, there was a significant age-dependent upregulation of genes encoding ribosomal protein, such as Rpl35, Rps27rt, and Rps28 (Supplementary Table 1). This age-dependent upregulation of Rps27rt in all clusters, including MG and Lym (Supplementary Fig. 8b), suggested aging-mediated changes in ribosomal complex composition within MG. Expression of two other genes associated with activated MG (Aif1 and Il1b) were also modestly increased with aging (Supplementary Fig. 8c, d, Supplementary Data file 3). We repeated the same analyses comparing sham to stroke in aged (Supplementary Figs. 9–11, Supplementary Data file 4) and young cohorts (Supplementary Fig. 12, 13, Supplementary Data file 5). Together, these data indicate a modest, but statistically significant effect of aging alone on Ifi27l2a expression in MG.

Regional Ifi27l2a expression with natural aging

While scRNA-seq showed that Ifi27l2a transcripts were enriched in MG and other cell types (e.g. Lym, VLMC) in both young and aged stroke brains, this approach did not provide any information regarding the spatial basis of these changes within the brain (e.g. within the primary injury in the cortex or within the secondary injury region occurring within the thalamus). Therefore, to determine if there were regional differences in baseline Ifi27l2a expression, the cortex and thalamus were extracted from the brains of naïve young (3 months, n = 4) and aged male mice (18-20 months, n = 4). Notably, Ifi27l2a mRNA was significantly upregulated in the aged thalamus (Fig. 2a, *p < 0.05). Ifi27l2a expression trended higher in the cortex of aged brains, but failed to reach statistical significance (p = 0.23). Other MG related genes, such as Il1b, Cst7, and Tyrobp, were also significantly increased in the thalamus of aged brains, further supporting an age-dependent increase in MG activation (Fig. 2b-d). Transcripts for C1qb and Lpl, two genes which associated with microglia phagocytosis, were not altered between young and aged samples (Fig. 2e-f). These findings align with our scRNA-seq data showing increased Ifi27l2a expression in the aged brain and further suggest regional differences in aging-related expression of Ifi27l2a, along with other genes associated with proinflammatory MG phenotype in the aged brain.

Fig. 2. Regional increases of Ifi27l2a expression in brain with normal aging and in post-stoke brains.

RNA was isolated from thalamus and cortex of young and aged brains for qRT-PCR analysis of a Ifi27l2a (n = 4 mice per group, *p = 0.041) and other genes associated with MG phenotype: pro-inflammatory genes b Il1b (n = 4 mice per group, *p < 0.020), c Cst7 (n = 5-6 mice per group, ***p = 0.0001, *p = 0.013) and phagocytosis related genes d Tyrobp (n = 4 mice per group, **p = 0.004), e C1qb (n = 4 mice per group), and f Lpl (n = 4 mice per group). Ctx: cortex, Th: thalamus, Y: young, Ag: Aged, Data presented as mean ± SEM. Two-tailed unpaired Student’s t-test. RNA was isolated from the cortex and thalamus of sham control or aged stroked mice at 3 days (3D) and 14 days (2 W) after stroke for qRT-PCR analysis. Summary of fold change in expression for g Ifi27l2a (n = 4-6 mice per group, 3D-ctx vs Sh-ctx; *p = 0.046, 14D-ctx vs Sh-ctx; *p = 0.038, 14D-Th vs Sh-Th; *p = 0.032), h Cst7 (n = 4-6 mice per group, 14D-ctx vs Sh-ctx; *p = 0.046), i Tyrobp (n = 4-6 mice per group, 14D-ctx vs Sh-ctx; *p = 0.056, 14D-Th vs Sh-Th; # p = 0.019). Sh; Sham. Data presented as mean ± SEM, two-tailed unpaired Student’s t-test. RNAscope assay shows regional and MG-specific expression of Ifi27l2a in aged brain after stroke. j A representative stitched image showing Ifi27l2a transcripts (red dots) in the peri-infarct area of aged brain at 2 weeks after stroke. Magnified image shows Ifi27l2a transcript (Red). scale bar: 20 µm. Double RNAscope experiment was performed with both Tmem119 probe (Green) and Ifi27l2a probe (Red) on sections from brains at two weeks post-stroke. k Tmem119 + (Green) microglia exhibit Ifi27l2a signals (Red) in the peri-infarcted area. l Magnified image shows Ifi27l2a transcript (Red) in microglia, identified as Tmem119+ cells (Green). White arrows indicate microglia expressing Ifi27l2a. scale bar: 25 µm, n = 4 mice. Source data are provided as a Source Data file. Created in BioRender⁴¹.

Regional and temporal expression of Ifi27l2a in aged stroke brain

To evaluate regional and temporal expression of Ifi27l2a and other MG-related genes following stroke, we analyzed aged thalamus and cortex by qRT-PCR at 3 and 14 days post-stroke. In this stroke model, we observe acute injury in the cortex and delayed injury in the ipsilateral thalamus. As predicted based on this injury progression, Ifi27l2a was significantly elevated at three days (cortex) and two weeks (cortex and thalamus) after stroke, compared to sham (Fig. 2g). We evaluated two other genes associated with MG activation (Cst7) and reparative phagocytosis (Tyrobp) that were elevated in our initial scRNA-seq analysis. While Cst7 and Tyrobp were not elevated 3 days after injury, both transcripts were increased in the cortex and thalamus at two weeks post-stroke (Fig. 2h, i). These findings indicate an early and prolonged involvement of Ifi27l2a in the primary injury as well as a role in the delayed secondary injury.

Resident MG represent a major cellular source of Ifi27l2a expression in the brain parenchyma after stroke

To provide spatial context to the alterations in Ifi27l2a expression detected by our scRNA-seq studies, we imaged Ifi27l2a mRNA transcripts in sections of mouse brains using single-molecule RNA fluorescent in situ hybridization (i.e. smRNA-FISH). Probing for Ifi27l2a in aged sham and stroke brains revealed elevated transcripts in both the infarct core and the peri-infarct area at 2 weeks post-stroke compared with sham-operated controls (Fig. 2j, Supplementary Fig. 14a). Combining Ifi27l2a smRNA-FISH with immunostaining for Iba1 (to mark MG/MΦ) confirmed that the majority of Ifi27l2a transcripts were present in activated MG/MΦ in the peri-infarct region of the aged brain (Supplementary Figs. 14b-15). Next, to profile the expression of Ifi27l2a specifically in MG, we performed smRNA-FISH with probes for both Ifi27l2a and Tmem119. This double staining approach revealed Ifi27l2a expression in Tmem119positive MG (Fig. 2k, l).

As an additional approach, we isolated MG directly from post-stroke brains via Fluorescent Activated Cell Sorting (FACS) to evaluate expression of Ifi27l2a selectively in MG following stroke. Tmem119 is a MG marker that distinguishes MG from other macrophages. Tmem119⁺ microglia from either the ipsilateral or contralateral hemispheres of aged brains following stroke (PSD 14) were isolated from single-cell suspensions after staining with CD45, CD11b, and Tmem119 antibodies. A gating strategy was employed to separate Tmem119⁺ MG, utilizing Tmem119 FMO to exclude the Tmem119^- population (Fig. 3a, Supplementary Fig. 16a). In FACS sorted Tmem119⁺ microglia, the levels of Ifi27l2a and inflammatory marker genes (Cst7, Il1b), were elevated in the ipsilateral hemisphere of aged mice following stroke compared to the contralateral hemisphere, as measured by qRT-PCR (Fig. 3b–d, *p < 0.05). This data shows increased expression of Ifi27l2a (and other inflammatory marker genes) specifically in Tmem119⁺ MG following stroke in aged brains.

Fig. 3. MG represent the primary source of Ifi27l2a upregulation following stroke.

a Tmem119+ microglia express Ifi27l2a in the aged brain following stroke (PSD 14). A gating strategy was employed to sort Tmem119+ microglia (MG), utilizing Tmem119 FMO to exclude the Tmem119- population. b Levels of Ifi27l2a (n = 5-6 biological replicates, *p = 0.027, Data presented as mean ± SEM, two-tailed unpaired t test with Welch’s correction, c Cst7 (n = 3-5 biological replicates, *p = 0.014, Data presented as mean ± SEM, two-tailed unpaired t test with Welch’s correction), and d Il1b (n = 4-5 biological replicates, *p = 0.048, Data presented as mean ± SEM, two-tailed unpaired t test with Welch’s correction) were measured by qRT-PCR. Upregulation of Ifi27l2a/IFI27L2 in stimulated primary MG, human MG and diseased human brain. Mouse primary MG were treated with TNFα (20 ng/ml) and IFNγ for 24 hours and 48 hours. e Ifi27l2a mRNA were increased in stimulated MG for 24 hrs (n = 6-7 from three independent experiments, ***p = 0.0005, two-tailed unpaired Student’s t-test. f Intracellular Ifi27l2a protein was induced by proinflammatory cytokine treatment for 48 hrs (n = 8-10 wells, ****p < 0.0001, two-tailed unpaired Student’s t-test). g HMC3 were treated with TNFα (20 ng/ml) and IFNγ (20 ng/ml) plus OGD (Stim). Induction of IFI27L2 mRNA in stimulated HMC3 cells were assessed by qRT-PCR (n = 5-6 from three independent experiments, *p = 0.036, two-tailed unpaired Student’s t-test.). Source data are provided as a Source Data file. h Representative images show expression of IFI27L2 in stressed human HMC. scale bar: 20 µm. i Human IFI27L2 protein expression in age-matched control and stroke human brain tissue collected from patients with no or mild neuropathology. IFI27L2 positive cells were increased in the stroke brain samples (n = 3) compared to age-matched controls (n = 2). scale bar: 75 µm. Red arrows indicate IFI27L2-postive cells. The left panel shows increased IFI27L2-positive cells in stroke human brain samples. The right panel shows IFI27L2 expression in IBA1-positive cells. Created in BioRender⁴².

Inflammatory stimuli induce Ifi27l2a/IFI27L2 expression in MG

We next used cultured MG to evaluate the potential for inflammatory mediators to promote Ifi27l2a expression. First, we used mouse primary MG collected from mixed glial cell cultures obtained from postnatal day 2 (P2) mouse pups. Primary MG were treated with the inflammatory cytokines TNF-α (20 ng/mL) and IFN-γ (20 ng/mL) for 24 hours (to measure Ifi27l2a mRNA) and 48 hours (to measure Ifi27l2a protein by ELISA). Both Ifi27l2a transcripts (Fig. 3e) and protein (Fig. 3f) were significantly increased following treatment with TNF-α and IFN-γ. Interestingly, Ifi27l2a expression followed Il1b expression, suggesting a positive correlation between Ifi27l2a expression and severity of inflammation (Supplementary Fig. 17a, r = 0.848). Co-expression analysis of Ifi27l2a and Il1b in MG and macrophage clusters corroborated this correlation in our scRNA-seq dataset (Supplementary Fig. 17b).

To determine whether these findings extended to a human in vitro MG model, we challenged human microglial cells (HMC3) by addition of pro-inflammatory cytokines (TNF-α [20 ng/mL] and IFN-γ [20 ng/mL]) in combination with oxygen/glucose deprivation (inflammation/OGD). This inflammatory challenge significantly upregulated IFI27L2 mRNA in HMC3s (Fig. 3g, *p < 0.05). In addition, IFI27L2 protein level was dramatically increased at 20 hours post inflammation/OGD (Stim) compared to control treatment (Control) (Fig. 3h, representative of n = 4).

Given these results, we next tested if IFI27L2 protein was increased in post-mortem autopsy samples from brains of patients featuring neuroinflammation. We first examined samples from patients who died from stroke with mild or no additional neuropathologies (n = 3), as well as those that died from other causes and without apparent neurological disease (n = 2) (Supplementary Table 2). Immunofluorescence showed significant IFI27L2 expression in stroke brains compared to low expression in age-matched control samples (Fig. 3i, representative of n = 2,3). Co-staining with IBA1 further revealed that the IFI27L2 expression largely correlated with MG/MΦ cell types. In another inflammatory context, immunofluorescence showed IFI27L2 expression was elevated in stroke patient brains featuring cerebral amyloid angiopathy (CAA) pathology and tauopathy compared to those without these hallmarks (Supplementary Fig. 18, Supplementary Table 3). As a further corroboration of our immunofluorescence findings, we utilized the open dataset GSE160936¹⁷ (snRNA-seq of AD postmortem samples) to evaluate the expression of IFI27L2 in Alzheimer’s disease (AD) samples. Our analysis revealed two major IFI27L2-expressing clusters in these data (Supplementary Fig. 19a). The samples from the entorhinal cortex showed an increase in IFI27L2 in astrocytes and microglia compared to age-matched samples (Supplementary Fig. 19b, **p < 0.01 and ****p < 0.0001). Together, these data suggest that inflammatory stimuli and aging induce Ifi27l2a levels in both mice and humans, and identify elevated IFI27L2 as a common biomarker in the brains of patients with multiple forms of neuroinflammatory disease.

Elevated Ifi27l2a expression is sufficient to promote MG activation

Given the induction of Ifi27l2a in MG in aged brains and following stroke, we sought to elucidate the functional role of Ifi27l2a in MG-mediated neuroinflammation. In particular, we asked if Ifi27l2a expression alone (without additional inflammatory mediators) could induce a pro-inflammatory phenotype in MG. We infected a murine microglial cell line (Sim-A9 cells) with a lentivirus in which the Cx3cr1 promoter drives expression of Ifi27l2a and an eGFP reporter, or eGFP alone as a control (Fig. 4a, Supplementary Fig. 20). At 5 days post-infection, quantification of cell morphology showed that Ifi27l2a expression increased the percentage of cells with a small, rounded shape (a more amoeboid morphology or de-ramification) compared to eGFP (Fig. 4b, c). Changes in microglial morphology are an early, quantifiable sign of inflammation in MG and altered MG functionality. Interestingly, MG with increased Ifi27l2a expression (using eGFP intensity as a surrogate maker) showed more dramatic morphological changes compared to cells that had low eGFP (Ifi27l2a) expression (Fig. 4d). Moreover, Ifi27l2a overexpression induced proinflammatory genes, such as Il1b, Il1a and Trem2 (Fig. 4e, f, Supplementary Fig. 21b). Expression of Tnfa was upregulated comparably following expression of eGFP and Ifi27l2a compared to baseline, suggesting a response to viral transduction itself (Supplementary Fig. 21a). Ifi27l2a expression did not change Tmem119 levels (Fig. 4g). These results provide direct evidence that Ifi27l2a alone can initiate MG activation, thus demonstrating the pro-inflammatory capacity of Ifi27l2a in the absence of other traditional inflammatory mediators.

Fig. 4. Ifi27l2a is sufficient for microglial activation and ROS generation.

a Lenti-Cx3cr1-Ifi27l2a-eGFP vector map and primer binding regions. b Overexpression of Ifi27l2a by lentivirus changed the morphology at 5 days after infection. scale bar: 100 µm c The % of cells that changed their shapes was significantly increased in Ifi27l2a-lentivirus infected cells, compared to control-lentivirus infected cells (n = 3 biologically independent samples, Data presented as mean ± SEM. *p = 0.025, two-tailed unpaired Student’s t-test). d Representative images show that the cells that express Ifi27l2a (eGFP as an expression surrogate) changed their morphology to round and amoeboid shapes. Red arrows indicate cells that express Ifi27l2a and show a round morphology; Green arrow indicates a cell which does not express Ifi27l2a and remains in the intact morphology. scale bar: 100 µm, 50 µm (zoom-in). Cx3cr1-driven Lenti-eGFP and Lenti-Ifi27l2a were infected into Sim-A9 cells for 2 days. The degree of inflammation was evaluated by qRT-PCR with primers for e Il1b (n = 4-5 biologically independent samples, *p = 0.014, **p = 0.003, ns=non-significant, Data presented as mean ± SEM), f Il1a (n = 3-5 biologically independent samples, *p = 0.022, ***p = 0.0004, Data presented as mean ± SEM), and g Tmem119 (n = 4-6 biologically independent samples, ns=non-significant, Data presented as mean ± SEM, ns=non-significant),1-way ANOVA with Bonferroni’s multiple comparison. h Ifi27l2a overexpression increased the intensity of CellRox dye in all cells (n = 4 biologically independent samples, *p = 0.012, Data presented as mean ± SEM, two-tailed unpaired Student’s t-test). i Ifi27l2a overexpression increased CellRox intensity within GFP expressing cells (n = 4 biologically independent samples). Mitochondrial ROS levels j. MFI, n = 3-6 biologically independent samples, *p = 0.013, **p = 0.003, Data presented as mean ± SEM, k. % of MitoSox+ cells, n = 3-6 biologically independent samples, *p = 0.026, **p = 0.004, Data presented as mean ± SEM) detected by Mitosox was increased in Ifi27l2a lentivirus infected cells, compared to control (*p < 0.05, **p < 0.01, one-way ANOVA with Bonferroni’s multiple comparison test). l Oxygen consumption rate after treatment with oligomycin, FCCP, and rotenone/AA was measured in Sim-9A cells infected with Lenti-eGFP and Lenti-Ifi27l2a (n = 6 biologically independent samples, Data presented as mean ± SD). Source data are provided as a Source Data file.

Ifi27l2a induces mitochondrial dysfunction and ROS production

Previous studies suggested Ifi27l2a localizes to mitochondria in non-CNS cells¹⁸. Based on this localization and our findings linking Ifi27l2a to pro-inflammatory phenotype, we asked if Ifi27l2a could mediate mitochondrial dysfunction in MG. Using lentiviral transduction, we evaluated the potential of Ifi27l2a to induce reactive oxygen species (ROS) production in MG. We utilized CellROX Red, a detector of general ROS species, to determine if Ifi27l2a expression induces ROS production in Sim-A9 cells under basal conditions. 5 days after viral delivery, flow cytometry revealed that Ifi27l2a overexpression alone significantly increased ROS production (Fig. 4h, as expressed in median fluorescence intensity, MFI; Ctrl: lenti-eGFP control, Ifi27l2a: lenti-Ifi27l2a-eGFP, n = 4, *p < 0.05). Next, we analyzed only GFP-positive cells (rather than all Sim-A9 cells), to confine our analysis to only those cells successful transduced with Ifi2712a-eGFP or eGFP. ROS levels were greater in Ifi27l2a-expressing cells compared to eGFP only control cells (Fig. 4i, Ctrl: lenti-eGFP control, Ifi27l2a: lenti-Ifi27l2a-eGFP).

To evaluate the role of mitochondria in this Ifi27l2a-induced production of ROS, we treated cells with an indicator of mitochondria-derived ROS, Mitosox dye. Ifi27l2a overexpression significantly increased mitochondria generated ROS (Fig. 4j) and elevated the overall percentage of Mitosox⁺ cells (Fig. 4k). Control eGFP cells were similar to the “no-virus” control (No), indicating that the effect was due to Ifi27l2a expression, and not the exposure to the virus. To determine the effect of Ifi27l2a on mitochondrial function, we performed the Seahorse Cell Mito Stress assay, which directly measures oxygen-consumption (Fig. 4l). Expression of Ifi27l2a in MG reduced mitochondrial respiration (maximal respiration and spare respiratory capacity), implicating the expression of Ifi27l2a in mitochondrial dysfunction (Supplementary Fig. 22). Together, these data demonstrate the potential of Ifi27l2a to target mitochondrial function in MG, leading to reduced mitochondrial respiration and increased ROS production in activated MG.

Hemizygous deletion of Ifi27l2a attenuates MG inflammation and reduces ischemic brain injury in mice

Inflammation is a significant contributor to brain injury following ischemic stroke. Given our findings that Ifi27l2a is increased in ischemic stroke and has the capacity to promote microglial activation, we asked if reducing Ifi27l2a expression could dampen microglial activation and attenuate brain inflammation and injury following stroke. We first evaluated the effect of Ifi27l2a reduction on the inflammatory response in primary microglia isolated from WT and Ifi27l2a^+/- (Het) brains. We intentionally used Het mice, as we reasoned that partial reduction of Ifi27l2a expression would reflect a more realistic therapeutic goal (for instance, with using anti-sense oligo mediated knockdown), and thus successful outcomes with Hets likely have increased translational relevance. We confirmed that Het microglia demonstrated reduced Ifi27l2a expression (68% reduction) by qRT-PCR (Supplementary Fig. 23). We used LPS challenge (20 ng/mL) as an inflammatory stimulus in primary microglia cultured from WT and Het littermates. 24 hours post LPS treatment, we quantified expression of classic inflammatory genes. LPS-induced Il1b, Il1a, and Tnfa in WT MG, whereas induction of each of these inflammatory genes was significantly reduced in Het microglia (Supplementary Fig. 24). These findings show that even partial reduction of Ifi27l2a can significantly attenuate inflammation in microglia.

To evaluate the effect of Ifi27l2a reduction in an in vivo pathological setting featuring neuroinflammation, we returned to the pdMCAO stroke model. At post-stroke day (PSD) 3, the primary cortical infarct volume was significantly reduced in Het compared to WT brains in male and female mice (Fig. 5a, b, Male: n = 5 or 6, Female: n = 4 or 5, *p < 0.05, ***p < 0.001). The area of activated MG (Iba1) was also reduced in the primary injury region (somatosensory cortex) at PSD 14 (Fig. 5c, d, n = 5 or 6, *p < 0.05). The pdMCAO model is a well-established model for evaluating delayed secondary injury, as occurs in the ipsilateral thalamus, following stroke¹⁹. We and others have shown significant gliosis and neuronal loss in the ipsilateral thalamus 1 to 2 weeks following stroke^8,20,21. Therefore, to evaluate the role of Ifi27l2a in secondary injury, we compared thalamic gliosis in WT and Het mice at PSD 14. Gliosis in the ipsilateral thalamus was significantly reduced in Het brains, for both microgliosis (Fig. 5e-f, n = 6, *p < 0.05) and astrogliosis (Fig. 5g, h, n = 6, **p < 0.01). Note that the reduced injury in Het mice was not due to differences in vascular developmental (e.g. via reduced MCA territory), as comparing the vascular territory between WT and complete Ifi2712a knockout (Ifi27l2a^-/-) animals revealed no difference (Supplementary Fig. 25, n = 6, p = 0.19). Behavioral analyses also showed improved performance by Het mice in tests of motor performance. In the foot fault test, Het mice showed significantly fewer foot slips at 3, 7 and 14 days (Fig. 5i, n = 9, *p < 0.05, **p < 0.01). In the Digigait test, Het mice showed reduction of contralateral hindpaw drag (left hindpaw) at 7 days (Fig. 5j, 1 wk, n = 8-10, *p < 0.05). A similar trend was seen at 14 days, though it failed to reach statistical difference (Fig. 5j, 2 wks, p = 0.224).

Fig. 5. Hemizygous deletion of Ifi27l2a is neuroprotective for ischemic stroke and promotes improved neurobehavioral performance.

a, b Brain infarction at PSD 3 was significantly reduced in Ifi27l2a^+/- (Het) compared to WT (M:Male n = 5-6 mice per group, F:Female n = 4-5 mice per group,***p = 0.006, F:WT vs F:Het *p = 0.019, M:WT vs F:WT *p = 0.020, Data presented as mean ± SEM, two-way ANOVA with Tukey’s multiple comparisons test). c, d Ifi27l2a deletion (Het) significantly reduced microgliosis in the peri-infarct cortex at 14 days following stroke (n = 5-6 mice per group, *p = 0.036, scale bar: 1 mm). Deletion of Ifi27l2a (Het) significantly reduced microgliosis e, f n = 6 mice per group, *p = 0.040, two-tailed unpaired Student’s t-test, scale bar: 100 µm) and astrogliosis g, h n = 6 mice per group, **p = 0.007, two-tailed unpaired Student’s t-test, scale bar: 100 µm) in the thalamus at 14 days post-stroke. i Foot fault tests were performed at 0, 3, 7, and 14 days after pdMCAO with WT and Ifi27l2a Het mice (Ifi27l2a^+/-), and deficits were examined by counting total foot slips on grids. The data was expressed as a percentage of total foot slips per total number of steps, n = 9 mice per group, PSD 3; *p = 0.012, PSD 7; **p = 0.007, PSD 14; **p = 0.009, ns=non-significant, two-way RM ANOVA with Fisher’s LSD. j DigiGait tests were performed at 7 and 14 days after stroke to examine the sensory-motor deficits following stroke. Left hind paw (affected paw) dragging was sustained at 1 week after stroke in WT mice, but not in Het mice. Right hind paw (unaffected paw) dragging did not change in WT vs Het mice. n = 8-10 mice per group, * p = 0.044, ns=non-significant, Data presented as mean ± SEM, Two-way ANOVA with Tukey’s multiple comparisons test. Source data are provided as a Source Data file.

In order to determine if Ifi27l2a reduction provides long-term improvement in stroke outcome, we assessed brain injury at 4 weeks after pdMCAO. To quantify brain morphology without the brain tissue loss and the brain expansion that results from removing the brain from the skull, brains were imaged within intact mice by iodine contrast-enhanced microCT (Fig. 6a). Brain atrophy was evaluated by determining brain midline shift (toward the injury side). At 4 weeks post-stroke, the Het mice showed significantly reduced midline shift, indicating a reduction in brain atrophy (Fig. 6b, c, n = 10, *p < 0.05). Together, these findings indicate that decreased Ifi27l2a expression attenuates microglial-mediated inflammatory responses, reduces primary and secondary stroke injury and long-term brain atrophy, and improves functional performance in motor-based tests.

Fig. 6. Hemizygous Ifi27l2a deletion (Het mice) reduces brain atrophy at 30 days following stroke.

a MicroCT image of mouse head indicating the location of coronal planes used to measure mid-line shift in stroke brains at 30 days post-stroke. b Coronal microCT images from +1.75 and +0.35 mm from Bregma following ex vivo contrast enhancement by Omnipaque perfusion. c Mid-line shift is calculated as the ratio of the distance from the inside of the skull to the longitudinal fissure (orange arrows) measured from the contralateral (L) and ipsilateral (R) hemispheres. Summary data from both brain locations is plotted as mean ± SEM for the WT and Ifi27l2a^+/- (Het) mice. Het mice show reduced mid-line shift, consistent with reduced brain atrophy (n = 10 mice per group, +1.75 mm *p = 0.048, +0.35 mm *p = 0.033, two-tailed unpaired t-test with Welch’s correction). Source data are provided as a Source Data file.

Discussion

We used scRNA-seq to explore the effects of aging and stroke at the cellular level in the brain. As a result of these studies, we identified Ifi27l2a as a gene that demonstrated significant age-dependent upregulation in the post-stroke brain. This initial finding led to further study related specifically to where and when Ifi27l2a was upregulated in the brain and to the functional role of Ifi27l2a in aging, stroke, and other neurodegenerative conditions. We now present the following major findings: 1) Ifi27l2a is mildly upregulated by aging alone and highly upregulated in MG following stroke, particularly in aged brain. 2) Upregulation of Ifi27l2a following stroke occurs predominantly in MG and macrophage populations. 3) Induction of Ifi27l2a expression drives MG to a pro-inflammatory phenotype, including mitochondrial dysfunction and ROS production 4) Reducing Ifi27l2a expression results in reduced MG activation and brain injury and improved functional outcome in an ischemic stroke model. When considered as a whole, we propose that inflammatory stress (caused by aging, ischemic stroke, or other pathological insult) initiates Ifi27l2a gene expression in MG, which then enhances and propagates inflammatory damage throughout the brain. Further, our data suggest that the level of Ifi27l2a expression in MG may serve as a molecular switch that triggers pro-inflammatory phenotypes in aging and following stroke.

Interferons and interferon-mediated signaling in neurodegenerative diseases

Interferons and interferon-mediated signaling were originally identified as antiviral²², anti-proliferative, and immunomodulatory mechanisms²³ induced by viral infection. These pathways play pivotal roles in host defense against viral infection. Accumulated evidence has also shown a critical role for interferon signaling (especially Type I IFN, α and β) in regulating neuro-inflammation in aging and diseased brains, such as in the AD and stroke brain^24–27. Recent studies suggest that type I interferon (IFN) signaling is a major contributor to brain injury, promoting neuroinflammation and further neurodegeneration by expressing multiple interferon-stimulating genes (ISGs). This emphasizes the pathological role of type I IFN-mediated signaling, especially in Alzheimer’s disease (AD) brains^25,28. However, beneficial roles for type I IFN in the brain have also been reported, such as IFNβ, which has neuroprotective and anti-inflammatory effects against ischemic brain injury, as well as the ability to confer a preconditioned state that can ameliorate brain injury^29–32. To understand why the same IFN mediates distinct effects in different disease conditions, some effort should be made to define the individual roles of ISGs in specific disease settings. Therefore, there is a need to define the individual roles of ISGs induced by type I IFNs and other inflammatory stimuli to dissect the complexities and nuances of these signaling pathways in regulating neuroinflammation, particularly in the young and aged brain.

Corroborating studies showing elevated Ifi27l2a expression in ischemic brain

We found significant expression of Ifi27l2a in our single-cell RNA sequencing (scRNA-seq) of 2-week post-stroke brains. This finding suggests that high levels of Ifi27l2a expression are sustained in the chronic stage of stroke. Interestingly, recent studies have also shown robust expression of Ifi27l2a in vitro cultured and sorted microglia from MCAO brains at the acute phase of stroke using microarray and bulk RNA-seq^9,10. These studies also suggested the possible involvement of IFN-I signaling in inducing ISGs, including Ifi27l2a, in the ischemic brain. However, neither study further investigated the specific role of each ISG in stroke or determined how each ISG acts in the activated microglia to initiate and spread neuroinflammation. Other major differences between our present work and the previous studies are the use of different MCAO models (pdMCAO vs. suture model of MCAO), the timing of stroke (2 weeks vs. 3 days), cell collection method (all cells for scRNA-seq vs. sorted MG), and different transcriptional analyses (scRNA-seq vs bulk RNA-seq or microarray). The fact that Ifi27l2a is upregulated in multiple studies employing different stroke models, conditions, and methodological approaches speaks to the rigor of the reported finding and to the justification of elucidating the specific role of this ISG in the context of aging, stroke, and other neurodegenerative diseases.

Functional role of Ifi27l2a

Outside of the CNS, a limited number of reports have suggested a role for Ifi27l2a in facilitating inflammation through its interaction with other cellular partner proteins. During conditions of inflammation, Ifi27l2a protein is rapidly expressed and interacts with nuclear receptor 4 A (NR4A) family members. The NR4A family is thought to support the expression of multiple genes involved in attenuating inflammation in various kinds cell types^33–35. Nuclear entry of Ifi27l2a enables binding between Ifi27l2a and NR4A, which triggers the export of NR4A to the cytosol and the subsequent reduction of NR4A-mediated anti-inflammatory gene expression³⁶. Although our studies did not specifically examine NR4A regulation, our data provided multiple lines of evidence showing the induction of a pro-inflammatory phenotype. Future studies will be required to determine if Ifi27l2a contributes to post-stroke inflammation through its disruption of anti-inflammatory NR4A signaling.

The other reported mechanism by which Ifi27l2a acts involves apoptosis. Studies in activated MG and other cell types showed that Ifi27l2a can be shuttled to the mitochondrial membrane, where it initiates a mitochondria-dependent apoptotic process^37,38. Our study supports this model, as we show Ifi27l2a initiates mitochondrial dysfunction by producing ROS. With regard to the potential to trigger inflammation and apoptosis, we speculate that with more severe or prolonged MG activation, the consequence of Ifi27l2a accumulation at the mitochondrial membrane might shift from inflammation to apoptosis.

Expression of Ifi27l2a drives MG to a proinflammatory phenotype

Expression of Ifi27l2a differed among the MG sub-clusters in the basal level of Ifi27l2a expression in aged sham brain and degree of Ifi27l2a upregulation following ischemic stroke. In sham brain, expression levels of Ifi27l2a were generally low, with slight elevation in the aged brains. Following stroke, however, there was a significant increase in activated MG subclusters demonstrating elevated Ifi27l2a expression. These data suggested a potential causative role of Ifi27l2a in driving the proinflammatory phenotype. To test this idea, we induced Ifi27l2a expression in MG and found morphological and transcriptional changes associated with proinflammatory MG. In addition, Ifi27l2a expression resulted in reduced mitochondrial respiration and increased ROS production. Our data suggest that induction of Ifi27l2a in MG may represent a previously unrecognized mechanism by which inflammation is initiated and amplified in the brain.

Reduction of Ifi27l2a attenuates ischemic injury and improves outcome following stroke

Given our data demonstrating significant induction of Ifi27l2a in stroke and the potential for Ifi27l2a to drive MG to a proinflammatory phenotype, we questioned whether reduction of Ifi27l2a expression could lessen inflammation and brain injury following stroke. As a proof-of-concept study, we used Ifi27l2a^+/- mice to determine if partial Ifi27l2a reduction could attenuate brain inflammation, reduce brain infarct, and improve functional recovery following stroke. Our decision to use het mice for these studies was driven by the idea that partial reduction of Ifi27l2a (versus complete reduction with Ifi27l2a^-/- mice) would have better translational potential. In other words, if partial deletion of Ifi27l2a was successful, then future therapeutic strategies aimed at lowering Ifi27l2a expression or disrupting Ifi27l2a protein function would have better chance of success. We selected the pdMCAO model since it allowed us to evaluate Ifi27l2a reduction on both early cortical infarct and delayed thalamic inflammation. The Ifi27l2a het mice showed an approximately 50% reduction in cortical infarct at PSD3 and significant reduction of brain atrophy at PSD30. At PSD14, het mice showed reduced microgliosis in the peri-infarct region of the cortex. In addition, het mice showed reduced astrogliosis and microgliosis in the ipsilateral thalamus. However, given that the size and location of cortical injury can influence the severity of secondary injury, our results cannot currently distinguish between a direct effect of reducing secondary injury versus a reduced cortical injury involving less thalamocortical fiber damage. Lastly, behavioral tests focusing on sensory-motor function also showed improved outcome in the Ifi27l2a het mice through 1,2 weeks after stroke. Based on these encouraging findings, future studies should be performed with aged animals and new strategies to target Ifi27l2a depletion or functional inhibition initiated at different points in the post-stroke phase.

Limitations

A limitation of our single-cell RNA-seq (scRNA-seq) study is the relatively low number of biological samples and the use of a single time point (14 post-stroke day) for evaluation. We incorporated duplicate stroke samples from both sexes (total n = 4) to account for biological variability in the injury/repair response. It should be noted that, even with this limited number of biological samples, we were able to cluster major cell types in the brain and identify the cell marker genes in each cluster. By further subclustering analysis, we were able to define subtypes of microglia and macrophages with the cell numbers that we recovered from the computational analysis. Ultimately, our scRNA-seq findings provided us with specific targets to corroborate and provided the basis for subsequent investigation of the molecular function of Ifi27l2a in microglia. Also, the demonstration of reduced ischemic injury used hemizygous deletion mice, in which Ifi27l2a expression or induction potential was reduced prior to the experimental stroke. Therefore, the translational value of this finding would be predicated on either prior treatment to limit Ifi27l2a expression or to treatment prior to the upregulation of Ifi27l2a expression following stroke. Future studies will be required to more precisely define the time course of Ifi27l2a induction following stroke and to then determine if targeting Ifi27l2a with reasonable delay after stroke (e.g. with pharmacological blockers) has similar benefit in reducing stroke injury or improving stroke recovery.

In summary, using unsupervised scRNA-seq, we have found a significant increase in Ifi27l2a expression in MG following stroke, with an exaggerated upregulation in the aged stroke brain. We present evidence for pro-inflammatory roles of Ifi27l2a in MG phenotype regulation, wherein Ifi27l2a acts as a molecular regulator of microglial function and phenotypical changes. Based on these data, we propose that elevated expression of Ifi27l2a contributes to a pro-inflammatory MG phenotype (producing more ROS and proinflammatory cytokines in MG). We ultimately tested the effect of reducing Ifi27l2a in an ischemic stroke model. Partial reduction of Ifi27l2a led to reduced brain injury and inflammation, preserved motor function, and reduced long-term brain atrophy. In total, these findings suggest that targeting of Ifi27l2a expression in MG could be a strategy for reducing neuroinflammation in aging, stroke, or other neurodegenerative diseases to limit disease severity and promote improved functional recovery.

Methods

Animals

All procedures were performed in accordance with NIH guidelines for the care and use of laboratory animals and were approved by the Institutional Animal care and use committee of the University of Texas Health Science Center (AWC 23-0052, AWC23-0035). Sperm from Ifi27l2a^−/− KO mice [Ifi27l2atm1(KOMP)Vlcg]¹² were obtained from the Diamond laboratory at Washington University in St. Louis and used for in vitro fertilization of WT (C57BL/6 J) eggs (Genetically engineered rodent models core, Germ core, BCM). The resulting heterozygous Ifi27l2a^+/− progeny were backcrossed to establish the Ifi27l2a^−/− colony. To maintain the same genetic background, Ifi27l2a + /- males were bred to Ifi27l2a + /- females (i.e. “het to het” breeding). This breeding generates roughly 25% WT, 50% het, and 25% full KO. A WT littermate control was used for comparison. Male and female mice in a C57BL/6 J background (11–14 weeks old: young, 18–22 months old: aged) were used for all experiments. All animals were housed in the animal care facility at the University of Texas Health Science Center and had ad libitum access to food and water and were maintained on a 12:12 light: dark schedule at an ambient temperature of 69–73 °F (with 30–60% humidity).

Permanent distal middle cerebral artery occlusion (PDMCAO) model

C57BL/6 J mice of both sexes were used for scRNA-seq at 11–14 weeks or 18–22 months of age. PDMCAO was performed by permanently disrupting flow via the right middle cerebral artery (MCA) by micro-cautery (Accu-temp)⁸. Mice were anesthetized with isoflurane (4% induction and 2% maintenance; 21% O₂/balance N₂) and body temperature was maintained at 37 °C by feedback-controlled heating pad and rectal temperature probe. Bupivacaine (0.25% at 1 ml/kg) was injected subcutaneously (s.c.), prior to any skin incision³⁹. The distal MCA was accessed via a craniotomy and permanently occluded just proximal to the anterior and posterior branches by electrocoagulation. Sham controls were generated with same procedure minus the electro-coagulation of the MCA.

Brain sample preparation for single-cell RNA sequencing

Sham and PDMCAO surgeries were performed on young and aged mice of both sexes. Brains were harvested at 14 days after surgery (PSD 14) (Table 1). At PSD 14 (or 14 days after sham surgery), anesthetized mice were transcardially perfused with heparinized PBS (10 U/mL). Brains were removed from the skull and sliced coronally into 3 mm-thick blocks (from a region spanning +1 to -2 mm from bregma), covering the cortical infarction and secondary thalamic injury site⁸. The brain slice was then minced with a razor blade and subjected to the brain tissue dissociation protocol (Miltenyi Biotec, Gladbach, Germany). Minced tissue was then incubated a collagenase/dispase mixture (150 µL of 1 mg/mL in 2 mL) for 30 minutes at 37 °C in a gentleMACS Octo Dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany) using the pre-installed program for adult brain dissociation. Myelin was removed using debris removal solution. Red blood cells were lysed and removed with red blood cell lysis solution. The final cell suspension was stained with trypan blue and live cells were counted using Countess II FL Automated Cell Counter (Thermo Fisher scientific, USA).

GEM generation, library construction, and sequencing

The 10X Genomics Chromium™, Single-Cell RNA-Seq System (10X Genomics, Pleasanton, CA) was used to prepare cells for scRNA-seq. Brain single cell suspensions were processed to generate barcoded cDNA libraries using GEM gel bead, Chip kit, and library kits (10X Genomics, Pleasanton, CA) as per the manufacturer’s instructions. Cells were partitioned with beads containing reagents (primers and RT) required for generating 10X barcoded cDNA in individual cell using Chromium^TM controller. The resulting cDNA libraries were sequenced with NextSeq500/550 Hi Output Kit v2.5 (75 Cycles, 20024906) on an Illumina NextSeq 500 System.

Sequencing data processing and analysis

The cell ranger pipeline (10X genomics, Pleasanton, CA) was utilized to map the sequences to mouse reference genome (mm10), and to process barcode containing sequence data, aligning the read and generating feature barcode matrices that could be further processed by the Seurat package¹³ using R. We also used cellranger aggr pipeline to combine outputs from multiple samples into one output file. AGGR1 (5616 total cells analyzed) was a combined population consisting of “sham brains of young male and young female” mice. AGGR3 (5073 total cells analyzed) was a combined population consisting of “sham brains of aged male and aged female” mice. AGGR2 (12,723 total cells analyzed) was a combined population consisting of “Stroke brains from young male and young female” mice. AGGR4 (total 7791 cells analyzed) was a combined population consisting of “Stroke brains from aged male and aged female”. Reads were mapped to the mm10 murine transcriptome (10X genomics, Pleasanton, CA). We used Seurat 3.1 and 5.0.1⁴⁰ to analyze scRNA-seq data for clustering, and DEG identification between clusters and between the two groups. The argument “min.cells = 5, min.features = 500” was used to filter out the cells. Then, log-normalization using NormalizeData was utilized. Feature counts for each cell were divided by total counts for the cell and multiplied by the scale factor (10,000). Then using log1p, data was natural log-transformed. The Uniform Manifold Approximation and Projection (UMAP) dimensional reduction technique (UMAP) was used for dimensional reduction and clustering was carried out using FindNeighbors and FindClusters with the resolution parameter either at 0.019 (generating 6–7 clusters) or at 1 (generating 25–27 clusters). Conserved cell type markers in each cluster were identified by using FindConservedMarkers. The name and level of genes that were differentially expressed in each cluster was determined using FindMarkers. Metadata and normalized read count data was extracted from Seurat objects and fed into Excel to further identify the critical genes (top 10 genes or top 50 genes) that were up- and down-regulated in each cluster and to find the correlation between levels of Ifi27l2a and other MG genes.

Single molecule in situ hybridization (smRNA-FISH)

The RNAscope fluorescent multiplex assay (Advanced Cell Diagnostics, Newark, CA, USA) was performed according to manufacturer’s instructions with 2 week post-stroke brains of aged mice (18–20 months) and aged shams to probe for Ifi27l2a and Tmem119 transcripts in brain cells. Mice were deeply anesthetized with 2.5% avertin (0.25 mL, i.p.) and then perfused transcardially with heparinized PBS (10 U/mL) followed by 4% paraformaldehyde (PFA). Brains were subsequently stored in 30% sucrose before sectioning using a microtome at 30-µm thickness. The murine Ifi27l2a probe and Tmem119 probe were designed by ACD Biosystems. Brain sections (PFA fixed, 30-µm thickness) from the 2-week post stroke brain and sham brains of aged mice (18–20 month old) were hybridized with Ifi27l2a and Tmem119 probes for 2 hours at 40 °C. At the same time, ACD 3-plex positive control and negative control probes were incubated on one brain section to confirm signal specificity. The probes were amplified according to the manufacturer’s instructions and labeled with Opal-480 and Opal-570 fluorophore (Akoya Biosciences, Marlborough, MA, USA). DAPI was used to label nuclei. Images were taken with a fluorescent microscope (Leica DMi8 fluorescence microscope system, Leica Biosystem, IL, USA) and a confocal microscope (Leica TCS SPE confocal system, Leica Biosystem, IL, USA). Multiple images were captured with the 10X objective covering the hemisphere and stitched to generate a single image (Leica LAS X software).

Cell culture

Human microglial cell line 3 (HMC3) cells were purchased from ATCC (CRL-3304, USA) and cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Thermo Fisher Scientific, Waltham, MA, USA) containing 10% fetal bovine serum, 20 ng/mL recombinant human M-CSF1 (Tonbo Biosciences, San Diego, CA, USA) and antibiotics (Pen/Step) in 5% CO₂ and 37 °C.

Brain processing and immunostaining

For detecting gliosis (Iba1 and Gfap) in the thalamus following stroke, we performed immunostaining as previously described⁸. Mice were deeply anesthetized with 2.5% avertin (0.25 mL, i.p.) and then perfused transcardially with heparinized PBS (10 U/mL) followed by 4% paraformaldehyde (PFA). Perfused brains were then submerged in 30% sucrose in PBS for 24 hours at 4 °C prior to sectioning at 30 µm thickness (Micron HM 450, Thermo Fisher Scientific, Waltham, MA, U.S.A.). Sections corresponding to - 2 mm from bregma, which contains hippocampus and thalamus, were washed with PBS, incubated with blocking buffer (10% goat serum, 0.3% Trion X-100 in PBS), and then incubated overnight at 4 °C with the following primary antibodies: Rabbit anti-Iba1 antibody (1:200) (Wako Pure Chemical, Japan), mouse anti-GFAP antibody-cy3 (1:500) (Millipore Sigma, MO, USA). We used either donkey anti-rabbit IgG-Alexa 594 or 488 (1:200, Thermo Fisher Scientific, Waltham, MA, USA) to recognize rabbit anti-Iba1 antibody. Sections were incubated with DAPI (4′, 6-diamidino-2-phenylindole) to label nuclei. Images were obtained using a Leica TCS SPE confocal system and a Leica DMi8 fluorescence microscope system (Leica Biosystem, IL, USA). Images were captured using a 10X objective for the composite stitched image and with 20X or 40X for higher magnification images of selected regions. To measure thalamic gliosis, ImageJ (threshold-analyze particles) was then used to calculate the area of IBA1-positive area in the given field.

Lentivirus infection and ROS measurement by flow cytometry

Control lentivirus (Cx3cr1-IRES-eGFP, initial titer-1.55×10⁸ TU/ml) and Ifi27l2a expressing lentivirus (Cx3cr1-Ifi27l2a-IRES-eGFP, initial titer- 1.07×10⁸ TU/ml) were generated (GeneCopoeia, Rockville, MD, USA) and went through in-house quality control and validation. Sim-A9 cells, a microglia-like cell line, was transduced with control lentivirus (eGFP alone) or Ifi27l2a expressing lentivirus at 5 MOI using polybrene (Millipore Sigma, St. Louise, MO, USA). Five days after infection, cells were incubated with CellRox (Thermo Fisher Scientific, Waltham, MA, USA) for ROS detection or MitoSox (5 $\mu$M) (Thermo Fisher Scientific, Waltham, MA, USA) for mitochondrial-derived ROS detection for 10 minutes at 37 °C. Cells were analyzed using a CytoFLEX S flow cytometer (Beckman Coulter Life Science, Indianapolis, IN, USA). For analysis, a gating strategy was applied first to remove debris using forward (FSC-A) and side scatter (SSC-A). Doublets were also excluded from analysis by FSC-height and width. CellRox Deep Red signal (excitation/emission; 644/665) was collected in the channel (BP 660/20) and the MitoSox Red signal (excitation/emission; 510/580 nm) in the channel (BP 585/42). Data were exported and analyzed with FlowJo software (FlowJo, Tree Star Inc., Ashland, OR, USA). The geometric mean of fluorescence intensities (MFI) and percentage of positive cells were calculated and expressed.

Mouse Ifi27l2a ELISA

To check the intracellular levels of Ifi27l2a protein in primary microglia, the murine interferon alpha-inducible protein 27-like protein 2 A (Ifi27l2a) was quantified by ELISA following the manufacturer’s recommendations (Abbexa, Cambridge, UK) after washing the cells with PBS two times and collecting lysates in RIPA lysis buffer.

Real-Time quantitative RT-PCR

To validate the findings of scRNA-seq data, we performed qRT-PCR. For qRT-PCR analysis, mice were anesthetized with 2.5% avertin (0.25 mL, i.p.) and transcardially perfused with heparinized PBS (10 U/mL) to remove blood. Brains from naive young (3 mons) and aged mice (18–20 mons), or brains from sham and stroke mice (PSD 3 or PSD 14) were harvested and dissected to obtain both cortex and thalamus. Total RNA was purified with TRIzol™ Reagent (Thermo Fisher Scientific, Waltham, MA, USA) using the RNeasy Mini Kit (Qiaqen, Germantown, MD, USA) according to the manufacturer’s instructions. The purity of RNA ( > 1.7 at 260/280) and concentration of purified RNA were measured by Nano-drop Spectrometer and 1 µg of RNA was used to generate cDNA with iScript™ Reverse Transcription Supermix (Bio-Rad, Hercules, CA, USA). The SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, Hercules, CA, USA) was used to detect newly amplified amplicons with a C1000 Touch Thermal Cycler CFX96 Real-Time System (Bio-Rad, Hercules, CA, USA). The PCR cycles were as follows: initial denaturation at 95 °C for 30 sec, followed by 40 reaction cycles of 95 °C for 5 sec, 56 °C for 10 sec, and 72 °C for 10 sec. To quantify relative gene expression, we used the ΔΔCt method using Ct values for the gene of interest normalized to GADPH. Data was expressed as fold change relative to control samples. All primers were purchased from Integrated DNA Technologies (IDT, Coralville, Iowa, USA). Primer sequences are provided in Supplementary Table 4.

Statistical data analysis

Statistical data analysis was performed using Prism 10 (GraphPad Software, San Diego, CA, USA) and R in Rstudio environment with p < 0.05 considered statistically significant. Data are presented as the mean ± standard error of the mean (SEM), and analyzed using an unpaired t-test (for two group comparisons) or a one-way ANOVA with Tukey post-hoc test for multiple comparisons.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

S.P.M., L.D.M., and G.S.K. conceived the experiments. G.S.K., E. H., J.M.S., M.C.G., S.K., H.B., A.B., A.C., J.L., A.D., L.V., Z.W., C.T., J.B-G., J. A., C.T., S.D.K., L.V., and T.W. performed the experiments. G.S.K. and S.P.M. analyzed the results. S.P.M., G.S.K., J.E.J., J.D.W., and L.D.M. discussed the results. G.S.K., J.D.W., and S.P.M. made the figures and wrote the manuscript. All authors reviewed the manuscript.

Peer review

Peer review information

Nature Communications thanks Bozena Kaminska and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The scRNA-seq data generated in this study have been deposited in the Gene Expression Omnibus database under accession code GSE261494. Other processed scRNA-seq data for DEG/clustering info generated in this study are provided in the Supplementary data/Source data file. The open dataset GSE160936 (human, snRNA-seq of AD postmortem samples) was utilized to evaluate the expression of IFI27L2 in Alzheimer’s disease (AD) samples. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-56847-1.