Inflammation and the emerging role of the Toll-like receptor system in acute brain ischemia

Abstract

Background and Purpose

Systemic administration of cytosine-guanine oligodeoxynucleotides (CpG ODNs) provides neuroprotection against subsequent cerebral ischemic injury. We examined the genomic response of leukocytes and brain cells following ischemia in the context of CpG preconditioning.

Methods

RNA was isolated from circulating leukocytes and ischemic cortex 3 and 24 hours after middle cerebral artery occlusion (MCAO) following CpG or saline pretreatment and subjected to microarray analysis. Genes uniquely up-regulated in CpG pretreated mice were examined for over-represented transcriptional regulatory elements (TREs).

Results

CpG preconditioning induced a novel response to MCAO within circulating leukocytes that was dominated by NK cell-associated genes and the GATA-3 TRE. Preconditioning also caused a novel brain response to stroke that was dominated by Type I interferon- associated genes and TREs.

Conclusion

CpG preconditioning invokes novel leukocyte and brain responses to stroke. In this, CpG may be a unique preconditioning agent, coordinating peripheral and brain responses to protect against ischemic injury.

Introduction

Bacterial non-methylated cytosine-guanine oligodideoxynucleotide motifs (CpG ODNs) alert the body to infection through activation of Toll-like receptor 9 (TLR9). In mice, TLR9 is expressed by B cells, plasmacytoid dendritic cells (pDCs), macrophages, microglia, and astrocytes. TLR9-activated cells produce the pro-inflammatory cytokines TNFα, IFNα, and IL-12. These cytokines further activate monocytes, neutrophils, natural killer cells (NK cells), and T cells, facilitating a coordinated inflammatory response to pathogen invasion.

Pre-exposure to CpG reprograms the cellular response to subsequent TLR stimulation. Unlike naïve cells, macrophages pre-treated with CpG do not generate TNFα in response to TLR4 stimulation, instead generating IFNβ ¹. Furthermore, systemic administration of CpG increases resistance to polymicrobial sepsis ². Hence pre-exposure to CpG redirects both cellular and systemic responses to subsequent TLR stimulation.

Systemic administration of CpG also protects the brain from subsequent ischemic damage³. Such ‘CpG preconditioning’ is time and dose dependent and requires TNFα. The precise mechanisms responsible for CpG preconditioning are not well understood, but likely involve both direct cellular processes and coordinated systemic responses that minimize ischemic damage.

We hypothesize that CpG preconditioning reprograms the response of the brain and the peripheral immune system to subsequent stroke. Here we provide evidence for such reprogramming and consider its potential neuroprotective consequences.

Results

CpG preconditioning induces a NK cell-associated peripheral response to stroke

We evaluated RNA from blood leukocytes 24 hours following MCAO using Affymetrix oligonucleotide microarrays. We found 422 genes to be differentially regulated in CpG pretreated animals relative to saline. We next identified over-represented transcriptional regulatory elements (TREs) in the genes uniquely increased in CpG preconditioned animals. In those genes for which upstream sequence was available for analysis (234) a single TRE, GATA-3, was over-represented with an adjusted p value = 0.118. A network depiction of interactions between GATA-3 and genes in the CpG preconditioned cluster is displayed in Figure 1. GATA-3 is linked to 53% of the genes within this up-regulated cluster (124 of 234). GATA-3 plays a critical role in the development of NK cells. Literature review identified 24 of the up-regulated genes as NK cell-associated: Klra5, Klra7, Klra8, Klra10, Klra18, Klra22, Klrb1a, Klrb1c, Klrb1f, Klrc1, Klrc2, Klre1, Klrg1, Klrk1, Rantes, Cma1, Eomes, Fasl, Gzmb, Il2rb, Ncr1, Ndg1, Prf1, and T-bet. CpG activates NK cells indirectly via IL-12 released from activated dendritic cells (DCs). Serum IL-12 levels were significantly increased 24 hours after MCAO in preconditioned animals (data not shown). Together, our data demonstrate that CpG preconditioning induces a novel, systemic NK cell response to stroke.

Figure 1. The GATA-3 TRE is over-represented in blood leukocytes 24 hours following stroke in CpG preconditioned mice.

A PAINT- generated Hypothesis Gene-TRE Network depicting the genes uniquely up-regulated 24 hours after CpG preconditioning that contain the GATA-3 TRE. Genes are depicted as ovals. P value threshold set at 0.2.

CpG preconditioning induces a Type I IFN-associated brain response to stroke

We evaluated RNA from ischemic cortex 24 hours following MCAO using Affymetrix oligonucleotide microarrays. We found 223 genes to be differentially regulated in CpG pretreated animals relative to saline. We next identified over-represented TREs in the genes uniquely up-regulated CpG preconditioned animals. In those genes for which upstream sequence was available for analysis (136) we identified 4 over-represented TREs with an adjusted p value < 0.1. Notably, each TRE is Type I interferon (IFN) -associated (IRF, IRF8, ISRE, HMG-1Y). A network depiction of interactions between the identified TREs and the genes in the CpG preconditioned cluster is displayed in Figure 2. The IFN-associated TREs are linked to 64% of the genes within this up-regulated cluster (88 of 136). Literature review identified 12 of the up-regulated genes as Type I IFN- associated: Oas1a, MHC class I (H2-D1, H2-K1, H2-L, H2-Q6), Ifi203, Ifi204, Ifi205, Ifi27, Isg20l1, Lmp7, and Psmb9. Thus an altered signaling cascade involving Type I interferons exists in the brain following stroke in CpG preconditioned mice.

Figure 2. Type I interferon-associated TREs are over-represented in the brain 24 hours following stroke in CpG preconditioned mice.

A PAINT- generated Hypothesis Gene-TRE Network shows the relationships between the uniquely up-regulated genes 24 hours after CpG preconditioning and the TREs shared in common. Genes are depicted as ovals. P value threshold set at 0.1.

Discussion

We report the first evidence that CpG preconditioning alters the genomic response to stroke in circulating leukocytes and in the brain. We demonstrate a distinct pattern of NK cell activity in the blood and a clear enhancement of Type I IFN signaling in the brain following MCAO. This pattern of up-regulated gene expression underscores a unique response to brain ischemia that may actively protect the brain from injury.

CpG preconditioning induced a novel genomic response in blood leukocytes that was evident 24 hours after stroke. Of those genes uniquely up-regulated in preconditioned animals, a majority contained the GATA-3 TRE, which is required for NK cell development. Additionally, 24 of the up-regulated genes were NK cell-related and serum IL-12 was increased at this time, supporting the notion of increased NK cell activity.

This unique systemic response may play a role in neuroprotection as NK cells have been shown to limit damaging neuroinflammation in experimental autoimmune encephalomyelitis (EAE) ⁴. Interestingly, administration of CpG ODNs prior to EAE induction also reduces disease severity ⁵. Furthermore, treatment with CpG inhibits inflammatory arthritis in an IL-12- and NK cell- dependent manner ⁶. Hence, CpG may also initiate a protective NK cell response to cerebral ischemia.

CpG preconditioning also induced a novel genomic response in the brain that was evident 24 hours after stroke. Of those genes uniquely up-regulated in preconditioned animals, a majority contained one or more Type I IFN-associated TREs. Moreover, 12 of the up-regulated genes were associated with Type I IFN signaling, further supporting a role for IFNs following stroke in preconditioned animals.

Microglial, astrocytes, endothelial cells and neurons all produce IFNβ. IFNβ can stabilize the blood-brain barrier ⁷, suppress inflammatory cytokines ⁸, and protect neurons from cytotoxic microglia ⁹. Systemic administration of IFNβ reduces infarct damage in several models of ischemic stroke ^{10, 11}. Hence an increase in Type I IFN signaling within the brain has the potential to be neuroprotective.

Our data supports a shift toward Type I IFN signaling following stroke in CpG pretreated animals. How might this shift occur? Mice lacking TLR4 incur significantly less damage from MCAO than wild-type controls ¹², indicating a damaging role for this receptor in ischemic injury. Pretreatment with CpG shifts the cellular response to subsequent stimulation of TLR4, leading to a suppression of TNFα and an increase in IFNβ. A similar series of events might occur following CpG preconditioning wherein pretreatment with CpG shifts the response of TLR4 to subsequent stimulation with endogenous ligands, such as HSP60, released after stroke, and potentially leads to suppressed cytotoxic TNFα and enhanced neuroprotective IFNβ.

Alternatively, the systemic increase in NK cell activity may explain the Type I IFN shift within the brain. NK cells promote the release of IFNα from pDCs in a CpG- or IL-12- dependent manner ^{13, 14}. Hence pretreatment with CpG may activate DCs to produce IL-12, thereby activating NK cells which, in turn, induce pDCs to produce IFNα.

We have shown that CpG preconditioning reprograms the peripheral and central responses to stroke. The appearance of novel NK cell and interferon genomic “fingerprints” after ischemia indicates that CpG preconditioning fundamentally changes the body's inflammatory response to stroke. This is consistent with our previous reports of reprogramming in which ischemic and LPS preconditioning induce novel, protective sets of gene transcripts following stroke ^{15, 16}. CpG appears to be a unique preconditioning agent, coordinating both systemic and central immune components to actively protect the body from ischemic injury.

Materials and Methods

Mice

C57Bl/6 mice (male, 8-10 weeks) were obtained from Jackson Laboratories (West Sacramento, CA, USA). All mice were housed in a facility approved by the Association for Assessment and Accreditation of Laboratory Animal Care International. The animal protocols met National Institutes of Health guidelines with the approval of the Oregon Health and Science University Institutional Animal Care and Use Committee.

Drug Treatments

CpG ODN 1826 (20-40ug; 200ul; Invivogen, San Diego, CA, USA) or saline was administered by intraperitoneal injection 72 hours before MCAO.

Surgery

Mice were anesthetized with isofluorane and ischemia was induced by MCAO as published previously¹⁷. Cerebral blood flow was monitored with laser Doppler flowmetry and temperature was maintained at 37°C. After surgery, mice were kept for 24 hours on a heating pad with access to soft food and water.

RNA isolation

Mice were anesthetized and blood was obtained via retro-orbital puncture. Animals were perfused with saline and, under RNase-free conditions, a 1 mm section was removed for infarct area analysis. The ipsilateral cortex region from the frontal 4 mm was snap frozen. Total RNA was isolated from the blood using the Qiagen PAXgene Blood RNA Kit and from the brain using the Qiagen RNeasy Lipid Mini Kit (Qiagen Inc.). RNA from individual animals was hybridized to single arrays.

GeneChip Expression Analyses

Microarray assays were performed in the Affymetrix Microarray Core of the Oregon Health & Sciences University Gene Microarray Shared Resource. RNA samples were labeled using the NuGEN Ovation Biotin RNA Amplification and Labeling System_V1. Quality-tested samples were hybridized to the MOE430 2.0 array and processed with Affymetrix GeneChip Operating Software (GCOS). Data was normalized using the Robust Multichip Average method. Normalized data was analyzed by multivariant ANOVA for each gene. P-values were adjusted for multiple comparisons using the Hochberg and Benjamini method. Significance was determined by p < 0.05 and fold change ≥ 2 for blood analyses and ≥1.5 for brain analyses.

Transcriptional regulatory network analysis

For our reference comparison group, we identified putative TREs in the 5000 bp upstream sequence of transcripts represented on the MOE430 Affymetrix gene chip using TRANSFAC PRO database version 10.4. We then determined the over-represented TREs in the uniquely up-regulated gene cluster compared to the reference group using Promoter Analysis and Interaction Network Toolset (PAINT) version 3.5.